Apex Optimization Strategies Operator

Capex Optimization Strategies: Operator Best Practices
for Making the Most of 3G Network Capex

Author: Houda Bostanji | Analyst
Editor: Kerem Arsal | AME manager
This report is part of Pyramid Researchs series of Africa and Middle East Insider Report | March 2014 Edition
2014 Pyramid Research | CAPEX Optimization Strategies: Operator Best Practices for Making the Most of 3G Network CAPEX
Executive summary
We are far from the days when telecom operators in developed markets showed
prodigious profit margins, outpacing the performance of traditional industries. Today,
things are more delicate: margins are shrinking and competition is becoming fierce.
These trends are now spreading to emerging markets, putting intense pressure on telecom
operators that need to build new networks, ensure quality of service and fund innovation.
An increasingly indispensable tool for operators contending with these challenges is capex
optimization.
To help telecom leaders make sense of capex optimization, Pyramid Research has
developed the MASI framework. Tailored to consulting audits and inspired by the Lean Six
Sigma principles, MASI is a methodology for generating techniques for reducing capex that
are adapted to each operators situation.
This Insider presents a selection of 17 techniques for managing 3G networks in a capexefficient way and thus freeing up resources for innovation and building next-generation
networks. These techniques have resulted from the application of the MASI framework to
the capex profiles of a wide variety of operators.
In this report, we explain each of these capex-effective techniques for managing
networks. Examples from the Middle East and Africa as well as from some developedmarket players are also provided. Operators can already harness these best practices to
control and reduce their 3G capex.
About the author
HOUDA BOSTANJI is an AME analyst at Pyramid Research specializing in the MENA region, where
she is responsible for producing in-depth analysis, for generating device, fixed and mobile
forecasts, and for contributing to consulting projects. She does research on networks, services
and innovative business models.
Table of contents
1. Introduction
2. Capex optimization for 3G networks
Optimizing the capacity planning & deployment process
Optimizing the network architecture
Leveraging network sharing
Implementing dynamic pricing and dynamic capacity allocation
3. Conclusion: Key findings and recommendations
Network capex optimization is not about cutting investments, it is about

allocating efficiently and profitably the available capital
In an environment where market and technological changes occur at an ever-quickening pace
and where operators continually need to roll out new network generations, fund innovation
and ensure quality of service to differentiate themselves and stay competitive, capex
optimization is about investing the right amount, at the right time and in the strategic, valuecreating spot. But how can this be made possible for each business unit manager?
Capex optimization
The right
amount
At the right
time
Invest
In the right
place
Source: Pyramid Research
In order to help operators maximize capex savings, we have developed our

own proprietary framework: MASI
MAPPING SPIDER
ASSESSING PYRAMIDS
Group the items

and tag for
optimization
those that have
both low
priority and
high capex
SORTING BLOCKS
Improve
Assign a priority
to each category
using two
criteria: service
profitability and
service
penetration
Sort
Chart capital
expenditure
categories and
items
Assess
Map
Business unit managers at telecom operators are increasingly under pressure to cut capex,
but often they are left with no structured approach to making those cuts. To help them sort
things out, we have developed MASI, a structure for analyzing and optimizing capex, inspired
by Lean Six Sigma and powered by strong analytical tools.
Eliminate,
combine, move
and reduce the
capex of the
items tagged as
low priority
IMPROVING TREE
5
Table of contents
1. Introduction
How to manage 3G networks in a capex-effective way

This section presents a number of strategies* produced by Pyramid Research's MASI
framework with the aim of reducing capex dedicated to 3G networks. The strategies are
illustrated with examples from implementations by regional and global operators.
Capacity planning & deployment
Transversal vs. siloed capacity planning
Comprehensive capacity planning process
Shortening the capacity deployment process
Network sharing
Network architecture
Differentiated capacity margin rules
Differentiated capacity investments
Heterogeneous networks
Policy control and fair usage
CDN, caching, compression
Passive infrastructure sharing
Single-RAN architectures
Active infrastructure sharing
Converged 3G/4G core network functions
4G as an overlay of 3G
Self-optimizing networks
Cloud RAN (C-RAN)
Dynamic pricing and dynamic capacity

allocation
Network yield management and dynamic

pricing
Dynamic capacity allocations and dynamic

customer experience management
* These strategies have been created by applying the

MASI framework to numerous capex profiles obtained
from operators with 3G networks at various stages of
maturity.
Table of contents
1. Introduction
Transversal vs. siloed planning to harness cross-layer synergies

Very often, capacity planning happens in a siloed approach where the different business units
responsible for each network layer plan their capacity needs separately, design the evolutionary
path of their network platforms and quantify the required capex accordingly.
In a siloed organization, consolidation meetings are then held to build an end-to-end view of the
capacity needs to be presented to CTOs, CFOs and other decision makers. This leaves little room
for optimization.
Introducing transversal technical coordination meetings right from the beginning of the process
can help define controlled evolutionary paths for the overall network and its capacity needs across
different layers and technologies, both fixed and mobile.
These transversal synergies result in less capex allocated to network capacity upgrades, as
alternative pathways are found to route and serve the additional traffic.
Transversal
Siloed
Capacity planning
Radio
2G
3G
Core network
4G
Radio
2G
3G
2G
3G
4G
Core network
4G
2G
3G
4G
Capex needs
Service
platforms
Transport
network
Service
platforms
Transport
network
Comprehensive planning from the equipment to the square meter

Capacity planning is not only about equipment, it should cover all infrastructure involved in the
capacity upgrade process. This includes elements such as:
Available racks within each equipments bay or cabinet
Available ports within each dispatching bay or fiber-optic distribution panel or frame
Available square meters within each regional or central physical site
Power supply and available electrical ports within each site
These elements, seemingly secondary, prove to be vital when it comes to installing and connecting
the extra capacity (module, server or full equipment) to the network.
This approach prevents last-minute deadlocks in the capacity implementation process where the
operation is delayed if, for instance, insufficient space is available and a new site needs to be
built, or if the dispatcher is saturated and a new one needs to be sourced and installed.
If such delays are avoided, the operator will be able to install and switch on the acquired extra
capacity directly. It will thus make full use of the capex involved, avoiding costly alternative
solutions.
Telecom
site
Telecom
cabinet
Distribution
cabinet
Optical fiber
distribution frame
Sources: Total Telecom Corp, Germes Online, Xternal Design, Longxing Telecom
10
Shortening the capacity deployment process to minimize overcapacity

To keep networks running well requires upfront capacity planning that anticipates future needs for
hardware, software, energy plugs, connectors, square meters and transmission infrastructure.
Some capacity upgrades, especially if additional space is needed, can require new technical
premises, extensive cabling, civil works and site development. This means the capacity
implementation processes can take up to two years.
In the meantime, disruptive events often interfere that drive exponential traffic growth, which
can exceed the planned capacity even before its delivery.
To minimize these risks, operators tend to opt for generous capacity margins, installing an
overcapacity of hardware, network operations center space and energy.
When the utilization rate of new capacity remains very low for a number of years, that capacity
can immobilize substantial capex without generating an adequate ROI. (Continues)
In long network deployment cycles, operators plan big additions of network capacity, which remain
unfilled for long periods:
Capacity deployment
Year 0
Year 1
Year 2
Planned capacity
Unused capacity
Carried traffic
11
Shortening the capacity deployment process to minimize overcapacity (contd)

A key practice is to focus on shortening the capacity deployment process and running frequent
planning cycles as needed instead of planning substantial overcapacity from the beginning. This
can be performed through actions such as:
Reducing the number of steps
Limiting the business units involved to those needed for a transversal process
Performing tasks concurrently
Considering build-to-demand technical architectures (e.g., small cells, C-RAN)
A shorter implementation process reduces the need for bigger capacity margins and offers more
flexibility to install capacity on an on-demand basis. Disruptive events can be more quickly
addressed, the level of uncertainty is reduced, and capex is freed up.
In short network deployment cycles, operators plan smaller additions of network capacity, maximizing
the network fill rate over time:
Capacity deployment 1
Year 0
Capacity deployment 2
Year 1
Year 2
Planned capacity
Unused capacity
Carried traffic
12
Table of contents
1. Introduction
13
Differentiating capacity margin rules based on network layers

Network layer sensitivity
Transport network
Core network
Sensitivity
Radio
Users
Traffic
SGSN
Applying uniform capacity margin rules for all radio,

transmission or core network modules invites
inefficient capex allocation.
It is still relevant to use a specific occupation rate
(e.g., 65%) at a core network or at a transmission
module to trigger capacity upgrade alerts. Triggers
can, however, be set higher for radio modules that
are closer to end users.
Tailored capacity margins will maximize the fill rate
of each network layer module and thus the return on
the invested capital, while securing QoS.
Capacity margins should be set according to various
parameters, including the node sensitivity.
For instance, the core network nodes are critical:
they concentrate substantial traffic loads, coming
from all access network cells, to which many users
are attached.
If a core network node fails, it disrupts all users,
whereas if a radio cell fails, it affects only the users
attached to it.
GGSN
++
14
RAN
Optimizing RAN capex: Differentiating capacity across cell sites
NodeB
Backhaul
SGSN
PCRF
Disparities in customers distribution and usage behaviors across different

geographical cell sites (e.g., between those situated in rural and urban
areas, or in high- and low-density districts) result in different traffic
patterns and peak levels.
RAN capex levels can be targeted on a site-by-site basis for increased
granularity of investments.
Investments should then primarily cover revenue-generating
congested sites rather than evenly covering all network cell sites.
These initial investments in capacity can then be used as a strong
foundation for further optimization by dynamic capacity allocation.
Transport
GGSN
Increased visibility is key: this can be achieved by building a dashboard of

indicators for cell sites to monitor the network QoS, the site utilization
rate and its related commercial performance.
Capex should then be allocated at sufficient levels to enhance
capacity in highly utilized, revenue-generating sites.
At other sites, capex levels can be adjusted downward to meet
minimum threshold requirements to serve the traffic and secure
the QoS.
15
RAN
Optimizing RAN capex: Heterogeneous networks, powered by small cells and

femtocells
NodeB
Backhaul
SGSN
PCRF
Transport
GGSN
In order to maximize bandwidth while reducing cost and optimizing the

user experience, radio access networks can be composed of different
types of base stations in addition to macro cells, namely small cells and
femtocells.
Small cells are proving to be an efficient alternative to Wi-Fi offloading,
depending on the end users actions.
Operators can incorporate small cells in big cities to tackle network
congestion. Femtocells ensure better signal penetration inside buildings,
and given that their backhaul traffic runs over the fixed network, they
offer higher bandwidth without overloading the mobile network.
As a result, 3G capacity capex is curbed, despite data traffic taking off.
Etisalat pioneered the deployment of femtocells in the UAE in 2010.
Vodafone resorted to the use of femtocells in public spaces in Qatar in
2010 and has also been conducting trials in Egypt.
Source: Pyramid Research, Etisalat, Vodafone

16
Optimizing the core network capex: Policy control and fair usage to maximize
and prevent abusive usage
Backhaul
PCRF
Transport
GGSN
Core network
SGSN
Transport
NodeB
Policies and fair usage rules can be implemented in the PCRF (policy and charging
rules function) to monitor and control bandwidth depending on usage.
This shapes better data traffic management and fosters usage-based and
progressive pricing.
Priority levels can be assigned to subscribers or applications through service-aware
QoS management assuring reliable service to high-paying segments while
remaining conservative toward peer-to-peer sessions.
Orange Tunisia offers different data card plans ranging from prepaid to unlimited.
These different user profiles introduce complexity in the network that can be easily
absorbed by the PCRF function.
It also enables Orange to control the bandwidth of unlimited subscribers and
prevent abusive usage as they reach monthly limits.
MTN Uganda manages packet traffic with dedicated bandwidth for subscribers who
purchase dedicated Internet volume bundles (e.g., 200Mb, 500Mb). Again, this too
prevents abusive usage.
Operators in the Gulf region in the Middle East have different uses for PCRF:
Throttling usage (e.g., Vodafone Qatar and Egypt, Nawras, Viva Kuwait, du,
Zain).
Charging premiums for usage beyond assigned usage caps (e.g., du).
Monitoring repeat abusive usage for caution (e.g., Ooredoo Qatar).
Source: Pyramid Research, Orange, MTN, GCC operators

17
Optimizing backhaul and core network capex: CDNs, compression, caching and
partnerships with OTTs to release pressure on network
Backhaul
PCRF
Transport
Core Network
SGSN
Transport
NodeB
Traffic generated by applications such as TV and OTT video is growing

rapidly, putting immense pressure on fixed and mobile networks.
Quite often, service platforms hosting such content are situated in a
central node in the core network, aggregating all access requests from
users and delivering the content.
Consequently, this centralized architecture overloads transport links all
the way from the central node to the user and causes high latency.
By using content delivery networks (CDNs) and network caching,
operators can move frequently requested content from the central node
down to the networks edge, closer to the end users. This releases
pressure on the network while enhancing QoS and QoE.
Capex-sharing opportunities can be found in forming partnerships with
broadcasters and content providers wherein the operator hosts the third
partys platforms on its network and complements them with a CDN and
caching.
GGSN
Saudi Telecom Company (STC) offers CDN services for national and
international content providers to manage bandwidth-intensive video
while guaranteeing QoS.
Zain uses data compression and caching to save bandwidth.
Source: STC, Zain
18
Single-RAN architectures generate capex synergies across technology families

when times to LTE rollouts and 3G upgrades overlap.
Spectrum
eNodeB
Backhaul
SGW
MME
Transport
PDN
GW
Converged or single-radio access networks (RANs) with multi-standard

antennas integrating 3G, 2G and 4G modules along with a software-defined
radio will deliver greater bandwidth at a lower cost over the long term.
This strategy can be especially useful when there is also a need to upgrade 3G
and 2G systems during the rollout of a 4G network.
In some cases, equipment prices can be so affordable that it makes more
sense to exchange all old 3G and 2G equipment for a converged RAN with LTE
than to just add 4G.
Etisalat Sri Lanka launched a 3.75G HSPA+ service in 2011 based on a
converged, LTE-ready RAN.
In 2011, STC deployed Huaweis SingleRAN technology and installed LTETDD while also upgrading its 3G infrastructure.
In 2013, NTT Docomo introduced a compact LTE base station with
simultaneous 3G capability, designed for low-traffic areas such as
mountains. It is an economically viable alternative to conventional LTE
stations.
Converged RAN
4G
3G
2G
Source: Pyramid Research, Etisalat, STC, NTT Docomo
19
LTE as an overlay on 3G can help control capex by sharing sites and related
infrastructure when 3G upgrades are not urgent
Spectrum
eNodeB
Backhaul
SGW
MME
Transport
PDN
GW
In a shared 4G overlay model, operators deploy LTE base stations

separately from their existing 2G and 3G base stations, although 4G towers
can be integrated with the existing infrastructure.
Capex savings are realized by the reuse of available 2G and 3G sites and colocating legacy and LTE technologies while maximizing shared hardware.
Depending on the phase of ROI on 3G, this method can provide an
opportunity to avoid acquiring new sites, to share backhaul equipment and
to make use of the available ancillary structures and equipment.
Zain KSA deployed its initial LTE network as an overlay on its 3G
infrastructure.
Verizon has initially deployed its LTE network as an overlay of
eNodeBs on its existing CDMA network, using existing cell sites and
backhaul infrastructure.
Telefnica Spain selected Alcatel-Lucent in September 2013 to
deliver a 4G LTE network overlay on top of its existing 3G network.
Sources: Verizon, Zain, Telefnica

20
Converged 3G and 4G core network functions

A convergent 3G and EPC (Evolved Packet Core) configuration will ensure higher capacity for the
3G network while adding minimal cost to the deployment of a new low-capacity LTE network. Over
the long term this can be a better investment than entirely new equipment.
Besides facilitating network migration, this architecture reduces signaling overhead and, through
shared session data storage, improves resource utilization. Operators can achieve this by using
their deployed 3G infrastructure. This results in significant capital and operational savings.
Converged 3G/EPC architecture

Deutsche Telekom deployed convergent
3G/LTE core network gateway nodes
Verizon Wireless adopted the convergent

3G/LTE core network model. The operator
combined a PGW node (packet data network
gateway), which usually provides access to
LTE, with an HSGW node (high-rate packet
data serving gateway), which provides
access to CDMA.
Source: Cisco, Deutsche Telecom, Verizon
21
Using self-optimizing networks to maximize network performance and delay

hardware investments
Software is playing an increasingly vital role in telecom networks.
Where hardware performance seems to have reached its limit, software is the alternative,
bringing more capacity, additional flexibility and performance while controlling costs.
Self-optimizing networks (SONs) is an automation technology enabling the dynamic selfconfiguration, self-optimization and self-healing of mobile radio access networks.
By making existing networks more efficient, SONs can postpone hardware deployments, thus
reducing capex.
MTN Uganda in 2013 selected JDSUs self-optimizing network technology to support its GSM, UMTS
and LTE networks. It draws on user location information as an input to RAN planning and
performance engineering, thus making more efficient use of existing network assets.
AT&T deployed SON technology across its network in 2012 with the help of Intucell, which was
acquired by Cisco in 2013. Under the deployed architecture, an overloaded radio cells traffic can
be offloaded in real time to neighboring cells without disrupting calls.
Source: MTN, AT&T
22
Cloud RAN: a network virtualization architecture pooling radio resources

Cloud RAN, or C-RAN, is a network function virtualization (NFV) technology; it offers a
collaborative radio and centralized real-time cloud computing and deployment. This enables base
stations to share their backhaul transportation, monitoring systems and especially their baseband
processing resources. In traditional implementations, however, resources are individual and
dimensioned based on traffic during busy hours, which for each base station can fluctuate
dramatically.
C-RAN maximizes the average utilization rate of these resources and their dedicated capex.
Designed primarily for 4G networks, when supported by additional virtualization a C-RAN system
can also consolidate 3G and 2G baseband processing functions in the cloud.
China Mobile developed its C-RAN program with Agilent Technologies, Alcatel-Lucent, Huawei and
ZTE.
Etisalat UAE also signed up in 2011 to support Alcatel-Lucents LightRadio architecture, which is
based on C-RAN technology.
Korea Telecom, in partnership with Samsung and Intel, launched a cloud-based LTE RAN in 2011,
connecting a distributed BTS network to a centralized cloud communication center (CCC).
KTs cloud communication center applies virtualization to manage network conditions
centrally. Radio resources can then be allocated dynamically depending on site, time and
volume of traffic.
KT says C-RAN has saved it money on real estate, physical equipment and operations
management, while doubling capacity in some cells due to better edge management.
Source: China Mobile, Etisalat, Korea Telecom

23
Table of contents
1. Introduction
24
Passive infrastructure sharing can reduce cell site capex

Passive infrastructure sharing is the sharing of the nonelectronic infrastructure within the radio site. It can
include:
Technical premises, shelters, switches, public utility
infrastructure, pylons, masts, towers, power supply,
generators, air-conditioners, cables, dispatching
frames or panels, cabinets, bays, security equipment
and transmission infrastructure.
Infrastructure sharing minimizes investments while
maximizing network coverage, especially in low-density
areas.
This optimizes site preparation costs and distributes the site
rental or acquisition expenses over the sharing parties.
In Africa, operators under network sharing agreements
often transfer their passive elements of their sites to
independent tower companies in order to rationalize and
pool the sites.
Sale-and-leaseback agreements are very common in
Sub-Saharan Africa.
In Uganda, an agreement between Orange and Eaton
Towers is expected to bring a capex gain of $32m over 10
years.
In Morocco, Meditel and Inwi agreed in 2008 to share
passive network infrastructure that includes engineering
works, access rights and towers.
Passive network elements
Source: ITU
Source: Pyramid Research, ITU, Orange

25
Active infrastructure sharing as a major capex saver

Active infrastructure sharing is the sharing of electronic infrastructure. It enables the pooling of
baseband processing.
This includes spectrum, switches, antennas, transceivers and microwave equipment.
It can also include full RAN sharing, backhaul sharing or core network sharing.
In Spain, an example of active network sharing is the RAN sharing agreement between Vodafone
and Orange during their 3G rollout. The operators shared no fewer than 4,300 base stations.
Savings of up to 40% per site are estimated by the completion of rollout.
The partnership has already generated a cumulative capex and opex savings of 160m over
the 2007-2011 period.
In Sweden, Tele2 and Telenor shared their backhaul and together built a shared fiber and
transmission network as part of their joint venture to build a nationwide LTE network.
In Bangladesh, mobile operator Robi Axiata signed a transmission network sharing agreement with
NTTN in 2012 to expand mobile service provision in areas uneconomical to serve on their own.
Other types of infrastructure sharing are also possible: Orange, for example, is pooling its service
platforms among various countries where the company is present.
Source: Pyramid Research, Orange, Vodafone, Tele2, Telenor, NTTN

26
Table of contents
1. Introduction
27
Harnessing network yield management and dynamic pricing to reduce capex
Network yield management
Network yield management and dynamic pricing are techniques derived from the airline industry
that aim to maximize the use and monetization of existing network assets, capacity and
bandwidth.
This can reduce and delay further capex, generate additional revenue streams from the
traditionally unused assets and thus boost ROIC and ROCE.
The goal of yield management, according to Enz & Withiam (2001), is to provide the right service
to the right customer at the right time for the right price.
Network yield management enables operators to increase revenue and profits mainly through
granular customer segmentation, dynamic pricing and improved capacity utilization.
Granular customer segmentation
High-end
Price-sensitive
Dynamic pricing
Per area
Congested/free
Per hour
Peak/off-peak
Per customer
High/low end
Tariff or access
adjustments to
services at offpeak hours for
price-sensitive
segment
Higher tariffs at
peak hours in
congested areas
Maximized
network fill rate
Traffic switched to
off-peak hours
Monetization of
unused capacity
Higher
FCF,
ROIC
and
ROCE
Reduced capex
FCF: free cash flow. ROIC: return on invested capital. ROCE: return on capital employed
Source: Pyramid Research, Telcordia
28
Network yield management requirements: Support from an advanced

OSS/BSS system
To deliver a successful yield management strategy, operators should use:
An accurate network information collection and reporting system:
It needs to identify and forecast peak and off-peak hours for each cell site.
It needs to collect specific information on each customer type and fine-tune the
segmentation and the offered plans and services accordingly.
A dynamic bandwidth capacity allocation system that does not affect QoS negatively:
These functions can be filled by the GGSN, PCRF and radio network to execute actions like
throttling or THP (traffic handling priority).
A convergent charging and billing system:
PCRF, OCS or mediation centers will store the network rules specific to each customer profile
and the related pricing strategies to apply. The rules can also be dependent on the time of
use and geographical location of each customer.
Dedicated communication and marketing channels:
These target end users with innovative services, real-time reductions and planned tariffs.
Source: Telcordia, Ericsson, Amdocs, Pyramid Research

29
A deployed example of yield management from AT&T

AT&T chose Amdocs and its convergent charging and billing system in early 2014 to support Drive
Studio, the operators connected car service.
The operator argues that the car is about to become essentially one more connected device on a
customer's shared data plan.
Amdocs system enabled AT&T to split the bill between the automaker and the end consumer,
charging the first for the connectivity and the second for the apps. Also, dynamic pricing supported
by promotional opportunities has been made possible for content and app providers as well as for
OEMs.
Differentiated & split services charging
Granular customer segmentation

Prepaid
Postpaid
Connected car
Premium
Mobile
Fixed
APPS
(navigation, entertainment)
CONNECTIVITY
(for car diagnosis, safety
services, car telematics)
CHARGED TO
(Connected car service)
AT&T Drive Studio
Amdocs system
Consumers
Automaker
Real-time and convergent:

Charging & billing
Customer relationship
management
Enabled promotional opportunities for

connected car ecosystem providers
Content
App
providers
providers
Example: free year of Internet music service
OEMs
Source: AT&T, Amdocs, Pyramid Research

30
A deployed example of yield management from MTN Uganda

Ericsson provided its Yield Optimization tool to MTN Uganda during the deployment of the MTN
Zone service in 2008. The tool provides dynamic discounting: as traffic on the network declines,
tariffs are reduced.
This has shifted traffic on the network from high-load to off-peak periods (see next slide for
further details).
Ericsson Yield Optimization system

Dynamic Tariffing Engine
MTN Zone
Rules: Optimal tariff, maximum/minimum discounts
Communication
Tariffs notified by: Cell broadcast, call setup notification, USSD
Charging and reporting

OCS
CDRs, revenue reports,
Sources: Ericsson, MTN, Pyramid Research

31
A deployed example from MTN Uganda, with quantified impact

Through yield management, MTN began to achieve not only more predictable and uniform traffic
loads throughout the day, but also higher revenue due to better targeting.
Revenue; market share
IMPACT
Traffic patterns; capex
Sources: Ericsson, MTN, Pyramid Research

32
Dynamic capacity allocation and dynamic customer experience

management makes efficient use of the available resources
Instantly aligning the allocated capacity to demand and shaping the customers experience to its
contextual data would align the capex spent on delivering the data with the revenue it generates.
In fact, dynamically allocating capacity to the customer depending on experience (usage, quality
of service, location, subscribed plan, network congestion, etc.) enables more efficient use of the
limited network resources, including bandwidth, and thus capex. Dynamic capacity allocation also
makes it possible to better manage the customer experience.
This can be performed by applying customers Big Data analytics to self-organizing networks.
User experience metrics can be collected across all intelligent machine nodes and employed
by the self-organizing network to adjust, in real time, the allocated capacity.
Dynamic capacity allocation can be viewed as a form of second-order network capacity
optimization, as initial capacity planning exercises at the physical level would already have set
the constraints within which dynamic allocations can be made.
Intelligent
network
nodes
User
metrics
Cloud &
analytics
Processed
data
Selforganizing
networks
Dynamic capacity &

customer
experience
management
33
Table of contents
1. Introduction
34
Key findings and recommendations

Capital expenditures are an immense drain on operators cash cash that could be better
used for revenue-generating assets and reducing capex has become a growing challenge
for operators.
Pyramid Research's MASI framework provides a structure for systematically planning and
optimizing capex.
Our methodology for developing techniques of saving on capex produced a number of levers
for controlling and optimizing capex dedicated to 3G networks:
Differentiated capacity margins, transversal comprehensive planning, fair usage, CDNs
and dynamic capacity allocations drive forward efficiencies in the capacity delivery
process, and as a result curb 3G capex.
Network sharing, converged or heterogeneous architectures, cloud-based services,
yield management, dynamic pricing as well as other services and technologies
produce capex synergies and thus further savings.
Operators such as Verizon, Orange, Zain, MTN, Etisalat, Korea Telecom, STC and Deutsche
Telekom are making use of some of the capex-saving techniques discussed in this Insider.
The defined techniques are built on a common-sense, time-tested foundation of aiming to
maximize profits through the more efficient use of available assets. This is done by
applying software- and hardware-based innovations in a structured manner and partnering
with other industry players, even competitors, whenever there is a solid business case.
With an insightful application of the MASI framework, doing more with less can become an
achievable goal, a reflex built into the operator's DNA, turning its modus operandi into a
lasting competitive advantage.
35
Glossary
Acronyms
AME
capex
CCC
CDMA
CDN
C-RAN
DOCSIS
EPC
FCF
FTTH
GGSN
HSGW
LTE
MASI
MME
NFV
Definitions
Africa & Middle East
capital expenditure
centralized cloud communication center
Code-Division Multiple Access
content delivery network
Cloud RAN
Data Over Cable Service Interface Specification is an international
telecommunications standard that permits the addition of high-speed data transfer
to an existing cable TV system
Evolved Packet Core is the IP-based LTE core network
free cash flow
fiber to the home
Gateway GPRS Supporting Node
high-rate packet data serving gateway
Long-Term Evolution
Pyramid Researchs capex optimization framework, aggregating four structured steps:
Map, Assess, Sort and Improve
Mobility Management Entity
network function virtualization
36
Glossary
Acronyms
OCS
OSS/BSS
OTT
PGW
QoE
QoS
RAN
ROCE
ROIC
SGSN
SON
THP
Definitions
online charging system
operation and business support systems
over the top
packet data network gateway
quality of experience
quality of service
radio access network
return on capital employed
return on invested capital
serving GPRS support node
self-organizing network
traffic handling priority
37
Sources and related research

Capex Optimization Strategies: Operator Best Practices for LTE and FTTH Deployments
Telecom Insider published December 2013
To overcome the challenges of capex optimization, Pyramid Research has developed the MASI framework, a methodology for generating ways of
reducing capex that are adapted to each operators situation. This Insider presents a selection of 16 MASI-generated techniques for capex
optimization in LTE and FTTH deployments: examples of fiber access and mobile broadband implementations come from the Middle East and
Africa, plus some developed-market players.
Big Data: Telecom Innovation through Analytics
Insider published October 2013
Big Data is important to telecom operators, but the task of implementation seems daunting. This Insider covers the possibilities enabled by Big
Data initiatives, guiding telecom operators through the role of market conditions (personalization, customer selection), some exciting forms of
Big Data analytics, the organizational challenges (availability of skills, differences between business units) and regulatory concerns. The report
contains case studies on Turkcells m-commerce analytics, Netflixs recommender system and Weve, an m-marketing and m-payments platform.
Mobile Data Forecasts
Forecasts updated quarterly
A premium version of the Mobile Demand Forecast, Mobile Data tracks current and future demand for mobile data services, within the context of
overall mobile demand. This forecast includes all the same tracking as the standard Mobile Demand Forecast but breaks down mobile data ARPS,
revenues and subscriptions at the service level.
Uganda: Mobile Market Consolidates as LTE Rollout Begins
Country Intelligence Report published March 2014
Ugandas telecommunications sector generated service revenue of US$858m in 2013; Pyramid Research expects it to reach $1.2bn in 2018. Mobile
data usage, the main growth driver, should increase at a 24.1% CAGR. The Bank of Uganda issued new mobile money guidelines, and mobile
money continues to grow in importance. Subscriptions to 3G networks will continue to grow steadily; we expect 35.0% of all subscriptions to use
the technology by 2018. Three operators launched LTE services in 2013, but there was no LTE coverage beyond greater Kampala in early 2014.
Smartphone Strategies: How Devices Can Revitalize the Role of Operators in the Mobile Ecosystem
Thematic Report published October 2013
This report presents strategies for mobile operators to use in response to the challenges brought by the new smartphone landscape. It examines
strategies such as device portfolios, smartphone distribution, subsidies, buy-backs, trade-ins, reverse subsidies, own-branded devices, handset
financing, data-sharing, tablets, OTT services. This report examines the device strategies of five mobile operators in depth: AT&T, MTS,
Telefnica, China Unicom and Orange.
CLIENT CONFIDENTIAL
2014 Pyramid Research. All rights reserved. No material contained in this report may be reproduced in whole or in part without the prior
written permission of the publisher. Neither the information nor the opinions expressed in the report should be construed as a recommendation
for the purchase or sale of any security. The information contained in this report has been obtained from sources we believe to be reliable, but
neither its completeness nor accuracy can be guaranteed. Opinions expressed are based on our interpretation of the available information and
are subject to change.
About Pyramid Research
ADVISING LEADERS IN THE GLOBAL COMMUNICATIONS INDUSTRIES ON EMERGING

MARKETS AND SERVICES OPPORTUNITIES
Pyramid Research (www.pyramidresearch.com) offers practical solutions to the
complex demands our clients face in the global communications industry. Its
analysis is uniquely positioned at the intersection of emerging markets, emerging
technologies and emerging business models, powered by the bottom-up
methodology of our market Forecasts for more than 100 countries a distinction
that has remained unmatched for more than 25 years.
HOW WE COVER COMMUNICATIONS MARKETS
Our highly structured analysis closely examines the performance, technology
trends, macroeconomic and regulatory environment, of service providers and
equipment vendors in each of the more than 100 countries we cover. Using a
bottom-up approach, we size market demand across services and countries, and
conduct end-user surveys to validate adoption trends and anticipate shifts in
demand. Our commitment to intensive, direct interviewing of regulatory
authorities, service providers and manufacturers complemented with end-user
data ensures the most reliable portrait of every market we examine.
39
Meet us in person
PLEASE CONTACT US AT:
Email: sales@pyr.com
Email: press@pyr.com
Telephone: +44 (0) 20 7406 6608
PLEASE ENGAGE WITH US VIA:

@pyramidresearch
OFFICE LOCATIONS:
EMEA, Asia Pacific
John Carpenter House
7 Carmelite Street
London, EC4Y 0BS
United Kingdom
Tel: +44 (0)20 7936 6530
Latin America, North America

179 South St., Suite 200
Boston, MA 02111
USA
Tel: +1 617 747 4100
Fax: +1 617 747 4133
To learn more about Pyramid Researchs product offerings and how they can be of service to your company, please contact sales@pyr.com or visit us
on the Web at www.pyramidresearch.com.
40

Apex Optimization Strategies Operator

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Apex Optimization Strategies Operator

Transféré par

Droits d'auteur :

Formats disponibles

Capex Optimization Strategies: Operator Best Practices

for Making the Most of 3G Network Capex

Optimizing the capacity planning & deployment process

Optimizing the network architecture

Leveraging network sharing

Implementing dynamic pricing and dynamic capacity allocation

3. Conclusion: Key findings and recommendations

Network capex optimization is not about cutting investments, it is about

In order to help operators maximize capex savings, we have developed our

Group the items

Optimizing the capacity planning & deployment process

Optimizing the network architecture

Leveraging network sharing

Implementing dynamic pricing and dynamic capacity allocation

3. Conclusion: Key findings and recommendations

How to manage 3G networks in a capex-effective way

Transversal vs. siloed capacity planning

Comprehensive capacity planning process

Shortening the capacity deployment process

Differentiated capacity margin rules

Differentiated capacity investments

Policy control and fair usage

CDN, caching, compression

Passive infrastructure sharing

Active infrastructure sharing

Converged 3G/4G core network functions

Cloud RAN (C-RAN)

Dynamic pricing and dynamic capacity

Network yield management and dynamic

Dynamic capacity allocations and dynamic

* These strategies have been created by applying the

2. Capex optimization for 3G networks

Optimizing the capacity planning & deployment process

Optimizing the network architecture

Leveraging network sharing

Implementing dynamic pricing and dynamic capacity allocation

3. Conclusion: Key findings and recommendations

Transversal vs. siloed planning to harness cross-layer synergies

Comprehensive planning from the equipment to the square meter

Shortening the capacity deployment process to minimize overcapacity

Shortening the capacity deployment process to minimize overcapacity (contd)

2. Capex optimization for 3G networks

Optimizing the capacity planning & deployment process

Optimizing the network architecture

Leveraging network sharing

Implementing dynamic pricing and dynamic capacity allocation

3. Conclusion: Key findings and recommendations

Differentiating capacity margin rules based on network layers

Applying uniform capacity margin rules for all radio,

Optimizing RAN capex: Differentiating capacity across cell sites

Disparities in customers distribution and usage behaviors across different

Increased visibility is key: this can be achieved by building a dashboard of

Optimizing RAN capex: Heterogeneous networks, powered by small cells and

In order to maximize bandwidth while reducing cost and optimizing the

Source: Pyramid Research, Etisalat, Vodafone

Source: Pyramid Research, Orange, MTN, GCC operators

Traffic generated by applications such as TV and OTT video is growing

Source: STC, Zain

Single-RAN architectures generate capex synergies across technology families

Converged or single-radio access networks (RANs) with multi-standard

Source: Pyramid Research, Etisalat, STC, NTT Docomo

In a shared 4G overlay model, operators deploy LTE base stations

Sources: Verizon, Zain, Telefnica