Vous êtes sur la page 1sur 93

the ITN community presents

Networks
Transformation
Guidelines

Evolutions in Wireless Access


Domain and Microwave

Name of the editor: Ronan Le Bras, Roman Łapszow

2017 edition
th
Version 1.0 date: 12 June 2017

Edited by: Written by:

Ric Bailey, Sébastien Béchet, François Brunet, David Cailliere, Maha Chouaref,
Alain Choumaker, Patrice Desmoulin, Vincent Diascorn, Jakub Domin, Maurizio
Fazzi, Colin Fraser, Christan Gallard, Jean-Pierre Gallen, Anna Maria Galindo
Serrano, Benoît Graves, Philippe Hamet, Eric Hardouin, Nathalie Jollivet, Steve
Jones, Serge Langouët, Arnaud de Lannoy, Ronan Le Bras, Fryderyk Lewicki,
Alioune Ly El Hadji, Roman Łapszow, Bernard Missir, Jerome de Murcia, Dan
Ronan Le Bras, Roman Łapszow
Musat, Laurent Mussot,Franck Payoux, Adrian Pike, Jose Pirlot, Yann Pitrel,
Glyn Roylance, Tomasz Rudy, Berna Sayrac, Alan Stidwell, Vanesa Tornero,
Ferdinand Tra, Tan Tran, Henk Tubbe, Bertrand Waels, Stefan Wendt, Jacek
Werner, Mark Woodgate, Carl Woolley.

-3- Orange confidential


Content

Content ............................................................................................................................... 4
1. Introduction ............................................................................................................................................................. 7
2. Spectrum ................................................................................................................................................................. 7
2.1 2.6 GHz ........................................................................................................................................................... 8
2.2 800 MHz .......................................................................................................................................................... 8
2.3 Re-farming of 900 MHz, 1800 MHz and 2100 MHz ......................................................................................... 8
2.4 Future bands (700 MHz, 3500 MHz…) ............................................................................................................ 9
3. 2G/3G Evolution .................................................................................................................................................... 13
3.1 Introduction – 2G/3G network life span ......................................................................................................... 13
3.2 2G/3G functional evolution ............................................................................................................................ 13
3.3 3G evolution – target configuration................................................................................................................ 14
4. 4G Evolution .......................................................................................................................................................... 15
4.1 Peak Throughput oriented features ............................................................................................................... 15
4.2 RAN Drivers to introduce time/phase synchronization .................................................................................. 18
5. Conversational Services........................................................................................................................................ 19
5.1 SMS .............................................................................................................................................................. 19
5.2 Voice ............................................................................................................................................................. 19
5.3 Video calls for LTE subscribers ..................................................................................................................... 22
6. Automation of network operations ......................................................................................................................... 22
6.1 SON features’ summary ................................................................................................................................ 23
7. LTE for Fixed Access ............................................................................................................................................ 24
7.1 Status of LTE for fixed services ..................................................................................................................... 24
7.2 Orange Group activities on LTE for Fixed services ....................................................................................... 25
7.3 Network aspects of LTE for Fixed services ................................................................................................... 25
7.4 Deployment Scenario & Devices ................................................................................................................... 26
7.5 Radio Aspect: Performance and capacity ..................................................................................................... 27
7.6 B2C services architecture ............................................................................................................................. 27
7.7 B2B services architecture .............................................................................................................................. 28
8. V2V and V2X ......................................................................................................................................................... 29
9. Professional Mobile Radio ..................................................................................................................................... 30
9.1 Limitations of legacy PMR networks .............................................................................................................. 30
9.2 LTE for next generation of PMR networks ..................................................................................................... 31
9.3 PMR: new business opportunities for Orange ............................................................................................... 33
9.4 On-going PMR activities within Orange ......................................................................................................... 34
10. Public IoT/LPWA/Mobile IoT Networks .................................................................................................................. 36
10.1 IoT at Orange ................................................................................................................................................ 36
10.2 LPWA Networks at Orange .......................................................................................................................... 37
10.3 3GPP solutions: Mobile IoT ........................................................................................................................... 40
10.4 Orange Strategy for IoT: LoRa and LTE-M ................................................................................................... 44
11. Small Cells (all 3G and 4G) ................................................................................................................................... 45
11.1 Outdoor Small Cells ...................................................................................................................................... 46
11.2 Indoor Solutions ............................................................................................................................................ 48
11.3 Small Cells for 4G advanced and 5G (Unlicensed bands LTE-U, LAA) ........................................................ 56
11.4 Wi-Fi .............................................................................................................................................................. 56
12. Antenna Systems .................................................................................................................................................. 59
12.1 Antenna Line Devices ................................................................................................................................... 59
12.2 Horizontal Antenna Sectorisation with Passive Antennas ............................................................................. 65
13.3 Active Antenna Systems Evolution ................................................................................................................ 66
13. Radio Access Network Architecture Evolution ...................................................................................................... 69
13.1 Introduction.................................................................................................................................................... 69

Orange confidential -4-


13.2 Virtualized RAN architecture ......................................................................................................................... 69
13.3 Virtualized RAN pro & cons ........................................................................................................................... 70
13.4 Next steps and conclusions ........................................................................................................................... 70
13.5 5G RAN Technology (as currently under definition in 3GPP) / vRAN ............................................................ 71
5G RAN architecture .................................................................................................................................................. 71
14. Other Wireless Technologies (CDMA, WiMAX)..................................................................................................... 71
14.1 CDMA-2000 Family ....................................................................................................................................... 71
14.2 WiMAX .......................................................................................................................................................... 71
15. Evolution of the microwaves domain ..................................................................................................................... 73
15.1 Microwave Equipment evolutions .................................................................................................................. 73
15.2 MW capacities and upgrades ........................................................................................................................ 74
15.3 Point to Multi-Point (PtMP) solutions in licensed frequency bands for fixed and mobile aggregation ............ 75
15.4 Sub 6 GHz Spectrum Band Low cost solution ............................................................................................... 76
15.5 Microwave Research and Development ........................................................................................................ 77
15.6 MW equipment Roadmap deployment recommendation ............................................................................... 78
15.7 Microwave Spectrum Regulatory Condition................................................................................................... 80
16. Design to cost Coverage solutions for rural areas in emerging countries .............................................................. 80
16.1 Technical Solution ......................................................................................................................................... 80
16.2 Geo-Marketing............................................................................................................................................... 81
16.3 Tools ............................................................................................................................................................. 81
16.4 Recommendations ........................................................................................................................................ 81
17. Green RAN ............................................................................................................................................................ 84
17.1 Background ................................................................................................................................................... 84
17.2 Energy Savings Features .............................................................................................................................. 84
18. On the road to 5G.................................................................................................................................................. 86
18.1 5G timeline .................................................................................................................................................... 86
18.2 What services will 5G offer? .......................................................................................................................... 87
18.3 What will be the 5G radio access technology? .............................................................................................. 88
18.4 Which spectrum for 5G? ................................................................................................................................ 89
18.5 Global initiatives towards 5G ......................................................................................................................... 91

-5- Orange confidential


1. Introduction
Wireless Network evolution is playing a key role in delivering on Orange Essential 2020 targets. The emphasis is not only
on increasing the performance of our mobile networks for faster mobile broadband experience but also to provide new
features enabling innovation in new areas of diversification for our business.
As a milestone on the road to 5G, LTE Advanced has enabled Mobile broadband data rate to reach several hundred of
Mbps in the most advanced networks of the Orange footprint. User expectation and daily usage is rising at a fast pace
thanks to LTE both at home and on the move. The use of and High-order MIMO and Carrier Aggregation with more than
three bands and potentially combining Unlicensed Spectrum in LAA scenario is raising the expectation even closer to the
1 Gbps speed limit.
Mobile networks are becoming an attractive alternative for fix services in MEA where Fix LTE is being roll-out in several
countries as a migration of Wimax opening opportunities for a wider adoption. Such a trend is also visible in Europe
where several affiliates are using LTE for Home Broadband and are considering even more aggressive use of LTE for fix
for B2B customer..
In addition, LTE Advanced features from Rel 13 onwards are opening new opportunities of services for Affiliates and
Orange Business Services, new verticals are increasingly being targeted to diversify our business, most notably PMR,
IoT and Automotive.
In the Private Mobile Radio domain, the replacement of obsolete networks based on TETRA in Europe represents a
significant opportunity for expansion covering both the Public Safety and Disaster Relief (PPDR) case where Orange is
active especially in France, and the PMR Services or Private LTE Networks for enterprise.
In the domain of the Internet of Things identified as one of the diversification areas of the Essential 2020 plan, Orange
has adopted a dual approach by entering the market with LoRa a non-cellular solution to be complemented with the use
of LTE-M that will address a wide range of use cases. Revenues on pure connectivity being modest, providing an E2E
service including IoT platforms is also a critical part of our IoT strategy.
In the automotive sector, LTE Advanced enabled this year some first tests of V2V & V2X use cases that will pave the
way to the autonomous car of tomorrow and increase safety for all. Even if these are only the first technical brick towards
real operational applications, they already show the impact the Mobile Networks evolution could have on our future.
Finally, the last part of the NTG 2017 Wireless Evolution is focused on the Road to 5G as it is clear that this will be the
focus of the years ahead until the first launch planned in 2020 and even afterwards. 5G is encompassing several major
evolutions in the radio domain and will rely on new CN and virtualisation, it is major transformation for mobile operators
both for their internal organisation as well as in the way they will handle existing and new services for customers and
verticals.

Contact: Ronan Le Bras

2. Spectrum
Licensed spectrum remains a major asset for mobile operators and forms the first building block towards being a “mobile
operator”. This principle has ruled mobile networks for the last decades, however recent publicity around unlicensed
spectrum, most notably the so-called “ Wi-Fi bands” at 2.4 GHz and 5 GHz, may have given an impression that one
could claim itself being a “mobile operator” by simply transmitting within the “free to use” bands.
In reality the licensed bands maintain a significant performance gap from their unlicensed counterparts, meaning
that whilst use of Wi-Fi could become a handy complement to licensed spectrum, it would by no means be a replacement
of the core 2G, 3G and 4G licensed bands, with 900, 1800, 2100 MHz and the additional 800 and 2600 MHz bands as
the main assets to provide operations of UMTS and LTE. In 2017 we should see the emergence of LTE in the unlicensed
5 GHz band with commercial launch of networks with standardized listen before talk (LBT) functionality. There is also
interest for TDD bands, in particular Band 40 (2.3 GHz) and Band 38 (2.6 GHz) which can be used to deliver LTE for
Fixed services in AMEA as a migration from WiMAX services.
nd
Besides, as the spectrum landscape constantly evolves, affiliates should keep an eye to the future availability of the 2
Digital Dividend at 700 MHz, which has been adopted at the WRC in 2015 and allocated in France, the 3.4 to 3.8 GHz
spectrum, seen as a capacity extension spectrum band for 4G or more likely 5G, the Downlink-only L-band in 1500 MHz
which could be used in “Supplemental Downlink” or SDL mode, aggregated with a paired spectrum band, i.e. 1800 MHz,
and also potentially the 2.3 GHz band.
The possible spectrum bands for LTE and UMTS, whether “new” or re-farmed may be obtained by two different means:
 From spectrum acquisition: 2.6 GHz, 800 MHz, additional 900, 1800 or 2100 MHz blocks
and later on 700 MHz and 3500 MHz (those bands are also candidates for 5G as stated below).
 Through re-farming of existing technologies: primarily GSM900, GSM1800 and UMTS2100.

Concerning 5G, it is expected that it will rely upon a larger spectrum portfolio compared to previous generations. Higher
frequency in the 24 to 86 GHz range bands could be suitable for ultra-fast broadband delivery in dense urban
environments. Frequency bands in these higher ranges are expected to be identified globally during the ITU World Radio
Conference in 2019. Orange supports such identification of higher frequencies, for exclusive use by 5G.

-7- Orange confidential


In line with this general analysis, at European level Orange fully supports EU 5G action plan based on the following
strategic options on spectrum:
 3400-3800 MHz is the primary band suitable for the introduction of 5G-based services
 700 MHz band to enable nation-wide and indoor 5G coverage
 24.25-27.5 GHz (so-called '26 GHz') band to be a pioneer band for Europe

2.1 2.6 GHz


The 2.6 GHz is a primary target band in LTE for high-speed and capacity but not deep indoor
The 2.6 GHz band represents a prime target for LTE, with generally the largest bandwidth available to boost the
maximum performance of LTE. With 2x20 MHz of spectrum, peak throughput of 150 Mbps is theoretically possible. The
main drawback of the band is the requirement of new antenna systems. Higher propagation and penetration losses at
this frequency also imply degraded coverage compared to lower frequencies. However other factors such as higher BS
and UE antenna gains as well better performance of MIMO & Rx diversity at higher frequencies tend to offset some of
these propagation losses.

Band plan
ECC DEC (05)05 FDD TDD FDD
UL DL
f
2500 MHz 2570 MHz 2620 MHz 2690 MHz

Figure 1 – European band plan for the 2.6 GHz band (ECC DEC (05)05)

2.2 800 MHz


The 800 MHz is a primary target band for LTE for deep indoor coverage
Most countries have aligned to the CEPT band plan:
 Reversed band plan: DL on lower part, UL on upper part,
 1 MHz guard band with DVB-T below 790 MHz,
 Duplex gap 821-832 MHz is used by PMSE (Radio Microphone).

Figure 2 – 800 MHz band plan

LTE at 800 MHz delivers much better coverage than either 1800 MHz or 2600 MHz in rural and suburban environments.
Overall a difference of around 10-12 dB may be achieved between 800 MHz and 2600 MHz. This has a significant
impact at cell edge. However in good radio conditions there is no obvious difference between 800 MHz and 2600 MHz
since performance is limited by interferences rather than coverage.
The downside of 800 MHz is the co-existence issue between LTE800 and DVB-T in the UHF band.

Figure 3 – LTE800 and DVB-T spectrum

LTE800 transmitter potentially interferes neighbour Digital TV receivers (Figure 3). This has led to thousands of filters
installed onto people’s home TV sets. MCOs should work with other local operators (e.g. joint venture) on the
interference avoidance when introducing LTE800 as done in France.

2.3 Re-farming of 900 MHz, 1800 MHz and 2100 MHz


All re-farming bands have the advantage to allow the re-use of existing antennas and feeders.
The 900 MHz band is seen as the target band to host legacy technologies, i.e. GSM and UMTS, with around 5 MHz
each. LTE in this band is supported by a number of devices, however eNBs do not support triple mode
GSM+UMTS+LTE, and spectrum assets are generally insufficient to cater for all three technologies. Recommended 900

Orange confidential -8
MHz usage is a split between 5 MHz kept for GSM as a “legacy” GSM layer to sustain long-term GSM users such as
M2M, roamers and low-cost voice-only devices, and the other 5 MHz re-farmed to UMTS.
The 1800 MHz re-farming has seen wide adoption of LTE1800, leading to progressive switch-off of GSM in this band, as
long as sufficient capacity remains in 900 MHz for GSM. Longer-term, 1800 MHz will be entirely re-farmed to LTE.
Until recently, UMTS was the only realistic technology that could be operated in 2100 MHz. We are now seeing a
growing eco-system supporting LTE2100, with base stations and most recent 4G smartphones supporting this band. This
can be an efficient solution for deployment of LTE on sites where no 1800 MHz was deployed, i.e. GSM900/UMTS2100
sites. If deployed as a complement to other LTE bands, one limitation is that Carrier Aggregation combinations with
LTE2100 are not supported as well as with more traditional LTE bands.
Multi-Standard Radio equipment in base stations allows some flexibility in migrating from one technology to the other
within the same band.

Figure 4 – Summary table of device support for LTE900 and LTE2100

Many Orange affiliates have deployed LTE in 1800 MHz, notably Orange Spain, Mobistar, Orange Luxemburg, Orange
Poland and in AMEA Sonatel in Senegal, Morocco, Botswana, Mauritius, Reunion Island, and Jordan, some being still
pre-commercial. A trial of LTE900 has even been completed in Orange Jordan. A trial of LTE2100 took place in Spain.

2.4 Future bands (700 MHz, 3500 MHz…)


Future bands allocations may be deployed with 4G or reserved for future 5G use.
One key question for any new band introduction in 4G will be the actual timing of 5G, if confirmed around 2020, but also
the nature of what 5G will be, as standardisation has not formally started.
The window of opportunity will narrow down for 4G as new bands available within the 2018/19 timeframe may be skipped
and targeted with 5G directly.

700 MHz
 Allocation
• 2x30 MHz FDD between 703 and 788 MHz with an additional service to be confirmed in the duplex
gap. Several options are proposed for the use of the duplex gap, ranging from:
 SDL (Supplemental Downlink)
 PMSE (Program Making and Special Events), e.g. wireless microphone/camera
 PPDR (Public Protection and Disaster Relief) for public safety services
 M2M (Machine to Machine)
No harmonisation has been reached so far in the duplex gap, however France already confirmed the
allocation of PPDR spectrum, with 2x5 MHz + 2x3 MHz allocation.

Figure 5 – 700 MHz band plan

 Availability
• First auctions have already taken place in some European countries, e.g. France, although real
availability may be later (France whole country in 2019).

-9- Orange confidential


• Early availability also potentially possible in AMEA, even before 800 MHz. Globally first countries
expected in 2018/19.
 Technical considerations
• It could be difficult to incorporate this band into multi-band macro BS antennas without increasing the
dimensions, or sacrificing performance, interference (PIM) between 700 & 800 MHz will require “side-
by-side” 700 & 800 MHz antennas.
As a general rule, 700 & 800 MHz should not be combined on to the same antenna column.
• A large proportion of filters protecting DVB-T TVs from LTE800 today will need to be replaced with new
filters to protect from LTE700 BSs and from devices.
 RAN support
• All RAN vendors already support the 700 MHz band
 Devices support
• It is expected to be widely supported in devices. 700 MHz is already supported in APAC but with
different specs (out of band emission rules).
• Support of Carrier Aggregation of 700 MHz and 800 MHz in a device is not guaranteed but proposals
have recently been presented to 3GPP, which may lead to devices circa 2018/19.
 Strategy
• The need of 700 MHz in 4G on top of 800 MHz is not obvious since it will not provide extra coverage
but only allow for more capacity in deep indoor.
• The 700 MHz band may well be reserved for future 5G deployments, as there should be a need for a
coverage layer in 5G.

2300 – 2400 MHz (Europe)


 Allocation
• The maximum band allocation is 100 MHz, but in many countries is at least partially occupied by
other users (e.g. military) and the amount of spectrum release by the regulator could be reduced. It is
also allocated to TDD, with no option for FDD.
• Significantly MNOs will only be able to obtain the spectrum via the Licence Shared Access (LSA)
mechanism, whereby the MNO would have ‘secondary’ rights to the spectrum, as agreed with
the ‘primary’ owner. This would mean the MNO agrees to deploy in the spectrum on either a time,
and/or geographical basis as agreed with the ‘primary’ owner. E.g. the agreement could be that
operation could be allowed for 10 years, and after that period (without re-negotiation) MNO operation
would have to cease.
• Orange is not keen on LSA based spectrum.

3400 - 3800 MHz


 Allocation
• 2 blocks of frequencies identified
 3400 – 3600 MHz
 With 2 options possible for regulators to choose from: TDD (band 42) or FDD (band
22), although eco-system is only developing in TDD, meaning the FDD option is
highly unlikely.
 3600 – 3800 MHz
 The only option is TDD (Band 43)

Figure 6 – FDD band 22

Figure 7 – TDD band 42

• In some countries (e.g. Romania), renewal of WiMAX licenses lead regulators to anticipate
auctions for allocations of 3.5 GHz bands. Given the uncertainty on European harmonization of the
3.4-3.6 GHz band on the choice of TDD and FDD, interim allocations may be defined. Romania opted
for an FDD band plan which could be re-allocated in TDD before 2020.
• Restriction of use to Fixed Wireless Access use is common, although it is anticipated that this
band will be allocated to Mobile Broadband short to mid-term

Orange confidential - 10
 Availability
• Already available in some countries, generally for FWA deployment initially
• Extension to Mobile Broadband possibly by 2018/19 in Europe, with variable conditions on FWA
depending on legacy of WiMAX operators.
• Earlier availability possible in AMEA.

 Technical considerations
• Mixing FWA and Mobile Broadband use on the same band should be possible (Huawei have eNB
capable of operating both FWA and LTE in TDD in the same band)
• TDD requires tight network synchronisation (2 ms) which would involve a network upgrade.
• Recommended rules for multi-operator deployment is to have no guard band and agree on:
 A common synchronization source (e.g. UTC provided by GPS)
 A common frame configuration (i.e. DL/UL ratio)
• Deploying TDD with 5 MHz guard band should be possible but is not within regulator plans. It may also
be challenging due to potential interference between operators due to blocking.
• RF can be easily incorporated into small cell products due to the small antenna size. Products already
in the roadmap, i.e. Nokia claim support of first 3.5 GHz TDD small cell, with potential support of an
FDD version later.
 RAN support
• All vendors support band 42
• Band 43 support is more variable
• No plans for Band 22 by any vendor
 Devices support
• Commercial devices today are all Fixed CPE, Routers or MiFi. No Smartphone to date.
• Only Band 42/43 supported today, no plans from vendors, even in the future, for Band 22
• Qualcomm announced the Snapdragon X16 LTE modem supporting B42/B43, therefore commercial
Smartphones are expected in 2017
Strategy: Both band 42 & 43 are essential bands for mid to long-term deployment and may be a “golden
band” for 5G. In the meantime earlier deployments for Fixed Wireless Access as a continuation of
WiMAX is possible. If so mixed modes allowing both FWA and LTE should be preferred.

LTE in 5 GHz unlicensed band – LAA (License Assisted Access)


 Allocation
• 2 modes are being standardized in 3GPP to make use of 5 GHz in a more effective way than standard
Wi-Fi:
 LAA (License Assisted Access) which requires aggregation of a licensed band, e.g. 2600
MHz, with LTE in downlink-only in the 5 GHz unlicensed spectrum.
 LWA (LTE Wi-Fi Link Aggregation) which requires aggregation of LTE in a licensed band with
Wi-Fi in 2.4/5 GHz bands. Aggregation is made at PDCP level, combined in the same S1
interface.
 LTE in 5 GHz is not being standardized yet as a stand-alone mode, although some US
operators are pushing for it (MulteFire).
• LAA versions will be with up to 80 MHz total, including 3x20 MHz of 5 GHz, leading to 600 Mbps total
speed (tested with Nokia on Qualcomm Snapdragon 820 MSM8996 – 1CA 270Mbps in OPL)
• Spectrum usable in 5 GHz follows Wi-Fi allocations. For Europe:
 5170 to 5330 MHz:160 MHz at max 200 mW
 5490 to 5710 MHz: 230 MHz at max 1W

Figure 8 – LAA Rel13: unlicensed band in DL-only mode

- 11 - Orange confidential
Figure 9 – LWA architecture

 Availability
• A first group trial has been completed with Qualcomm. Tests with vendors (Nokia and Ericsson)
accomplished successfully with LTE-U. Commercial launch expected in 2017 with “Listen Before Talk”
(LBT)
• Co-existence of LAA with Wi-Fi is managed through LBT mechanisms required by standards
 Technical considerations
• Due to low power, this is not a macro network technology, only for small cells
 Devices support
• Strong support from Qualcomm, expected Snapdragon 835 on X16 platform supporting LBT
 Strategy
• Two scenarios seem of interest for Orange:
 Enterprise through indoor small cells such as Pico-cells. Co-existence with WLAN is the main
challenge, which may see reluctance from venues operating Wi-Fi
 Outdoor small cells in hot-spot mode
• Interest for DAS systems to boast capacity, although not yet planned by vendors

The Orange Group priorities for bands 450-5600 MHz has been presented on Figure 10 below. Main spectrum bands
and recommendation are listed on Figure 11 :
450 700 800 900 1500 1800 2100 2300 2600 3500 3700 5600
2G
3G CDM A

4G
5G

In-Use Possible Unlikely


Figure 10 – Summary table of Group priority bands per techno

Figure 11 – Main spectrum bands and recommended technologies

Orange confidential - 12
Spectrum band priorities for UMTS, LTE and NR radio networks: Group Recommendation 2—1
 800 MHz: primary band for LTE with lower spectrum bandwidths than 2.6 GHz and aiming at rural and deep
indoor coverage.
 900 MHz: target re-farming band for UMTS, with a target to use 900 MHz as the only remaining band for GSM
/ UMTS legacy users.
 1800 MHz: prime band for LTE deployment and a good trade-off between capacity and coverage, with re-use
of existing antennas where GSM traffic allows spectrum re-farming.
 2100 MHz: currently used exclusively for UMTS however short to mid-term re-farming to LTE could be
considered.
 2600 MHz (FDD): primary target band for LTE, aiming at peak LTE performance and capacity.
 2300/2600 MHz TDD (band 38/40) seen as a target for “LTE for Fixed Wireless Access” as a migration from
WiMAX, typically in AMEA.
 3500 MHz: eco-system developing only in TDD (Band42) meaning FDD options is unlikely (Band22). Seen as
a potential key band for 5G. Earlier use as a FWA + LTE mixed band is possible.
 3700 MHz TDD (Band43): eco-system not as strong as band 42, to be used short-term for FWA and later on
with 4G or 5G.
 700 MHz: new band seen as future complement to 800 MHz (2018/19), potentially to be reserved for 5G
 5 GHz (unlicensed): potential use for LTE in LAA (License Assisted Access) being assessed with trials as
LTE-U and potential commercial deployments in 2017.
Contributors: Glyn Roylance, Roman Łapszow Contact: Ronan Le Bras

3. 2G/3G Evolution
3.1 Introduction – 2G/3G network life span
Investment in GSM & UMTS infrastructure continues to decline from both supplier and operator perspective with
increasing focus on the roadmap for switch-off combined with strong drivers to re-farm and re-utilise the multi-mode
infrastructure, particularly as LTE coverage becomes even more prevalent and 5G begins to become a reality.

Data usage continues to shift from UMTS to LTE with more voice traffic being supported via VoLTE with fall back to
GSM, leaving UMTS the prime candidate for switch-off and re-farming of frequencies to LTE. The increasing maturity of
IoT strategy also drives operators to consider leveraging on GSM investments, presenting further arguments in favour of
a pre-emptive UMTS switch-off.

Ultimately, and without specifying any fixed timeline, it is anticipated that all operators will engage in a process which
delivers graceful migration to new technologies with ample advance planning and notification towards the GSM & UMTS
subscriber base. While a “cleaner” UMTS shut-down roadmap is envisaged, the story for GSM is complicated by the
legacy M2M installed base. Partial re-farming of the 2G frequencies may cause some decrease in the capacity of GSM-
based packet data services, but this should have minimal effect on voice service & M2M (due to the relatively low data
bandwidth requirement). Beyond 2020, it may even be possible to retain only 2.4 MHz of 900 MHz spectrum for GSM
(sufficient for 1+1+1 TRXs per site). Furthermore, some operators are even considering sharing a common 2G network,
e.g. by national roaming, to liberate yet more 900 MHz spectrum for 3G or LTE services.

Generally, the overall voice, data and IoT strategy will drive the investment decisions in GSM & UMTS, however, from
the Orange Group perspective, there is a clear recommendation to limit investment in both technologies as far as
possible. The key issue for the time being is to secure compatibility between radio technology layers; Huawei & Nokia
have adopted the SRAN release strategy while Ericsson continues on the path of individual RAT release format. The
relevant release/features need to be deployed which maximise performance and secures interworking and on-going
compatibility between technology layers.

3.2 2G/3G functional evolution


From the GSM perspective, RAN modernisation programmes (including the recent Nokia-ALU project) have secured all
necessary evolution requirements including support for IP RAN and implementation of energy saving solutions, with the
introduction of future proof SRAN infrastructure. Unless supported by conclusive business case, further evolution of the
GSM networks is anticipated to be extremely minimal.

Regarding UMTS, most of Orange countries have already deployed HSPA+ technology with Dual Carrier HSDPA,
allowing a peak rate advertised at 42 Mbps. No significant evolution is planned beyond that step, since only minor
features may be relevant as a complement.
In countries where 4G has been launched a few years ago, 3G traffic rise has now gone flat, leading to 4G carrying
more than 50% of the traffic in all European countries except ORO and OMD. Need for extra 3G capacity is only seen in
countries where 4G has not been launched yet, although even for these countries, the number one target is now to
launch 4G, while reducing investments in 3G to the strict minimum to cope with the handover period before 4G off-load is
effective.

- 13 - Orange confidential
3.3 3G evolution – target configuration

The target 3G network configuration should be Dual-Carrier HSPA with 64QAM in Downlink allowing 42 Mbps and
HSUPA 2ms TTI with QPSK allowing 5.7 Mbps in the Uplink.

Carrier Aggregation: Dual Carrier HSDPA is P0, all other aggregation scenarios are not recommended.
Aggregating two 5 MHz contiguous downlink carrier is already supported by a wide range of terminals. It will enable to
double the user throughput and enhance its Quality of Experience.
Due to lack of ecosystem and a common UE platform between 3G evolution & 4G, it is not recommended to upgrade
RAN network to support Multicarrier HSDPA (3 and 4 carriers) as well as Dual Band Dual Carrier and Dual Carrier
HSUPA.

Higher Order Modulation: 64QAM DL is P0, 16QAM UL is no more recommended


DL 64QAM is considered as default feature which is widely deployed both at Device and RAN sides. 64QAM Downlink
performance is improved only in cell centre.
Due to lack of ecosystem, additional pre-requisite features (3G uplink interference cancellation) and limited gain (cell
centre only), it is not recommended to introduce 16QAM in the Uplink.

MIMO 2x2 for 3G is not recommended


Despite important introduction of 2Rx in the devices with the arrival of LTE, it not recommended to introduce MIMO 2x2
in 3G due to performance impact on legacy customers, high hardware investment and incompatibility with DC HSDPA.

3G smartphone signalling features: Bundle with fast dormancy is P0


In order to reduce the amount of 3G smartphones’ signalling it is recommended to introduce fast dormancy, PCH state
(CELL or URA) & direct up-switch from PCH towards DCH.

2G/3G network evolution: Group Recommendation 3—1


Orange Group foresees the continued utilization of GSM beyond 2020 with 5 MHz of spectrum maintained in GSM900
for legacy devices, roamers and M2M services.
 A GSM re-farming strategy should be established, re-allocating spectrum capacity towards UMTS & LTE
bands as applicable
 Minimise (if not suppress) investment in GSM & UMTS
The target 3G network configuration should be Dual-Carrier HSPA with 64QAM in Downlink allowing 42 Mbps and
HSUPA 2 ms TTI with QPSK allowing 5.7 Mbps in the Uplink.
The Group recommends deploying software only features with no impact on hardware in order to limit the investment
in 3G.
 Dual-Carrier HSPA coupled with 64QAM (DC-HSPA) at 42 Mbps has been widely deployed in the Group and
is recommended for deployment in all countries.
 Fast dormancy coupled with direct up-switch from PCH state is high priority for MEA and Europe country in
case 3G smartphones’ signalling is increasing

2G/3G SW strategy: Group Recommendation 3—2

Based on Radio Access Skill Centre’s 2016/2017 budget recommendations document, MCOs should consider the
below software strategy guidelines when planning the evolution of the GSM & UMTS networks.
 Alcatel Lucent:
2G:
LR14.3G software release is the target release for 2017 because it is Alcatel-Lucent parking release for 2G.
3G:
LR15.3W software release is the target release for 2017 because it is Alcatel-Lucent parking release for 3G.
4G:
LR16.2L software release is the target release for 2017 because it is Alcatel-Lucent parking release for 4G.

 Ericsson:
GSM:
G11B is the BSS recommended parking release with extended maintenance until 2015. G15B can no longer be the
next parking release because it will not be supported by the new OSS platform ENM. Next parking release should be
G16B and maintenance of G13B should be extended until end of 2016 but negotiation is still ongoing with Ericsson to
consolidate this scenario.

GSM & UMTS:

Orange confidential - 14
 Huawei (SRAN)
Huawei’s strategy is the delivery of software loads in Single RAN stream. SRAN is a solution which combines 2G,3G
& LTE SW releases. It is thus not possible to de-correlate 2G, 3G & LTE software releases. For 2017 & 2018,
introduction of GBSS19.1 & RAN19.1 (as part of SRAN 12.1 ) is recommended. Green light for roll out planned by Q3
2017. Previous recommended version (GBSS17.1 & RAN17.1 as part of SRAN10.1) is supported till Q1 2018.

 Nokia (SRAN)
Single RAN 16.10 is the target for all affiliates which have compatible equipment.
 All former ALU affiliates:
They have or will have brand new Nokia equipment which are compatible with SRAN
 Orange Poland and Egypt:
They have legacy equipment which are not compatible with SRAN
They will have a mix of SRAN and SRAT
However single RAN has been delayed a lot in 2016. It was not ready when the first rollout operations started in the
MEA affiliates back in 2016. The first rollout operations have been done with SRAT

SRAN 17
The software strategy for 2017 is being worked out by the Nokia skill center together with Orange France teams. 2
releases are of high interest for Orange in 2017 : SRAN 17A and 17ASP
 SRAN 17A: CP (Pilot) is due in June, 2017, C5 (General availability ) is due in August, 2017
This software release has a first set of Category M IoT features, some user throughput enhancement features
compared to SRAN 16.10 (mainly enhanced carrier aggregation schemes), and has a better way of managing the
base station configurations
 SRAN17A SP (Service Package): CP is due in October 2017, C5 is due in December 2017
This software release is also of interest for Orange since it comes with the Airscale Radio units
Contributors: Mark Woodgate, Carl Woolley, Adrian Pike Contact: Sébastien Bechet

4. 4G Evolution
Thanks to mature ecosystem, adoption of 4G by customers has been very fast since early deployment in 2012. The 4G
ability to provide much superior data speeds and usage comforts compare to 3G/2G enables a democratization of mobile
data broadband in the world. Across the group, 4G network is now mature in Europe and progressing very quickly in
MEA region (Morocco, Jordan, and Tunisia). As a symbol, Orange group’s 4G data traffic overpass 3G one in 2016.
In parallel of the commercial network deployment, 4G carry on evolve in standardisation. New features are now available
in order to support marketing throughput race towards 1Gbps as well as improvement of user experience and network
capacity.
The group recommendation is to deploy 4G and some of the evolution to benefit from its higher spectrum efficiency and
lower latency. The driver for investment in new spectrum (2.6GHz, 800MHz or 1.8GHz refarming) and LTE equipment is
currently the need for extra capacity to carry growing data traffic.

4.1 Peak Throughput oriented features

4.1.1 LTE carrier aggregation


Carrier Aggregation allows network operators to extend 4G operational bandwidth. This can be achieved by aggregating
two or more carriers. For both downlink and uplink, the carriers may be contiguous or non-contiguous belonging to the
same or different frequency bands. Most of inter and intra band combinations available in Orange countries are already
standardized for 2, 3 and 4 Carrier Aggregation in the downlink and 2 carrier aggregation in the uplink.
The peak data rate achieved by a carrier aggregation capable device is the sum of the achievable throughputs on all the
aggregated bands in single carrier mode. For example a category 9 device is able to achieve 450 Mbps in the downlink
when 3 carriers of 20 MHz are aggregated as it receives 150 Mbps on each band.
In the Downlink, Carrier Aggregation increases the user throughput everywhere in the cell as far as the considered
frequency bands are available. On top of that, carrier aggregation allows a very fast and efficient load balancing between
carriers.
In the Uplink, according to field tests, Carrier Aggregation advantage is different depending on the position in the cell:
 In cell center, when the maximum output power of the device is not reached, the peak throughput is increased
by the addition of throughputs on both frequency bands.
 In cell edge, the throughput can be increased thanks to a better management of uplink resources: the resources
are allocated on the frequency band with the best quality.
 In cell middle, degradation can be observed when 64 QAM is not activated. It is then highly recommended to
deploy carrier aggregation jointly with 64 QAM in uplink.
All RAN vendors already support up to 4 Downlink Carrier Aggregation and 2 Uplink Carrier Aggregation. The impact on
hardware is low and limited to some baseband configurations.

- 15 - Orange confidential
3 Downlink Carrier Aggregation is already implemented in several commercial devices and some devices supporting 2
uplink carrier aggregation are about to be available in 2017.
In downlink there is no impact on the device battery life whereas in uplink this still needs to be investigated.
Downlink Carrier Aggregation is recommended, once several frequency bands are already deployed, as it improves
significantly user experience and enable a better management of the frequency bands.
Uplink Carrier Aggregation recommendation will be finalized with the device battery life study and must be deployed only
with uplink 64 QAM.

4.1.2 Higher-order MIMO


Using more antennas on both the transmitter and the receiver sides allows reaching higher peak data rates. Adding
antennas in reception is more beneficial than adding antennas in transmission as the receive diversity gain is much more
significant than the transmit diversity gain.
Upgrading the RAN to MIMO 4x4 requires:
 Antennas upgrade: 2 additional antenna ports are needed, either dedicated to the considered frequency or
duplexed with another frequency (on the same or distant radomes).
 Radio unit upgrade: 2 additional power amplifiers are needed.
MIMO 4x4 on the RAN enables the following features and improvement:
In the uplink:
 4 receive diversity (4 receive antennas at base station) which enhances significantly user throughput (up to
100% in cell edge) and coverage (between 3 and 5dB) for all devices.
In the downlink:
 MIMO 4x4 for compatible devices (4 transmit antennas at base station and 4 receive antennas at device), which
improves significantly user throughput from cell center (up to 30%) to cell edge (up to 40%). The gain in cell
center depends on base station antennas configuration and site topology. MIMO 4x4 can also improve the
capacity when the number of compatible devices in the network is high.
 MIMO 4x2 (4 transmit antennas at base station) for legacy devices, which provides a slight gain in cell middle
and cell edge.
 MIMO 2x4 for compatible devices (4 receive antennas at device), which improves significantly user throughput
in cell edge (up to 40%), in the areas where the RAN is configured in MIMO 2x2.
The figures have been obtained during FDD field tests but the same trend should be observed in TDD.
MIMO 8x8 on the RAN offers new features (Multi-user MIMO and horizontal beamforming) but their benefits still need to
be assessed.
MIMO 4x4 is already supported by RAN vendors and the first devices are available at 1800 MHz and 2600 MHz. Low
frequency bands such as 800 MHz won’t be supported.
The maximum configurations in combination with carrier aggregation which are possible with devices today are:
 3 downlink carrier aggregation with MIMO 4x4 on 1 band
 2 downlink carrier aggregation with MIMO 4x4 on 2 bands
MIMO 4x4 RAN upgrade costs prevent from a massive upgrade but MIMO 4x4 has to be considered:
 In case of green field or RAN renewal especially for countries with limited spectrum.
 On few sites for demonstrations / marketing messages / hot spots capacity because on top of 3 Carrier
Aggregation this is the only way for many countries to increase more 4G peak throughput.
 For LTE for fix as customer premise’s equipment (CPE) can easily embed more antennas than smartphones
(even 8 antennas in TDD but gain still need to be proven).
 On some sites in order to increase devices penetration which will improve user experience in cell edge even in
the areas without MIMO 4x4 and enhance capacity where MIMO 4x4 is deployed.

4.1.3 Higher-order Modulation


In downlink baseline LTE networks are using 64 QAM as modulation scheme for the highest radio conditions while 16
QAM and QPSK modulations are used respectively for medium and poor radio conditions. In uplink the same principle
determines the usage of 16 QAM and QPSK as baseline modulations.
New modulations have been introduced in order to increase peak data rates in good radio conditions:
In downlink the 256 QAM modulation allows a peak throughput increase up to 20% but its usage is limited to very small
areas where the radio conditions are extremely good. Its support necessitates a software upgrade but is also linked to
the power amplifier generation and the available power headroom.
In uplink the 64 QAM modulation can bring up to 40% increase on peak throughput and it is used mainly in cell center.
Both uplink 64 QAM and downlink 256 QAM are supported by all RAN vendors and several commercial devices are also
available.

Orange confidential - 16
Downlink 256 QAM is recommended as a nice to have to advertise a higher peak throughput.
Uplink 64QAM is recommended as it improves the throughput in good to medium radio conditions and is mandatory to
activate uplink carrier aggregation.

4.1.4 Massive MIMO for 4G capacity


A ‘rough’ definition of M-MIMO is an eNB antenna where the number of independently controlled antenna ports is
significantly larger than the number of simultaneously active UEs. Massive MIMO technology requires active antenna
(Power amplifiers & Antenna dipoles gathered in single equipment) as well as device compatibility. It enables to increase
capacity on the network for both uplink and downlink.
Massive MIMO supports enhanced beamforming techniques with narrow beams which enables an increased network
capacity in both directions (uplink & downlink). Main features are:
- Downlink 3D beamforming : steering beam in the UE’s preferred direction (in 3D)
- Downlink & uplink Multi user MIMO : reusing radio resources (time/frequency) in spatial directions
- High order Uplink Diversity
The following diagram illustrates an Massive MIMO example for a 4 x 8 xpolar array connected as a sub-array
architecture with 4 vertical elements per TXRU (power amplifier).

Figure 12 – Massive MIMO deployment principles

In 2016/2017, Massive MIMO is only available in TD-LTE in a commercial version. Indeed, most of TD-LTE commercial
devices already support Massive MIMO. Field measurements achieved jointly with China Mobile shows Massive MIMO
solution can take much more traffic than conventional TD LTE antenna (x3). It looks like a promising technique for future
5G deployment which could be deployed in TDD band (3.4 - 3.8GHz band)
Some evolutions are planned by the standard to provide a single version of Massive MIMO for both FDD & TDD so
called Full Dimension MIMO (FD-MIMO developed in 3GPP Release 13). Penetration rate of compatible devices will be a
key factor to enable deployment. Moreover, Chinese RAN vendors (Huawei & ZTE) are currently working on FDD
solution that can be used with legacy devices. Nowadays, it is too early to make any recommendations about usage of
FDD Massive MIMO.
Several hurdles are still to be overcome and further studied to enable Massive MIMO commercial deployment. It is
currently expensive to buy (high cost) and to operate (high power consumption) but is likely to reduce as economies of
scale and potential to re-use device components start to influence costs.
The integration of active antenna into existing site solution still needs to be further studied. In early deployment, it might
be separate from existing antennas on site. Antenna manufacturers are hardly working to adapt their antenna portfolio to
Massive MIMO in order to limit number of antennas per sector.
The picture shows one possible option being studied by Vendors. A Massive MIMO array may be added inside an
existing 2.7m radome by shortening the vertical height allowed for all the other bands. The loss of vertical beamwidth
may result in some coverage loss or might be compensated by adding 4T4R configuration to these bands.
Existing regulatory rules (security perimeter, EMF exposure limitation) has not been created to support Massive MIMO
and might need to be adapted at least from measurement methodologies point of view.
OLN will carry on working on Massive MIMO maturity and its deployment feasibility in the coming years. Due to
expensive price, it looks to be unsuitable to FWA service in a first stage (especially for MEA countries). FDD Massive
MIMO maturity will be digging further in 2017 as well TDD solution in order to prepare the arrival of the 5G.

4.1.1 Cells cooperation features


Uplink Coordinated Multi-Point is high priority for the group (P0)

- 17 - Orange confidential
Intra-site UL COMP is already available in all RAN vendors. The feature enables to improve UL user experience by 33%
in overlapping area and capacity by about 20%. All 4G devices can take benefit from the feature. This recommendation
is valuable for both MEA & Europe region.

Intra-Site Downlink Coordinated Multi-Point is medium priority for the group (P2)
In short term, intra-site coordinated scheduling feature is now available in all RAN vendors. By improving cooperation
between intra-site cells, interference would be mitigated. This feature could be useful in case network suffers from high
downlink load & important overlapping area. It should help to improve cell edge user experience by about 30%
maximum. For other situation, there is no need to use it.
In average term, more advanced schemes (Dynamic Point Selection, Joint Transmission) should be supported by RAN
vendors. However, ecosystem & maturity remains quite uncertain at this stage. The Features’ efficiency is linked with
device penetration (TM10 capable devices).

(f)eCIC : (further) enhanced Intercellular Interference Coordination is not recommended by the group
Enhanced inter-cell interference coordination enables to coordinate macro cell & small cell (across time) aiming at
extending temporarily the small cell coverage in order to boost its offload efficiency. This feature needs time
synchronization as well as compatible terminal. From Orange group estimation, even with 100% compatible device, the
expected small cell offload boost should not be significant (from 33% to 39%). Moreover, RAN vendors also highlighted
some feature interactions issues (LTE Carrier aggregation and VoLTE) in small cell coverage extension area which is not
tackled in standardization. The poor level of maturity and ecosystem as well as limited gain justify to not recommending
the feature.

4.2 RAN Drivers to introduce time/phase synchronization


It was identified that several new services and 4G network evolutions require the introduction of a new type of
synchronisation (phase/time synchronisation) that is not massively deployed in the Orange group network. Several
technical solutions exists in order to introduce time/phase synchronisation, all of them needs important deployment effort
and has to be anticipated.

Figure 13 – Identified RAN Drivers for phase/time synchronization

Phase/time synchronisation nationwide deployment would be necessary if regulator oblige operators for some nationwide
actions for mobile broadband: such as developing geolocation service based on 4G to create a European “911” or
coverage obligation for new introduced spectrum.

In 2016, some local deployment is ongoing to enable LTE for fix services based on TDD spectrums in Europe & MEA. In
addition, no RAN 4G features (even cumulated) or new services can justify a massive and expensive introduction of
phase/time synchronisation in mobile networks. It is likely that roadmap time synchronisation introduction is phased with
the 5G commercial deployment in TDD Spectrums (such as 3.5GHz and 3.7GHz at a first stage).

Once time synchronisation is introduced for TDD, co-sited FDD spectrum can also be time synchronised in order to
benefit from 4G RAN features. Trial performed in 2016 in France confirmed that time synchronised FDD network does
not impact network QoS.

Orange confidential - 18
4G evolution: Group Recommendation 4—1
For 4G evolution feature for both Europe and MEA :
 Downlink 256 QAM : P1
 Uplink 64 QAM : P1 but mandatory for Uplink Carrier Aggregation
 Downlink Carrier Aggregation : P0 (for multi-band sites)
 Uplink Carrier Aggregation + Uplink 64 QAM : P1 (device battery life to be checked)
 MIMO 4x4 : P1 (when the spectrum is limited, for LTE for fix, otherwise on hot spot areas)
 MIMO 8x8 : P2 (for LTE for fix, evaluation ongoing)
 Massive MIMO: no recommendation yet for mobile broadband (it is not recommended at this stage to use
Massive MIMO for LTE for fix due to expensive product
 Intrasite UL COMP : P0
 Intrasite DL COMP : P2
 eICIC & feICIC : not recommended
Massive time synchronization introduction in network (RAN Drivers) in Europe:
 P0 for 5G introduction in TDD band (network preparation needs to be anticipated to minimize its cost
introduction)
 Not recommended to enable only advanced RAN features
Contact: Sébastien Bechet

5. Conversational Services
Transformation to all IP is ongoing, especially since the arrival of LTE, and we can see in the operator world a move of
conversational services over IP: SMS over IP, VoIP over LTE (VoLTE), VoIP over Wi-Fi (VoWiFi) and to a lesser degree,
video calls over IP.

After a first unsuccessful attempt, IMS-based RCS (Rich Communication Suite) is coming back to the front stage thanks
to partnership with Google. It is expected to bring additional services on top of the voice (instant messaging, file sharing,
video sharing…).
WebRTC is a web technology that enables easy establishment of voice and video communication between two web
browsers (and web application on smartphones). It is gaining more and more interests from operators due to the facility
of deployment and its low cost.

5.1 SMS
SMS are supported on LTE by reusing the SMS centre used in 2G/3G thanks to a new interface between the MME and
the MSC.
Next step is to move to SMS over IP (SMS are encapsulated in SIP messages), but it requires an IMS core and a new
equipment to handle SMS messages (IP-SM-GW), as well as the interfacing to the current SMS center. SMS over IP can
be envisaged especially together with VoWiFi, to deliver SMS even without cellular access
No specific RAN feature is needed.

5.2 Voice
The target solution for voice service on LTE is Voice over IP steered by IMS (VoLTE). VoLTE deployment within Orange
has started in Orange Romania in 2015.
But before the arrival of VoLTE, voice services are ensured on LTE networks, by using CSFB (Circuit Switched Fall
Back) procedure. CSFB consists in a fallback on 3G circuit network or 2G circuit network of all voice calls performed from
or to an LTE smartphone being under LTE coverage.
CSFB remains mandatory after VoLTE launch to support voice call from roamers (VoLTE roaming not expected before
years), and non-VoLTE capable UEs.

5.2.1 Circuit Switched Fall Back (CSFB)


CSFB is supported by all infrastructure and devices vendors from first LTE releases and is identified as the mandatory
solution for LTE launches and roaming by NGMN. It permits to redirect LTE users to 2G or 3G voice calls.
Impacts of CSFB are:
 some network equipment have to be upgraded. In particular, MSCs shall support the SGs interface (new
interface between MME and MSC).
 voice call setup times increase: additional delays compared to a native circuit switched call in 3G/2G, depending
on target network, used procedure and type of call (MTM, MOC or MTC).

- 19 - Orange confidential
 need for an accurate mapping between LTE Tracking Areas (TA) and 2G/3G Location Areas (LA) to avoid
Location Updates during the fallback, which would result in longer call set up times. MSC in pool is seen as a way
to mitigate the worst consequences of mapping inconsistencies.
Several CSFB procedures are described in the 3GPP standards:
 CSFB based on release and redirection R8 (basic procedure): a ‘break before make’ procedure without prior
knowledge about which cells the mobile device is redirected to. All system information blocks (SIB) are read before
connection re-establishment.
 CSFB based on release and redirection R9: same type of procedure as R8 but with prior knowledge about the
cells the mobile device is redirected to.
Up to 3.5s could be gained theoretically for the longest call set up times corresponding to a CSFB to 2G MTM call.
 CSFB LTE to 3G based on Packet Switched Handover: a ‘make before break’ procedure.
The gain is mainly in the interruption time for the data session, which should be strongly improved (500ms instead
of 6s to 10s with RRC R8). Additional delay in call set up time should be likely comparable to RRC R8 procedure.

CSFB R8 & R9: Group Recommendation


CSFB R8 and R9 are recommended and have been deployed from 2012 with success in Orange countries.
In order to speed up the return to LTE after completion of the call in 3G or 2G; “fast return to LTE” feature is
recommended: it permits to return to LTE in 5s, instead of 20s to some minutes.
 In case of CSFB LTE to 3G, RAN features are recommended
 In case of CSFB LTE to 2G, proprietary device feature is recommended, available with Qualcomm chipsets.

5.2.2 VoLTE support on RAN


All RAN vendors provide dedicated VoLTE features, both in FDD and TDD, which permit to enable VoIP calls
controlled with IMS.

Only a few features are mandatory to permit VoLTE calls


 Management of multiple bearers
 Specific prioritization of voice packet on QCI 1 bearer
 Modification of the bearer QoS to enable call on hold, 3 party conference, IMS preconditions.
But 20 to 25 additional features can be activated to ensure a good voice QoS, a good service coverage and a
satisfying quality in mobility, while permitting to have a large number of simultaneous voice calls.

QoS VoLTE QoS is ensured whatever the cell load and traffic mix on the customer terminal. This is a
strong difference with OTT voice services, which can strongly be impacted by concurrent traffic.

Capacity With Robust Header Compression feature, the rough number of simultaneous VoLTE calls is in the
order of 200 per 5MHz. Consequently, there is no capacity issue with VoLTE in initial deployments.

The VoLTE coverage in terms of coupling loss (attenuation between the eNodeB and the UE) is
Coverage
similar to the 3G CS coverage. VoLTE permits to benefit from the coverage of 800MHz.
VoLTE coverage appears also to be better than OTTs (Skype) on LTE.
PS HO ensures the intra LTE mobility for voice calls.
Single Radio Voice Call Continuity (SRVCC) permits to move from VoIP in LTE to legacy 2G/3G
Mobility CS when LTE coverage is lost.
SRVCC in alerting and pre alerting phase (a and bSRVCC) are not RAN features (features in UE,
IMS, MSC), but it is highly recommended to have them in order to decrease the call failure ratio
(Indirect consequence on mobility radio settings)
The DRX feature aims at decreasing the battery consumption, in order to have similar consumption
UE battery
as in 3G CS. Despite it is known to cause small degradation of the audio quality, DRX is required
consumption
by a well know device manufacturer. VoLTE does not permit to be “greener” than 2G or 3G CS.

The document “VoLTE Radio Guidelines” is summarizing the list of mandatory and optional features.

VoLTE KPI (call setup success rate, call drop rate…) from first affiliates having deployed commercially in Europe are
satisfying, being as good as in 3G and sometimes better. Radio optimization phase is needed, especially for SRVCC.

Orange confidential - 20
New codec for higher quality and better radio resource usage

VoLTE and VoWiFi permit to improve the audio quality thanks to AMR-WB 23.85 kbps (especially in noisy
environments), codec rate which is not available with 2G/3G.
The activation of this new codec rate (in the IMS) is recommended since Codec / Rate Theoretical MOS
coverage loss in respect to AMR-WB 12.65 kbps is low and impact on
LTE capacity is low (VoLTE capacity is in the order of 200 users per cell
and consequently capacity is not seen as a major issue). Nevertheless in AMR-WB 23.85 ~4,1
some specific cases (few LTE spectrum), it can be advised to limit the
codec rate used in VoLTE call to 12.65 kbps in order to have gain on AMR-WB 12.65 ~3,8
capacity and data performance (limitation to be applied in IMS, but with a
cost).
EVS 24.4 SWB ~4,4

From S2 2017, the codec EVS (Enhanced Voice Service) could be EVS 13.2 SWB ~4,1
introduced in order to further increase the peak quality or keep same
quality with less radio resource usage.
EVS 9.6 SWB ~3,7
Some EVS modes like EVS 13.2 kbps Channel Aware Mode are
specifically interesting because they could provide coverage gains and
better quality in lossy environments such as on Wi-Fi access. It is currently under tests.

As it is the case already in 2G/3G CS, voice rate adaptation is appearing in VoLTE/VoWiFi, with both RAN features (in
2018) and autonomous adaptation in some UE chipset.

Cost of the introduction of VoLTE at RAN level


 The cost of RAN features depends on each contract with vendors.
 The impact of VoLTE on the LTE capacity is considered as negligible in normal deployments Nevertheless it will
provoke an increase of the number of simultaneous connected users, for which Orange may pay some licences.
Even if it is only a fraction of the needs for the LTE data, it is not negligible when considering the VoLTE cost
separately.

VoIP for 4G fix deployment


In case of 4G fix offers with VoIP service (LTE Flybox with VoIP line), the constraints are quite different:
 No mobility is expected
 Coded used by fix phones is generally G711 @ 64kbps, compared to AMR 12.2 kbps, 12.65 kbps or 23.85
kbps.
The technical solution to support voice depends on many parameters: fix or mobile IMS, availability of a PCRF,
dedicated APN or not, one or more VoIP lines per Flybox, etc. …

Most of the time, the CPE is not VoLTE compatible but only VoIP compatible, meaning that VoIP calls can be established
based on IMS, but some VoLTE RAN features for coverage and capacity optimisation cannot be used. Nevertheless,
VoIP calls can be prioritized over the radio similarly to VoLTE calls. Exact settings are depending on the use of TDD or
FDD and the presence of mobile VoLTE users on the same technology.
To be noted:
 VoLTE coverage in TDD is about 7 dB shorter than in FDD
 QoS strategy between mobile VoIP, fix VoIP on 4G and other services is to defined case by case based on
marketing assumptions.

5.2.3 VoWiFi
VoWi-Fi can be enabled on certain devices that are VoLTE compatible. It permits to extend the voice coverage by
provide voice service with Wi-Fi when cellular is not available.
An IPsec tunnel is created between the device and the mobile Core Network (ePDG) that permits to secure the voice
transmission over both trusted and non-trusted Wi-Fi accesses.

QoS aspects
VoWIFI quality depends on Wi-Fi QoS and Wi-Fi channel occupation by other services. Thus, use of Wi-Fi QoS (i.e.
priorisation of traffic) is recommended (i.e. tagging voice packets with QoS Access Class “Voice”). However, segments of
the network may not be managed by Orange, Wi-Fi devices are not controlled by Orange and furthermore the use of a
VPN tunnel from the VoWIFI device to the e-PDG makes translating correctly the IP QoS to Wi-Fi QoS difficult. So
VoWIFI needs a very good Wi-Fi network design and it is to note that quality can never be fully guaranteed as it is a
license free band.

- 21 - Orange confidential
Consequently, QoS is not full time ensured with VoWiFi: When no QoS enablers are activated, tests have shown that
voice quality is badly impacted as soon as there is some other data activity on the device and/or on other devices
attached to the Access point.

Mobility aspects
Seamless mobility is offered between VoLTE and VoWiFi, but not between 2G/3G CS and VoWiFi. The key point is that
mobility parameters (especially thresholds on the signal strength in LTE and Wi-Fi) are proprietary and managed in the
device.
Some requirements are put in the OGDR (Orange group device requirements), requiring mobility based on signal quality
and not only power.
Orange strategy for the voice service is today Wi-Fi preferred (vs Cellular preferred). It is partly due to the need to
promote the use of service seen as new and innovative by the client. Some affiliates have chosen to let the choice to the
user between Wi-Fi and cellular preferred. Anyway in all cases the user is able to activate/deactivate Wi-Fi calls.

5.3 Video calls for LTE subscribers

Standardized Video calls on LTE and Wi-Fi (ViLTE/ViWiFi) are feasible, requiring mainly a device firmware upgrade. But
due to the low penetration of ViLTE devices, the reachability of video calls would remain low. In 2017, thanks to the
partnership with Google on RCS, it is expected that many android devices could be used to perform RCS based video
sharing (video not synchronized with voice) or video calls. Consequently the Orange strategy is no more focused on
ViLTE but on RCS video calls.
The platform used for RCS will be often the Google one, which means that video calls are likely to be in best effort mode,
without associated network QoS and hence without specific additional RAN feature. QoS will be ensured thanks to
appropriate video bitrate adaptation mechanisms in the application.

WebRTC technology is also part of the strategy for 1-to-many calls, in complement to RCS video sharing. It is completely
independent from the RAN, requiring no specific RAN feature.

Voice services for LTE subscribers: Group Recommendation 5—1


 CSFB is the first step for voice services for LTE subscribers and shall be activated at LTE launch and maintained
after VoLTE introduction for roaming purposes and Non-VoLTE capable LTE UEs.
• fast return features shall be activated to enable a quick return to LTE after voice call completion
(feature in RAN in 3G and in UE in 2G).
 Dedicated VoLTE RAN features are mature with more than 20 RAN features available to optimise the mobility
(e.g.SRVCC), QoS, UE, battery savings and coverage. VoLTE RAN KPIs are satisfying.
 VoWiFi is recommended to extend the voice coverage where cellular network is not available (e.g. deep indoor
at home)
 Orange strategy for the voice service is today Wi-Fi preferred but some affiliates have chosen to let the choice to
the user between Wi-Fi and cellular preferred.
 Despite Wi-Fi QoS enablers are existing and can be activate, VoWiFi quality is not fully ensured (segments of
the network not managed by orange, Wi-Fi devices on the access point not controlled by Orange, etc …)
 Seamless mobility between VoWiFi and VoLTE (not with 3G/2G CS) is managed by the device in a proprietary
way.
 New Codec rate for VoLTE and VoWiFi (“HD+” is AMR-WB 23.85 kbps) is recommended for marketing purpose
and QoE. EVS codec could be introduced from H2 2017 to provide even higher quality or better radio resources
usage.
 Strategy on video calls is no more ViLTE but RCS video sharing thanks to Google. No need for specific RAN
feature.
 End to end competencies are mandatory for radio engineers in order to troubleshoot and optimise VoLTE at
radio level (e.g. SRVCC efficiency). Transversal work on VoLTE/VoWiFi is strongly recommended
Contact: Franck Payoux

6. Automation of network operations


The main solution to automate network operations is SON (self-organising networks) functionalities to:
 master the increasing network complexity,
 enhance the end-user Quality of Service,
 minimize deployment and operational costs and delays

Orange confidential - 22
The first SON features are deployed in the RAN networks to simplify engineering tasks related to network planning,
deployment, operations and maintenance and optimization.

6.1 SON features’ summary


SON features can be classified in three main pools:
 Self-configuring: automate manual steps required for planning and deploying a wireless radio network. These self-
configuring features encompass Plug and Play configuration of sites, creation and optimization of neighbour lists.
 Self-optimizing: optimize automatically and continuously the network. These self-optimizing features are focused on
the modification of Radio parameters to find optimized usages of network layers (balancing of traffic, interference and
optimization of coverage and energy usage…)
 Self-management & Self-healing: reduce the engineering tasks necessary for managing and operating a network.
Those features take actions to counteract network element failures based on alarms.

SON features aim to sustain two main goals of the operator:


 Network Management Efficiency:
With the deployment of new technologies, Radio networks’ size is exploding and becomes too complex. Automation
relaxes this complexity and allows teams to concentrate on other key operational domains. OPEX and CAPEX savings
are expected when activating SON features, whatever these savings concern reduction of engineering tasks to operate
the network, reduction of network troubles to be solved, reduction of dimensioning needs, increase of traffic/user
absorption by existing networks. In the building of business models for SON features activation, such impact on OPEX
and CAPEX saving must be taken into account.
 Improvement of Quality of Service:
Always-on monitoring and reactive optimization of RAN networks helps improving the customer perceived QoS. Also,
enhanced optimization per cell will ease operations which are difficult to do manually today.

6.1.1 Impact of SON features’ activation on the network’s operations


By automating several activities previously managed by operational teams, the SON features have a strong impact on
the following topics:
 Operational processes: as many manual operations are to be replaced by automated methods, the impact on
operational processes is significant. This assessment of operational processes’ impacts is expected with the introduction
of SON solutions, in order to ensure smooth transition from full manual behaviour to automated behaviour.
 Integration of SON in RAN outsourcing operations: In the context of RAN management operations outsourcing,
SON features will serve two usages: the Managed Service Partner will benefit from SON features to reduce the efforts in
Build and Optimisation domains while the Operator will be able to use SON features in Open loop to audit MSP
Optimization and Build activities (out of classical Performance KPI audits).

6.1.2 SON solutions’ architecture


Two architectures are possible when deploying a SON feature on mobile networks:
 Distributed architecture: the SON feature is implemented in the base stations which are communicating one with
each other to operate modifications in their parameters setting.
 Centralized architecture: the SON tool is implemented on top of OSS equipment to benefit from large amount of
information (even if it is less reactive) to perform efficient SON optimizations.
These two architectures can be complementary deployed to allow reactivity and centralization, forming hybrid
architecture.

Figure 14 – Distributed & Centralized SON solutions

6.1.3 Global status of Group recommendations about these SON features


Concerning distributed features provided by the RAN suppliers in 4G, the Group recommendation is that LTE SON
features such as Auto Configuration, ANR (Automatic Neighbour Relationship) or MRO (Mobility Robustness

- 23 - Orange confidential
Optimization) can be activated in Orange LTE networks as they are sufficiently mature. The other distributed features
are not recommended today as they enter in concurrence with equivalent centralized features.
Concerning the centralized tools, the result of 2016 Group RFQ jointly with Deutsche Telekom and BUYIN was to
source a single product for both Groups: the Eden Net solution of Nokia.
Following this RFQ closure, the Group recommends the introduction of this Centralized SON tool in all Orange
countries. Already 11 countries have invested on this solution, and it is now necessary that all countries follow this
trend.
To go along with this recommendation, OLN has completed the following actions:
 A specific Skill Center is launched to support the introduction and maintenance of C-SON tool in countries,
 Negotiations are performed with Managed Service Partners to get savings in outsourcing contracts linked to the
usage of C-SON tool. Such savings will ease the investment in this solution,
 Optimization guidelines are built to help the initial configuration of C-SON tool.
 A Custom module factory is structured at Group level to develop Orange optimization scripts in the framework of
Nokia tool (via an Open API) and generate new automation use cases in optimization domain.
 Innovation programs are ongoing with Nokia to prepare the evolutions of Nokia Eden NET roadmap for Orange
Group.

Centralised SON deployment: Group Recommendation 6—1


 The Group recommends the generalization of Centralized SON solution introduction in the networks by the end of
2017.

Contact: Vincent Diascorn

7. LTE for Fixed Access


Due to its superior performance in terms of throughput / latency / native IP support, LTE is seen as the first wireless
technology able to provide a real alternative to wireline (e.g. ADSL) or wireless (e.g. WiMAX, PtoP link) to provide
Broadband Internet for Residential (B2C) & Business users (B2B).
Several operators have since launched LTE for Fixed Broadband services either as a regulatory obligation (Germany) or
as a viable solution to complement FTTx deployment in area poorly served by Copper or as a differentiating offer target
specific segments.
Orange has been using LTE for Fixed services in several countries to address both B2B and B2C Residential customer
since 2013, the interest for has been intensified with the arrival on the market of LTE Advanced or even pre-5G solutions
marketed by suppliers under the “Wireless Fiber” concept. Increased in capacity and customer data throughput using
carrier aggregation capabilities have attracted the attention of operators.
However “LTE for Fixed services” usage, where customers usage may be more than 10 to 20 times higher than on
Mobile Broadband (i.e. 50 GB per month rather than a few GB), may disrupt the mobile network. Customer demand
cannot be served infinitely in a cost effective manner, therefore limiting usage is a must have. Firstly, Home TV services
cannot be offered by LTE. Secondly, reasonable caps and density of customers are required according to the radio
resources and the offers portfolio. To mitigate the possible disruption, progressive open in limited areas and to perimeter
the access in authorized location should be considered.
Fixed services are also identified as one of the first use case of 5G with early deployment in the US from Verizon.
However, whatever the wireless solution, the viable business case is still to be found for mobile operators and also, a
frequency spectrum plan should be done before any deployment to pay attention to 5G deployments in future Band42.

7.1 Status of LTE for fixed services


Both modes of LTE (FDD or TDD) can be used to provide Fixed broadband services, the choice by operators to focus on
one particular mode are mainly linked with the cost and availability of spectrum in the target markets.
In Region 1 (Europe / MEA) the use of TDD spectrum reserved for Fixed services may be associated with lower
spectrum fee than for a FDD spectrum, however this is not a worldwide view as in Asia and US, mobile services have
been using TDD spectrum (China , Japan, US).
 Global FDD spectrum possibilities
o Europe/MENA : FDD 800, 900, 1800, 2100, 2600
 Global TDD spectrum possibilities:
o Europe: TDD 2600 [B38], TDD 2300 [B40], TDD 3.5/3.6 [B42/B43] (Romania)
o MENA: TDD 2600 [B38], TDD 2300 [B40] (Mali), TDD 3.5 (B42) (Morocco/Tunisia)
LTE for Fixed services may be an opportunistic way to use spare LTE FDD capacity in rural, it is not recommended to
deploy in purpose extra sites or spectrum bands to tackle specifically LTE for Fixed residential customer usage. On the

Orange confidential - 24
other hand, LTE for Fixed services could be envisaged as an LTE TDD alternative solution, especially for answering the
case of end of life of WIMAX networks used for business services. In this case, the deployment of specific LTE TDD
dedicated network layer for fixed B2B customers makes sense as they generate a significant high ARPU.

When considering the use of LTE for Fixed Services several strategies can be engaged:
 LTE used to provide a Fixed Business broadband service (B2B): consisting of “basic internet access”, “internet
guaranteed”, “VoIP” and “L3VPN” with a certain level of SLA in the form of guaranteed Bit rate, mainly targeted
for business customers and envisaged to replace WIMAX networks. Some affiliates started in 2016 with “basic
internet and VoIP” offers as RD Congo /Ivory Coast. New comers are going to deploy bundles of advances
services bundles as “basic internet access”, “internet guaranteed”, “VoIP” and “L3VPN” as Romania (B43),
Morocco (B42), Tunisia (B42), and Mali (B40/41).
 LTE used to provide a Fixed Home broadband service (B2C): consisting of a Single Play / Dual Play offer
(“basic internet access”, “VoIP”) with no access to VOD / Live TV service and no guarantee of service,
particularly targeted for residential customers. This is the case of Belgium, France, Romania Botswana and
Spain.

7.2 Orange Group activities on LTE for Fixed services


An OLN anticipation project “LTE for Fixed” was launched in 2013 to assess the key aspect of running fixed services on
a network initially designed for Mobile use.
In 2015, the project move to a delivery phase focusing first on the case of WiMAX migration for affiliates in MEA.
 Major work was conducted to provide a solution able to target B2B customers in RD Congo and O Ivory Coast.
 In addition several case studies were completed with interested MCO (Mobistar in 2013) and Orange Spain in
2014 to assess the economics behind such service.

In 2016, a transverse delivery project was launched to industrialise the approach and enable countries in both MEA and
Europe countries to add Fix services to their portfolio.

7.3 Network aspects of LTE for Fixed services


Network elements to support LTE for Fixed services are presented on Fig 15 below. The main impacts are highlighted in
following list:
1. to consider the spectrum availability (FDD or TDD mode) and frequency spectrum plan
2. to select the good CPE (4G Flybox) in relation with the spectrum and the service expected (outdoor
CPE are preferred)
3. to evolve the radio associated considering the FDD or TDD mode
4. to control the traffic thanks to the core equipment’s:
o the PCRF to set-up the policy Control for Fixed usage and to implement cell lock feasibility
o the IMS to establish the VoIP services
5. to reflect the need of VPN Concentrator for implementing L3VPN services

Figure 15 – E2E network impacts to implement LTE for Fix broadband access

The main actions to anticipate and to perform the good deployment are:

- 25 - Orange confidential
1. To make services assessment (specially coming from WIMAX technology) and to define the bundles of services
according to marketing areas
2. To consider the new elements to be introduced in the network (VPN concentrator, PCRF capable to manage
dedicated bearers).
3. To select the cost effective CPE according to the service defined and mode.
4. If TDD mode is chosen, need to guarantee the camping strategy by means of the co-existence implementation
between FDD/TDD Networks ( PLMN ID, TAC, access class, SPID, to guarantee the E2E synchronization
(specially for TDD mode) and also, to evolve radio network evolution according to the TDD band selected)
5. To assess the dimensioning considering type of services and spectrum.
6. To define the E2E QOS policy according to the bundles of services and throughputs expected
7. If there was an existing WiMAX network, customer migration would need to be ensured, especially if the same
TDD band is used for WIMAX and TD-LTE. Refarming analysis needs to be considered.

7.4 Deployment Scenario & Devices


For Fix usage, two types of devices can be used for either home or business applications:
 Indoor CPE – with Orange product called Flybox 4G.
 Outdoor CPE installed on the wall or on a small mast and equipped with a directional high antenna.

.
Figure 16 - Two scenarios for fixed broadband access

Mounted outside the building, the outdoor CPE (Customer Premise Equipment) also called Outdoor Unit (ODU) contains,
the antenna part, LTE Modem and SIM. The connection with the Indoor Unit or Gateway is often done using Ethernet
Cat 5/6 cable with PoE bringing power as well as data connection. Coax cable is also a possibility but no favoured due
to cable loss and cabling constraint and the choice of an adequate transport mechanism for the IP link over the RF cable.
Outdoor CPEs require a professional installation and the knowledge of the position of the best serving Cell site. Self-
pointing antenna are still topics under study to facilitate the deployment by non-professional. The use of outdoor
antennas for the CPE increases significantly the network capacity and coverage. The CPE outdoor will extend coverage
and increase capacity (could double capacity of fixed LTE users)

Indoor CPE Outdoor CPE

Huawei Flybox B310/B68L + TP Link MR200 Huawei B2338-168

Lower device cost (50-60€) Higher device cost (120-150€)

Requires professional installation: on a pole on the


Simple installation, can be given to customer
roof of the building or on building wall.

Usually installed in indoor window or deep indoor


Usually installed in Line of Sight with Base Station.
conditions. Poorer radio conditions and
Better radio conditions and performance; stable radio
performance caused by in-building penetration
link to guarantee high bitrate services.
losses and fading.

Smaller antenna gain Higher antenna gain (6-9 dBi)


Requires good coverage Large cell radius (less sites)
Deployed: Spain, Belgium next deployment : Romania/ Morocco

Orange confidential - 26
7.5 Radio Aspect: Performance and capacity
A fix home user generates, for internet +VoIP services (without TV), 10 to 20 times the monthly traffic of a mobile user
with a large portion of traffic during the evening Busy Hours (typically few GBytes per month on mobile and 20 to 30
GBytes per month on fixed).

To sustain this additional traffic two main resources need to be dimensioned: radio and baseband (the processing power
at the base station).

LTE cell capacity of some commercially launched LTE FDD networks with fixed offers is roughly 23 Mbps for 15 MHz
channel (and 2/3 of this figure in a 10 MHz channel), providing an average user throughput around 3 Mbps. Typical
capacity of such cell is 1300 mobile customers, or 105 fixed customers, or a combination of both.

Some solutions, especially outdoor antenna, increase significantly the cell capacity.

7.6 B2C services architecture


The figure below highlights the main E2E B2C architecture technical aspects to consider:

Figure 17 – E2E B2C architecture design

Affiliate should contact OLN to evaluate their local context and define a viable LTE deployment strategy including a good
dimensioning considering for residential market the following items: limited number of customer by cell, progressive open
in limited areas and the application of fair usage policy.

- 27 - Orange confidential
7.7 B2B services architecture
The figure below highlights the main E2E B2B architecture technical aspects to consider:

Figure 18 – E2E B2B network architecture

Affiliate should contact OLN to evaluate their local context and define a viable LTE deployment strategy including the
migration of B2B services already supported on WiMAX. The affiliate should anticipate this strategy considering that
camping strategy should be ensured, spectrum refarming will happen if TDD/LTE with WIMAX in the same band, to
assess the guaranteed bandwidth/QOS policy and the E2E services architecture strategy. A transversal team is
mandatory to success.

LTE for Fixed Broadband in AMEA : Group Recommendation 7—1

Use of LTE to serve the corporate market and potentially replace WiMAX or PMP technologies is now a reality with
devices targeting the B2B market and features such as QoS and GBR supported.
Both LTE TDD and FDD can be used for fixed services, the cheaper price and wider availability of unpaired
spectrum among Fixed Wireless Access operators has helped the TDD mode to develop a stronger ecosystem for
Fixed Outdoor CPE until now. Deployment of LTE network dedicated for Fixed usage should be assessed carefully
based on local context and target markets.
 Frequency spectrum plan should be done before any deployment to pay attention to 5G deployments in
future Band42.
 Make services assessment (specially coming from WIMAX technology) and to define the bundles of services
according to marketing areas will be the starting point.
 Make assessment of dimensioning considering type of services and spectrum band/mode.
 Consider the new network elements to be introduced in the network (VPN concentrator, PCRF capable to
manage dedicated bearers)
 Select the cost effective CPE according to the service defined and mode. CPE outdoor are preferred to
increase capacity/coverage efficiency
 To define the E2E QOS policy according to the bundles of services and throughputs expected
 To define the E2E services architecture according to the service defined
 If TDD mode is chosen, need to guarantee the camping strategy, to guarantee the E2E synchronization
(especially for TDD mode) and also, to evolve radio network evolution according to the TDD band selected.
 If there was an existing WiMAX network, customer migration would need to be ensured, especially if the
same TDD band is used for WIMAX and TD-LTE. Refarming analysis needs to be considered.

Contact: Vanesa TORNERO (RNM), Alioune LY El Hadji (RNM), Franck PAYOUX (RNM), Dan Musat (RNM), Ferdinand
TRA (RNM), Ronan Le Bras (RNM), Philippe HAMET (OLPS), David CAILLIERE and Jose Pirlot (CNC)

Orange confidential - 28
8. V2V and V2X
3 GPP has standardized LTE for V2X ecosystem in Release 14 expected to be finished by June 2017 (commercial
products by end 2018 /beginning 2019?). A Partnership between Orange, PSA, Ericsson and Qualcomm has been
settled in 2017 in order to plan evaluation on the field of early prototypes and assess the potential of the technology
through some specific use cases in the coming months.
Moreover in the US, the DOT (Department of Transportation) is considering mandating support of V2V communications
on any new vehicle starting 2019/2020.

V2X modes – possible interactions between the vehicle and its environment
V2X (Vehicle to everything) defines any communication involving a vehicle as a source or destination of a message.
Depending of the nature of the other communication endpoint, several special cases exist:
 V2V: Vehicles broadcast information that can be used by other vehicles to enhance autonomous driving. This
information could be things like speed, direction of travel, braking information.
 V2I: Vehicles broadcast information to roadside Infrastructures, for example traffic lights and traffic sensors.
Similarly, roadside infrastructures can broadcast information to vehicles (such as speed limit information) that
can be used to alert drivers etc.
 V2P: Connectivity between vehicle and pedestrians, for example for some road safety applications.
 V2N: Connectivity through a network to service platforms, either MNO platforms (e.g. big data platforms), Car
OEM platforms (e.g. remote diagnostics), or other parties’ service platforms (e.g. eCall, insurance, highways).
This also includes connectivity to the Internet.

Figure 19 – V2X communication modes

LTE V2X has been designed for road safety and non-road safety services
Both road safety and non-road safety use cases are tackled. They can be split into 4 categories with the following
examples:
 Road safety
• See next table
 traffic efficiency
• See next table
 Mobility and comfort
• See next table
 Digitisation of transport & logistics
• Remote sensing & data collection (vehicle-to-vehicle & vehicle-to-backend)

- 29 - Orange confidential
High level LTE V2X features
Main features are:
 Introduction of ITS 5.9 GHz band of operation [5855 MHz, 5925 MHz]
 V2X service in E-UTRAN coverage and out of E-UTRAN coverage
 V2V Direct mode with enhanced sidelink D2D
 2 enhanced radio resource allocation modes: mode 3 (=network controlled) and mode 4 (= autonomous)
 Doppler handling for high speed mobility
 Enhanced MBMS for V2X (Multicast/Broadcast)
 New functions in the network: V2X control, V2X UE application and V2X application server
To better understand this upcoming LTE-V2X ecosystem: launch of a dedicated Anticipation project in 2017
Main objectives:
 Analysis and synthesis of the LTE-V2X Release 14
 Understanding of security/privacy/authentication mechanisms defined in LTE-V2X
 Study of deployment scenario and first recommendations (spectrum, actors’ model, for Orange, etc.)
 Understanding of the manufacturers roadmap and impact on our network (software or hardware upgrade)
=> preparation of future tests in our operational network
Contact: Laurent Mussot

9. Professional Mobile Radio


Professional Mobile Radio (PMR) networks are radio communication systems used in market segments listed below:
 Public safety, such as police, fire brigades, ambulance. These professional networks are commonly called
PPDR (Public Protection and Disaster Relief).
 Vital Infrastructure Operators (VIO), such as energy, transports companies.
 Private companies, such as industries, manufacturers, taxis, civil work, etc.

9.1 Limitations of legacy PMR networks

PMR legacy standards:


 Legacy PMR systems are based on analog technologies (MPT1327, APCO 16) or on digital technologies
(TETRA, TETRAPOL, APCO25, DMR, etc.).

Orange confidential - 30
 In particular TETRA (Terrestrial Trunked RAdio) was developed by European Telecommunications Standards
Institute (ETSI) as an open standard for trunked PMR systems for being an official European digital PMR
standard. Followed by its significant success also outside of Europe, it has moreover become an international
standard. TETRA is mainly used by public safety and vital infrastructure operators.
 As an example, in France, around 500 000 PMR subscribers are using PMR voice services (push to talk),
mainly on TETRAPOL networks. They are distributed between Public Security Forces (police and rescue
missions of the State) and civilian purposes (business, administration).

Frequency bands:
 Frequency bands used for legacy PMR networks are currently not harmonized.
 As an example in France, 400 MHz band is used. This band is divided into three frequency groups:
 The Defence Ministry has got 37.15 MHz for its own purposes
 The Interior Ministry has got 7 MHz for its own purposes
 ARCEP has got 29.4 MHz for covering the needs of other PMR users

Characteristics and limitations of legacy PMR networks:


 Legacy PMR networks are based on dedicated infrastructures, dedicated frequencies and dedicated
terminals, as standards are specific, which implies high costs for deploying and operating such systems.
 Resilience and security on these networks are usually very high as some mission critical services are
performed.
 These technologies are only “voice centric”: ability to perform voice group call (push to talk), to prioritize calls
and to establish direct call between terminals without going through a base station (Device To Device call).
 Legacy PMR technologies are not suited to offer data broadband services, such as pictures sharing, push to
video services or professional databases accesses. So PMR customers can not benefit from services already
available for mass market customers.
 In some cases, these PMR technologies become obsolete and maintenance costs are very high.
 In addition, PMR frequencies bands being not harmonized between countries, no roaming is possible.

9.2 LTE for next generation of PMR networks

LTE is chosen to support next PMR networks generation as LTE is able to transport, with a low latency, real-time
flows such as voice or video, high amount of data and to ensure efficient remote accesses to professionals’ databases.
LTE being fully standardised (no proprietary solutions) and commonly used in the world, PMR customers will benefit from
a dynamic ecosystem and from economies of scale.
Decision to use “PMR over LTE” for public safety usages was already taken in several countries such as USA, UK,
Australia, South Korea, each of these countries having adopted different strategies in terms of migration path between
their legacy PMR network and next generation of PMR network.

PMR features standardized in 3GPP to answer public safety needs:


Data services and applicative Push-to-X (including push-to-talk) solutions are sufficient for most professional
customers outside of public safety scope, can be proposed right now to PMR customers.
But for fully answering public safety needs in voice domain, and so for offering services equivalent to the one available
on legacy PMR networks (especially on Tetra/Tetrapol networks), some new features are standardised in 3GPP LTE
releases 12 and 13.
 Direct mode communication between close terminals (D2D), called “Proximity based Services (ProSe)” in
3GPP:

• The mobile should be able to discover other mobiles and communicate in D2D mode outside of LTE
coverage (no connection to the base station).
• The feature is included in 3GPP release 12 (March, 2015), some enhancements are done in release
13 (2016).
 Group communications in voice, video or messaging (One-to-many group calling / Talk group management):

• Continuity of these types of calls shall be ensured in case of mobility between different cells.
• Group calls enablers to deliver an IP flow to a group of terminals using eMBMS (LTE multicast) and/or
unicast bearers.

- 31 - Orange confidential
• These features are included in 3GPP release 12.
 Mission Critical Push-To-Talk (MCPTT) service and evolution to support multimedia group communications:

• Specific performances requirements, interface between application server and network.


• This feature is included in 3GPP release 13 for voice services (2016).
 High level of availability and resilience:

• Availability, if the connexion with the core network or backhaul is lost, to provide locally Push To Talk
services under the eNodeB coverage area (isolated eNodeB operation mode).
• This feature is included in 3GPP release 13 (2016).

According to the standardisation progress, these public safety requirements will be available from 2016/2017.
Release 12 features are already available and tested for some suppliers.

Frequency bands for PMR over LTE:


Dedicated LTE spectrum for public safety and some Vital Infrastructure Operators (highly dependent on
countries decision):
In the context of the evolution of PMR systems over LTE, an ECC decision at CEPT level aiming at harmonising the
spectrum in Europe for public safety broadband PMR has been published in June 2016 with several options to be
chosen at national level by administrations (concept of “flexible” harmonisation):
• Option 1: up to 2 x 5 MHz of spectrum in the 400 MHz band (450-470 MHz). May be revised later by
adding the 410-430 MHz band.
• Option 2: up to 2 x 8 MHz of spectrum in the 700 MHz band outside of the 703-733 MHz / 758-788 MHz
band plan. This decision will allow also use of the 703-733 MHz / 758-788 MHz for administrations wishing
to do so.
In France, decision was taken in July 2015 to allocate to French Interior Ministry: 2x3 MHz (733-736/788-791 MHz)
included in standardized 3GPP APT band 28 AND 2x5 MHz (698-703/753-758 MHz) not included currently in a
standardized 3GPP band (standardization process supported by French public safety bodies and some suppliers). It has
to be noticed that there is a risk to have no ecosystem (radio infrastructure and terminal) available from the suppliers in a
short/mid-term in this band, as the standardization is not completed and as only France attributed so far such
frequencies to public bodies.
In UK, no dedicated frequency band is allocated to public safety bodies (Home Office). EE hosts public safety users on
its LTE frequencies and network.

Spectrum for customers having no “mission critical” function:


Outside of public safety and Vital Infrastructure Operators having mission critical functions, public and private
companies will have probably no access to dedicated LTE spectrum harmonised at CEPT level (even though
there is a work item at CEPT level aiming at looking for the possibilities to use the 450-470 MHz for LTE that could
address the broadband PMR needs besides broadband PMR for public safety). Therefore, if public and private
companies wish to perform their “no mission critical” PMR applications over LTE, they may have to use public operators’
frequencies. Nevertheless, some organisations defending PMR interests (e.g. Agurre in France), looking for more
spectrum for their future PMR broadband needs, asked, through a French public consultation, to have the possibility to
use some frequency bands (e.g. 2.6 GHz TDD) for these PMR broadband needs.

Possible models for PMR over LTE networks:


Several networks’ models are possible, strongly linked to the availability or not of LTE spectrum dedicated to the PMR
customer:
 Deployment of dedicated “PMR over LTE” networks as today, when dedicated spectrum is attributed
(typically for public safety). This implies very high costs and needs a long time in case of national deployment.
Public operators could have the responsibility to think, build and run such dedicated networks in this model.
 Use of public operators LTE networks for supporting PMR applications, dedicated spectrum being
available or not. This makes sense as LTE networks are already deployed/under deployment, as operators
have LTE spectrum and experience in think, build and run processes of such technology. As an option, some
resources could be reserved for PMR customers on operator LTE network.
 Mix of both scenarios, with a combination of dedicated elements and synergies with public operators LTE
networks, typically:
• dedicated services, HSS, core network, but radio shared with public operator (RAN sharing model).
• dedicated services, HSS and part of core network, but radio, MME and S-GW provided by public
operator (MVNO model).

Orange confidential - 32
• dedicated services platforms connected to public operator LTE networks.

Depending on customer’s needs and requests, the swap of legacy PMR networks to “PMR over LTE” networks,
will be fast or not.
For instance in UK, public safety body chose to migrate their PMR applications to LTE from first semester 2017 and to
switch off Tetra network quickly.
In other countries, a progressive migration to LTE is envisaged, with a transition phase during which both networks
(legacy and LTE) will co-exist, voice applications (PTT) remaining on legacy PMR networks and data applications being
deployed over LTE.
For customers (such as big companies, production plant, industrial area) using no specific PMR network today,
but having their professional applications supported by various technologies such as DECT, WIFI, etc., adoption of LTE
could be fast as soon as technical and commercial agreements are concluded with public operators.

9.3 PMR: new business opportunities for Orange

Legacy PMR networks are managed by historical PMR actors such as Thales, Motorola, Airbus D&S. Public
mobile operators are not involved as legacy PMR technologies are not among the ones managed by them.
Evolution towards LTE offers the opportunity for public mobiles operators such as Orange to be involved in
PMR business, as they already offer large LTE coverage and benefit from an E2E expertise for thinking, building and
running LTE networks.
Orange, as a major LTE public operator, is a credible partner to propose LTE-based technical and commercial
solutions for PMR customers; these new business opportunities shall be identified.

3 types of PMR customers are targeted by Orange:

Typical customer’s needs:


 Guaranteed radio access: the access and QoS on the network shall be guaranteed for safety and production
segments, even in case of congestion.
 Differentiated quality of service: specific QoS (typically for Push-to-Talk service and for some real time video
services) shall be allocated for priority services of the 3 PMR segments.
 Voice and messaging group services: applicative push to talk and push to message solutions, with a specific
QoS, shall be available for the 3 segments. For public safety and some VIOs, these solutions should be able to
include evolutions specified by 3GPP in R13 (MCPTT).
 Video services: professional real time video services in group mode (push to video) solutions shall be available
for safety and production segments.
 Resilience and security: key request from safety and production customers segments.

- 33 - Orange confidential
Despite Orange strengths, it will be necessary to complete our solutions portfolio to answer
PMR customer’s needs and so to make them utilizing our LTE networks.
In particular, it is key to:
 Include additional quality of service/prioritization/pre-emption mechanisms, to guarantee voice and data
services of PMR customers in any situation.
 Integrate new functionalities to provide similar voice-centric services as the one proposed on current PMR
networks.
 Extend and tune the radio coverage in some specific areas (such as industrial areas), in accordance with
customer’s needs.
 Evaluate with PMR customers their needs in terms of network resilience and propose the relevant solutions.
 Integrate regulatory constraints if any.

9.4 On-going PMR activities within Orange

Anticipation activities:
Main objective of OLN PMR anticipation project, ongoing since 2015, is to contribute to Orange technical strategy
definition for the 3 targeted customers segments presented previously.
Two PMR trials - one in France with Nokia and one in Poland with Huawei - aiming at identifying and testing network
and service enablers answering PMR customers’ needs, were conducted and completed in 2016.
Some technical solutions were evaluated during several tests campaigns on each PMR trial. The following tests
domains were experimented:
 Dynamic radio resources reservation for PMR usages: this feature enables the reservation of radio
resources for PMR traffic, PMR users being managed on a specific PLMN id, when a single 4G network is
used for PMR and non PMR usages (tested on Nokia and Ericsson RAN).

 Local resilience in case of macro network failure (IOPS-like behaviour): this architecture enables a
local resilience for the most critical PMR services (e.g. push-to-talk) in case of macro network failure. This
local resilience was performed, on PMR trial in France, using a micro EPC/HSS/PTT application equipment
provided by Athonet.

 RAN sharing configuration (MOCN) on a commercial Orange France eNodeB to manage PMR traffic
on a specific core network: this architecture enables the local management of PMR flows on separated
core network and applications servers, these equipment being installed for instance on the customer site.

 LiTRA solution: this Huawei application server, experimented on PMR trial in Poland, provides advanced
Push-to-X services (voice, video, distress call, private calls, geolocation, etc.) associated to a good user
experience. In particular, new QCIs (65, 66, 69, 70) defined in 3GPP release12 for providing a specific QoS
to PMR services, are supported from E2E on Huawei equipment, including LiTRA application server.

The successful outcomes of PMR trials tests showed that 4G networks are technically able to deliver PMR services
and that necessary network tuning to answer PMR customers’ needs can be put in place.

In 2017, some remaining anticipation activities will be conducted on new 3GPP standardized features (if available) and
some implementation studies of features and architectures experimented during PMR trials on Orange 4G networks will
be also completed.

Delivery activities with Orange France:


Three PMR streams are currently on-going with Orange France:
 PMR trials with customers
 PMR commercial offers preparation
 Answers to French Interior Ministry RFQ for their future radio network

a) PMR trials with customers in France


Two types of PMR trials are conducted with some Orange big account customers. Trials objectives are:
 to prove to customers that Orange is already able to provide PMR services on its commercial network, while
waiting for commercial offers
 to collect feedbacks from customers about the selected Push-to-X solution, in order to improve possibly the
solution
 to test in real environment the selected network enablers/E2E architecture

Orange confidential - 34
 to help to prepare commercial offers
st
1 type of trials is based on the Push-to-X services supply, this application being provided by Nokia/Genaker. Such
trials started in February, 2017.
nd
2 type of trials is based on the Push-to-X services supply and on the guarantee of quality of service for PMR
customers, by reserving dynamically some resources for PMR traffic on the eNodeB(s) covering the customer’s site.
Such trials will start in summer 2017.

b) PMR commercial offers in France


Some commercial offers are under preparation to address the French PMR market for each customer segment
presented previously (green, yellow and blue segments).
For each segment, the technical solutions under study are based on the deployment of new services and network
enablers for PMR customers:
 Deployment of a Push-to-X application.
 Some guarantees for the PMR customer on its site, whatever the congestion level on the radio equipment due
to non PMR customers.
 Specific Quality of Service for PMR services.
 Some solutions for increasing resilience in case of macro network failure and/or for enabling autonomous
behaviour on a customer site.

First PMR offers should be launched in France beginning of 2018.

c) French Interior Ministry (MI) RFQ


French MI launched « PC STORM » project, objective being to eventually replace their 2 existing PPDR Tetrapol
networks by “one” network based on LTE.
Current MI RFQ is related to a first phase of 4 years. MI wants to validate in this first phase a new technical solution
with a “Proof of Concept”, in order to perform the generalization in phase 2. MI requests are based both on the use of
dedicated equipment on their dedicated spectrum (2*8MHz are already attributed to MI in 700MHz band) but also on
public mobile operators networks use. In phase 1, a very low volume of users is concerned (6000 users), when 300k
to 500k users will be concerned in phase 2.
MI RFQ is composed of 7 lots. Orange answered to 2 lots beginning of Q2 2017. MI decision should take place by
end of the year.

- 35 - Orange confidential
PMR over LTE: Group Recommendation 9—1
 LTE is the technology selected to support the future broadband PMR, as a replacement or complement of
legacy PMR networks.
 Orange is a credible partner to propose PMR over LTE solutions to the 3 PMR customers segments (public
safety, production and standards). New business opportunities shall be identified with Orange business
teams.
 Two PMR trials were completed by end of 2016 for identifying and experimenting network and services
enablers which could be deploy on Orange commercial LTE networks.
 Some PMR experimentations with big accounts in France started in February 2017. PMR commercial offers
are under preparation for a launch by Orange France beginning of 2018.
 Orange answered to French Interior Ministry RFQ in Q2 2017, RFQ related to a first phase of 4 years.

Contact: Nathalie Jollivet

10. Public IoT/LPWA/Mobile IoT Networks


10.1 IoT at Orange
IoT / M2M and LPWA networks became a hot topic in 2015 in the industry and within Orange with a strong focus on the
Connectivity part of the IoT value chain targeting the low end of the IoT markets. M2M / IoT is one of the two
diversification areas of the Orange Essential 2020 strategic plan.
In addition to the current Cellular M2M offers and services using mainly 2G/3G Mobile networks (8 millions M2M SIM
Cards around the Orange footprint); mobile operators are in a position to extend the range of addressable IoT use cases
by using several alternatives IoT Connectivity such as LPWA for the Low end segment of the IoT.

B2C

B2B2C

B2B

However the extension


to new verticals where mobile operators are not seen as legitimate will imply that they are fully equipped to a new
paradigm:
 IoT is an End to End business meaning that all aspect of the solution must be available to fulfil the promises.
 Connectivity only capture a small portion of the global Market forecast for IoT (5 to 8 % according to analyst)
 The Value in IoT is in the services to be developed on top of connectivity
 The adoption of IoT by some verticals can be slow as product cycle is longer that mass market electronics

Orange is addressing the IoT markets through its affiliates and Orange Business Services / Orange Application for
Business is developing several of the technical bricks (Datavenue – Live Object - MALIMA) necessary to fully cover the
customer needs from connectivity to services for both B2B and B2C markets.
On the B2C market, Orange is working to develop this new business of connected objects distribution and associated
services, with a current focus on « Wearables », « Smarthome » and « automotive » verticals.
On the B2B side, through OAB and the local M2M Business units & Orange Mobile Enterprise, the focus is on high
priorities verticals: Smart Industry, Smart Cities, and Transport & Logistics.

On B2B market, affiliates are currently focused on M2M connectivity and seeking volume, but they aim at developing the
revenue from services. IoT covers very different use cases requiring characteristics and expected SLA; Orange will
address the different cases with a complete portfolio of IoT Connectivity solutions

Orange confidential - 36
10.2 LPWA Networks at Orange
Within Orange and at a larger scale in Europe, 2016 marked the move from anticipation and experimentation of the first
LPWA solutions to delivery project and commercial Launches. Several large players including Cable and Mobile
operators have announced deployment of Unlicensed LPWA networks based on LoRa (Comcast, Softbank, and Tata) or
agreement with SigFox. (Telefonica, Kyocera)
The buzz around LPWA networks created by the arrival of new players and solutions (Sigfox, LoRa, Ingenu) did not fade,
it was in fact reinforced by the announcement of the completion of the first version of Mobile IoT solutions specified by
3GPP in June 2016 and heavily marketed by some of main promoter. (Vodafone, Huawei for NB-IoT, AT&T, Verizon for
LTE-M).

Need for LPWA networks


Within the IoT, some new needs have emerged to enable various use cases and devices in the Low end part of the IoT
market such as sensors, smart meters (water, energy), trackers (PET, cattle, asset) …
The key characteristics of these uses cases are that they imply:

 low cost solution (module + connectivity) to enable an implementation in high numbers of devices
 low consumption solution as many devices would be running on battery and not easily rechargeable
 low data exchange from a few packet per day (metering) to packets per hour (smart parking)
 long range allowing to address deep indoor devices (metering) without repeaters

LPWA networks can be of two kinds:


- Unlicensed LPWA networks dedicated to IoT using unlicensed spectrum (LoRa, Sigfox, Ingenu)
- Licensed LPWA networks or Mobile IoT build over existing cellular networks. (LTE-M, NB-IOT, EC-GSM-IoT)

See below: IoT covers all types of connected objects from high end such as Video Camera to Low end. Only a subset
requires a new LPWA connectivity, connected objects using standard 3G and LTE will continue to develop.

LTE-M

NB-IoT

EC-GSM-IoT

10.2.1 Alternative Unlicensed LPWA Solution:


Several LPWA solutions using unlicensed Spectrum are available. There is still currently no single standard for
LPWA. These solutions are designed for License exempt bands such as the SRD band in 868 MHz band in Europe,
as a consequence their technical characteristics are much more constrained that cellular system.
 Low transmission power (typically: 10 to 25 mW), 10x lower than GPRS (250 mW)
 Strict Duty Cycle (1% for End points, 10 % for Gateway) according to the Sub –band used.
As a direct consequence of these constraints, the following characteristics:
 Low Data rate from 100 bit/s to 250 Kbit/s
 Low power consumption / Long Battery life
 Low Cost (claimed less than 3$)
The major LPWA (Low Power Wide Area) solutions on the market Worldwide are:
 Sigfox (Ultra Narrow Band)

- 37 - Orange confidential
 LoRa Semtech
 Ingenu (Homerider)
 Qovisio (FR) ( UNB),
 Weightless (UK), IEEE 802.15.4k (For utilities)

Orange launched a review of the major solutions in 2015 and selected LoRa as the LPWA solution for the Group.

The choice was made on several criteria including an Open


Model, a growing ecosystem and support of bidirectional
link.
LoRa is built on a proprietary physical interface own by
Semtech, the LoRa alliance was launched in Q1 2015
aiming at developing an open ecosystem around the
LoRaWan specification that is compatible with an open
operator model. The LoRa model allows deployments of
multiple networks operators per country in both public IoT
and private network modes for a specific B2B verticals.

Orange joined the LoRa alliance launched in Q1 2015 develop


the open ecosystem and ensure interoperability; device
certification and roaming become a reality.
The alliance has now attracted more than 400 members
including large mobile operators such as SKT (Korea), KPN
(Holland), Tata (India), Softbank, NTT (Japan).

Strength & Limits of Unlicensed LPWA solutions

The use of unlicensed spectrum is interesting in the initial phase of IoT however it brings some constraints that should be
understood when considering deployment:

 Unlicensed LPWA such as LoRa are well suited for delay tolerant, infrequent
exchange of small packets mainly in the Uplink
 LoRa support bi-directional use but not simultaneously (Asynchronous mode), it
is not suited for use cases requiring real time transaction such as POS – Payment
 LoRa and other unlicensed LPWA solution shared spectrum with several
devices and applications, QoS cannot be guaranteed, use cases linked with security
and safety should not rely only on it.
 LoRa and other unlicensed LPWA are more resilient to Jamming as devices do
not need to synchronized before transmission, the use as a secondary LPWA solution to
a first Cellular solution is an interesting capabilities ( Home Security)
 LoRa can also be deployed indoor using Small Pico-Cell / Nano-gateway to
address local deployments for Smart Home / Smart Building and Smart Industry use
cases.
 It support bi-directional link in asynchronous mode used to implement a slow
power control and rate adaptation and to send data to the end points.
 LoRa support a Geolocation by using the TDOA enabled Uplink Macro diversity
and precised Timestamping by GPS synchronized Gateways

10.2.2 Current Orange activities on Unlicensed LPWA

Orange has stopped its cooperation with Veolia in the joint venture M2OCity which is delivering M2M Smart Metering
operator services for in the water metering sector.
LoRa Pilot in Grenoble – IoT Territory
Within the 3L project in 2015, was the selection and the deployment of a First IoT Territory in Grenoble in France based
on LoRa. The pilot done with Orange Smart Cities and OLPS/ BIZZ has covered both technical and business aspects of

Orange confidential - 38
the LoRa solution. The network deployed in two months on 14 sites, it covers a 40 km x 40 km area in Grenoble and its
surrounding was open to friendly Business Users in May 2015. A first version of Datavenue is used as the Data
Management platform for this IoT Territory. The pilot is used as a tool to develop the LoRa ecosystem. It covers a wide
range of use cases with start-ups and large groups. A process to test the Interoperability between LoRa Module and
Infrastructure has been set up.

Orange France LoRa Network deployment and activities outside France.


Orange has capitalised on the expertise developed in the support to M2Ocity and the Grenoble pilot to select LoRa as
the most suitable LPWA solution for the group in mid-2015. The choice was made on several criteria
including the openness of the LoRa ecosystem.
A LoRa Skill Center was created to support the deployment of Orange France and the other activities
in the Orange footprint including OBS prospective projects.
The first activity of the Lora SKC was to lead the selection of a group supplier for the complete
LoRaWan Network; a RFP was launched in September 2015 leading to the selection of the solution
from Cisco and Actility. The LoRa SKC is currently in the phase of validation of the solution that is
being rolled out in France.
A second RFP was launched in 2017 and selected a new supplier for Gateway only: Kerlink with
Kapsch.
Activities in Affiliates:
The commercial opening in France took place in summer 2016 after a first phase of pilots to validate radio planning and
deployment processes.
The first objective was reached with the Top 18 Urban Areas covered by End 2016 with around 1500 Gateways. The
current target is to cover the Top 120 Urban Areas by mid 2017, including 2600 cities with 1000 extra gateways. Further
extension towards a National coverage was announced in June 2017 in order to enable the commercial team of OBS to
compete against competition (Objenous Lora & Sigfox) and address use cases around tracking that represent a
significant part of the current commercial pipe for LoRa and LPWA networks. (Asset Tracking in the Industry &
Agriculture, Logistics).
In France, LoRa is tightly linked to the Orange Live Object platform that is necessary to collect and expose the Data to
the customer Server and IT system using API. As LoRa does not support IP to the device, interworking functions needs
to be implemented by the Network Server and by an IOT platform.
In addition to the roll-out in
France, a solution suitable for
small size deployment by
other Orange affiliates in
Europe and by OBS outside
the Orange footprint is under
study based on a mutualised
solution LoRa Anywhere.
For any inquiries regarding
Lora and unlicensed LPWA
solutions, It is advised to
contact the LoRa Skill Centre
from the very early stages of
business case definition as it
can assist MCOs in their case
analysis.

Orange activities at OLN are focusing on supporting the LoRa delivery project and the extension to other affiliates using
the LoRa Anywhere concept. (Below)
 A first POC of LoRa is ongoing at Orange Slovakia using some element of the LoRa Anywhere solution.
 Orange Romania and Orange Moldavia are evaluating LoRa technology with a local partner.
 Several affiliates in MEA are planning to evaluate LoRa as a first LPWA solution capitalising on the Group LoRa
Anywhere technical solution.
 OBS will include a full LoRa + Live Object offer to its portfolio.

It is advised to contact the LoRa Skill Centre from the very early stages of business case definition. The LoRa Skill
Centre evaluated most of the new comers technologies or propositions and can provide guidance to affiliates as well as
provide the right Orange Lab expertise on any LoRa network topic.

- 39 - Orange confidential
Unlicensed LPWA: Group Recommendation 10—1

 Orange has selected LoRa as its first LPWA solution for the Low end IoT markets (mainly B2B)
 Affiliates in Europe should study the business opportunity of LoRa in their local markets
 A mutualised solution is being tested LoRa Anywhere to lower the entry cost.

Contact: Bertrand Waels & Jerome de Murcia (SKC)

10.3 3GPP solutions: Mobile IoT

10.3.1 Standardisation Status

Orange is active in 3GPP standardisation activities on Machine Type Communications (MTC) with the objective of
improve the support of Low End M2M / IoT in a way that enable upgrade with the minimal impact on the existing network
base and support for true interoperable ecosystem.
Both RAN and Core Networks have been improved since release 10/11. However not all features have been
implemented as they prepare networks to connect large amounts of IoT devices that will only materialise in a few years.
The LTE-MTC work targets the optimisation of the control and usage of resources of LTE/Evolved Packet Core and
UMTS/GPRS networks for M2M devices.

Release-12 specification completed in 2015, focused on power consumption and signalling reduction for small data
packets and aims at provision of low-cost LTE MTC UEs targeting a modem cost similar to GSM/GPRS.
• PSM (Power Saving Mode) feature is introduced for both the UE and CN Network and equally applicable to
GPRS, UMTS and LTE
• For LTE new categories of UE (Cat 0 and Cat 1) specified in Rel 11/12, were expected on the market in 2016.
The trend in the industry is to focus only on Cat 1 device (10 Mbps max) and then skip the Cat 0 to focus on Rel 13. Cat
1 Module are available since 2016 for testing.

Release-13 specification published in June 2016 marked a turning point as under pressure from new players using
unlicensed solution, 3GPP RAN accelerated its process and defined three evolutions paths designed to address IoT &
LPWA connectivity requirements taking into account the diversity of the existing infrastructure (2G & 4G) and the need to
address some case with a Low Cost simple solution by design.

 2G evolution: EC-GSM-IoT is an evolution of GSM / GPRS supported in the standardisation phase by Ericsson &
Nokia with active support from Orange and 2G Chipset / Module Maker. It was aimed at capitalising on the large
footprint of GSM 900 around the world to address quickly LPWA use cases with a Low cost entry.
 EC-GSM-IOT was specified in GERAN (Now RAN 6) in Rel 13 June 2016; it addresses the LPWA requirements with
an evolution of the legacy (E)-GPRS that could be applied by Software Upgrade. It introduces:
 new EC-GSM-IOT logical channels designed especially for EC-GSM-IOT devices
 Allocation of Radio resources (Time Slot) to EC-GSM-IoT traffic on a dynamic basis
 Extension of coverage using Repetition schemes of L2 and L3 activated according to the need of coverage
extension and to the logical channel considered.
 Removal of support of CS domain to lower device complexity.
In addition, it will benefit from generic RAN and Core network features specified for the support of Power optimisation
features:
 PSM Power Saving Mode
 eDRX

 4G evolution: LTE-M (eMTC/ Cat M1) & NB-IOT


The work started within the LTE-MTC workstream in earlier release was complexified by the transfer from
GERAN to RAN of the WI on NB-IoT; it created a dual evolution path for LTE evolution.
 LTE-M / eMTC evolution of LTE including the support of MTC category of Device Cat 1, Cat 0 and Cat M1 device
on an existing LTE Carrier. It is named LTE-M when Cat M1 is used with other IoT feature (PSM, eDRX)
 NB-IoT a new approach based on the cleaned slate approach brought to 3GPP by Neul-Huawei and later
merged with Ericsson NB-LTE proposal. NB-IoT uses category of device NB1 (used to be call Cat M2) in Rel 13.

Orange confidential - 40
Figure 20 – The 3GPP path to NB-IoT
Note: The introduction of NB-IoT to 3GPP under pressure from Huawei & Vodafone met opposition from the US players
(operators and chipset makers). It led to several blocking point in 3GPP RAN in late 2015. Although a working agreement
was reached to avoid a fragmentation of the NB-IoT solution with two uplink modes, there is still some uncertainty on the
level of interoperability between key vendors and chipset.
In addition, there is still some intense lobbying in 3GPP and outside to impose NB-IoT as the de-facto 3GPP backed
LPWA solution to the detriment of LTE-M based on Cat M1. Improvement of both solutions in Rel 14 and Rel 15

LTE-M key features:


 Cat M1 devices are reduces in capabilities compared to Cat 0
 Reduction of bandwidth to 1,4 MHz (1.08 MHz / 6PRB)
 support of Half Duplex & Full Duplex Mode
 single RX/ TX antenna chain
 Reduced Data rate:
 Full Duplex: 1Mbps UL / 800 kbps DL
 Half Duplex: 375 kbps UL / 300 kbps UL
 Enhanced coverage: up to 15 dB with two Modes A & B using repetition scheme.
 Mode A (5-10 dB)
 Mode B (10-15 dB) (Further extension up to 18 dB possible according to simulation)
 LTE-M (Cat M1) can be activated on any FDD or TDD LTE Band
 LTE-M (Cat M1) devices & traffic to be multiplexed with other types on the same LTE carrier (10 MHz / 20 MHz)
 LTE-M (Cat M1) inherit from LTE the support of Idle & Connected Mode Handover
 LTE-M (Cat M1) is supporting some level of VoLTE although not in extension coverage Mode B or in Handover
 LTE-M support SIM, eSIM/eUICC
 Also LTE-M support a new UE power class (+20 dBm) for IoT and the current +23 dBm class

NB-IOT key features:


 NB-IOT is a new Sub RAT of LTE with the following PHY features
 support new Category of device NB1 (M2 name was used also but no longer valid)
 Narrow band support of 180 kHz (1PRB)
 Support three mode of deployment Inband LTE, Guard Band and Standalone

 Inband scenario, NB-IoT supported within an existing LTE Carrier, the real impact is more than 200 KHz. At the
moment this is the only “allowed scenario” in Europe
 Guard Band: NB-IOT uses the edge of the LTE Carrier. Regulation is not clear due to CEPT co-existence rules.

 Standalone using GSM 900 Spectrum. Regulation not clear yet due to co-existence and also the change of use
from 2G to 4G is only possible is Spectrum is subject to Neutrality.
In addition,
 Supports two modes for uplink

- 41 - Orange confidential
 Single tone with 15 kHz and/or 3.75 kHz tone spacing
 Multiple tone transmissions with 15 kHz tone spacing
 No support of Turbo code for the downlink
 Single transmission mode of SFBC for PBCH, PDSCH, PDCCH
 New narrowband channels: NPSS, NSSS, NPBCH, NPDCCH, NPDSCH, NPUSCH, NPRACH
 Radio protocol have been also simplified compared to standard LTE to reduce signaling
 NB-IoT Significantly reduced broadcast system information
 NB-IoT support different mode of Data transmission: Data over NAS, Non-IP and IP.
 NB-IoT support SIM, eSIM / eUICC

 Note: NB-IoT was introduced as a very radical evolution compared to LTE-MTC, it combine several new
concept in both RAN and Core within a very short period and under pressure from some key players to be
included in Rel 13.
 Orange adopted a neutral position in 3GPP and worked to ensured that analysis was not biased and that a truly
3GPP interoperable solution would be available.
 The current analysis is that NB-IoT radical approach will require a longer time to reach an acceptable level of
maturity.

In addition, Core network features specified in SA include for LTE (Cat M1 and NB-IoT/Cat M2) the support of Power
optimisation features:
- PSM Power Saving Mode
- eDRX
- Support of non IP data transmission / Data over NAS

Further improvements to even better optimise the performance of NB-IoT and LTE-M (Cat M1) are included in the Rel 14
that is being finalised. It focuses on improving the power consumption and coverage as well as geolocation.

These three 3GPP solutions are promoted by GSMA through the name of Mobile
IoT. GSMA is positioning the Mobile IoT as Operator Managed “Licensed LPWA”
solutions Mobile IoT includes the support of a 3GPP network solution and also
consider the security aspect such as support of eSIM (eUICC) and also consider
the IoT as an End to End solution facing Standard based IoT Platforms using
established protocols (LWM2M, One M2M)
Orange is active in the Mobile IoT initiative
• Chair of the EC-GSM-IOT group from 2015 to 2016
• Vice-chair of the LTE-M Taskforce. (AT&T Chairman)
The NB-IoT Forum is chaired by Vodafone.

10.3.2 Current Orange activities on Mobile IoT

Anticipation:
3L project: In 2017 the project 3L was renamed IoT Strategy & 3GPP ecosystem.
3GPP Radio evolution for IoT
In 2015/6:
 NB-CioT Huawei proposal to RAN was tested in Shanghai. (Lab and Field)
 LTE Cat 0: Rel 12 of LTE evolution for IoT/M2M was tested both in labs and on the field in France. These tests
showed that Half Duplex mode developed for Cat 0 has no impact on legacy traffic.
 LTE Cat 1: First chipset were certified based on Qualcom Chipset, some modules are used in IoT demo

In 2016 OLN started the evaluation and of the three 3GPP Mobile IoT solutions, validation is planned for 2017:
Trial of EC-GSM-IOT:
 With Ericsson in France both in labs and on the field between end of 2015 and February 2016.
o Partial implementation by Ericsson showed a coverage extension of 18 dB during lab trial.
o Test were done using a prototype chipset from Intel
 A Lab Trial with Nokia took place in Poland in late 2016 with prototype module from Sierra Wireless
implementing only the Uplink Coverage extension. It showed good results and allowed a full Uplink extension
gain of around 20 dBm as well as transmission of IoT Traffic using data frame in COAP until a server.

Orange confidential - 42
 The field trial planned in Poland is currently on hold due to lack of chipset implementing the full solution.

LTE-M: CAT-M1:
- With Ericsson:
o Lab Trial took place from Q3 2016 in Cevennes using Ericsson RAN & CN with Chipset from Qualcom in
the 800 MHz
o First Call and field test took place at Orange Gardens using prototype device from Sony-Altair and an
updated Node B from Ericsson. Device instability prevented from running the full set of test.
o Performance test are ongoing in the Lab on the latest Ericsson SW 17.1 focusing on extension of
coverage Mode A, Uplink and Downlink Data rate.
o Several Chipset & Module from Qualcom, Altair, Sequans are now available for testing (Q2 2017) on 800
MHz.
- With Huawei:
o Lab trial took place in Q2 2017 Orange SKC Lab in Romania to complete the de-risking of Huawei RAN
SW 12.1 to be used in Orange Belgium pilot.
o Field Trial is conducted in Belgium on a limited number of site including one Indoor Site at OBE
HeadQuarter dedicated to IoT
o Chipset from Qualcom was used for the testing as IoDT testing with other chipset have not yet been done
by Huawei on European Band.
- With Nokia:
o Lab and Field test took place in France in Q2 2017 on Ex-ALU line of product to validate the SW version
for future use in Orange France pilot later this year.
NB-IOT:
- With Huawei
o To support the planned deployment of NB-IoT & LTE-M in Belgium by end of 2017, NB-IoT is tested in
Lab in Romania and in field in Brussels using SW 12.1 supporting both NB-IoT / LTE-M.
o Chipset and module for NB-IoT is available from Neul-Huawei.
Contact: Clélia Bichon-Daire (3GPP for IoT Evolution) / Ronan Le Bras (IoT Strategy & 3GPP ecosystem)

Roadmap from Network vendor for Mobile IoT: Skill Center


The full impact in term of Hardware of Software of implementing the different Mobile IoT solutions is being investigated
with the RAN vendors. A country by country analysis is necessary as it depend on the existing HW & SW.
 Upgrade to support LTE-M is a Software only upgrade on all Vendors.
 Upgrade to support NB-IoT may necessitate Hardware upgrade depending on existing capacity (BBU, PA)
 NB-IoT is not planned on Ex-ALU line of product.

- 43 - Orange confidential
LTE-M RAN product and Skill centers validation roadmap for the key features
 The basic set of LTE–M features including Cat M1 + eDRX + Ext Mode A will be available from Huawei and
Ericsson by mid-year and validated by Q4 2017
 Some questions remain on Nokia and Ex-ALU support for eDRX.
 Ext Cov Mode B is planned to be tested but there is no certainty on the validation yet.
Contact: Y Pitrel (IoT benchmark), B Bourgeois (Huawei), JB Landre (Nokia), JM Chauffray (Ericsson)

10.4 Orange Strategy for IoT: LoRa and LTE-M


For the lower segment of the IoT where the emergence of the LPWA use case have emerged, Orange has adopted a
pro-active strategy by adopting LoRa as a first LPWA solution available since 2016 in France and for pilot elsewhere.

LTE-M

For Mobile IoT, Orange Support LTE-M as it is an easier, less costly and
more versatile Mobile IoT solution able to cover a wider range of use
cases with its better support of mobility, higher data rate.
LTE-M is supported by all RAN vendors as a Software upgrade only.
In the initial phase before full national coverage is in place for LTE, the use
of 2G as a back-up or main connectivity solution using minimal IoT feature
(PSM, eDRX) will be considered.
In MEA where 2G GSM will continue to be a pervasive networks for some
years Orange is supporting an approach based on EC-GSM-IoT.

NB-IoT will be fully evaluated in Belgium and compared to LoRa and LTE-M for the some key use cases.
The maturity of the Mobile IoT solution ecosystem (LTE-M, NB-IoT) will be monitored including the interoperability
aspects.

Orange has selected a combination of


technical solution to address the IoT
and LPWA markets. The objective of
the 2017 is to continue the development
of the LoRa ecosystem by expending
the coverage in France and abroad
through the LoRa anywhere project.
In addition, Orange will continue its
effort to address the IoT market by
launching the first LTE-M networks in O
Belgium and O Spain.

In order to accelerate the ecosystem of LTE-M device in European bands ( 800 /1800) , Orange is opening a LTE-M
Open IoT Lab at Orange Gardens near Paris using Ericsson dedicated Network in the Box solution, to host object
makers and Start-ups willing to develop their object and services on LTE-M before the commercial opening in the Orange
Footprint.

Orange confidential - 44
Public IoT / IoT Connectivity: Group Recommendation 10—2

 Orange support LoRa as first LPWA solution for the B2B use cases in France and through LoRa
Anywhere solution outside
 Orange Support LTE-M as the group solution for Mobile IoT using 2G as back up if needed.
 In MEA, the use of 2G and EC-GSM-IOT is an option until LTE coverage is significant
 Affiliate should engage with the group to evaluate the impact of Mobile IoT in their network
 IoT is a transverse subject and required a TTM project.
Contact: Ronan Le Bras

11. Small Cells (all 3G and 4G)


A Small Cell within the scope of this document is defined as a transmitter whose transmit power and coverage is smaller
than a macro cell. Small Cells may include 2G/3G/4G technologies such as Micro Cells, Pico Cells, or Enterprise Femto-
cells, optionally combined with Wi-Fi access points, and are managed by the operator. Note that Wi-Fi-only APs are not
considered as small cells. Limited scale deployments have taken place so far in Orange Group but also there are lower
than expected worldwide deployments.
Small Cells are likely to be deployed in moderate numbers in 4G, and should be a major cornerstone to support
5G deployments.

At this stage focus on 4G investment is rightly targeted on macro deployment. Once this initial coverage layer is rolled
out and traffic uptake is substantial, the first need for 4G small cells will be as a complement of coverage.
A first 4G small cell trial has been completed with success in 2014 in France. One major question mark being the
feasibility to operate 4G small cells in a multi-vendor environment, where the macro network is provided by another
vendor. The recent implementation in Poland provides positive operation feedback although limited to the case when a
dedicated carrier is assigned to small cell. Maintenance issues with different layers of macro (Huawei) and small cell
(Nokia) resulted swap of the small cell. It seems more reasonable to remain with a single RAN vendor approach for
outdoor small cell deployment, to ease operational management. With indoor deployment, flexibility is possible, as
solutions may range from Femto-cell to Pico-cells possibly from another vendor than the macro network.
3G small cells are still relevant options for AMEA countries where 4G has either not been launched or still not off-loaded
sufficient 3G traffic. Successful deployment in Moldova has shown that 3G small cells were a cheap alternative to macro
networks to solve localised coverage issues or boost capacity in congested dense urban areas where traffic comes
predominantly from wireless ADSL users at home.

3G micro cells for offload are recommended to be deployed with the same RNC architecture and the same
vendor as the macro layer. 3G outdoor small cells on Femto architecture have insufficient QoS to be deployed easily in
urban areas with high overlap with the macro network.

The main hurdles to deploy outdoor small cells today are access to street furniture and backhaul. Wholesale agreements
with street furniture companies such as JC Decaux, lamppost and electricity companies are being trialled to ease fast-
track deployment of small cells in wider scale. Small Cell Forum is actively working on supporting operators with
suggestions that can be adopted in individual countries to overcome deployment barriers. Unfortunately local
authorities are often decreasing EMF limits; in some places the levels are as low as 5V/m Paris, 6V/m Luxembourg or
7V/m Poland which requirements cause difficulties in deployment sites in the street furniture.
For backhaul, recommendation is to prioritise fibre, even when civil works are required, as long as pricing remains
reasonable, i.e. below 5 k€. Points of Presence (PoP) of fibres are getting more and more ubiquitous in cities where
FTTx deployments in cities have taken place, meaning short distances (up to 10 m) can be found between street
furniture and fibre PoPs. Once deployed, fibre is future-proof and delivers predictable bandwidth.
Reversely, Microwave links are complex to deploy for small cells, less future-proof and less predictable in terms of
performance. Non Line of Sight Microwave links are not recommended. Trials have shown complexity in terms of
design, and limited performance, e.g. up to 50 Mbps, not in line with 4G requirements. This is due to obstacles between
the small cells and its next hub (a macro site or another small cell), such as buildings or trees. Line of Sight (LOS) MW
links are possible, and a good option when fibre is not cost effective.

3G small cells are also a way to provide cheap coverage in white zones. A single outdoor small cell may provide
light indoor coverage within a 1 km radius. Deployment costs in a village may be minimal, by re-using ADSL and site

- 45 - Orange confidential
infrastructure from the customer (e.g. the council). This concept has been successfully demonstrated in Poland and
France with ALU Metro 3G. This should be considered more as an ‘outdoor Femto’ for white zones or rural zones.

The main priorities for delivery in countries remain Femto Cells indoors to increase voice coverage for home and
enterprise, small scale 3G & 4G micro-cells outdoors to increase coverage, Wi-Fi access points for B2B offers and small
scale marketing driven operator Wi-Fi deployment. Anticipation projects will focus on completion of 4G outdoor small cell
trials with Ericsson, Huawei, and Nokia, while ALU outdoor products will be discontinued. For indoor deployments, focus
is on the introduction of the Nokia (ex-ALU) Femto solution in multi-standard 3G/4G as a complement to the 3G only
solution for enterprise, which remains the preferred option compared to Pico-cells from the competition.

Anticipation trials are focus on LTE in the unlicensed band, with further testing with RAN vendors on their LAA (License
Assisted Access) products. Trial has been accomplished with Ericsson Pico-cell in France by end 2016. Improved
coverage and throughputs were observed vs Wi-Fi (tests on 5 GHz). In the next step a trial with LBT (Listen before Talk)
functionality is planned in Belfort in H2 2017. LTE-U Nokia with 2CA (licensed 1.8 GHz and unlicensed 5 GHz) both 20
MHz BW was tested in Poland – achieved throughputs exceeded 270 Mbps.

In details:
 Indoor Femto-cells:
• Introduce the Nokia (ex-ALU) 3G Femto solution in Morocco, Cameroon and Romania
• Migrate the management system to SAM platform to allow also 4G
• Consolidate commercial services for residential customers in France, and enterprise globally (France,
Poland, Belgium, Luxemburg and the Caribbean)
• Consider extension of Nokia Femto solution to new countries. Introduction costs on platform requires
precise business case analysis. For new countries sufficient business prospect is mandatory to
balance the substantial “entry ticket” of around 600 000 euros for a 3G solution.
 Trial LAA 5 GHz unlicensed solutions with RAN vendors, in particular Ericsson but also with Nokia (LTE-U)
 3G outdoor micro-cells for capacity with RNC architecture: Small scale deployments are recommended to
sustain capacity needs in urban areas before off-load of 3G by 4G is significant enough. Micro-cells should
rely on the same architecture as the macro network (RNC / Iub interface) using the same vendor as the
macro. All vendors (expect ALU) have such products:
• Huawei: the 3902E to be replaced by the dual-mode 3G+4G 3911E product. Note that the 3803E is not
recommended
• E///: 3G micro RBS and micro RRU, T2 since 2014
• Nokia 3G FlexiLite, T2 since 2014
• ALU: No micro-cell products exist with RNC architecture, now replaced by Nokia FlexiLite.
 3G outdoor micro-cells for coverage with either RNC or Femto architecture:
• The ALU 3G Metro Light Radio has been deployed with success in small volumes, although Last Buy
Order was in June 2016. Only for white zones.
• Ericsson, Huawei and Nokia should all support xDSL backhaul on RNC for their micro cells product,
however testing does not prove the E2E feasibility with the xDSL backhaul so far
 4G outdoor micro-cells for coverage:
• Nokia (ex-ALU) MCO Compact is replaced by FlexiZone product and it requires the NetAct OSS
• Nokia FlexiZone formal T2 validation to be scheduled
• Ericsson micro RBS 4G and mRRU 4G to be validated
• Huawei 3G+4G micro 3911E being trialled in Luxemburg, installed in Poland and to be validated with
SRAN 12 availability
 Trial street furniture solutions from several companies, at least JC Decaux for bus shelters, advertisement
panels and Phillips with Ericsson for “ZeroSite” lampposts.
 B2B Wi-Fi: Orange may respond to bids from business customers such as city councils or venues (e.g.
shopping malls, hotels, airports, train stations, stadium) to provide Wi-Fi services on behalf of these customers.
Although the Wi-Fi operator will be the city or the venue itself, Orange may negotiate the use of an “Orange
SSID” to provide direct connectivity to its own customers.
 Orange operator Wi-Fi: Wi-Fi access points are only recommended as tactical complementary solution to
3G/4G if driven by marketing needs to increase visibility of Orange as a brand. This is not a solution for
cellular traffic off-load.

11.1 Outdoor Small Cells

11.1.1 Outdoor Small Cells: the Key Drivers


The main priority to increase coverage and capacity remains macro network deployments, small cells are viewed as
complementary solutions part of the overall tool-kit. Small cell deployments are first considered in Orange group to off-
load 3G data traffic in non 4G countries and extend 4G coverage in countries with mature 4G roll-out.

Orange confidential - 46
Data Off-load
Small cells deployed on targeted hot-spots of traffic shall relieve network load when economical. High concentration of
users in specific locations, such as main streets, market squares, or even residential buildings in the case of wireless
ADSL markets would benefit from dedicated “small coverage, high capacity, low-cost” solutions.
Capacity needs today are first in 3G in non 4G countries, although overall volumes for 3G small cells within the Group
should remain low as the window of opportunity before off-load of 3G by 4G has now dropped in many countries. Needs
for 4G capacity should still take a few years, since 4G networks have been launched recently and most countries have
substantial spectrum assets to further increase capacity on macro networks first.
Coverage extension
Difficulties in site negotiation have at times exacerbated the coverage challenge in indoor locations. Growing numbers of
High Environmental Quality (HQE) buildings are already degrading the indoor coverage as the additional loss from HQE
structures is around 10 to 20 dB compared to traditional buildings. Deployment of low frequency bands such as UMTS at
900 MHz or LTE at 800 MHz remain the main recommended solutions for a homogeneous footprint of deep indoor
coverage, and highly valued indoor locations such as airports, underground or stadium may still be served by traditional
DAS systems in a multi-operator context. Residential and Enterprise Femto can alleviate local coverage issues for
customers. Beyond dedicated indoor solutions, outdoor small cells can be a cost effective solution to compensate for
poor macro coverage.

11.1.2 Outdoor Small cells: the Key Challenges


The key challenges facing small cells at this time are:
 Access to street furniture with wholesale agreements is not mature. Trials with JC Decaux in France,
Luxemburg and Spain, validates the technical feasibility but also the economic model.
 Backhaul remains a challenge, however solutions exist for small cells backhaul:
• Fibre - if no fixed backhaul is available, i.e. on street furniture, deploying a dedicated fibre is feasible in
most cities thanks to the growing availability of Points of Presence of fibres in the vicinity of the street
furniture. For distances below 10 meters, costs of the civil works are below 5 k€ and remain a better
option than investing in a Microwave solution. Fibre is future-proof in terms of bandwidth, predictable,
and lower OPEX than Microwaves.
• Line of Sight (LOS) MW links are possible, and a good option when fibre is not cost effective.
• Non Line of Sight Microwave links are not recommended. Trials have shown complexity in terms of
design, limited performance, e.g. up to 50 Mbps, not in line with 4G requirements. However the active
antenna technology with M-MIMO might bring solution based on mmWave for E2E Small Cell solutions
more robust than tested so far.

11.1.3 Outdoor Small Cell Trials


The key findings of the E2E Small Cell trials:
 10 to 20% off-load of the macro layer can be achieved with 1 small cell per macro sector. This is significant
enough to justify investment on a small cell site. Off-load varies widely depending on site location, which
requires good knowledge of traffic geo-localization.
 For off-load, 3G architecture choice is key: micro cells operated on traditional RNCs with Iub interface
are better than those on Femto Gateway with Iuh interface. Indeed, using the same RNC architecture and
same vendor as the macro network (i.e. Huawei micro 3G in a Huawei macro 3G area) gives better
performance, better off-load and is far easier to integrate in the existing network in single vendor mode.
Recommendations vendor per vendor:
As a general rule, outdoor small cells should be deployed with the same vendor as the macro layer to
maintain a homogeneous network.
 Huawei:
• 3G recommendation: was the 3902E until end 2015, now evolving to the 3911E (3G+4G product)
 3902E – T2 green light granted in 2013, but Last Buy Order end 2015
 Trialled in Orange Moldova in Q2 2013 and deployed successfully in Moldova and Senegal.
 Note that the 3803E is not recommended and has not been validated
• 4G recommendation: 3911E 3G/4G solution successfully installed and operating in commercial
networks (Poland) considered by Luxembourg and Moldova, not validated (SRAN 12 dependency).
 Ericsson:
• 3G recommendation: the micro RRU and micro RBS 3G products have been trialled successfully in
France, with T2 green light granted in 2014.
• 4G recommendation: the micro RBS 4G has been trialled successfully in Spain, with T2 green light
granted in Q1 2016. The micro RRU12 is being validated but still facing issues with antenna connector
alarms.
 Nokia:

- 47 - Orange confidential
• 3G recommendation: the FlexiLite 3G has been trialled successfully in Armenia, with T2 green light
granted in 2014.
• 4G recommendation: the FlexiZone 4G is being trialled, T2 is planned in June 2017

 Nokia/ALU:
• Following the take-over from Nokia, it has been decided that the ex-ALU portfolio for outdoor small
cells will not be maintained. The Nokia FlexiLite 3G and FlexiZone 4G product will be recommended as
soon as the Nokia OSS NetAct is in place in the country.
• In the interim period, ALU products may still be deployed in small volumes in a “PoC” mode.
• 3G recommendation: Metro Light Radio has been granted T2 and is deployable only in countries with
an ALU Femto-GW and restricted to rural areas with no macro coverage. This now only applied to
France. Recommendation is to migrate to the Nokia FlexiLite 3G as soon as possible.
• 4G recommendation: the MCO Compact 4G is only deployable for “Proof-of-Concept” and will not be
officially validated. Recommendation is to migrate to the Nokia FlexiZone 4G.

11.2 Indoor Solutions

11.2.1 Indoor Solutions: Market Overview


There are a variety of indoor market segments with different user needs, costs, capacity requirements and deployment
constraints. In addition to enhancing coverage and capacity, the possibility of introducing new locally-targeted business
services must also be considered. Each market segment requires a different business approach by Orange.

Indoor Market Segments Indoor Market Needs

11.2.2 Indoor Solutions: “Toolbox of Solutions”


There are many different indoor coverage solutions to address the various market scenarios.

Pico and Femto cell solutions


Femto-cells are available from Nokia/ALU, to boost voice coverage and 3G data services for the residential (Home – low
power ~20 mW; up to 8 users) and enterprise (~250 mW; up to 16/32 users) markets. There are currently no 4G LTE
Femto products for the residential market. A new 3G/4G access point for the enterprise (Medium/large enterprises – 50
to 125 mW – up to 32 users in 3G and 64 users in 4G) will be available in Q3 2017.
Femtocells use the ‘Iu-h’ architecture. Although the unit cost of Femto products is very small (50 € for residential 3G, 250
to 2 k€ for Entreprise 3G / 3G+4G), the architecture requires investment in a network Femto-GatewayThe cost (both
initial CAPEX and yearly OPEX) for these network elements is very high (~600 k€ CAPEX + 250 k€ yearly OPEX).
Consequently it is not economical to deploy small numbers of Femtocells; the strategy must be to use them for ‘mass’
deployment. For enterprise Femtos deployment, the breakeven point is around 1000 units, depending on the Femtos
configuration.
Femto deployments do however use cheap xDSL backhaul and the business process for deployment is simple and
relatively cheap (customer ‘plug and play’).

Orange confidential - 48
Picocells use the ‘Iu-b’ architecture with (normally) a direct connection to an RNC. However this architecture leads to a
higher unit cost and also a higher backhaul cost unless a way can be found to use DSL. They are available with 250 mW,
1 W, 5 W & 10 W depending on product.

Repeaters
Indoor coverage may be provided by simply ‘densifying’ the macro/micro outdoor network, e.g. adding a microcell close
to the building to be covered. Repeaters can also improve coverage and throughput for users who have a suitable
outdoor radio signal but do not have an Ethernet (eg. DSL) connection (so a Femto-cell is not an option). Users choosing
Mobile Broadband as an alternative to Fixed but with poor indoor radio conditions may see a great benefit in such
products. But it should be emphasised that if there is no outdoor coverage at the building location, then repeaters are not
an option !

Network Repeaters (Operator installed)


A number of medium/high power repeaters (2G, 3G and 4G) have been approved from 2 suppliers: Commscope
(formerly Andrew) and Selecom having a full portfolio of repeaters covering all frequencies. Selecom products are mainly
used in France and French speaking African countries whereas Commscope is the preferred choice of European
countries.

- 49 - Orange confidential
High power repeaters are not recommended for mass-market deployments owing to their potential to interfere with the
macro network, particularly during fault conditions. However remote OMC systems are now available and mature (e.g.
Commscope’s AIMOS platform) or Selecoms’s VisiOmc (which is also offered as a cloud service). The only drawback is
that each supplier will have its own OMC platform and this makes it impractical to have a mix of suppliers.
All repeaters that have been approved for use by Orange are included in a Common Pricelist that forms part of the joint
Orange/DT contract. This is updated when any new products are approved.
Costs of installation can be similar to the costs of installing a BS and small passive-DAS (construction, site
acquisition, etc.) although there is no associated networking cost; this is a major factor against the installation of
operator-installed repeaters in most countries.

Personal (“Home”) Repeaters


Low-power “Personal”/“Home-repeaters” are considered as Customer Premises Equipment and “plug and play”, i.e. the
user can install it without operator intervention and without the need for a strong technical background. These products
are also closely followed by Orange competitors to improve QoS; in fact the drive is from marketing, to reduce churn.
The cost of deploying these “home-repeaters” is low; the hardware only needs to be sourced and distributed where it is
needed (no installation costs or planning on Orange side). The specifications for a stable, mature and fail-safe “home-
repeater” are quite strict as the introduction of this product should never risk compromising the established macro
network. A number of products from a supplier called Nextivity; Cel-Fi Pro home repeater, have proved to be the only
product currently to meet this need. The Cell-FiPro was granted formal approval at the end of 2014 and this product
covers up to 4 frequency bands simultaneously. Different variants exist : 900/1800/2100/2600; 800/900/2100/2600 ,
etc.…; (contact ALD SkC).

2G Home Repeaters
In terms of 2G Home Repeaters; mainly targeted at 900 MHz voice and data support, no viable option has been
proposed to date although Orange still continues to monitor the market for opportunities. There are 2 products that have
been assessed as suitable for local deployment purposes. They are the VB 900 product from Coiler and the R17PRO
from Remotek. However, it should be noted that these are not “plug and play” products and that some engineering
knowledge is required to correctly install and commission them; they need an external coverage antenna and the filter
ranges need to be correctly chosen. Indeed the Remotek product only offers filters that are tuned in the factory and
cannot be changed in the field. They are only advised to be deployed where 2 or 3 band-segments are needed, as in this
case the higher-power Commscope solution is too expensive.

Orange confidential - 50
Distributed Systems (Passive DAS / Active DAS / LampSite / DOT)
Distributed solutions include:
 Passive DAS
 Active DAS (of which there are many varieties of products)
 The ‘distributed radio system’ concept (BBU + Distribution Hub + Remote Radio Units); examples of this are
the Huawei Lampsite and Ericsson RadioDOT.

Passive DAS
Traditionally, most indoor deployments have used a passive “Distributed Antenna System” (DAS). Passive-DAS consists
of a central BTS/Node-B/Repeater connected to a network of co-axial cable, splitters and where necessary filters,
feeding many small antennas, or occasionally radiating cable (“leaky feeder”), in order to distribute the RF signal to all
the required areas of the building. Design Guidelines exist for Passive-DAS, see the OLN Sharepoint.
Passive-DAS size is limited by the cable losses and so only medium-sized buildings can be covered with this
technology. The exact size limitations depend on the detailed design considerations.
The high RF power levels used in the input combining network and the fact that RF distribution components & feeder
may be easily damaged means that there is a risk of generating high levels of passive intermodulation (PIM). This is
another reason why passive DAS is less favourable for use in a multiple wide-band, multi-operator, multi-technology
environment.
Overall, DAS can be quite wasteful of energy & so is not a “green” solution.

Active DAS
Active-DAS also uses a centrally-placed BTS/Node-B, but the distribution system uses active components and either
optical fibre or in some cases CAT5/6/6A ‘Ethernet’ cable between the hub and remote units. Extremely large and
complex, Active-DAS installations are possible with multiple operators, bands and technologies all configured and
managed remotely via the DAS-vendor’s OMC. Active DAS may be analogue or digital:
 Analogue active DAS. In this case the RF signal from the BS is modulated onto either an optical carrier (fibre-DAS)
or an IF (100 – 400MHz) carrier (‘Ethernet’-DAS) and re-transmitted in analogue form to the remote antenna units
(RAU) (& vice versa).
 Digital active DAS. In this case the RF signal from the BS is digitally encoded for transmission between the BS and
the RAU. This system allows very long distances to be spanned and a large number of remote units. Transmission
interfaces may or may not comply with CPRI standards depending on the product.
Secondary distribution hubs may be used to split the signals further; antennas may be integrated with the RF-heads; RF-
heads may need a local power-supply or this may be fed down the same cable as the optical fibre.

- 51 - Orange confidential
Specialised antennas may be required for indoor coverage and especially for stadium coverage applications where the
precise coverage pattern can be very important (contact ALD SkC for details).
Most equipment vendors offer a turnkey design and installation service to assist smaller operators who may not have
sufficient expertise.
All this means that a DAS can be very expensive and the costs need to be carefully assessed. However they can
accommodate multiple operators which is sometimes a requirement of the building owner or event organiser.
Many Active-DAS systems allow only one technology and a single band to be distributed; if multiple bands, multiple
technologies or multiple operators are required then each system needs separate equipment installed ‘in parallel’ which
multiplies the overall cost by the number of installed systems. However there are some wide-band systems which are
capable of carrying more than one band and/or technology (e.g. UMTS + LTE) via the same equipment. These systems
may have relatively lower power limits per channel in order to preserve amplifier linearity. They may also be initially
more expensive than single band systems but may be more cost effective in the ‘long term’ where multiple bands
or technologies will be deployed as costly upgrades can be avoided.
Fibre active-DAS is ideally suited to very large multi-operator, multi-technology installations, such as football stadia. At
the POI there is normally a high degree of flexibility with vendors offering either RF or digital switching of signals so
that each BS/sector/technology can be connected to specific antennas/sectors within the building or stadium. So different
operators with different capacity and coverage needs can share the same system.
Also, for active-DAS, each RF Tx signal is individually attenuated before being connected to the active DAS hub and so
unlike the passive DAS solution, the RF power levels within an active DAS system are quite low and kept separate
from the other RF signals.
Active DAS RFQ
During the early part of 2016, Orange were made aware of a RFQ that had been launched only by DT. As Orange were
looking to deploy some DAS projects in 2017 especially in the Paris metro area, it was therefore requested that this
became a joint DT/Orange RFQ.
The ADAS RFQ was concluded in August 2016 and the main recommendations are as follows:-
1. For medium and low power solutions, the choice of supplier product will be either Commscope (ION-U or ION-
E) or Kathrein K-BOW.
2. For high power deployments, the choice was either Commscope (ION-U) or Huawei (Single DAS 2.0).
As of April 2017, Kathrein has been chosen for some enterprise solutions in France while Huawei was chosen as the
main supplier for the Paris Metro project.
These products have been incorporated into global or local agreements as required. Commscope: signature of an
amendment to the CSC 05CG126 to update the pricelist; Huawei: signature of a dedicated amendment to the RAN CSC
09CG252 and Kathrein, signature of a dedicated TSA to the existing BuyIN FPC.

An overall technical summary of the architecture of each solution is given below.

Commscope
ION-E series - Unified wireless infrastructure delivers a more flexible network:
Universal access point (UAP) data and power through Category 6A twisted pair cabling. Supports gigabit Ethernet for Wi-
Fi, IP cameras or other devices in addition to wireless over a common cable.
Central Area node (CAN). Server level control and primary signal distribution. 2U and 4U subrack options available.
Transport expansion node (TEN). The secondary distribution point connected to a CAN using multi-mode and single-
mode fibre.
ION-E point of interface (e-POI) attenuates high power RF signals from their source to the CAN across all frequencies.

Orange confidential - 52
ION-U series
Auto discovery of Network components.
Intra system cabling assistance
Auto setup, levelling and commissioning
Integrated measurement of PIM and external interference
Full visibility of all network resources
Remote re-sectoring
Service level alarms
Continuous self-testing
Detailed uplink/downlink analysis

Huawei Single DAS 2.0

- 53 - Orange confidential
Kathrein K-BOW

Distributed Radio Systems


Two products are currently being evaluated by Orange Group. These are:
 Huawei Lampsite
 Ericsson RadioDOT
They are similar in utilising a central BBU, distribution hub and remote antenna units. However the implementation is
different. This is illustrated in the diagrams.
The products are not suitable for multi-operator use and initial analysis indicates that the installation costs may be
similar to DAS. However there may be other benefits from central resource management.
Trials have been completed successfully for both products but formal validations remain in stand-by due to
limited number of requests from countries.

Orange confidential - 54
Indoor solutions

Figure 21 – Product vs market segments for indoor solutions

 Home
– 3G Femto-cells are the only affordable active solutions for home users. Cost of a 3G Femto is
less than 50€. Although the unitary price is low, entering the residential market requires a significant
overall investment in terms of Femto platforms and Information System. To date, Orange France and
Orange Belgium have launched such service. Nokia (ex-ALU) is the sole supplier for Femto-cells the
group.
– 4G home Femto-cells are not available (Nokia HomeCell 4G). Interest for home users is being
assessed, as a complement to Wi-Fi for Data services, and VoLTE continuity of service.
– Voice over Wi-Fi is seen as a viable alternative to Femto-cell at home, once launched. However this
will only serve a restricted number of devices supporting the feature initially.

 Enterprise (small / medium / large)


– The main request is “basic voice / data connectivity at reasonable cost” for employees, not
necessarily visitors.
– 3G remains a priority to serve “all users” and ensure CSFB for non-VoLTE 4G devices
– 4G is requested to boost data performance, ensure VoLTE support but also ensure indoor
connectivity to support 4G PMR services.
– Evolution to LAA / LWA may be envisaged to maximise use of unlicensed 5 GHz spectrum
– Mono-operator solutions are of interest for Orange to develop operator stickiness
BUT many companies request multi-operator (or operator agnostic) solutions to serve users from all operators,
or at times to get flexibility for swapping operators in the future through the same indoor infrastructure.

Recommended solutions:
 For Small Offices requiring only one small cell
– Femto or Pico-cells tech-eco analysis conclusions:
– For country volumes > 1000 small cells (accurate breakeven point depending on the Femtos
configuration), ALU/Nokia Femto cells (including Femto GW) are cheaper than Pico-cells.
Solution is commercial today.
– For country volumes < 1000 small cells, pico-cells are more affordable. However uncertainties
remain for connectivity to xDSL / FTTx affordable backhaul. Impact on transport architecture
needs local studies, as SeGW, DHCP relays, IEEE1588v2 synchronisation and IP routing of
RNC are required. Such configurations are being trialled (Luxemburg, Morocco, Romania…).
– Eventually backhaul availability may be a key criteria when choosing a Femto or Pico
solution:
 xDSL backhaul is more adapted to Femto-cells
 Leased Lines are more suited to Pico-cells
– Voice over Wi-Fi is an alternative in a similar fashion as for home users

- 55 - Orange confidential
– Repeaters if no active solution is available in country or customer backhaul not sufficient

 For Medium / Large Offices


– Femto / Pico-cells
 Up to 10 small cells may be deployed to serve a building. Orange France has launched an
offer called “Orange Confort” which is charged to the customer, as a package with up to 10
Nokia Femto-cells. Beyond 10 units, optimization becomes more complicated due to lack of
Scrambling Codes or PCI, neighbour cells declarations, and interferences between cells.
Besides, for larger buildings companies are less inclined to pay for a single operator solution
and prefer to have a multi-operator solution.
– DAS Systems
 Passive DAS Systems for smaller buildings
 Active DAS Systems for larger buildings or long distances for cabling
– Repeaters if customer backhaul is not sufficient

 Public Venues (small / large)


Fundamental requirements:
– Multi-operator is a must have
– Voice & Data are of interest, including 3G, 4G and Wi-Fi
Recommended solutions:
– Passive & Active DAS Systems
– Wi-Fi may be served by a dedicated Wi-Fi system

11.3 Small Cells for 4G advanced and 5G (Unlicensed bands LTE-U, LAA)
The ecosystem for LAA is changing and expected availability of Snapdragon 835 Qualcomm chipset mid 2017 might
have significant influence on Small Cell product portfolio. Nokia, Huawei and Ericsson claims to be compliant with
Release 13 LAA standards including dynamic channel selection, LBT, SDL.

Figure 22 – Nokia Small Cells supporting LAA solutions (now and EOY)

LAA is under tests and can provide relatively low cost additional feature to boost capacity of deployed small cells.
However LAA is UE dependent and penetration of mobiles supporting LAA will be critical for deployment decisions.
Solution available from Nokia is expected end of 2017 should allow 4CC: 1 licensed 20 MHz, 1.8 GHz and up to
3x20 MHz unlicensed 5 GHz. Solution attractiveness strongly depends on entire ecosystem development – Qualcomm
chipset availability and mobile penetration supporting LAA. Expected time frame for commercial LAA launch is end of
2017 and UE availability in the next 6 to 12 months.

MulteFire is a new LTE-based technology but with one big difference — it is deployed in unlicensed spectrum and does
not require any licensed spectrum. This means that more entities can deploy MulteFire and can create threat for
incumbent operators. The MulteFire Alliance push the concept to be suitable for any spectrum such as the global 5 GHz
unlicensed spectrum and 3.5 GHz shared spectrum in the U.S. the US market could significantly speed up evolving
ecosystem. Question stays open will it be LTE technology or rather 5G New Radio.
Works on standardization started on fully autonomous (no operator license frequency anchor) unlicensed 5G.

11.4 Wi-Fi
Wi-Fi access points have been largely deployed for many years with various contexts and various successes, primarily
addressing the laptop market. Making use of Wi-Fi to complement 3G and 4G networks is a more recent approach, made
possible by the wide availability and use of Wi-Fi within Smartphones.

Orange confidential - 56
Operator Wi-Fi
Operator Wi-Fi should not be confused with Home or Community Wi-Fi:
 Home Wi-Fi: connectivity to the home ISP box (e.g. Livebox). Widely used today by many types of devices:
smartphones, tablets, PCs, set top box for video services, IoT devices…
 Community Wi-Fi: connectivity to ISP boxes from other members of the Community, with limited bandwidth
allocation (typically 1 Mbps allowed to the community) and variable quality of experience, i.e. acceptable
indoors in vicinity of the ISP box, challenging outdoors due to low coverage and high interference.
 Operator Wi-Fi: connectivity to access points deployed and managed by the operator. Enhanced quality of
service can be achieved through high bandwidth backhaul, advanced radio features and seamless
authentication features (SIM authentication, PassPoint)
Defining the right business model of Operator Wi-Fi is a major cornerstone in enabling wider deployments of Wi-Fi
access points. Three business models are foreseen in order to create value:
 Premium Wi-Fi, based on Wi-Fi operator sites. This differs from Home or Community Wi-Fi, which are
based on ISP boxes (Livebox). It will rely on the same premium business model as H+ or 4G, valorisation will
be included in the offer or as an option. Volumes of traffic or local abundance will depend on the size of the
infrastructure.
 Roaming Wi-Fi: Business model with two wholesale possibilities, either a stand-alone by minutes or coupled to
cellular contract (MBytes). Retail bundle (3G/4G + Wi-Fi) is becoming a must as now Europe is roaming free for
cellular data, it should be applied as well to Wi-Fi.
 Business Wi-Fi: Traditional business model. WLAN Infrastructure cost will be paid by the business customers
(e.g. Hotel, Airport,…). Orange may then give B2B customers a discount on the infrastructure when Orange
customers can as well use this private infrastructure or when Orange can install small cells on site. Orange
customers may then by granted a “Premium Wi-Fi” experience through dedicated backhaul and SIM
authentication.

Operator Wi-Fi or “Carrier-grade Wi-Fi” is not seen as an Off-load solution. Wi-Fi is a very effective solution at Home for
residential customers to move away from Mobile Broadband with an “all you can eat” home usage not limited by the
Mobile Broadband monthly caps.
In terms of performance Wi-Fi is best in indoors locations. Typical use cases will include shopping malls, hotels,
universities, train stations, restaurants,…
The coverage in outdoor environments is limited due to the lower power of Wi-Fi, especially in UL. As such it is not
recommended to deploy Outdoor Wi-Fi access points to cover indoor buildings.
2 vendors have been selected by the Group to provide low cost Wi-Fi access points with both indoor and
outdoor products: Cisco and Ruckus.

In term of technical solutions, the following trends have emerged recently:


- since beginning of 2015, all the products offer 802.11ac technology.
- controllers are nowadays either physical or virtualized. The virtualized version allows for example to sell Wi-Fi
as a service offer to the B2B market
- unleashed Access Point : for small deployment (for example a hotel with 5 to 20 Access Point), the controller
will be replaced by an Access Point that will play the role of controller.
- 802.11ac Access Points have been introduced in the Orange portfolio solutions, whereas virtualized controllers
and unleashed Access Points in under validation. More information available by contacting the Wi-Fi skill center

Voice over Wi-Fi


“Voice over Wi-Fi” or “Wi-Fi Calling” is an alternative to Femto or small cells for customers to enhance their indoor
coverage. It is foreseen to be adopted more easily for SOHO and residential first. Its success will greatly depend on its
QoS and ease of use.
VoWiFi requires some End to End enhancements (UE, Network, IS, IMS platform) and the specific introduction of a
specific node called ePDG. Network selection between Wi-Fi and 2G/3G/4G remains a challenge as devices have
relative autonomy to select Wi-Fi over 3GPP without operator control, and may disrupt overall voice experience when
connected to poor quality Wi-Fi. Wi-Fi connection managers are therefore evolving through additional user controls such
as network preferences (Wi-Fi or Cellular) for the Voice service, independently of the Data service. Within the Enterprise
it is expected that if VoWiFi becomes a reality, it will still take a few years to materialise. Ensuring a seamless voice
service over an existing WLAN will be challenging, extra access points will generally be required to reach the signal
strength and quality required for real-time voice, and costly optimisation is to be expected. Besides a number of
companies not in favour of opening access of WLAN to Smartphones will still require 3GPP solutions.

Small cells: Group Recommendation 11—1


Outdoor Small Cells
 3G/4G outdoor small cells are seen as solutions for coverage extension and future densification of the
network to cope with future 4G congestion. Longer-term, small cells should be a key pillar of 5G

- 57 - Orange confidential
deployments.
Outdoor 3G micro-cells used in urban areas should be deployed using a traditional RNC / Iub
architecture in a mono-vendor environment, i.e. with the same vendor as the macro network. Several
products have been validated by the Group, Huawei micro 3902E, E/// micro RBS and Nokia FlexiLite
3G. There are trial implementations of 3911E 3G/4G, not validated (SRAN 12 dependency).
 Outdoor 3G micro-cells for coverage, i.e. deployable in white zones, can be served either by RNC
based solutions as above, or also rely on a Femto / Iuh architecture. The ALU Metro Light Radio 3G is
end of support
 Outdoor 4G small cells should be deployed using the same vendor as macro cells. Ericsson
mRBS and been validated, and ALU MCO Compact has been replaced by Nokia FlexiZone.
 Access to street furniture remains a challenge in Western Europe. Trials of solutions from several
companies are on-going, with JC Decaux for bus shelters and advertisement panels and Phillips with
Ericsson for “ZeroSite” lampposts.
 Small Cell backhaul should rely first on fixed solutions when available, or when civil works for the last
tens of meters is affordable. Microwave solutions are an alternative, but only in “Line-of-Sight”. “Non
Line-of-Sight” products are not recommended due to complexity of design a limited performance.
Indoor solutions
 DAS remain the best solution for multi-operator deployment in public buildings, (e.g. train station,
stadium). However these can be costly deployments and so a full assessment of the relative costs &
benefits for each system must be considered at the outset. An assessment of €/M² for different
scenarios from the recent ADAS RFP has been done and can be provided on request.
 For new installations, an ADAS (Active DAS) should be deployed if the number of bands is 3 or
higher ( pim avoidance and better design for LTE)
 Active DAS is recommended for LTE with co-located MIMO 2x2 from day one for double capacity vs.
SISO.
 Repeaters: for users without fixed broadband (like ADSL) and no Femto 3G offer in country,
alternative solutions such as 3G home-repeaters can be considered. A low-cost ‘home’ repeater is
available for the “SOHO” market (Nextivity Cel-Fi). This is recommended where a good external signal
exists and additional capacity is not required.
 3G only and 3G/4G Femto cells are recommended as a Voice coverage solution for residential
users and as a Voice and Data coverage solution for Enterprise with up to 10 units per building.
Business case to justify investment on Femto Gateway architecture is challenging for small market
size. Nokia/ALU is the sole supplier selected by the Group.
 4G Femto-cells may be evaluated upon country request. 3G/4G Pico-cells with traditional RNC/EPC
architecture are recommended for small scale deployments when Femto Gateways are not affordable,
or for high-end deployments with leased line backhaul.
 Distributed Radio Systems (Huawei Lampsite and E/// Radio DOT) have been trialled successfully
in Poland and France but have not been validated so far due to the limited deployments planned.
These solutions being mono-operator, DAS systems with native multi-operator support are generally
preferred.
Licensed Assisted Access
LAA is under tests feature to boost capacity of deployed small cells with the anchor LTE frequency
and unlicensed 5 GHz. Expected throughputs up to 600Mbps. Solution UE dependent. Expected
time frame for commercial LAA launch end of 2017. Works on standardization started on fully
autonomous (no operator license frequency anchor) unlicensed 5G.
Wi-Fi
 Wi-Fi is a mandatory/must have technology in Home environment (Livebox, Set Top Box, …)
 Except Home, Wi-Fi roll-out is primarily driven by B2B offers (e.g. Airport, hotels), where Orange may
respond to bids launched by venues when an effective business model is found. Stadiums are part of
this business.
 Operator Wi-Fi deployment should be driven mainly by Marketing interest in country. This is not an off-
load solution as usage remains minimal compared to cellular in nomadism.
 Potential virtualization of the controllers has to be taken into account for any new project.
 Cisco and Ruckus are the selected suppliers for low cost Wi-Fi Access Points.
Contributors: Colin Fraser, Roman Łapszow, Bernard Missir, Stefan Wendt
Contacts:
 Outdoor Small Cells: Roman Łapszow
 Indoor, DAS, Repeaters: Bernard Missir, Colin Fraser
 LAA: Roman Łapszow
 Wi-Fi: Stefan Wendt

Orange confidential - 58
12. Antenna Systems
12.1 Antenna Line Devices
Antennas Line Devices“ (ALD) are defined as all the ancillaries necessary to enable RF signal reception and emission
on a radio site to cover and service the end-user. ALDs are the closest elements and most sensitive link to the customer
in the radio-chain.

Figure 23 – Antenna Line Devices

The product groups of ALDs can be divided into the following different product families:
 Passive Basestation Antennas
o (Extended) Legacy products, designed for 800 / 900 / 1800 / 2100 bands - CSC families 1.x
o Ultra-Wideband products, covering frequencies up to 2690 MHz - CSC families 2.x
o Additional families to cover LTE 700 and SDL 1400 – CSC families 2.9.x, 2.10.x, 2.11.x, and 2.12.x
 Remote Electrical Tilt (RET), to control the electrical downtilt of the antenna remotely via OSS - CSC families 3.x
 Tri-sector antennas and self-supporting structures - CSC families 6.x and 7.x
 Mast Head Amplifiers (MHA) or Tower Mounted Amplifiers (TMA), to improve the UL sensitivity - CSC families 8.x
 Multiplexors (MUX), for combining or splitting multiple bands onto common feeder cables with an emphasis on DC
auto-sense and auto-bypass functionality as well as frequencies now down to 690 MHz - CSC families 9.x
 RF coaxial feeders, connectors and flexible jumper cables – in specific CSCs
 Repeaters and indoor coverage systems (Distributed Antenna Systems : DAS) – see also chapter 12.2

For Orange Group and Buyin, there is a strict technical shortlist with all validated ALDs in the CSCs, which is updated
every semester with new products. This list is mandatory; in other words all Orange affiliates are obliged to select from
this shortlist their ALD equipment (whether running in “RAN turnkey” or direct sourcing). The document is distributed to
all identified technical and purchasing persons within the Orange affiliates. Detailed data-packs per supplier are
maintained and also provided on a regular basis to the affiliates.
The actual suppliers for ALDs in Orange Group are: Amphenol-Jaybeam, RFS, Kathrein, Huawei, Radiodesign,
Commscope. The selected suppliers for Antennas in Orange Group are : Amphenol-Jaybeam, Kathrein and
Huawei. As the global ALD market is rapidly changing, we strongly recommend to get in contact to obtain the latest
version if you have any doubt for your deployment strategy.

- 59 - Orange confidential
Evolution or revolution ?
Over the last few years, the ALD eco-system has undergone much evolution; from mainly dualband antennas in 2010 to
tripleband in 2011, quadband from early 2012, triggered by first LTE2600 roll-outs. Since 2013, there has been a real
exponential growth of complexity and band-permutations. Already in 2017 we will see, “heptaband” antennas with 7
different physical arrays (14 connectors and all these antenna inputs allow independent RF connectivity to the base
station for separate beam and electrical downtilt control). Side-By-Side quadband or hexaband antennas are mostly
future-proof for MIMO 4x4. From 2017 onwards, Orange is driving the suppliers’ roadmaps to extend their designs into
“Mega-WideBand” arrays, allowing stable and consistent radiation patterns over a complete frequency range 1400 –
2700 MHz. On the other end of the spectrum, the industry is designing for “Ultra-Lowband” arrays, which are able to
cover all bands between 698 – 960 MHz. Note that antennas as a transducer are “technology-agnostic”, i.e. modulation
does not matter for the wave propagation properties.

2017 Penta / Hepta 2018 : Octaband

Figure 24 – Antenna complexity increased exponentially over the last years

Laws of physics always apply !


We must never forget that antennas are dominated by the laws of physics. Radiating arrays need their physical
dimensions in function of the given frequency wavelength to achieve the required performance for gain, side lobe
suppression, beam width, cross-polar discrimination, port isolation, etc.…
Therefore, antennas cannot be reduced in physical dimensions without making compromises on RF performance.
All Orange requirements are specified in a master document “Base station antenna specifications” (document available on
request).
Antennas for 3-sector-site deployments have a typical horizontal beam width (-3dB) around 65° +/- 7°. Antenna physical
dimensions are dictated by the lowest frequency band(s), in most cases the 700-900MHz band is mainly responsible for
the antenna radome width. The antenna gain is fully related to the array length; arrays in the “10-lambda length” provide
about 18dBi gain. Whereas Lowband arrays require an internal reflector of roughly 25cm width, High band arrays require
reflector widths of about 15 cm. Multi-band antenna platforms have incorporated sophisticated radiating element coupling
and building techniques for “Stacking” or “Side-by-Side” (TWIN) configurations inside their radome; strong patents apply
per specific supplier. Antennas with LowBand “Side-By-Side” array configurations are highly recommended due
to the higher gain possibilities and being more future-proof, avoiding any internal filters on the low-band ports.

Popular antenna models, grouped per radome size:


1: Ultra-Wideband X-polar panel 1695-2700 MHz ( 2 connectors )
2: Side-By-Side 2X-polar 1695-2700 MHz ( 4 connectors )
3: Side-By-Side 3X-polar 1695-2700 MHz ( 6 connectors )
4: Legacy tripleband 790-960 // 1710-2170 // 1710-2170 MHz stacked highbands ( 6 connectors )
5: High gain tripleband 698-960 MHz with 2x Ultra-Wideband 1695-2700 MHz Side-By-Side ( 6 connectors )
5: Quadruple band 698-960 MHz with 3x Ultra-Wideband 1695-2700 MHz Side-By-Side ( 8 connectors )
6: Filtered Lowband 790-862 // 880-960 // 1710-2690 // 1710 -2690 MHz ( 8 connectors )
7: Hexa Ultra-Wideband 698-862 // 880-960 // 4 x (1710-2690 MHz) - with 12 connectors
8: Quad Side-By-Side 2X-polar 698-960 MHz + 2X-polar 1695-2700 MHz ( 8 connectors )
8: Hexa Side-By-Side 2X-polar 698-960 MHz + 4X-polar 1695-2700 MHz ( 12 connectors )
8: Hepta Side-By-Side Filtered Lowband 690-803 / 880-960 // 698-960 MHz + 4XP 1425-2700 MHz (14 connectors )
9: Octaband Side-By-Side Filtered Lowband 690-862 / 880-960 // 698-960 MHz + 4XP 1695-2700 MHz
+ XP 1425-2700 MHz (16 connectors )

Orange confidential - 60
Group 1 2 3 4 5 6 7 8 9
Figure 25 – Typical base station antenna dimensions

Deployment of only 1 antenna per sector is not recommended:


As a matter of fact, multi-band antennas are deployed to reduce clutter on site, obviously by having more technologies
all inside the same radome (thus 1 antenna per sector). This is indeed attractive for a quick Time-To-Market and
reduced civil works on site, however it also comes with important disadvantages:
 a higher cost per antenna (average cost 1600 € for hardware only)
 with any antenna failure, it must be completely replaced impacting all services
 more complex RET nomenclature and band identification for I&C and OAM/OSS
 a higher weight and wind-load due to increased size of the radome ( probably need a crane )
 less future-proofing for 5G which will most likely require a new separate antenna system
 difficult or impossible FDD/TDD co-habitation (isolation between systems to be respected, contact ALD SkC)

When using multi-band antennas, it is recommended to anticipate the near-future (< 4 years) deployment upgrades and
to select antennas with sufficient input ports, appropriate for the upcoming new band-planning (700? , 800? , 2600? ,
4RXDiv ?). Unused ports are simply water-proofed and cause no harm to the overall system performance.

To take into account being ready for MIMO 4XN where sites costs can be justified, it is recommended that only
antennas with “side by side arrays” are deployed, as stacked arrays such as legacy antennas (family 1.7.1)
cannot be used for MIMO 4x4 !

We consider the tripleband legacy antenna (CSC family 1.7.1 with 6 connectors – size group 4) as the ‘basic’
antenna for all green-field deployment, when only UMTS2100 band is involved. When LTE1800 and/or LTE2600 is
used or anticipated, then the minimum antenna model considered is a Side-By-Side tripleband (CSC family 2.8.1
with 6 connectors – size group 5).

Deploying 2 antennas per sector is highly recommended:


Orange affiliates who are pushing to obtain building permits for installing 2 physical antennas per sector
have clear advantages:
 a lower complexity on site and cabling error avoidance
 much more cost-efficient, with a lower cost per antenna
 a higher degree of technical freedom for site-(re)engineering and upgrades
 better antenna performance per individual band
 much easier for MIMO 4x4 (more antenna choices)
 more straight-forward RET nomenclature and band identification for I&C and OAM/OSS
 FDD/TDD co-habitation: for TDD is always recommended to have a separate antenna for higher order MIMO (8T8R)
 easier PIM avoidance by design (see further note)

- 61 - Orange confidential
For scenarios with 2 antennas per sector, it is advised to separate 4G-LTE and GSM900 in another physical radome,
especially when LTE2600 is deployed (to allow the possibility of MIMO 4x4). When defining an antenna site upgrade,
always keep in mind that for Carrier Aggregation between the (future) bands, the antennas should have the same
azimuth direction at boresight for the aggregated frequency bands.

PIM or not to PIM ?


Passive Inter Modulation is a physical phenomenon which is created when mixing 2 or more high power RF carriers, in
whatever form or modulation type. The outcome is a set of strong interferers with multiple orders, we speak of PIM3,
PIM5, PIM7, PIM9, … The effects can be really catastrophic when these interferers (mostly wideband) fall into Orange
used uplink receiving bands, going from increased intermittent bad mobile voice quality to noticeable lower data rates.
PIM is strongly dependent on the transmitted RF power (instantaneous traffic) and used higher modulation (LTE and
UMTS are more sensitive to create PIM than GSM only). Although PIM is heavily influenced by the build quality of the
network chain and the ageing effect of components (for example: old feeders reuse or multiplexing); in some cases, it is
even impossible to build a reliable site without having impact of PIM. The site in this case is performing “sub-optimal”,
which is clearly a waste of CAPEX and prolonging Return-On-Investment.

A few examples:
1) To ensure PIM is avoided when deploying 700, 800 and 900 low bands in a single antenna radome, the 700 and 800
bands should always have a dedicated port and should never be combined. Example antennas for this kind of
application include antenna families 2.9.6 and 2.11.x. This type of antenna also includes the new SDL 1400
frequency in a large wideband array for possible future use.
2) PIM is very commonly created when mixing 2 low bands with 1 or more high-bands. Make no mistake by thinking
these frequencies are far apart and will not have an effect, on the contrary ! For any multiplexing of 2 or more bands
on a feeder or antenna port, it is highly recommended that a PIM simulation is run with your specific band allocations
using the Group approved tool from Quintel. This can be bought via a (cheap) negotiated Group licence or a request
made to the ALD skill centre.

Figure 26 – Output of a simulation showing PIM impact on 900/1800/2100 bands due to LTE800 activation

Additional guidelines on antenna selection:


Basestation antennas with high gain (>17dBi) are strongly recommended for deep indoor planning because of their
narrow vertical beamwidth, allowing better coverage and a lower interference due to a good side lobe suppression. For
these reasons, antennas with physical heights shorter than 2meter are always strongly discouraged for LTE800
deployments.
For LTE800 and GSM/UMTS900, it is also recommended to have separate electrical downtilt, allowing
optimisation when the LTE800 network is loaded. Use Twin “Side-By-Side” or filtered lowband 800//900
antennas when deploying both technologies in the same sector. Twin antennas “Side-By-Side” are preferred for
future-proofing and flexibility but require more attention for the building permit due to its larger radome width (group 8/9).

Orange confidential - 62
In case LTE800 wants to be deployed by reusing the existing Extended Low-band antennas (790-960 MHz), and thus the
electrical down tilt of LTE800 layer is identical to the GSM900 layer, the planning has to be very carefully done as this will
lead to higher interferences in LTE800 band, thus capacity and throughput loss on the LTE800 layer. In some cases, it
will lead to uncontrollable coverage, degradation of a total LTE800 cluster or missing CSFB (CS Fall Back) towards the
GU900 layer at cell-edge !
A simple rule is advised: make sure the down tilt of an LTE800 sector is never smaller than 4° as the vertical beam
width at 800 MHz is very wide (-3dB = V10° for a 2.7meter height antenna). Please contact the ALD Skill centre for more
technical details.
Multi-band passive base station antennas are still covering wide-scale deployments and we do not see the ‘real’
Adaptative Active Antenna Systems coming to site before 2020-2022 (see chapter 13.3). Variations with embedded or
integrated radios and hybrid antennas will come to the market but will, in short term, not eliminate passive antennas as a
building-block for site-engineering on the (high-mobility coverage) macro network.

Remote Electrical Tilt (RET):


The electrical downtilt of the vertical beam can be controlled due to phase-shifting technology inside the antenna.
Historically, this sector tilting was done with manual adjustable tilt (MET) on site whereas today almost all antennas have
RET incorporated, allowing to do this remotely from the OSS. Antennas with “embedded RET” are becoming the de-facto
standard and we anticipate to see “MET” disappear from the European market within the next few years. Remote
Electrical tilt is controlled via the 3GPP/AISG IuAnt protocol which is implemented in (e)NodeB software. Therefore, we
strongly recommend to use RET antennas when the LTE layer is deployed. All base station manufacturers also
allow the AISG RET control feature from 3G NodeB OSS. However, it is important to notice that the OSS implementation
of AISG control is completely different per vendor and not all vendors have the same maturity for RET features. Rolling-
out of RET on the network requires a strict naming procedure and inventory management of serial numbers
during I&C on site ! RET implementation for GSM900 is not recommended, unless UMTS900 or LTE800 is considered.

Mast Head Amplifiers (MHA/TMA) and Multiplexors (MUX):


Both ALD products are being boosted with the LTE upgrade strategies, mainly due to the attractive low-cost and fast
upgrade path when re-using the existing site infrastructure. MHA and MUX products have undergone a tremendous
technology step and a positive price–erosion drives the business case. Re-using the existing RF feeders for multiple
bands are possible by using RF multiplexers. Multiplexors are the preferred option for combining 800+900 bands and
1800+2100 bands when a quick Time-To-Market is required for LTE roll-out. For other multiple frequency band
combining, we advise to always perform a thorough analysis for the risk of Passive Inter-Modulation.
Since 2013, dualband MHAs have been developed to make tower-top installations simpler, lighter, cleaner and with less
clutter. Important to mention is that dualband MHAs are dominantly used on the Orange networks in combination with
“DC Autosense Bypass” Multiplexors. It is strongly recommended to use these “DC Auto-sense Bypass MUX” when
using Mast Head Amplifiers in the network, in combination with RET deployment. Both ALDs are contributing to most
cost-efficient LTE upgrade evolution, these must be considered as a feasible option.

RF coaxial feeder and flexible jumper cables:


Only the cable manufacturers “RFS” and “Eupen” are validated for network build today. The list is mandatory; in other
words all Orange affiliates are obliged to choose cables, connectors and jumpers from these 2 suppliers (whether
running in “RAN turnkey” or direct sourcing).
A lot of elements were taken into account for this vendor selection: of course technical performance and compliancy,
long-term reliability, supplier competence, field-support and training, local stock (logistics) and services in the countries.
We also strongly recommend to use the suppliers’ installation tools for cable preparation, as inappropriate
connector installation practices is the main source of site performance issues like VSWR, water-ingress, bad
Passive Inter-Modulation (PIM), VSWR, corrosion, etc...

For flexible 1/2” jumpers, only ‘factory moulded jumpers’ are allowed in the field ; it is strictly forbidden for the
installers/riggers to make ½” Hi-Flex jumpers on site to length because of severe risk of PIM. Different factory
lengths are available in the CSC to fit your application. Moreover, besides the big risk, the RF jumper cost is so low that
manual connector installation on jumpers is at least 2-3 times more expensive in material cost and labour manpower !

DIN 7/16” RF connectors must always be torqued with a spanner of 25-30 Nm. If not, risk of Passive Inter-Modulation is
very high and may degrade your network quality considerably. For more details on PIM and scenarios and solid risk
assessment on frequency band combinations, please get in contact with the ALD Skill Center.

The NEW 4.3-10 connectors for all Antenna Line Device products
Advantages:
 Higher accuracy of passive intermodulation (PIM) compared to 7-16 connector
 40% smaller dimension than DIN 7-16 connector
 minimum center-distance between 4.3-10 connector ports is 40 mm, compared to minimum 70 mm for 7-16”
 largely facilitating on-site installation while simultaneously increasing connector density for multiband antennas.

- 63 - Orange confidential
Figure 27 – Short historic overview of connectors for mobile communication systems

The key technical improvement of the 4.3-10 connector is that the PIM performance of the connection is no longer
dependent on the torque applied to secure the nut of the male connector. Due to the new design, the PIM
performance is assured by attaching the female into the male connector. The nut only serves to mechanically
secure the connection and waterproofing.

Migration Strategy
• All new antennas start to use the 4.3-10 connector from now onwards.
• Legacy antennas will switch to the 4.3-10 connector according to customer requirements.
• Pure TDD antennas will not switch immediately to the 4.3-10 connector.

The female 4.3-10 connector fits 3 possible male jumper cables, always with 1 / 2” Super flex

Figure 28 – The 4.3-10 connectors

It is crucial to inform the engineering teams and installers of this new connector technology as soon as
possible. We strongly encourage you to launch an internal communication and/or seminar about the 4.3-10
connector. Please ask for 4.3-10 connectors and jumper cable samples for your own evaluation as well as
for showing to your riggers and installation companies.

Take care to ensure that when antennas arrive on site, the correct jumpers for the connector type will be available
also. Adaptors from DIN 7/16” to 4.3-10 do exist but are highly discouraged to use because of increasing PIM risk.
These should never be used on outdoor sites; the proper jumper cable mating MUST be used.

Finally, we note that antenna vendor Amphenol-Jaybeam is validated by the Group to offer the new 4.3-10 jumper
cables as an accessory to their antennas with contractual short lead times and an Orange buffer stock imposed.

Orange confidential - 64
Antenna Line Devices: Group Recommendations Recommendation 12—1

 Consult the “Orange ALD Technical Shortlist” for any upgrade scenario on radio-engineering and antennas.
 For affiliates with only a building permit allowing for one antenna per sector, a tripleband (900/1800/2100)
antenna family (CSC 2.8.1) should be considered as the ‘minimum’ configuration.
 However, when 2 antennas per sector can be installed, this remains the preferred scenario for future flexibility.
 For heavy loaded sites with limited installation space, hexaband antennas with 12 connection could be
considered for deployment.
 Design the antenna sectors with identical azimuth, ready for Carrier Aggregation between different bands.
 Antennas with high gain (>17 dBi) are recommended for general indoor coverage and best interference control.
 Antennas with physical heights shorter than 2 meters are strongly discouraged for LTE800 deployments.
 Do not combine 700 and 800 ports into the same connector or feeder due to risk of PIM. Deploy each technology
onto a separate connector.
 Select antennas with separate electrical downtilt for LTE800 and GSM/UMTS900 within the same sector to allow
control of LTE coverage which typically requires more downtilt; if not possible (cost-reduction or reuse), make
sure the tilt of LTE800 is more than 4° to avoid too much interference and capacity degradation for LTE800.
 Implement 3GPP/AISG Remote Electrical Tilt (RET) when LTE is deployed on the network.
 RF Multiplexors (with DC autosense bypass) and dualband MHAs are the most cost-efficient upgrade path for
LTE upgrade evolution.
 For RF cables, connectors and jumpers, very strict rules apply for correct installation practices.
 Follow the recommended migration paths for 7-16 DIN to 4.3.10 connectors.

Contact: Henk Tubbe , Colin Fraser, Alain Choumaker

12.2 Horizontal Antenna Sectorisation with Passive Antennas


The multi-sector (4-5-6 sectors) passive antenna solution is a radio capacity upgrade option - not a coverage
solution. It needs to be highlighted this is a short term solution which is generally not recommended for wide
deployment. As the typical traffic growth predictions might consume quickly the possible gains of sector split.
There are number of drawbacks that need to be concerned before upgrading the site:
1. Multi-sector configuration cause increased number of antennas on site as well as number of RRUs this might
have direct impact on rental cost on OPEX increase as well, it limits the further site evolution with additional
bands in future (multi-bands with high-band split-beam are rare on the market and extremely expensive)
2. The typical 6-sector split can add practically 40-70% in selected bands which might be not enough in long term.
Capacity demand doubles every 1.7 years (at current traffic growth rate), the 6-sector solution can consume the
additional capacity within 1 year (often comparable to environment permission process for new antennas) =>
the multi-sector solution is a short term alternative
3. Multi-sector configurations are typically limited to high-bands as low-band stays in legacy configuration are
problematic for carrier aggregation implementations; new SW releases allowing multiple secondary channels
are required
From the business case point-of-view, the cost of site reconfiguration for multi-sector (and neighbour cell RF re-planning)
might be not adequate to the capacity gain which will be quickly consumed.

Deployment
 Intended only for highest traffic sites/sectors in the network. (4 - 5 sectors can be more common than 6 sectors)
 It works by increasing the cell density and is an alternative to building new sites for capacity (or deploying new
technologies, spectrum, etc.).
 Depending on local MCO conditions, sites should normally be first upgraded to the maximum 3-sector carrier
capacity configuration before being converted to higher sectorisation.
 Sites with 6-sectors of 2G are well suited for conversion to 6-sectors of 3G /4G.

Site Configuration & Equipment


 Horizontal 33˚ beam width antennas can be used in specific locations to minimise the increase in softer
handover when upgrading to 6-sectors from legacy 3-sectors sites. Note: H45° antennas give inferior
performance. Some antennas with “dual-beam” 2 x H33° inside the same physical radome have been validated
and are available in the ALD Corporate Sourcing Contracts (see section “Antenna Line Devices”).
 Increased downtilt (~1-2 degrees) is required to maximise the capacity gain, minimise soft handover, and
prevent service area extension / interference.
 MCOs should consult the Group RAN Skill Centres for detailed information regarding the upgrade process & the
equipment support available for the 4-5-6-sector configuration.

- 65 - Orange confidential
 Due to the increased EIRP, radio license and EMF exposure limits should be checked (MCO dependent).
 Transmission: E2E network traffic dimensioning need to be considered as MW backhaul might a bottle neck.

4-5-6-Sector Performance
 The overall network performance of 4-5-6-sector sites is very similar to 3-sector sites.
 Received signal level coverage with 6xH33° is increased by approximately 3dB over the 3xH65° (~ 2 to 6dB
depending to the site & sector conditions).
 Downlink capacity for a 3- to 6-sector upgrade increases from 40% to 70% depending on radio design and
interferences.
 For 3G there is a 0-10% increase in Softer Handover (intra-site HO) with H33˚ antennas and no increase in Soft
Handover (inter-site HO).
 Interference does not increase unacceptably when using 6xH33° and additional down-tilting is recommended.
 It needs to be mentioned the performance results for 4G sites are mainly based on simulations there were no
Group trials for such implementations as demand for such deployments is very limited

Horizontal Antenna Sectorisation with Passive Antennas: Group Recommendation for EU 12—2
 Generally the evolution of heavy loaded sites to multisector configuration architecture with separate antennas
is not recommended for wide deployment as it is only a short term capacity solution limited to selected
bands.
 4-5-6 sector sites is an option to consider increasing capacity where multi sectors have already been
deployed or could be envisaged depending on local deployment flexibility. Capacity gains of up to 70% (for 6
sectors) have been achieved in both 2G and 3G field trials on isolated sites.
 The concept of multisector sites are evolving towards dynamic sector/beam forming supported by Massive
MIMO technology which seems to be more adequate for capacity extension than traditional passive antenna
with static sectorisation.

Horizontal Antenna Sectorisation with Passive Antennas: Group Recommendation for AMEA 12—3
 4-5-6 sector sites is an alternative to support increasing capacity where multi sectors have already been
deployed or could be envisaged depending on local deployment flexibility. Capacity gains of up to 70% (for 6
sectors) have been achieved in both 2G and 3G field trials on isolated sites.
 Even the multisector reconfiguration is not generally recommended for those MNOs where limitation of
installed antennas on site is not so restrictive as for matured markets a short term solution with passive
antenna sectorisation can be considered. However the TCO needs to carefully calculated as life cycle of new
antennas might be too short to justify the investment.
Contacts: Roman Łapszow, Henk Tubbe

13.3 Active Antenna Systems Evolution

Active Antenna Systems (AAS) refer to systems where the transceiver is embedded within the antenna radome and can
create “actively” beams. There is no need for separate RRH, and the radio transceiver is “hidden” within the antenna. It
brings number of issues as passive band availability of multi-band AAS and maintenance issues. The current AAS
roadmaps indicate that two active bands can be supported (E///, Huawei, Nokia) which allow reducing site space
requirements however flexibility of future site reconfigurations would be limited when AAS is installed.
There are two kind of AAS: semi-integrated antennas with no functionality of beam-creation for DL (provided by E/// AIR
products and Huawei AAU) and Massive-MIMO (MM) arrays dedicated to LTE2600 and 3500MHz (LTE or 5G). While the
old vertical sectorisation feature that splits the cell into inner and outer cells is no longer on the roadmaps, the MM
technology with full flexibility of beamforming in vertical and horizontal plane is a promising technology for the next
generation antenna products. However similar to past vertical sectorisation MM antennas are complex HW products that
impacts the cost and implementation business case.

Semi-integrated can reduce site complexity and limit number of external connectors when needed to migrate to
2T4R/4T4R - at this stage: not recommended for rollout
The site with semi-integrated antenna would lose reconfiguration flexibility and vendor dependency would increase
 option for high bands with 4T4R when „spaghetti configuration” in not-implementable
 a candidate for bands 42 & 43 with 8T8R / 16T16R (with TDD MCIC cluster connectors)

Orange confidential - 66
Semi integrated antennas systems are currently available with various configurations of passive bands. The wide
portfolio of products is provided mainly by Huawei. The main drawback of the solution is an increased dependency on
selected vendor. The total cost of passive antenna and RRH are close to semi integrated solution at this moment. For
those cases antenna system is no longer a separate part of the network that is vendor agnostic but an active part of
access system which is HW & SW dependent. The high cost of future upgrades need to carefully calculated in the
individual business case. A potential swap of RRH - which in such case is part of AAS - will require time consuming
permission process.
In case the legacy passive system platform is not sufficient – the semi integrated product could be installed taking under
consideration limitations mentioned above. Even for passive antennas their short lifecycle due to more frequent need for
upgrades with new bands and technologies is a challenge.
It need to be highlighted the industry of semi integrated antenna portfolio is very unbalanced – it is limited to Huawei
family with AIR antenna system available from Ericsson and Nokia that is targeting towards MM products.

Figure 30 – Huawei AAU5953 1800 4T4R+2600


Figure 29 – Huawei AAU3970 2600 8T8R, 4T4R+ 700-800P + 900P + 1427-2170P (Q3’17)
AAU5972 1800 8T8R + 2300-2600P 8T8R (Q3’17)

For site configurations with no requirements above 2T2R, passive antenna system is always a recommended solution.
For multiband configurations with 4T4R (on 1 or 2 high-bands), high complexity of connections and large number of
RRH might be not be acceptable for specific site requirements, the semi-integrated antenna could be considered.

Massive MIMO Active Antenna Systems (MM AAS) are possible candidates for bands 42 & 43 but less likely for L2600;
technology dedicated mainly to TDD with higher complexity and lower gains for FDD; promising for 5G:
 option for 16T16R in EU region (no high-rise buildings and lower TCO), expected to be
possibly integrated with L2600 passive array (4T4R) in the next step; the concept of full
integration with low bands is considered but could be very difficult due to weight, size limits as
well as EMF limitations
 a must for 32T32R / 64T64R; current visibility only as a standalone product; most likely wall-
mounting option required in EU

There are early – standalone – solutions for MM AAS. The configurations starts from 8T8R/16T16R and 32T32R/64T64R
arrays for 2600 and 3500 bands – both available for TDD only but it is envisaged FDD products also might be available.
The configuration of 64T64R based on 8x8 array matrix is the most common proposal by the industry at the moment.

Figure 32 – Nokia AirScale 2600 64T64R 128


Figure 31 – Ericsson AIR 6468 2600 64T64R 128
elements
elements

Generally for a higher order MIMO than 8T8R, migration to fully embedded AAS is a must due to the large
number of connections between the radiating elements and active components (PAs). For bands at 3500 MHz,
due to propagation loss, higher in-building penetration requirements and challenges related to the budget link, higher
antenna gains need to be available. The band 42 and 43 can be achievable on the same macro grid with MM.
There are number of unknowns for antenna system for bands 3500 MHz. Adding a new band to the legacy antenna
system is not trivial. EMF (health & safety) issue is still open as a major point as well as out of band emissions.

- 67 - Orange confidential
For implementation several options are identified:
Separate MM AAS: if possible add a standalone antenna on site
 solution secures flexibility and mitigates risk of impact on legacy system
 deployment options of this option is limited as the main focus for 42, 43 bands is for
capacity sites which are already heavy loaded by existing antenna systems

Highband antenna upgrade: swapping to a “hybrid” highband MM AAS integrated with legacy bands passive antenna.
Solution for dual antenna per sector (high and lowband antennas) with two options:
 2600 MHz 4T4R passive and 3500 active MM (16T16R / 32T32R / 64T64R*)
(*note that MM active 32T32R / 64T64R is highly unlikely in this antenna)
 2600 MHz and 3500 MHz BOTH based on MM; if 2600 MHz stays on FDD
=> low tech-eco efficiency (low feasibility, risk of being too heavy and complex)

Single antenna per sector: split the existing antenna system into two blocks:
 700-2600 MHz upper part of the radome
 3500 MHz active system below the passive arrays
Pack all bands into the size of legacy radome of low band 2.7m passive antenna but this is highly UNLIKELY to see
this antenna developed.

The products for 3500 MHz are still under trial and their design are not matured. Studies on requirements on out of band
(OBB) emissions for band 42 and 43 are ongoing. The complexity of filters that are varied for specific market could
significantly impact the HW antenna cost.

Figure 33 – OBB emission requirements (depends on entity – ECC most restrictive below 3400)
and early cavity filters solution for Japan market (ZTE)

Active Antenna Systems Evolution: Group Recommendation 12—4


 Active Antenna Systems are not recommended for wide implementation at this stage.
 In case the legacy passive system platform is not sufficient – the semi integrated product could be installed
taking under consideration high cost of future upgrades and overall TCO for the individual case.
 Industry of semi integrated antenna portfolio is very unbalanced – it is limited to Huawei family with AIR
antenna system available from Ericsson and Nokia that is targeting to MM products.
 The 3500 MHz bands brings opportunity for antenna technology “jump”. Legacy macro passive antenna
systems reconfiguration for AAS (even in case TCO would be positive for AAS swap) is difficult mainly due to
a need of combining it with RAN renewal process plus a need of considering time consuming site permission
at the same time. New high bands (in the future also mm-waves for 5G) require additional, separate antenna
system. This opens a potential for compact, advanced antenna solutions either based on M-MIMO or/and
with beamforming / beamstearing functions. Not expected earlier than 2018/19.
Contributors: Ric Bailey, Colin Fraser, Henk Tubbe

Orange confidential - 68
Contact: Roman Łapszow

13. Radio Access Network Architecture Evolution


13.1 Introduction
In 2016, the RAN ecosystem has moved to the virtualization, as precursors like Altiostar and Nokia were joined by all
others suppliers including Ericsson and Huawei, who have now virtualized 4G RAN product in the roadmap, with globally
a general availability of these 4G solutions in 2018/19.

Even if suppliers will also propose traditional RAN solution for 5G, they all have virtualized RAN solutions, which will be
available in 2018 for first trial activities.

13.2 Virtualized RAN architecture


The virtualized 4G RAN architecture is defined on following principles
 A split of the BBU between the RLC and the PDCP, which defines

o A Distributed Unit (DU) managing the layer one and RLC/MAC protocols.
o A Centralized Unit (CU) managing the PDCP, the signalization (RRC and S1/X2) and the OAM.
o A new interface called F1 between CU & DU, characterized by
 A latency requirement of 5ms between the low BBU and the vBBU, inducing a maximum
distance of around 200km in best cases.
 A throughput proportional to the end user throughput (+20%).
 This interface is under standardization at the 3GPP, with the objective to interconnect a CU
from vendor A with a DU from vendor B.
 The DU will be implemented at the radio site level or at a first level of aggregation (< 15km)
o In a first step the DU will reuse current BBU hardware, thus without virtualization. Orange has now the
confirmation that almost all equipment deployed in Orange networks are compatible with vRAN.
o Some suppliers like Altiostar also propose to manage the DU function in the RRH (called iRRH) to
minimize the number of equipment at the radio site level.
o In a second step, when the new generation of Intel processors including advanced acceleration
capabilities will be available, the DU will be virtualized. This evolution will allow the distribution of small
compute at the radio site level, which will be controlled by the metropolitan NGPoP as a remote
compute and could be used by others network functions like MEC of vEPC in some specific cases

 The CU will be implemented on a virtualized infrastructure in a central location, defined by Orange and called
metropolitan NGPoP
o Due to CPU and real time constraints, the vRAN NGPoP will require some specific acceleration
functions like DPDK.

- 69 - Orange confidential
o The CU management in a NGPoP is now agreed by main suppliers, but terms of the contract have still
to be negotiated, to define the SLA between the Orange infrastructure and the supplier VNF.
This new architecture will be mainly implemented as described in this picture:

2G & 3G RAN virtualization:


 At this stage, there is no real plan to virtualize 2G and 3G BBU.
o Main suppliers have or announce virtualized controllers (BSC & RNC). There is no gain, which can
justified a proactive migration to vBSC and vRNC, but in case of swap of existing solutions due to end
of life or RAN renewal, the opportunity of virtualized controlled shall be studied according to suppliers
roadmap and NGPoP availability, as these solutions will provide:
 Gains when the 2G & 3G load will decrease,
 A future proof solution,
 Synergies with the virtualization of others network functions.

13.3 Virtualized RAN pro & cons


Even if some optimizations are expected thanks the centralization, due to the RRH dominance in the CAPEX & OPEX,
significant gains are not expected.
Virtualized RAN gains are linked to best KPI, new services and network
flexibility:
 Best KPI:
o The centralization of the PDCP will provide best KPI of the
4G/5G dual connectivity in all deployment scenarios.
o The implementation of applications at the NGPoP level or
the antenna site, when the DU will be virtualized is a real
opportunity to reduce drastically the latency between the
user and applications.
 New services: in 5G the network slicing is a corner stone to deploy new services with end to end quality of
experience. As the radio interface is always a main bottleneck, the RAN slicing is perceived as mandatory and
is linked to the virtualization.
 Network flexibility:
o These new services will induce new call profiles, impacting the network load in unforeseeable ways.
The RAN virtualization will allow to manage dynamically this load thanks the pooling and scaling of
CPU resources.
o The standardization of the f1 interface is seen as an enabler to increase the RAN ecosystem, with the
emergence of new CU suppliers.

13.4 Next steps and conclusions


In 2016, this architecture was analysed during a live trial in Poland performed in collaboration with Altiostar, which has
confirmed main principles and gains:
 The feasibility of the BBU split,
 The robustness of the f1 interface,
 The compatibility of the NGPoP infrastructure with vRAN VNF,
 The flexibility linked to deployment and hardware resource management,
 QoE improvements linked to applications deployed at the edge of the network.
As the RAN ecosystem is globally moving to the virtualization, the main problematic is to define the Orange vRAN
strategy, in term of
 Gains: as described above, there are some vRAN gains, which shall be quantified to drive the deployment,
 Plan : when and how the RAN will be virtualized, according to network constrains, 5G deployments,
 NGPoP and vRAN: work vRAN requirements on NGPoP, NGPoP constraints on vRAN VNF and impacts on
network operations, to have a full compatibility between the RAN and the Orange infrastructure, allowing vRAN
deployments.
These topics will be worked by the vRAN project in 2017 & 2018, thanks studies and trials.

Orange confidential - 70
13.5 5G RAN Technology (as currently under definition in 3GPP) / vRAN
5G RAN architecture
The virtualized RAN architecture defined for the 4G RAN is fully compatible for the 5G (see Radio Access Network
Architecture Evolution) and all benefits associated to the RAN virtualization will fully benefit to the 5G, especially to
manage the load induced by new services and the slicing.

At this stage, major suppliers RAN suppliers have traditional and virtualized RAN architectures in roadmaps.

Affiliates will have to select the appropriate architecture according to the planning of the 5G deployment and especially
the availability of the virtualized infrastructure, which could be provided by the supplier (AirFrame for Nokia for example)
or specified by the Orange (NGPoP), which is the Orange recommendation.

RAN Architecture: Group Recommendation: Group Recommendation 13—1


The distributed RAN architecture is still the recommended architecture.
The virtualized RAN architecture is seen as the architecture of the next generation of RAN, especially the 5G RAN,
thus in the 2020 timeframe.
In the meantime
 The virtualized infrastructure shall be deployed in Orange networks
 Virtualization skills shall be developed to prepare these vRAN deployments
 The transport network shall evolve to prepare 5G & vRAN deployments

Contributors: Maha Chouaref, Ric Bailey, Alan Stidwell


Contact: Arnaud de Lannoy
References
OREP: virtualized RAN
vRAN Project: vRAN 2016 deliverables

14. Other Wireless Technologies (CDMA, WiMAX)


14.1 CDMA-2000 Family
The CDMA technology is outdated and not adapted anymore to any of Orange operators. Existing networks and use
cases if any should be migrated to FDD or TDD LTE.

Contact: Bertrand WAELS

14.2 WiMAX
WiMAX technology is now obsolete and cannot compete technologically with LTE in terms of features, spectral efficiency
peak bitrates or traffic capacity. Moreover, all WiMAX suppliers – including Telrad, formerly Alvarion – ceased the
development of WiMAX technology few years ago offering only limited sales and support to existing deployments. There
is a high risk that Telrad announces end of support and of sales for all WiMAX equipment by end of year 2017.

Because of these limitations and obsolescence of WiMAX technology, migration to LTE is strongly recommended for all
WiMAX affiliates. LTE offers an evolution path and has the advantage of a richer ecosystem and better performance
compared to WiMAX. Telrad is not selected by Orange as LTE supplier and their LTE solution is not proven to be
functional.

14.2.1 WiMAX Technology Risks and Migration to LTE


Back in 2006, Orange selected Alvarion as unique supplier of WiMAX technology within the group. At the beginning of
year 2013 Alvarion sold to Telrad Networks its Broadband Wireless Access (BWA) division, which includes WiMAX.
Orange continued the WiMAX business with Telrad and the Corporate Sourcing Contract (CSC) was renewed in July
2013 for duration of 4 years.
In February 2013, Telrad communicated the “maintenance mode” for WiMAX products:
- No further WiMAX product development, new releases or bug fixes starting with February 2013
- End of WiMAX sales will occur during next 2-7 years depending on product, i.e. end of year 2019 the latest
- End of WiMAX support in 5 years’ time, i.e. February 2018

- 71 - Orange confidential
At the beginning of year 2017 we estimate the lifespan of WiMAX technology to 1-2 years. There are some important
risks in operating WiMAX technology:
- Impossibility to order new equipment: 16d is discontinued by Telrad since June 2015, 16e end of sales was not yet
announced by Telrad but might follow soon. At the moment of writing, Telrad could not tell which 16e parts can still
be ordered.
- Telrad support (SLA) and repair services are limited and might cease at short notice during year 2017. In the best
case, Telrad WiMAX support can be available until February 2018.
- CSC will expire in July 2017 and will not be renewed, with exception of basic support services.
- WiMAX cannot cope with increasing data rates and traffic volumes requested by subscribers.

In small networks or to avoid the cost of a major swap, WiMAX technology could be operated beyond years 2017-2018
provided that:
- An alternative fix access solution is developed in parallel to WiMAX
- There is no need for growth in terms of sites, traffic capacity or subscribers
- Local knowledge is available to compensate for lack of manufacturer support
- Stock of spares is available to replace faulty equipment

Figure 34 – Estimation of Telrad (Alvarion) WiMAX Products Lifetime

The following Telrad WiMAX products and system releases are validated for deployment:

WiMAX System Release Base Station Terminals (CPE) End of Sales Date
th
16d TDD R4. 7 (4.7.0.201) Macro BS BMAX 1000 (PRO or Si) All 16d products: End of Sales 19
16d FDD R3.7 Macro BS/ uBS BMAX 1000 (PRO or Si) June 2015.
BMAX 1000 (PRO or Si) CPE BMAX 1000: End of Sales
th
19 June 2015
CPE BMAX 3000 (outdoor) CPE 3000 & 4000: End of Sales
R3.5M
CPE BMAX 4000 (indoor) November 2015.
16e (TDD) NPU 3.0.20.69 Macro BS
KZ-Tech Airstream 1100 Airstream 1100 - Not available for
AU 3.0.20.39
USB dongle ordering
CPE 7000 Outdoor (dual Not communicated yet by Telrad.
mode WiMAX / TD-LTE)
Note: Products marked in red are end of sales / discontinued and not available for purchase.

14.2.2 Brief Overview of WiMAX Technology and Standards


WiMAX stands for Worldwide Interoperability for Microwave Access and is a Broadband Wireless Access (BWA)
communication system based on IEEE 802.16 family of standards. WiMAX technology enables provision of high speed
data over a wide area using connectivity between a subscriber station and base station. Originally designed for fixed
terminals, 802.16 standard has been later developed to allow ‘nomadic’ and even fully mobile operation.
The first WiMAX standard was issued in October 2001. 802.16-2004 also known as 16d or fixed WiMAX, has been
approved in June 2004. To allow mobile and vehicular applications, a new revision, 802.16e-2005 – the 16e or mobile
WiMAX – was completed in December 2005. The current and latest WiMAX standard is IEEE 802.16-2012. Among
others, it contains 16m enhancements, was published in June 2012 but never supported by the industry.
A list of IEEE 802.16 standards is available on IEEE website: http://grouper.ieee.org/groups/802/16/published.html.

802.16-2004 802.16-2005 802.16-2009 802.16-2012


802.16-2001
(superseded)
16d 16e – Rel. 1.0 16e – Rel. 1.5 16m
(superseded) (superseded) (superseded) (active)

Oct 2001 Jun 2014 Dec 2005 May 2009 Jun 2012
Figure 35 – Evolution of WiMAX 802.16 Standard

Orange confidential - 72
WiMAX 16d exists in both TDD (Time Division Duplex) and FDD (Frequency Division Duplex) profiles in 2.5GHz and
3.5GHz frequency bands. Possible channel bandwidths are 5MHz in TDD variant and 1.75, 3.5 or 7 MHz in FDD mode.

WiMAX 16e has the advantage of an open and richer ecosystem and benefits from an improved radio interface. It exists
only in TDD profiles in 2.3GHz, 2.5GHz and 3.5GHz frequency bands. Supported channel bandwidths of 5MHz or
10MHz together with specific radio features such as 2x2MIMO and more efficient modulations allowed increased sector
capacity and peal data rates compared to 16d system.
Evolution to 16m could have improved further WiMAX radio performance due to features such 4x4 MIMO, beam-forming
or larger radio channel bandwidths (20MHz). The 16m standard was not implemented by WiMAX manufacturers.

WiMAX: Group Recommendation 14—1


 Migration from WiMAX (16e or 16d) to LTE is strongly recommended for all WiMAX affiliates before end of year
2017, when WiMAX technology might be end of vendor support and end of sales.
 Cease maintenance contract with Telrad and use Telrad professional services when needed. High cost of Telrad
maintenance package does not reflect in their service quality!
 Telrad – formerly Alvarion – is not selected to supply LTE technology.
 WiMAX networks can possibly be operated beyond year 2017 by re-using dismantled equipment and local
knowledge. However expect cease of vendor and Skill Centre support beyond this date.
 Investments in WiMAX technology shall be limited and minimised.
 For new fixed access networks (FWA), LTE technology shall be considered instead of WiMAX.
Contact: Dan Musat

15. Evolution of the microwaves domain


15.1 Microwave Equipment evolutions
LTE-A feature deployment and the consideration of 5G requirements challenge the ability of MW products to support
higher capacity, lower latency services. However MW Vendors remain convinced that MW has a future that will play a
significant part in the implementation of 5G networks and are developing MW features capable of this.

For decades there has been continual development of the radio interface to meet growing capacity demands and this
trend continues with vendors talking about 8192QAM and 16384QAM modulation states combined with 224MHz channel
bandwidths. However we are tending toward the limits of this type of development: in recent ZTE development tests
8192QAM has shown only a 3% capacity increase over 4096QAM; finding 112MHz in the majority of traditional bands
has always been problematic for planners and seldom achieved.

Other developments have been seen in channel aggregation methods which have larger aggregation advantages as we
advance toward an all IP transport structure. Existing methods such as XPIC and MIMO, using Adaptive Modulation
techniques, now give better QoS for priority services due to more efficient LAG type protocols and hierarchical QoS
mechanisms with a broader granularity/operation. Vendors are deploying frequency band combining schemes also,
where a traditional band (say 23GHz) is combined with a lower availability Eband link of the same distance. The high
capacity provided by the Eband link can be used for lower priority traffic whilst higher priority traffic can be assured
through QoS mechanisms over the remaining capacity. We are currently testing 10Gbps links in Eband products with
20Gbps on the roadmap (2018); for instance NEC are claiming to have trialled an E Band link in Russia which achieved
8.34Km @ 10GBps but the availability was only 97.24% (which is about 10days a year unavailable time). Aggregating
this with a traditional channel can give us the required high availability for some services (~99.995% in ~535Mbps) and
an additional 5 > 10Gbps for lower priority traffic. The key issue is the higher capacity E Band link will fade a lot more
quickly than the traditional band.

Proposals for use of higher frequency mmWave frequency bands are progressing with many MW vendors testing
prototype RF modules in the W Band (92 > 114GHz) and D band (130 > 175GHz). D Band currently has the highest
priority within the ETSI mmW ISG and CEPT SE19 and as such sub bands and channelisation have been proposed.

- 73 - Orange confidential
It is expected that most MW vendors will have product available in D Band in 2020 to support 5G implementations.

Higher layer (IP/MPLS) transport protocols are now well implemented in some Vendors’ products and will easily migrate
into the denser meshed networks required for 5G. It is, however, still unclear whether early implementations of 5G will
require L2 or L3 transport protocols; it is expected that this will be determined by existing local network configurations
and network services to be deployed.

With 5G comes the requirement to support network slicing where the transport network will be divided into layers, of
virtual transport services, flexible enough to give varied capacity, latency and availability, on the fly as required. This will
be achieved using SON systems to optimise the transport network using SDN/NFV features implemented in the transport
elements.

15.2 MW capacities and upgrades


The following table is an example of available capacities according to channel spacing and modulation format (source
NEC, Ethernet layer 2, 64 bytes VLAN tagged frame).

Traditional bands 6 - 42 GHz E-Band (80 GHz)


Capacity
Mbps 28 MHz 56 MHz 112 MHz 250 MHz 2000 MHz

128 QAM 153 308 602 1300 10000


256 QAM 175 352 690 - -
512 QAM 196 394 771 - -
1024 QAM 216 432 864 - -
2048 QAM 239 479 957 - -
4096 QAM 260 525 - - -

The capacity of the MW link can be increased by:


1. increasing the modulation format,
2. increasing the channel bandwidth,
3. increasing the number of radio channels,
a. on the same frequency on the two polarisations with XPIC,
b. on the same frequency band, by adding one or several frequency channels,
c. in different frequency band(s) .
rd
In the 3 solution (a, b, c), the multiple radio links can be aggregated with a layer 1 mechanism allowing a full GE pipe or
partial 10GE pipe. In case of radio failure, priority traffic can be saved by using QoS.
Of course depending on the vendor’s solution the implementation and engineering of such capacity upgrade can be
different, i.e. ODU change or not, new board required or not, antenna change or not, release change or not. For any use
case or project do not hesitate to contact the MW skill centre.

Microwave equipment: Group Recommendation 15-1 / 2


 Hybrid Ethernet Microwave radios should be deployed to reduce TCO for network upgrades.

Orange confidential - 74
 For Greenfield rollouts full packet MW solutions should be considered if RAN interfaces support this.
 Local spectrum licensing should be considered carefully in coordination with upgrade scenarios to ensure
cost optimisation.
 E-band solutions should be deployed for high capacity links in urban areas dependant on local conditions
i.e. potential for interferers in light licensing regimes
 New packet MW products should be considered as a migration path when upgrading long haul systems.
Contact: François Brunet

15.3 Point to Multi-Point (PtMP) solutions in licensed frequency bands for fixed
and mobile aggregation
Point to Multipoint Microwave radios are FDD/TDMA transmission systems where one node can transmit and receive
from multiple sites and are based on the same technology in the physical and MAC layers – which is 802.16-2001 LMDS
(Local Multipoint Distribution Service).

(source: Cambridge Broadband Networks)


Figure 36 – PtMP Microwave system
The suppliers of PtMP solutions offer solutions in 10.5 GHz, 26 GHz, 28 GHz, 32 GHz and 42GHz frequency bands,
however the range is limited to up to 5 km. The maximum sector throughput is 540Mbit/s (for Intracom) while each
terminal in the sector can achieve up to 270 Mbit/s.
Their use is suited to urban environments in countries where excessive microwave spectrum licence fees apply as PtMP
reduces the need for multiple licenses and represents important license cost savings.
One of the key PtMP radios features is support for hitless adaptive modulation adjusting to current weather conditions
(up to 9 states from 4QAM/QPSK to 1024QAM), triggered by BER measurements. Another important feature
characteristic for PtMP MW systems is statistical multiplexing. This feature can lead even increase of the effective
throughput by four times with no additional cost thanks to dimensioning procedure that can be done to fit the mean traffic
values, not the peak values.
In addition, current PtMP MW equipment often implements algorithms that allow efficient compression of Abis data (e.g.
reduction from) as well as IP optimization.
The potential benefits of Point to Multipoint systems include:
 Opex reduction thanks to:
o lower spectrum licensing fees
o less rooftop space needed (at the Hub site)
 may solve frequency availability issue
 Reduction in number of deployed terminals thanks to limited number of radio units (for the same number of CPEs
connected only N+1 radio units needed for PMP while 2N needed for PtP MW)
 fast to deploy thanks to:
o fewer installations
o only one visit of technician per one new link
o faster spectrum licensing
 bandwidth dimensioning is done at the Hub level (it takes into account the whole sector)
 support for multiple network interfaces (PDH, SDH, Eth)
 more gain from statistical multiplexing

- 75 - Orange confidential
 opportunity to deploy PMP in emerging markets dense areas
 opportunity to use for enterprise solutions

However some limitations of PtMP technology guide its applications in countries and these are:
 limited reach (up to 5 km) – suited only for urban and sub-urban areas
 potential problems with site visibility – very dense areas can suffer from NLOS (Non-light-of-sight)
 only 5 frequency bands (10.5 GHz, 26 GHz, 28 GHz, 32 GHz and 42GHz)
 current capacity limited to ~540Mbit/s per sector (not suited for mass LTE introduction yet)
 only star topology with possible relay (Intracom)
 potential regulatory issues (regulatory approach to PMP is not always clear in countries)
 poorer bps/Hz than for traditional Point-to-Point MWs
 not always suitable for macro cells backhaul, only for low capacity sites

Note that the PtMP solutions have already successfully been deployed in Kenya and are available through the NEC
sourcing agreement. However the bulk of the savings with these systems is based on the lower spectrum license costs.
The business cases have always suggested that at least four cell sites should be located within a PMP sector for the
solution to be viable.
Microwave Skillcentre team, in the OLN/RNM LTE for fix project framework, will start a point to multipoint
products market survey with the objective to define the use cases where PtMP solution could make sense, from
technical and economical perspective, compare to LTE for fix solution.

PMP: Group Recommendation 15—1


 It is recommended that Point to Multipoint systems be deployed in urban areas with low mobile backhaul capacity
requirements and/or enterprise applications in affiliates where the microwave spectrum licence costs are high as
these solutions only offer significant savings in this case.
 For any use case or project it is recommended to contact the MW skillcenter and/or the LTE for fix project leader
(Vanesa Tornero)
Contact: François Brunet

15.4 Sub 6 GHz Spectrum Band Low cost solution


The Sub 6GHz spectrum bands are very useful for long distance and near or non-line of sight communications as well as
line of sight short hops. The bands currently used for this purpose include spectrum in the 2.4 GHz, 3.65 GHz & 5.8 GHz
Sub6 solutions are generally TDD transmission systems. So compared to legacy system the capacity is divided by 2 for
the same channel spacing and modulation format.
Recently Affiliates have expressed strong interest in low cost solutions to connect B2B customers, in particular
highlighting low cost solution for last mile with low SLA. Currently MCOs are using traditional licensed microwave bands
in specific cases, but this resulted in a high cost of installation, therefore no longer competitive for this specific market
applications. The request to Orange corporate Skillcentre by the MCOs is to propose and find alternative solutions in the
sub 6GHz bands, solutions that are very competitive in their simplicity to install, limited to only some specific features
supported and most importantly low price.
In the last year or so, some trials have been carried out with vendors such as RADWIN, Blinq Networks and Dragonwave
at sub 6GHz to investigate performance of this equipment at these frequencies bands 5.8GHz and 3.6GHz but the
solution still proved to be costly.
Therefore the Project launched by the Microwave skillcentre in Q1/2016, based also on the ongoing RFI in Orange
Poland in cooperation with the Skill centre, is to asses & investigate the potential low cost solutions, and at the same
time still effective to deliver the level of performance request for B2B customers. Out of the vendors participating to the
RFI, Cambium equipment currently seem to be the best suited for these low cost applications.

To be compete Microwave Skillcentre team, in the OLN/RNM LTE for fix project framework, will start a point to
multipoint products market survey with the objective to define the use cases where PtMP solution could make
sense, from technical and economical perspective, compare to LTE for fix solution.

However the use of unlicensed bands such as 5.8GHz means that there will be risks from interference. This risk cannot
be quantified until links are deployed and, as this is unregulated spectrum, field trials are mandatory. It is also true that
the new generations of these types of equipment such as the latest cambium 450i will be capable to support interference
cancellation mechanisms to reduce this risk further.
Cambium tests:
TSC MW tested in 2016 entry-level products Cambium: PtP 450, PtMP450, PtP450i, PtMP450i. The Ethernet bridging is
implemented on TDD. The bandwidth in both directions is tuneable (50/50...20/80....). The radio characteristics are at

Orange confidential - 76
expected level, the spectrums are compliant (freq tested 5.5GHz, 5.8GHz) with ETSI EN301 893 V1.8.1/ETSI EN
302 502 V1.2.1.
The “450” family main advantages is its low price. But the performances are poor (bandwidth, 10kpacket per second).
“450i” family has better performances (45 kpps for PtP, 32 kpps for PtMP). “450/450i” has limited QoS feature
capabilities (only 2 queues high/low prio).
The big advantages of these products are the unlicensed frequency bands they used, the easiness to install and operate
and their low price.
The main drawbacks are the limited performances; low bandwidth, core architecture based on CPU, (means that when
device operates under high rate pps, the overload condition on cpu will lead to unexpected behaviour, ex: QoS bugs).
The pps limit (packet per seconds) is important (will have effect with traffic with numerous small packets)
In conclusion “450/450i” families are interesting in the context of connecting enterprises, or very temporary operator links,
with low SLA requirements.
The middle range product PTP650 offers more performances but at higher price.

MW Sub6: Group Recommendation 15-3


 Sub 6 GHz low cost solution for B2B customers in under evaluation at Corporate level with the aim to provide a
group recommended equipment such as Cambium, for these applications, once the lab and Field tests have
proven to be successful. Target completion Q3/16.
 For any use case or project it is recommended to contact the MW skillcenter and/or the LTE for fix project leader
(Vanesa Tornero)
Contact: Francois Brunet, Tan Tran

15.5 Microwave Research and Development


Current evolutions to 4.xG and future 5G requirements will impose new criteria on the transport network. As stated
already these criteria will be focussed on the provision of flexible high capacity, low latency and variable reliability, virtual
transport services that can be adapted to requirements as needed. Changes to the RAN architectures will mean
significant changes to the access transport topology and architecture with the introduction of Centralised and Cloud RAN
architectures. For MW the main topology implementation is in the Mid-haul and Fronthaul; although support for CPRI
transport is limited due to the high capacity requirements for each cell. Development in the functional RAN architecture
split will allow more efficient use of Fronthaul transport and MW vendors are involved in standards development for
Ethernet Fronthaul solutions (such as IEEE 1914 NGFI).

3GPP functional split options


Complexity of networks and services will be rationalised in parallel with the introduction of 5G networks. This will be
achieved with the introduction of SDN and NFV. The performance of the network and its service provision will then be
orchestrated using SON functions.
SDN will centralise the control plane and uncouple services from support by dedicated physical elements allowing ease
of network configuration and service provision changes. NFV will separate the logical transport functions from the
physical elements and allow virtualised transport layers to seamlessly support a variety of services over the one physical
network.
SON will allow the real-time monitoring of network performance and optimisation. There is still some debate as to
whether this should be in a centralised manner or whether some transport functionalities will require distribution to
improve network reactivity.
Vendors are working on solutions but the implementation in MW technologies is still unclear. Some PoC trials are being
programmed (i.e. ORO and Nokia Bell Labs – SON for MBH) but it is evident that vendors have still not integrated their
BUs to support these activities.
At a recent Industry forum it was noted by operators and vendors that there is still a bewildering complexity in the area of
telecom virtualisation due to the discontinuity of the overall strategic objectives and the multifarious software solutions
emerging.

- 77 - Orange confidential
Higher capacity has been identified as a key requirement for 5G services and MW vendors are identifying a mixture of
existing feature developments and new feature research. The following list shows the areas in which vendors are
working:
OAM (Orbital Angular Momentum) has been talked about in the industry for a few years now and has mainly been
research projects in a number of Universities. The view among vendors seems divided as to whether it has any future;
some claiming that it is really only a specialised form of MIMO that will not significantly improve throughput by
comparison and will have too many practical limitations in deployment. Other vendors are still heavily involved in
research.
Full Duplex radio was developed a number of years back but for MW industry existing spectrum usage and regulation
meant it would be very difficult to introduce. It enables a doubling of spectrum usage. With the opening of new spectrum
for 5G however this can now be considered as a possibility; the added factor of the higher frequencies involved means
that smaller antennas can be used and Tx and Rx antenna isolation can be more easily achieved. It is unlikely that we
will see any solutions until at least 2020.
Other developments vendors are working on involve the improvement of Forward Error Correction methods and better IF
interface performances plus the use of higher symbol rates (in higher frequency bands) to improve latency performance.
ZTE are currently working on a feature called FTN (Faster Than Nyquist) whereby they can enable the use of symbol
rates above the aliasing threshold by implementing a new technology that overcomes the intersymbol interference
created and increasing spectrum usage by 20-40% it is claimed.

Beam forming Antennas are key area of research and many vendors are working on solutions for higher frequency band
solutions. For 60GHz deployment in Small Cell scenarios (i.e. lamp posts) it will enable the optimal commissioning and
operation performance to be constantly achieved and for P2MP solutions better directivity will mean lower interference
potential and higher capacity per user.

Contact: Steve Jones

15.6 MW equipment Roadmap deployment recommendation


Vendor Type Equipment Release Validation status
9500MPR indoor R7.1 Done
MSS1/MSS4/MSS8/MSS1c/M R8.0 Target Nov 2017
PRe
9500 MPR outdoor R7.1 Done
MPT-MC/HC/XP/HQAM/GM R8.0 Target Nov 2017

MPT-HL (longhaul) R7.1 Done


R8.0 Target Nov 2017
Microwave 9600LSY EoL announcement
Nokia (ex. July 2015
Alcatel-Lucent)
ODU300 EoL announcement
March 2012
9500MXC EoL announcement
March 2012

9400AWY EoL announcement


March 2012
1353NM (EML) NR8.1 PL3 (EoL July
NMS 2014)
5620SAM (EML&NML) 14.0R7 Done
NSP (EML&NML) 17.6 Target Nov 2017
Mini-Link TN R5.4FP Done
R6.0 Target June 2017

Mini-Link LH (longhaul) 1.6FP Done


Ericsson Microwave 2.0 Target June 2017

Mini-Link CN510 R2.1 Done


outdoor PT6020/2020 R2.6 Done

Orange confidential - 78
outdoor 6363/6352 R2.5 Done
R.2.7 Target June 2017
SoEM & IPT NMS 16A R1B Done
16A R1D Target June 2017
Contact: Jakub Domin, Nokia & Ericsson Microwave TSC Leader

Vendor Type Equipment Release Validation status


RTN 900 series release
Microwave RTN905 1E/2E- RTN910A V100R800C10 Done
V100R900C10 Target June 2017
LH Full Indoor RTN980L V100R800C10 Done
RTN380 V1R6C00 Done
E-Band RTN380H (10Gbps) V1R6C00 Done
Huawei RTRN380H V1R8C00 Target Nov 2017
RTN310 V1R1C01SPC100 Done
All outdoor RTN320 V1R6C00 Done
RTN320 V1R8C00 Target Nov 2017

NMS U2000 V2R15C60 Done


V2R16C60 Target June 2017
iPasolink 100E R3.00.08 Done
iPasolink 200 R3.00.22 Done
R3.00.51 Mar 2016
iPasolink 400 R3.70.13 Done
R3.72.14 Mar 2016
iPasolink 1000 R3.51.09 Done
Microwave
VR 2 R 2.0 Done
R 4.0 Target June 2017
VR 4 R 2.0 Done
R 4.0 Target June 2017
VR 10 R2.0 Done
NEC R4.0 Target June 2017
Ex R3.0 HW f/w V 3.10.09 Done
Ex R3.0 HW f/w V 3.7 Done
E-Band
Ex Advanced (10Gbps) R1.0 Done
Ex Advanced (10Gbps) R2.0 Target Nov 2017
iX R1.0 f/w V1.0014 Done
All outdoor
iX R1.8 Target Nov 2017
PNMSj R19 Done

NMS MS5000 R11.1.12 Done


R12.02 Done
R12.06 Target June 2017
UNMS R2.1 Target Dec 2017
Contact: Maurizio Fazzi, Huawei & NEC Microwave TSC Leader

- 79 - Orange confidential
15.7 Microwave Spectrum Regulatory Condition
A database collecting information with “License fees” is under constant development and is available in the Share Point
under link http://shp.itn.ftgroup/sites/TIDonline/Shared%20Documents/Forms/AllItems.aspx .

Contact: Fryderyk Lewicki

16. Design to cost Coverage solutions for rural areas in emerging


countries
16.1 Technical Solution
The present recommendations rely on a theoretical study in order to provide trends for a radio network mobile
deployment in rural area.
Here are the typical use cases that were studied:
 Very isolated villages and isolated villages,
 Groups of spread villages and gathered villages,
 Mine or tourist spots,
 Road axis.
Given the geographic distribution of the operator’s settlement in AMEA, rural areas have been ruled out in a first step of
the network deployments.
The main reasons are:
 Reduced financing for the project, and prioritization of high density areas for a guaranteed return on investment.
 Environment hostility and complexity of the deployment of the solution: relief, high distance to the closest PoP of
the existing network, lack of energy.
 Requirement for a specific strategy in order to cut down the global cost particularly high regarding site infrastructure
(including energy): the purpose is to determine the most relevant balance between site deployments,
coverage/capacity target, and TCO.
The main technologies to provide the global solution has been considered: the solar energy and the pylon, the
microwave and the satellite, and 2G and 3G services.
Three main parts are contributing to the global cost (CAPEX and OPEX) of the rural solution:
 The site infrastructure: it is the most expensive item for a radio mobile network deployment. It includes the solar
energy (ORYX solution) and the pylon. Cost is highly linked to the height of the pylon. The civil engineering aspect
depends highly on the country e.g. logistic for transportation of the material, type of ground; Updated AMEA Site
guidelines have been published in April 2011 [RAN Technical SharePoint: Site Engineering].
 The radio access network (RAN): it includes the base station 2G (and 3G), the antenna and the feeders. The
capacity and coverage are the main drivers for the definition of the solution e.g. number of sites to be deployed,
location of the sites and the type of base station. The coverage is calculated roughly for flat terrain, and outdoor
coverage.
The frequency band is 900 MHz for GSM, and 900 or 2100 MHz for 3G (depending on spectrum availability).
 The backhaul last mile which aims at connecting the new target area to the existing network (backhaul, backbone).
o Comparison between Satellite and Microwave is based on the distance to the PoP and the volume of traffic
to transport. High volume of traffic compromises the satellite solution while long distance compromises the
microwave solution (mainly due to the infrastructure cost impact). For low capacity and long haul, satellite
backhaul is the best suited solution.
o Point to point Microwave solution provides the same data rates on up and downstream while satellite
solution allows supporting different data rates on up and downstream.
o Considering Microwave backhaul, low frequency is used for long haul (7 GHz) while higher frequencies (13
GHz even 23 GHz) are used for shorter haul.
o RAN and MW backhaul radio equipment are conditioning the minimal pylon height which is required to
guarantee the global announced performance.
o Considering Satellite backhaul, two types of orbits is used: stable or inclined (the cost of bandwidth in
inclined orbit is cheaper than that in stable orbit).
According to the business model, the technical approach of the rural coverage is different. Different business models for
rural coverage are:
 “Standard” model: the operator deploys and finances its own investments (high CAPEX, OPEX).
 “Site Sharing” model: the operator shares with another operator the passive infrastructure (pylon and energy). The
CAPEX and OPEX (per operator) of pylon and energy are reduced but the potential revenue will be also impacted
by the presence of the competition.
 “TowerCo” model: the operator rents the pylon to install its own RAN and backhaul (the energy is supplied by the
TowerCo).The CAPEX is reduced and is converted in OPEX (by a rent). The potential revenue could be impacted if
the TowerCo rents also the site to competing operator.

Orange confidential - 80
 Today other “alternative” solutions appear on the market of the rural coverage:
o These solutions take into account the rural site (infrastructure), the RAN (Pico or micro base station only),
the controller and the transmission by satellite (up to core network). These solutions address mainly the very
isolated villages and possibility the isolated villages (small coverage).
o The business model can be: “Full OPEX” with or without initial fees, “Revenue Sharing” or a combination of
both. These business models are proposed by consortiums with financial partners and are customized for a
considered country (they could be reused for others countries or not). The rural site, RAN and the backhaul
are belonging, deployed and operated by the consortium.
o Depending on the cost level, the risk could be limited (no CAPEX and sometimes no OPEX).
o As these solutions are not referenced in the catalogue, they will have to be validated. The technical solution
is in the hands of the consortium. Specific attention shall be paid to contractual constraints (e.g. minimum
number of sites, capacity evolution, and contract duration). SLA shall be present in the contract and reflect
Orange affiliate requirements in terms of availability, performance (throughput, coverage) for the provided
rural sites.

16.2 Geo-Marketing
A Geo-marketing approach has been defined to help emerging countries planning their network deployments in rural
areas. It is founded on a five steps methodology:
o Structuration of a relevant data base with the country,
o Identification of uncovered areas,
o Identification of addressable population,
o Mapping of the studied typical use cases with addressable population in the country thanks to an optimized
algorithm,
o Possible addition of marketing strategic criteria to refine prioritization of the areas proposed to be covered.

The idea is to propose a list of potential new sites which can be prioritized by key drivers, in particular topology,
population to be covered and remoteness from existing network,

16.3 Tools
Two tools are being designed to help rural new deployments decision-making by providing sites prioritization and
technical and economical best choices.
o The Geo-marketing tool is based on a geographic information system (today, based on MapInfo proprietary
solution). It allows identifying highest potential sites to be deployed in rural areas (either for isolated largest
villages or populated groups of villages) and provides a sites prioritization taking into account key drivers.
Expecting to be automated in 2017, this tool is available on request as expert support.
o The “Design-To-Cost” tool on the Web interface is under study for a secured access. This tool provides the
complete estimation costs of the rural site, the complete cost comparisons between the microwave and the
satellite for the typical use cases or a defined specific use case in AMEA rural area.

16.4 Recommendations
The recommendations consist in guidelines in order to generate the most adapted solution for radio access network,
backhaul and site engineering aspects at the lowest global cost.
The present recommendation concerns technical solutions/suppliers already included in our catalogue for each part
(Pylon and solar Energy, RAN and Backhaul).
 The recommendations are given for different use cases within rural areas: isolated villages, group of villages,
industrial or tourist spot, road axis.
 In a first step multiple technical alternatives have been considered. The recommendations of the present study are
mainly based on an economic quotation of these alternatives.

Solutions for rural areas in AMEA: Group Recommendation 16-1


 The location of sites and the number of sites must be optimized in order to enable (and limit) the rural coverage
of a village or a group of villages.
 The top height for RAN antennas is 60 meters. Indeed over this limit the reception of the radio signal cannot be
guaranteed at the bottom of the pylon.
 This height has then to be challenged with constraints related to the deployment of the backhaul technology,
which are generally prevailing.
 For the backhaul the required height of the pylon depends on the technology to be deployed:
o In case of microwave: depending on the distance to the nearest PoP of the existing network, there is a
minimal required height of the pylon to ensure a reliable transmission link as a function of relief, ground

- 81 - Orange confidential
cover, environmental conditions... If this calculated height of the pylon exceeds the value calculated for
the one RAN part, then it will be the final height of the pylon.
o In case of satellite: there is no constraint on the height of the pylon which will be determined by the RAN
equipment constraints.
 The civil engineering should take into account the ground requirements:
o Energy supply should preferably rely on renewable technologies and specifically solar energy through
ORYX solution.
o Energy saving features should be activated if possible and available: recommended functionalities are
(1)
listed in the document RAN Power consumption strategic plan .
 In case of high CAPEX constraints, alternative solutions (“Full OPEX”, “Revenue Sharing”) can be ; in this case
the infrastructure, RAN and backhaul are provided, deployed and operated by one consortium, the contract must
be deeply studied (number of sites, performance, capacity evolution, duration clauses) in order to ensure a good
Quality of Service. These solutions address mainly the very isolated villages and are customized for the
considered country. Today, none of these solutions are referenced in the catalogue.

16.4.1 Isolated villages


This refers to isolated village with a low traffic and located near (< 50 km) or far (≥ 50 km) from existing network PoP.

Isolated villages: Group Recommendation 16-2


 The location of sites should be optimized in order to enable (and limit) the coverage of the village with a single
site.
 The capacity should not be over sized but should be upgradable for a potential growing village.
 For a small village in terms of traffic and very isolated i.e. located above 50 km from the closest PoP of the
existing network,
o A 2G only coverage should be considered.
o A micro base station should be deployed. The base station transmission power should be configured by
adapting its level to the village area to be covered.
o The mobile backhaul network should rely on satellite type of connection for an expecting traffic below 4
(2)
Mbps (3 Mbps downstream and 1 Mbps upstream as example) on the transmission link.
(2)
However, if the expected data rate on the transmission link exceeds 4 Mbps , then microwave alternative
should be studied and compared to satellite based solution.
 For a larger village, not far from the closest PoP of the existing network (below 50 km),
o A 2G/3G coverage should be deployed.
o A distributed base station should be deployed. The base station transmission power should be configured
by adapting its level to the village area to be covered.
o For a distance to the PoP of the existing network between 35 km and 50 km, and an expecting traffic
below 2 Mbps (1.5 Mbps downstream and 0.5 Mbps upstream) on the transmission link, the mobile
backhaul network should rely on satellite type of connection. Below 35 km or above 2 Mbps, the
microwave alternative should be studied and compared to satellite based solution.

16.4.2 Group of villages


This refers to group of villages located near (< 50 km) from existing network PoP.
The principle is to determine the most adapted solution for radio mobile network coverage of an area handling a set of
villages. These villages are of different importance with major villages surrounded with secondary villages and hamlets.
2 different cases can be considered:
o Villages are gathered within a limited area, and inter village distance does not exceed 10 km.
o Villages are spread and located within a bigger area. The inter-village distance could be up to 15 km.
 The determining factor for the deployment of mobile network is the number of sites; this is linked directly to the cost
of the civil engineering for the pylon to be installed.
o Thus the most adapted solution should imply the deployment of a couple (2-3) sites only in order to reach
the lowest cost.
o The radio mobile network deployment solution has to be designed as a global solution for the area to be
covered and not village by village to be covered.
 The coverage can be either partial (for 3G) or complete (for 2G). This depends on the deployment strategy i.e.
coverage of hamlets or inter village area or focus on the major villages.

1
Reference to the chapter 17 of NTG document.
2
Data rate is strongly linked to hypothesis which are detailed in [Reach/ Design-to-Cost – Word document]

Orange confidential - 82
Group of villages: Group Recommendation 16-3
 The location of sites and the number of sites should be optimized regarding the position of the major villages in
order to provide to the most part of the population the best quality of service.
 The services based on 3G technology should - as a priority - target the main villages, and thus are driving the
location of the sites within the area to be covered.
 Considering a configuration with gathered villages (inter village distance less than 10 kilometres, and an area to
be covered of ~ less than 700 km²):
o A single site deployment is recommended with a high transmission power.
o The microwave is the recommended technology on the transmission link.
o RAN solution should be upgradable as the capacity may concern a couple of villages and the traffic
forecast may potentially increase.
o A distributed base station should be deployed. A macro base station can also be considered.
 Considering a configuration with widespread villages (inter village distance less than 15 kilometres, and an area
to be covered of ~ 1600 km²):
o For an area to be covered over than 35 km sidelong, a single site solution can’t be considered. A multi-
sites solution should be deployed.
o A multi sites deployment is recommended with a 2 sites configuration.
o The microwave is the recommended technology on the transmission link.
o The highest pylon (deployed to reach the next PoP through microwave) should be located where radio
coverage requires a high position of the RF RAN equipment.
o RAN solution should be upgradable as the capacity may concern a couple of villages and the traffic
forecast may potentially increase.
o A distributed base station should be deployed to answer to capacity and upgradability constraints.

16.4.3 Hot spots e.g. mine, tourist location.


This refers to isolated industrial site (e.g. mine or forestry site) or tourist location (e.g. lodges) located far (≥ 50 km) from
existing network PoP.
The purpose is to provide a radio mobile network in a spot location where:
 The capacity needs are significant as the targeted market segment consists in business men or tourists. Added
value services are expected with 3G.
 The coverage area is limited to a couple of km², and should not drive the choice of the solution.
 The site location is isolated.

Mine or tourist locations: Group Recommendation 16-4


 The location of sites should be optimized in order to enable the coverage of the spot with a single site. As the
target area to be covered is not widespread, a small site infrastructure should be considered taking into account
environment constraints.
 The 3G based services should be available within all the covered area.
 As the traffic forecast may increase the solution capacity should be upgradable and the capacity of the
installation should be sized largely.
 A distributed BTS should be deployed to answer to capacity and upgradability constraints.

(2)
The mobile backhaul network should rely on satellite type of connection for an expecting traffic below 4 Mbps
(downstream) on the transmission link.
(2)
However, if the expected data rate on the transmission link exceeds 4 Mbps , then microwave alternative
should be studied and compared to satellite based solution.

16.4.4 Road Axis


This refers to a road axis between an existing network PoP and a crossroad without considering other traffic than the one
from road users.
Due to regulatory constraints depending on the considered country, road axes have to be covered.
 The coverage of the target zone is the main purpose of the solution to be deployed.
 The capacity is not an issue for this use case.
Local environment constraints have to be taken into account case by case; the present recommendations do not assume
specific obstacle (e.g. hill) on the road axis.

Road Axis: Group Recommendation 16-5


 2G only based solution should be deployed.

- 83 - Orange confidential
A distributed BTS should be deployed to answer to capacity and upgradability constraints:
 In order to enable the road axis coverage, sites will be chained all along the road axis. The distance between the
sites should be optimized in order to balance the number of sites and the height of the pylons to be deployed,
and to minimize the number of sites.
 Considering only the road axis coverage the microwave is the recommended solution for the transmission link
mainly if any village or an interesting place is located along the road axis (these facilities should be integrated
into the design of the solution – in this case, the capacity to be transported includes the neighbouring villages’
traffic). However in case the ground profile raises some relief issues then the satellite alternative should be
considered.

Contact: Serge Langouët, Patrice Desmoulin, François Brunet, Tomasz Rudy

17. Green RAN


17.1 Background
Orange has taken on commitments towards the financial community, regulatory bodies and our customers to reduce our
carbon emissions and decrease significantly our energy consumption while improving our customer experience due to
sustainable and reliable energy.
Over the last years, within ITN perimeter of consolidated Orange countries, we have been facing a huge increase of
energy consumption and spending, due to usage booming, growth of our customer base and energy costs rise.
In this context and taken into account that RAN equipment’s power consumption are for most of affiliates one major
contributor to ITN energy OPEX, several initiatives are launched or extended in 2017:
 Measure power consumption for each new hardware introduced in Orange networks
 Push energy metering embedded in RAN products (available for Huawei, to be deployed with Nokia through
Valkyrie/Mosaic renewals, and introduced from 2017 for Ericsson with Radio22 new generation)
 Deploy Energy Saving Features available in RAN suppliers portfolio (5 to 10% energy savings can be reached
according to trials performed with Huawei in Moldova & Morocco)
 Use C-SON Green features to automate the best trade-off at cell level between QoS and energy.
 Develop 5G advanced sleep modes in 3GPP standard bodies

17.2 Energy Savings Features


The table below provides a “per RAN Supplier” recommendations of Energy Savings Features available in RAN
supplier’s portfolio. It is important to notice that these features implementation should be studied on a case by case basis
based on HW compatibility (MCPA vs. SCPA …), network configuration (GSM900/1800 colocation …) and business
cases (as these features are optional and might trig a specific commercial fee).

Energy Savings Features – Per Supplier recommendations


 Former Alcatel-Lucent (now Nokia)

Validation status & energy savings Comments


2G
Compatible Hardware : SCPA (Legacy TRX & Twin
TRX dynamic power saving (DPS) Green Light granted (10% savings)
TRX)
Compatible Hardware : MCTRX, MC-RRH, TRDU2x60-
Dynamic Power Voltage Adjustment (DPVA) Green Light granted (10% savings)
09 & TRDU2x60-18
3G
Green Light targeted in LR15.3W in Q2-17 Compatible Hardware : TRDU2x80-21, TRDU2x60-09,
PA bias
Baseline feature RRH60-21C, RRH2x60-21/A, RRH2x40-09
Green Light in LR13.3W (8% savings) LR13.3W: applicable on 2 PA
Automatic Carrier Shutdown (ACS)
Green Light targeted in LR15.3W in Q2-17 LR15.3W: applicable on single PA
Autonomous modem power down upon ACS Green Light targeted in LR15.3W in Q2-17
4G
Dynamic PA Switching Green Light granted (7-16% savings per RF module) Compatible hardware: RRH-800-A & RRH2600-4R-A
Automatic Cell & Carrier Shutdown Green Light targeted in LR16.2L in Q3-17

 Ericsson

Orange confidential - 84
Validation status & energy savings Comments
2G
Green Light granted (5-15% savings) but issue Additional tests planned in OFR in Q2-17 to
BCCH power saving
faced during roll-out in OFR understand traffic loss seen during roll-out
improvement in G16B to support Mixed Mode
MCPA TX Power saving Green Light granted (5-15% savings)
configuration
3G
Green Light granted (3-5% savings) but issue Investigation on-going in OSP and additional tests
Traffic Aware Power Save
faced during roll-out in OSP planned in OFR in Q2-17
4G
Micro Sleep Tx Green Light granted (10% saving) HW constraints for 4G (RUS02/RRUS12 or Radio
RRUS12, not compatible with mixed-mode Highly recommended for roll-out 2217, not compatible with mixed-mode)

Cell Sleep mode To be tested in OFR & OSP in Q2-17

Improvement planned in L17 Q1 (validation not yet


MIMO Sleep mode Not recommended in L16B (service interruption)
scheduled)

 Huawei
Validation status & energy savings Comments
2G
Enhanced BCCH Power Consumption
Green Light granted (1,2%-4% savings)
Optimization
Multi-Carrier Intelligent Voltage Regulation Green Light granted (2-6% savings)
SDCCH congestion seen in few networks (not
Dynamic Cell Power Off Green Light granted (6-7% savings)
systematic
DBS3900/BTS3900AL/BTS3900A with APM30 and
PSU Smart Control Green Light granted (1% savings)
battery (applies for 2G/3G/4G)
TRX PA Intelligent Shutdown Green Light granted (3% savings) if BCCH TRX in baseband frequency hopping
TRX Power Amplifier Intelligent Shutdown on
Green Light granted (6% savings) old double transceivers (MTRU,DRRU,DRFU)
Timeslot Level
3G
Features not compatible
Multi-Carrier Switch off Based on Traffic Load
Green Light granted (1-4% savings) Multi-Carrier Switch off based on QoS expected to
Multi-Carrier Switch off Based on QoS provide better savings
Energy Efficiency Improved by PA voltage
Green Light granted (1-2% savings)
adjustment
DBS3900/BTS3900AL/BTS3900A with APM30 and
Intelligent Power Management Green Light granted (1% savings)
battery. (applies for 2G/3G/4G)
RRU PA Efficiency Improvement Green Light granted (8% savings) Only applicable to RRU 3838
4G
Symbol Power Saving Green Light granted (10-20% savings)
Green Light granted (up to 5% savings after
RF Channel Intelligent Shutdown
optimization)
Intelligent Power-Off of Carriers in the Same Green Light granted (up to 5% savings after
Coverage optimization)
Two modes available for this feature
Adaptive Power Consumption Green Light granted (1.8% savings) A. 1,8% gain for Adaptive power adjustment mode
B. 13-16% gain for cell regular sleep mode
DBS3900/BTS3900AL/BTS3900A with APM30 and
PSU Intelligent Sleep Mode Green Light granted (1% savings)
battery (applies for 2G/3G/4G)
RRU PA Efficiency Improvement Green Light granted (1% savings) Only applicable to RRU 3268 2.6G

 Nokia
Validation status & energy savings Comments
2G
Intelligent MCPA TRX shutdown Validation in SRAN16.10 planned in OFR in 2017
BCCH TRX Energy Saving Mode Validation in SRAN16.10 planned in OFR in 2017
Energy Optimized TCH Allocation Validation in SRAN16.10 planned in OFR in 2017
3G
Power Saving Mode for BTS Validation in SRAN16.10 planned in OFR in 2017
4G
Validation in SRAN16.10 planned in OFR in H2-17
Power saving Micro DTX (symbol shutdown) Lab tests on-going in OEG
Validation in SRAT16 planned in OEG in Q2-17
Power saving - MIMO switch off Validation in SRAN16.10 planned in OFR in H2-17
Lab tests on-going in OEG
Load-based Power Saving with Tx Path switching off Validation in SRAT16 planned in OEG in Q2-17
Power saving – multiLayer N. with cell switch off Validation in SRAN16.10 planned in OFR in 2017

Contact: Yann Pitrel.

- 85 - Orange confidential
18. On the road to 5G
5G is the next generation of mobile cellular technology, which is expected to be deployed around 2020. It will not only
provide mobile data/broadband access with higher data rates and increased capacity, but also provide opportunities for
supporting new innovative (machine-type) services and above all, a flexible software-based way of deploying, operating
and managing the networks. This section provides an overview the 5G landscape in terms of services, requirements,
standardization, radio access technology, spectrum, regulatory aspects and the global initiatives together with the
associated timelines.

18.1 5G timeline
The 5G ecosystem is developing fast through: (1) standardization in 3GPP, (2) tests, demonstrations and trials by the
industrial actors, (3) collaborative research (4) discussions in various industrial fora (NGMN, GSMA, 5GAA etc.) and (4)
regulation (ITU, CEPT, FCC, etc.).
ITU-R, the worldwide regulatory body, is setting the scene by (1) identifying the 5G candidate frequency bands that will
be rendered official in the World Radio Conference (WRC) of 2019, and by (2) fixing the main requirements of 5G this
year which will serve as a benchmark to evaluate the candidate technologies after 2019. The details of the spectrum
aspects can be found in the dedicated section below.
3GPP, the main standardization body for mobile networks, is working on a candidate technology at a fast pace and
driving the timeline for 5G, with 2 main phases planned for delivery. The associated delivery dates are as follows:
 Dec 2017 - Specification of “Pre-phase 1” : Dual connectivity New Radio (NR) with 4G and 4G Core
 June 2018 - Specifications of “Phase 1” : Standalone NR for early deployments and 5G Core
 Dec 2019 - Specifications of “Phase 2” : “Complete” 5G
The details of these three deliveries can be found in the dedicated section below.
Pre-phase 1 is the version available for commercial launch starting in mid-2019, Phase 2 in 2020 and Phase 2 in 2021-
2022. Several operators have already announced their 5G deployments. Among them are:
 Verizon and AT&T in the USA, who will pilot Fixed Wireless Access in 2017. Verizon will pilot 5G FWA in the 11
cities on the 28/39GHz band whereas AT&T’s deployment will be in the 2 pilot cities on the 39GHz band is
available in the USA.
 Korea Telecom (KT) who will trial a “5G pilot” during the winter Olympics of Feb.2018.
 Japanese operators (DoCoMo and KDDI) who will deploy phase 1 NSA 5G in 2020 (and showcase the new
technology during the summer Olympics the same year).
 Orange has announced that the 5G deployments will start from 2020.
With a 5G ecosystem evolving according to this pace set by 3GPP, more and more trial announcements are being made
by different actors worldwide. For Orange, our current intentions include the following trials:
 An E2E large scale 5G field trial in 2018 in France (followed by RFI and RFP)
 5G pilot zones in Orange European countries during the Euro2020 World Cup
Possibility of having additional trials is open and depends on the vendor equipment roadmaps.

Figure 37 – 5G Timeline

Orange confidential - 86
18.2 What services will 5G offer?
5G aims to offer a wide variety of services that can be grouped into 3 classes:

18.2.1 eMBB
5G will offer enhanced Mobile BroadBand (eMBB) services, which is an evolution of the mobile data services
offered by 4G. The use cases envisioned so far, e.g. pervasive video, smart offices, HD Video/Photo Sharing
in Stadium/Open-Air Gathering, remote computing…, will come with new or largely improved Key
Performances Indicators in 5G, compared to 4G, like high data rates, high user density, high user mobility
(trains, airplanes…), highly variable data rates, deployment, and coverage (arenas, transports, indoors/outdoors…).
3GPP has already proposed a set of KPI and associated requirements for high data rate and traffic density scenarios as
depicted in Table 7.1-1 of 22261-f00: 3GPP TS 22.261 V15.0.0 (2017-03)
These service-level requirements can be translated in the following requirements from a RAN level point of view (3GPP
TR 36.913):
- minimum throughput: 50 Mb/s “everywhere” (cell edge)
- higher capacity: spectral efficiency. x10 compared to LTE-A
- higher peak rate: 10 Gbps in higher frequency (so called cm/mm Waves); 1Gbps in 3.5/3.7 GHz; 2Gbps with
4G+5G aggregation.
In parallel, ITU has started to communicate on early key requirements (and the definition of the related KPI) related to
the minimum technical performance of IMT-2020 candidate radio interface technologies in DRAFT NEW REPORT ITU-R
M.[IMT-2020.TECH PERF REQ] . The work in ITU will be completed in Nov. 2017.

18.2.2 URLLC
Ultra Reliable and Low Latency Communications (URLLC), also called critical applications, require ultra-reliability, very
low latency and very high availability. In other words, the 5G network should provide full reliability and high availability
with a real-time responsiveness. URLLC services are expected to expand into new market segments to facilitate the
digital economy in areas such as energy, e-health, automobile, Multimedia Priority Service (MPS) – safety and security,
factory 4.0 and advanced communications. In each of these areas many use cases have been identified but it is sure that
many more will arise once URLLC technology will be available. To give some examples:
- Energy: control systems for smart grids.
 eHealth: tele surgery, vital sign monitoring, pre hospital emergency.
 Automobile: platooning, see-through, autonomous driving, collision avoidance, road hazard warning.
 MPS – safety and security: assistance on disaster situations.
 Factory 4.0: Process and factory automation, manufacturing in hazardous environments, inspection and
maintenance with aerial robots.
 Advanced communications: Virtual Reality, Advance Reality, Gaming with haptic return.
The requirements in terms of latency and reliability of each one of these use cases will be characterized by the nature of
the communication i.e., tactile internet, control and automation or teleoperation. To give an order of magnitude, latency
-3 -9
will go from 1 ms to 20 ms and reliability from a Block Error Rate (BLER) of 10 to 10 .

18.2.3 mIoT
3GPP started a massive effort during the release 13 cycles to greatly improve the way IoT connectivity is handled by
mobile network. It focused mostly on enabling mobile network operator to compete with non-3GPP LPWA operators like
SigFox or LoRa and consequently targeted mostly delay-tolerant and low volume IoT applications (smart metering,
environmental monitoring). These efforts have leaded to: (1) the specification of a new radio platform dedicated to MTC
traffic (NB-IoT), (2) the addition to the LTE platform of a specific IoT capacity with a dedicated feature (LTE-M – coverage
extension) and (3) a new low-cost & low-power device class (Cat-M1). Next, release 14 and 15 are progressively
improving NB-IoT and LTE-M by additional features (positioning), capacity and flexibility. LTE-M is especially trying to
widen is application field.
Concerning 5G, there is a consensus within 3GPP that, as pre-5G releases do provide solutions for the IoT that are just
hitting the market right now and are far from overloaded, it would be premature and even counterproductive to start the
standardisation process for a new Massive IoT (mIoT) feature in 5G now. Rel-15 LTE-M and NB-IoT will both evolve with
the aim of meeting the 5G requirements. Therefore discussions on a new Massive IoT access technology in 5G have
been postponed to 5G Phase 2. It can be assumed that, in the next coming years, including first years of the 5G era,
3GPP will rather focus on continuing to improve existing IoT technology without breaking the backward compatibility.
However, as both NB-IoT and LTE-M concepts are tightly bound to the assumption that MTC traffic is delay-tolerant and
of low volume at the same time, intrinsic limitations will remain. We can recall for example that 5G requirements still rely
on Cellular-IoT traffic model defined by GERAN in 2014 and assuming a payload of 50 to 200 bytes sent once a day to
once every two hours, with an acceptable latency of 10 seconds (for alarm traffic). However, Massive IoT in 5G is a
concept that goes beyond just the number of low-power, low-cost and low-traffic connected devices. “Massive” means
also a massive number of uses cases and application and some of them will most probably imply shorter latency and/or
higher traffic volume. Answering such needs would be the purpose of 5G mIoT. It is unclear if LTE-M or NB-IoT can be
evolved that way or if a specific 5G-IoT connectivity will be needed. However some the required technical blocks, such
as NOMA (Non Orthogonal Multiple Acces) are already part of 5G Phase 1.

- 87 - Orange confidential
18.2.4 FWA
Fixed Wireless Access (FWA) is the first category of 5G services envisioned for deployment. FWA includes the delivery
of services to customer’s premises: TV (with the generalization of 4K definition), internet access, cloud-based services,
VoIP… FWA can also be an interesting option for B2B use cases, delivering high throughput to enterprises, provided
services can be guaranteed.
With the upcoming 5G, wireless connectivity can be competitive in terms of throughputs, in comparison with fiber (or
other wireline solutions), thanks to the capacity increase enabled by this new generation, mainly obtained via the use of
high frequency bands (a.k.a. mm- or cm-wave bands). This part of the radio spectrum is highly sensible to obstacles and
vegetation; Line-of-Sight (or near Line-of-Sight) conditions are usually required, what impacts the types and positions of
access points and CPEs (Customer Premises Equipment). The cost of 5G equipment in such bands and the
densification of radio sites (needed to improve coverage) require a case-by-case analysis to assess the viability of this
use case in a 5G perspective.
Despite these challenges, US and Asian (Korea, Japan) operators are considering 5G FWA deployments, for mostly
urban and suburban environments, as fiber has not been massively rolled out in these countries.

18.3 What will be the 5G radio access technology?


3GPP is currently defining the 5G radio access technology, with 2 phases, whose details can be found below:
 5G Phase 1 (June 18, early drop Dec 17)
While the full “5G phase 1” specification will be delivered as initially planned in June 2018, an earlier drop will be made
available as soon as Dec 2017 following pressure of operators in a hurry to deploy commercial services, mostly in the US
(AT&T) and Asia (NTT DoCoMo). This earlier version will be focused on eMBB with “Non-Stand-Alone” (NSA)
architecture, meaning dual connectivity of LTE and NR 5G simultaneously, and connectivity to the legacy EPC. FWA and
Low Latency communications will also be supported by this Phase 1 “early drop”.
The full phase 1 release will complement the early drop with specification of the Next Gen Core Network (NG-CN),
completion of all architecture variants in NSA and SA (Stand-Alone) modes, and the addition of Ultra Reliability to Low
Latency Communication to provide (URLLC) services on top of eMBB, which has been strongly supported by Orange to
develop the vertical market in Europe. Major innovations brought by the 5G phase 1 release include a “New Radio” (NR)
interface, new RAN architecture options, and new spectrum bands and techniques.
The New Radio will be based on new channel coding and waveforms, more Energy Efficient features allowing for
“Advanced Sleep Mode” in base stations (e.g. sleep duration of 20 ms in 5G vs. 1 ms in 4G) leading up to 80% power
saving at low load, very low latency below 1 ms, ultra-reliability up to 99,999%.
New spectrum bands should initially cover the 3.4 – 3.8 GHz range, a 5G version of the 700 MHz band, and several
mmWave bands including the 24.25 – 27.5 GHz band targeted in Europe. Besides, large bandwidths (e.g. 100 MHz for
the 3.4-3.8 GHz range) will be possible. For the first time ever it will be possible to aggregate the new radio with the old
(i.e. 5G bands with legacy 4G bands), thanks to the dual-connectivity functionality, ensuring higher throughputs to 5G
users compared to 4G.
Novel spectrum management techniques could also allow for flexible re-farming of 4G towards 5G, with a tight co-
existence of 4G and 5G users within the same band plan, avoiding the need to dedicate spectrum blocks (e.g. 10 MHz
each) to 4G and 5G. Pairing of low UL band with high DL band is also considered, which could lead to “hybrid bands” in
5G such as “UL 800 + DL 3500”.
On the RAN architecture front, Orange has strongly pushed for an open midhaul concept, with a new functional split
between “centralised units” (e.g. in NG-POP) hosting Virtualised Base-Band-Unit (vBBU) and “decentralised units” (e.g.
on radio sites) maintained for “low BBU”. Such interface, named “midhaul” within Orange, as a middle ground between
traditional backhaul and fronthaul transport interfaces, will be based on a PDCP / RLC split, and aimed to be “open” and
allow multi-vendor deployments. A work item is now under progress in 5G, while its 4G equivalent is now being studied
for specification within the same 3GPP release as 5G.
Orange is also pushing for extreme long range communications as a basis for possible low-cost deployment solutions for
large rural areas (typically African countries) and also energy efficiency as a means to optimize our OPEX.
 5G Phase 2 (Dec 19)
Beyond 5G phase 1, 3GPP has also agreed a package of Study Items for Rel15, which will pave the way for the delivery
of 5G phase 2 planned for Rel16 and due by December 2019.
Notable additions include a study on “NR unlicensed”, which should allow 5G unlicensed base stations in stand-alone
mode, i.e. not requiring any licensed band. The 5 GHz Wi-Fi band is considered, as well other unlicensed spectrum
bands in mmWave (37, 60 GHz). This is primarily aimed at serving the B2B market, where customers may by-pass
mobile network operators and source infrastructure directly from RAN vendors. Although this represents a clear threat for
Orange, opportunities may also arise in a similar fashion as Wi-Fi and LoRA.
Other topics studied for 5G phase 2 include Non-Orthogonal-Multiple-Access (which could enable massive MTC), 5G for
satellite & aerial connectivity, integrated wireless backhaul, which will optimise the use of access and backhaul within the
same cell, and further Cellular V2x evolution.

5G radio access technology enablers, as currently under definition in 3GPP, can be summarized as follows:
1. Flexible numerology, notably the Sub Carrier Spacing (SCS) can range from 15 kHz to 120 kHz [240 kHz]. A high
subcarrier spacing means a low OFDM symbol duration which is well adapted for high mobility or/and high frequency
mmW (phase 1)

Orange confidential - 88
a. Note that the LTE numerology is part of NR which allows optimized co-existence with LTE
2. New waveforms based on filtered CP-OFDM allows a better co-existence between numerologies (phase 1)
a. Note that the waveform is spec transparent and can be received with a legacy LTE receiver which allows
optimized coexistence with LTE
3. One symbol mini slot duration that can occur in any OFDM symbols of a slot (7-14 symbols), i.e., URLLC
PDSCH/PUSCH can last 1 OFDM symbol (phase 1)
4. Fast retransmission with configurable (PHY layer/semi-static) timing relationship between PDSCH and ACK, UL
Grant and PUSCH at UE (phase 1)
5. Flexible TDD configurations where the DL and UL switching point can be semi-statically or dynamically chosen per
slot (downlink only, uplink only, DL (phase 1)
a. Dynamic TDD is complex to handle due to interlink-interference, low bands necessitates coordination between
cells
6. Cross link interference mitigation between TDD cells (phase 2)
7. URLLC and eMBB multiplexing capability (phase 1)
a. Downlink: the URLLC service can pre-empt an ongoing eMBB transmission, the impacted radio resources can be
indicated to the eMBB UE and retransmission can be performed on the transport block that have been preempted
b. Uplink : grant free transmission, i.e., URLLC user have dedicated UL resources where they can transmit without
waiting for an UL grant
8. MIMO toolbox which is a superset of the MIMO techniques developed so far in LTE until Rel. 14 (Full Dimensional
MIMO) (phase 1)
a. Below 6GHz the massive MIMO techniques (64T64R) may be very similar to the ones developed in FD MIMO
Rel. 13 and 14 where the digital precoding is performed both in the elevation and azimuth domain, the number of
antenna ports is significantly increased (up to 32) and the antenna port to TXRU (RF chain) virtualization is
possible by precoding the CSI-RS
b. Above 6GHz: beam management will allow to beamformed also the broadcast channel (PSS/SS/PBCH) so that
narrow beams can fight the severe mmW path loss.
c. In TDD full channel reciprocity exploitation and beam correspondence are building blocks of 5G MIMO
9. Non Orthogonal multiple access will allow to optimize the capacity of eMBB, URLLC and mMTC. From an mMTC
point of view it will allow efficient contention based access which significantly decreases the control signaling (phase
2)
10. Stand-alone unlicensed NR, this was a very controversial topics but was finally accepted (phase 2)

URLLC services have to coexist with the other 5G services which have different requirements. To allow an optimal
service multiplexing the concept of slicing is being developed.
Then, as the name of this group of 5G services implies, URLLC communications have two essential requirements, i.e., a
very low latency and a high reliability. To achieve these demanding constraints, the following techniques are being
investigated.
Low latency enablers: Short TTI, eTC, fast HARQ feedback scheduling puncturing, dynamic TDD, resource reservation
for E2E path, handover mechanisms for low latency and Mobile Edge Computing (MEC) which is becoming an important
enabler for applications that demand low-latency operations and potentially improved quality of service.
High reliability enablers: LDCP and polar codes, Multi-connectivity, MIMO for spatial diversity, large subcarrier spacing
for vehicular applications to face the Doppler effect.
One of the most distinguishing characteristics of 5G radio access is its being partially virtualized. The virtualized RAN
architecture defined for the 4G RAN is fully compatible for 5G (see Radio Access Network Architecture Evolution)
and all benefits associated to the RAN virtualization will fully benefit to 5G, especially to manage the load induced by new
services and slicing.
At this stage, major RAN suppliers have traditional and virtualized RAN architectures in their roadmaps.
Affiliates will have to select the appropriate architecture according to the planning of 5G deployment and especially the
availability of the virtualized infrastructure, which could be provided by the supplier (AirFrame for Nokia for example) or
specified by Orange (NGPoP), which is Orange recommendation.

18.4 Which spectrum for 5G?

18.4.1 Regulatory aspects


 Additional spectrum for IMT to be identified at WRC-19
ITU-R has started successfully the preparatory process for WRC-19 Agenda Item 1.13 with the completion of the
document on <Spectrum needs for the terrestrial component of IMT in the frequency range between 24.25 GHz and 86
GHz> and the document <Characteristics of terrestrial IMT systems for frequency sharing/interference analyses in the
frequency range between 24.25 GHz and 86 GHz”>. They should be used for sharing and compatibility study on the
candidate bands for IMT-2020 under WRC-19 AI1.13. The agreed IMT-2020 parameters are based on the proposals
from mobile industry (3GPP, GSMA, etc).

- 89 - Orange confidential
There is no prioritisation at this stage of the candidate bands under studies at ITU-R. The whole range between 24.25
GHz and 86 GHz will be studied with the following split:
 24.25-27.5GHz, 31.8-33.4GHz.
 37-40.5 GHz, 40.5-42.5 GHz, 42.5-43.5 GHz, 45.5-47 GHz, 47-7.2 GHz, 47.2-50.2 GHz, 50.4-52.6 GHz.
 66-76 GHz, 81-86 GHz.
The sharing and compatibility studies are planned to be finalised in January 2018.
 CEPT
CEPT has commenced to prepare regulatory framework for successful 5G deployment in Europe according to the EC
Mandate on “5G pioneer bands” and according to the actions defined in “CEPT roadmap for 5G”. This includes the work
on technical conditions and availability of the 3400-3800 MHz band for 5G as well as on the guidance to administrations
for defragmenting the 3400-3800 MHz band. The work on this “5G pioneer band” should be completed by June 2018.
In parallel, CEPT has started the preparatory process for WRC-19 Agenda Item 1.13 with the aim to obtain identification
of harmonised spectrum for IMT. CEPT develops the regulatory framework for the second “5G pioneer band” with the
intention to harmonise the 26 GHz band in Europe for 5G before WRC-19 and thus influence other Regions’ decisions at
WRC-19. In addition to the 26 GHz band, CEPT has prioritised the 31.8-33.4 GHz and 40.5-43.5 GHz bands for studies
to prepare the Conference. CEPT is also preparing the European Common Position (ECP) to be used as the basis for
negotiations at ITU World Radio Conference in 2019. Orange is contributing to this work as has been and is still
supporting 5G deployments in 3.4-3.8 GHz, 700 MHz, 26 GHz, 32 GHz and 42 GHz (long term).

18.4.2 Technology aspects


 5G spectrum in 3GPP:
Licensed spectrum:
3GPP allows for flexibility to introduce NR bands in a release independent manner (from Rel-15) at every RAN plenary. A
large number of NR bands (both sub-6GHz and above 24GHz) have been proposed so far as part of phase 1, but no
particular frequency band is explicitly prioritized over the others as the specification work is contribution driven. However,
the following frequency ranges have support from a large number of operators from different regions and it is expected
that their specification would be finalized at an earlier stage (pre phase 1) motivated by commercial deployments.
 3.3 – 4.2 GHz range (Asia + Europe): either a single band 3.3-4.2 GHz or 2 bands (3.3-3.8 GHz & 3.6-4.2 GHz)
 24.25-29.5 GHz (Asia & US + Europe): single band has been excluded due to technical issues. Two options are
under consideration:
o Two bands: 24.25 - 27.5GHz & 26.5 - 29.5GHz
o Three bands: 24.25 - 27.5GHz & 26.5 - 29.5 GHz & 27.5 GHz–28.35 GHz. The latter was pushed by Verizon for
US operation.
 37-40 GHz (US)
 4.4-4.99 GHz (Asia)
Besides 3.4-3.8 GHz and 24.25-27.5 GHz included in the above ranges, Orange is also interested in 700 MHz which is
supported by other European operators (except DT) and Huawei, but not so far prioritized by other BS vendors and
UE/chipset makers. Lobby is on-going towards vendors to speed-up standardization and also product roadmaps to
support early trials of 5G in 700 MHz within Europe.
No particular band has been included in the frequency range 6-24GHz, mainly because of lack of regulatory visibility on
this range. However, this is not excluded from the scope of Rel-15.
Unlicensed spectrum:
5G operation in unlicensed spectrum would be possible in phase 2 (Rel-16) in combination with and under assistance
from licensed spectrum (LAA), as well as standalone with no support from a licensed carrier. In case of NR-based LAA,
the unlicensed cell may connect to an LTE or NR anchor cell operating in licensed spectrum. Both below and above
6GHz (up to 52.6GHz) frequency bands are in the scope. Example frequency bands are Wi-Fi 5GHz band, 37 GHz, and
60 GHz. Operation of 5G in stand-alone unlicensed mode is a potential game-changer for the industry, especially for B2B
since customers may deploy their own infrastructure without a license. They could then by-pass mobile operators by
having RAN vendors to deploy and operate directly RAN infrastructure in their premises.
 New Spectrum Concepts for 5G in 3GPP:
Multi connectivity & Carrier aggregation:
This implies that a device has simultaneous connectivity to multiple frequency bands. In the scope of Rel-15, multi-
connectivity can be achieved in different ways, i.e., LTE-NR Dual Connectivity (DC) or NR-NR Carrier Aggregation (CA).
In DC, the aggregation is done at PDCP level and tight interworking between NR and LTE is possible if a common PDCP
layer is used for both, which benefits from co-localized 4G and 5G PDCP functions, either on local radio sites, or in
Centralized Units (such as NG POP). CA is even more demanding on the interaction between the involved cells and
typically has tight latency requirements on the backhaul between sites.
Similarly to the NR bands, 3GPP allows for flexibility to introduce LTE-NR DC and NR-NR CA combinations in a release
independent manner throughout the release. Several LTE-NR DC combinations have been included so far in the Rel-15
scope, among which LTE 4DL/1UL B 1-3-7-20 + NR 1DL/1UL 3.4-3.8GHz (including fallback modes) is of high interest
for Orange. NR-NR CA combinations will be specified later during Rel-15. The scenario of high interest for Orange is NR
700MHz + NR 3.4-3.8 GHz. LAA CA combinations would be specified in phase 2.

Orange confidential - 90
LTE-NR coexistence:
LTE-NR spectrum multiplexing, or in other words, NR-LTE coexistence in the same band has been considered in 5G
Phase 1. When NR is deployed in sub 6GHz, this could bring several benefits: efficient and more dynamic refarming,
maximizing the whole network capacity, etc. In DL, it would be possible by using MBSFN sub-frames (initially defined for
Broadcast in LTE), or at symbol level (for example through scheduling by using mini-slot). In UL, it would be possible by
reservation of sub-carriers per technology.
Such a technique, if truly flexible and dynamic, could avoid traditional re-farming technique, which rely so far on static
allocations per technology (e.g. 20 MHz split in 10 MHz on LTE / 10 MHz on NR). However technical hurdles are not fully
overcome in 3GPP. Compatibility with legacy hardware is also critical to its success.
Pairing of SDL/TDD frequency band with FDD UL in lower frequency:
LTE-NR UL sharing is also introduced in 3GPP phase 1, including the following scenarios:
 SDL NR spectrum + UL spectrum (shared with LTE)
 TDD NR spectrum + UL spectrum (shared with LTE)
This would allow to extend the UL coverage of a high TDD band and make a better use of the FDD UL spectrum. One
scenario of high interest for Orange is LTE UL 800MHz + NR 3.4-3.8 GHz. This is extremely useful in the context of
Massive MIMO in 3.5 GHz, which is expected to lead to a 10 dB gap between UL and DL. Eventually the DL 3.5 GHz
coverage may equal the one of low bands (700/800/900) when paired with a low UL band, which could be reached
thanks to UL pairing, or with traditional CA (e.g. 700 + 3500).
Flexible duplex:
Both paired (FDD) and unpaired (TDD) spectra will also exist in 5G with FDD being more dominant in the lower part of
sub-6GHz spectrum. Besides, flexible duplex TDD and FDD have also been included in the 5G scope, but not prioritized
for phase 1.
Flexible TDD brings the possibility to dynamically assign the transmission resources to DL and UL directions depending
on the instantaneous traffic conditions.. While this is appealing to better adjust DL / UL ratio to real traffic asymmetry, this
leads to challenges for inter-site synchronization and interference management (from BS to BS and UE to UE). It seems
therefore unpractical for macro cell deployment and more suited for small cells / indoor deployment.
Flexible FDD brings the possibility to use the paired spectra indifferently in both UL & DL directions.
Channel & transmission bandwidth:
Minimum and maximum channel bandwidths larger than in LTE would be possible in 5G: Minimum channel bandwidth
will be 5MHz for sub 6GHz and 50MHz for above 6GHz. Current assumption on maximum channel bandwidth is 100MHz
for sub 6GHz and 400MHz for above 6GHz. However, forward compatibility will be ensured for adding new maximum
channel bandwidths in future releases.
Besides, 5G would allow a new concept of flexible channel bandwidth at both BS and UE, with bandwidth resolution
down to a single PRB in both DL and UL (i.e. 0,2 MHz), instead of limited operation with a finite set of channel
bandwidths (i.e., 1.4/3/5/10/15/20 MHz) as in LTE. With this new concept, the UE can access to a 5G cell regardless of
system bandwidth in the cell, i.e. adapt to any bandwidth allocation (e.g. 11,4 MHz), as long as it doesn’t exceed its
maximum total bandwidth handling capability. This will be particularly suited to make full use of odd spectrum allocations,
which may be the case in legacy bands such as 900 or 1800 MHz bands.
Furthermore, it is expected that 5G allows for higher signal transmission bandwidth (i.e., better spectrum utilization) than
in LTE. In LTE, maximum spectrum utilization for a 20MHz channel is 90%. It means that for a 20 MHz channel
bandwidth, transmission bandwidth would be in practice 18 MHz (100 PRBs). Currently, for 5G, spectrum utilization of up
to 95/97/99% is being discussed for different channel bandwidths and sub-carrier spacing. It is particularly interesting for
sub-6GHz where the spectrum is scarce.

18.5 Global initiatives towards 5G


European initiatives:
o Action plan / Trial plan
Following the “5G Manifesto for timely deployment of 5G in Europe” co-signed by a number of European company
CEOs, including Stéphane Richard for Orange, the European Commission has issued in September 2016 the “5G for
Europe” Action Plan. This plan aims at fostering a coordinated approach in the European member states in order to
ensure commercial large scale introduction of 5G in Europe by the end of 2020. Beyond this objective, there is also a
willingness of the EC to put Europe in a good position with respect to the other regions in the race to 5G. Among the
different actions of the plan, the most important ones are:
 Promoting preliminary trials, under the 5G-PPP arrangement, to take place from 2017 onwards, and pre-
commercial trials with a clear EU cross-border dimension from 2018.
 Encouraging Member States to develop, by end 2017, national 5G deployment roadmaps as part of the national
broadband plans.
 Ensuring that every Member State will identify at least one major city to be "5G-enabled" by the end of 2020 and
that all urban areas and major terrestrial transport paths have uninterrupted 5G coverage by 2025.
 Agree by end of 2017 on the full set of spectrum bands (below and above 6 GHz) to be harmonised for the
initial deployment of commercial 5G networks in Europe.

- 91 - Orange confidential
 The Commission will work with the industry and the EIB Group (made up from the European Investment Bank
and the European Investment Fund) to identify the objectives, possible configuration, and modalities for a
venture financing facility.
Following this action plan, a working group involving industry, academia and verticals stakeholders, including Orange,
has been set up to define a common roadmap towards pan-European trials. This roadmap is due in May 2017.
o Outcomes of the 5G PPP phase 1 projects (source: 5G PPP white paper of MWC)
The European Commission created a Public-Private partnership in the area of 5G in 2013, the 5G PPP. The objective is
to create “an open platform that helps us reach our common goal more coherently, directly, and quickly”. 19 projects
have been selected following a first call that started in 2015 and will end in 2015. This first phase laid the foundation for
the 5G design, including physical layer, RAN design, backhaul/fronthaul and network management. From RAN
perspective, the waveform options and massive MIMO schemes have been discussed and evaluated and a RAN design
proposed, integrating LTE-Advanced evolutions. This RAN design has been tailored towards serving simultaneously the
5G service classes: eMBB, URLLC, and mIoT. The performance levels ensure an unprecedented experience for end
users including high data rates, reduced end-to-end latency, massive connectivity, ultra-reliability and support for very
high mobility. The 5G vision emerging from 5G PPP phase 1 projects goes far beyond what is announced for early 5G
deployments in the US and Asia: for eMBB service, the integration of mm-wave and frequencies below 6 GHz, along with
ultra-dense networks and nomadic nodes, ensure the targeted performance levels with ubiquitous coverage and in high
mobility scenarios, in contrast with standalone deployments of mm-wave networks, suitable for fixed usage. The
innovations related to the transport network allow also translating the peak throughputs available at the air interface into
perceived user experience at affordable deployment cost for operators. The 5G concept developed in 5G PPP phase 1
projects and exposed in [MWC16] is not limited to the RAN; it covers the end-to-end path and allows the 5G network to
act as a secure, reliable and flexible orchestration platform across multiple domains. A second phase of 5G PPP projects
will start in June 2017, and will focus on demonstrators and proof of concepts.
North America does not have a comprehensive research program towards 5G.However, 5G Americas, a US industry
trade organization composed of leading telecommunications service providers and manufacturers, is the main actor in
the framework of American activities towards 5G and has already published several whitepapers on 5G. US
communicate on an aggressive roadmap to release high frequency bands (above 6 GHz) before 2020 and on the
importance of cyber security in the framework of 5G. AT&T and Verizon announced 5G field trials in 2016; Verizon is
member of the of the 5G Trial Specification Alliance, together with KT, NTT DOCOMO and SK Telecom, whose goal is
“to develop an aligned 5G trial specification that would serve as a common, extendable platform for different 5G trial
activity around the world, focused on technical fundamentals and promoting a more inclusive, open, and collaborative
approach to the development of 5G trial networks”.
China, Japan and Korea are very active in 5G research and massively communicate on planned trials. Huawei
announced they will invest at least $600 million in 5G research over the next four years (with already 200 researchers
involved in 5G). South Korea should globally invest $ 1.5 billion on 5G. South Korea announced they will launch a 5G
trial for Winter Olympic Games in 2018. Japan announced they will have 5G trials for Olympic Games in 2020. In the
framework of Chinese activities towards 5G, the main actor is the “IMT-2020 (5G) Promotion Group”. In the framework of
Japanese activities towards 5G, the main actor is “The Fifth Generation Mobile Communications Promotion Forum”
(5GMF). In the framework of Korea activities towards 5G, the main actor is “5G Forum”. All these Associations have
issued whitepapers on their 5G vision in terms of requirements but also on technological and architectural aspects. All
regions seem more or less aligned on roadmap and scope except South Korea which want to go faster and focus only on
enhanced mobile broadband.
 NGMN is pursuing its 5G work programme through a wide variety of work items, such as spectrum, end-to-end
architecture, security, test and trials, requirements etc. The aim is to impact the 5G ecosystem (standardization,
regulation, industry fora …) through agreements between the member operators.

5G priorities and recommendations 18-1


5G is the technology we will operate in the 2020-2030 decade. 5G is the future Internet as a whole, including Radio
Access Networks (3GPP RANs, WLAN/Wi-Fi …) as well as a convergent core network between fixed access and
radio access(es). 5G will deliver more than connectivity: a native design with a full software approach will transform
networks into one programmable and unified infrastructure integrating networking, computing and storage resources,
paving the road to new business models and value propositions.
5G standardization process started end of 2015 on 3GPP side: first phase of 5G specifications around mid-2018 and
the second phase around end-2019, for submission to ITU-R for evaluation as an IMT-2020 technology. First
commercial deployments could start in 2020.
5G offers the possibility to offer a wide variety of services:
- Mobile Broadband (eMBB) will remain a core business for Orange by 2020 and beyond.
- Ultra Reliability Low Latency Communications (URLLC) will be an enabler for new business/wholesale services (e.g.
tactile Internet).
- Massive IoT is a ramp-up of today’s IoT.
- 5G Fixed Wireless Access (FWA) emerges as a new option to be compared with fiber & copper (e.g. vDSL, Gfast,…)
for the last mile.
Main Orange 5G milestones:
- 2017/18: trial regarding key enablers for 5G = M-MIMO + vRAN/vEPC + cmWaves

Orange confidential - 92
- 2019 : E2E 5G pilot
- 2020: starting 5G deployment and NGCN tests
Recommendations:
Start getting prepared for the 5G roll-out on the following points:
- spectrum acquisition:
- 3.4 – 3.8 GHz for capacity, reserving 700 MHz for coverage and identifying cmWaves bands (24.25 –
27.5GHz) for Fixed Wireless Access
- lobby the public authorities to clear the 3.4 – 3.8GHz from existing services including Wireless Local Loop,
microwave and satellite services in order to be able to launch 5G services by 2020.
- the need for tight synchronization required for the TDD duplexing
- the E2E architecture comprising NGPoP and transport (fronthaul/mid-haul/backhaul)
Contact: Christian Gallard, Berna Sayrac (RNM/TSY), Eric Hardouin (OLR)

- 93 - Orange confidential

Vous aimerez peut-être aussi