Académique Documents
Professionnel Documents
Culture Documents
1. Capacity Contents
2. Coverage Contents
3. Interference Contents
4. CoBCCH Contents
5. Dual-Band Networking Policies Contents
6. Location Area Contents
7. Neighboring Cell Optimization Contents
8. Frequency Replanning Contents
Capacity Wang Xi, Tang Xiaoli 61537, 49854 See the phone book.
Coverage Shi Yuan, Xu Binbin 146349, 232322 See the phone book.
Interference Wan Yin, Chen Zhe 181609, 182362 See the phone book.
Capacity
X Measures for Capacity Planning and Optimization
Principles for Capacity Evaluation and Optimization
Equipment
Um Interface
Transmission
Cases
How to Use the Capacity Evaluation Tool
Standard Operating Procedure and Deliverables
Smart phones
are the key drive
Develop
customized
Feature
Capacity expansion and
Network capacity evaluation optimization solutions capacity
and forecast
1. To increase the dual-band optimization
1. Improvement of the utilization
network capacity:
of resources (UR) solutions based on
• Dynamic adjustment of the half- • CoBCCH
1. Traffic forecast rate threshold (HR/AMR).
• Linear forecast 2. When the frequency spectrum live network
• Adjust dynamic PS thresholds
• Secondary index regression and the soft capacity are
limited:
evaluation and
prediction 2. Inter-cell load sharing
• Traffic model-based prediction • Adjustment of dual-band network • IBCA development
thresholds 3. When capacity expansion is
2. E2E network resource • Parameter adjustment for inter- forecast.
cell reselection limited:
evaluation
• VAMOS
• Um resource evaluation
3. Accurate capacity expansion Develop dedicated feature
• Equipment resource evaluation
guidance
• Transmission resource • TRX expansion solutions.
evaluation • Site addition
Formulate optimization
Identify network capacity solutions.
bottlenecks.
Capacity
X Measures for Capacity Planning and Optimization
Principles for Capacity Evaluation and Optimization
Equipment
Um interface
Transmission
Cases
How to Use the Capacity Evaluation Tool
Standard Operating Procedure and Deliverables
Capacity
X Measures for Capacity Planning and Optimization
Principles for Capacity Evaluation and Optimization
Equipment
Um Interface
Transmission
Cases
How to Use the Capacity Evaluation Tool
Standard Operating Procedure and Deliverables
Capacity
X Measures for Capacity Planning and Optimization
Principles for Capacity Evaluation and Optimization
Equipment
Um Interface
Transmission
Cases
How to Use the Capacity Evaluation Tool
Standard Operating Procedure and Deliverables
Bearer or 70%
bandwidth Transmission Bandwidth/Allocated Transmission Bearer
Bandwidth
IP path or 70%
bandwidth [T7411: Mean Transmit Rate of an IP Path at the IP
Layer] /Configured IP Path Transmission Bandwidth
Interface
[Bandwidth Sum of All Physical Interfaces of Interface
board 70%
Boards]/[Physical Interface Bandwidth]
bandwidth
Transmission resources on
In Gb over FR mode, the bandwidth usage of the BC is used for
the Gb interface:
monitoring the traffic load on transmission resources on the Gb
The Gb interface is between Expand Gb
interface.
the BSC and SGSN. transmission
Capacity expansion threshold: Average PDCH Rate
Insufficient Gb transmission resources and add
[RL9608: Uplink Bandwidth Usage of the BC] > 80%
resources lead to low licenses.
or
GPRS/EDGE service rates or
[RL9610: Downlink Bandwidth Usage of the BC] > 80%
even access failures.
In Flex Abis mode, all Abis TMD timeslots of each BTS are
dynamically assigned. When a PS or CS service request is
Transmission resources on Average PDCH rate:
initiated, Abis timeslots are assigned in real time according to
the Abis interface (Flex RR2752: Congestion
actual requirements. In this mode, Congestion Rate of Dynamic
mode): Rate of Dynamic
Resource Assignment is used for monitoring the traffic load on
The Abis interface is between Resource Assignment Expand Abis
transmission resources of the Abis interface.
the BTS and BSC. (16 kbit/s) transmission
Capacity expansion threshold:
Insufficient Abis transmission RR2751: Congestion resources.
RR2752: Congestion Rate of Dynamic Resource Assignment (16
resources lead to poor user Rate of Dynamic
kbit/s) > 3%
experience and access Resource Assignment
or
failures. (8 kbit/s)
RR2751: Congestion Rate of Dynamic Resource Assignment (8
kbit/s) > 3%
In non-Flex Abis mode, Abis timeslots, except idle timeslots,
always map TCHs and PDCHs of the BTS one to one. When the
Transmission resources on capacity of TCHs and PDCHs is expanded, the capacity of Abis
the Abis interface (non-Flex timeslots is expanded simultaneously. Therefore, you can use the
mode): idle Abis timeslot application failure rate to monitor the usage of
Average PDCH rate: Expand Abis
In non-Flex mode, insufficient idle timeslots in non-Flex Abis mode.
The percentage of high transmission
Abis transmission resources Idle Abis Timeslot Application Failure Rate (Non-Flex Abis Mode)
coding rates for EDGE resources.
lead to low PS service rates = [R9109: Number of Unsuccessful Application Attempts of Abis
services is low.
and difficulty in improving the Timeslot Because of No Idle Timeslot]/[R9101: Number of
coding rate. Some MSs even Application Attempts of Abis Timeslot]
cannot access the network. Capacity expansion threshold:
Idle Abis Timeslot Application Failure Rate (Non-Flex Abis Mode)
> 3%
Capacity
X Measures for Capacity Planning and Optimization
Capacity Evaluation and Optimization
Equipment
Um Interface
Transmission
Cases
How to Use the Capacity Evaluation Tool
Standard Operating Procedure and Deliverables
same. CPU1 and CPU2 of XPUMs in subracks 0 and 2 1 XPUT 3 31.138 56 64 -2 0 25 379
system assigns
carry SS7 signaling links at the same time. As a result, TRXs of the sites
XPUM 0 36.633 57 0 0
their load is rather heavy. The maximum CPU usage to CPU1/CPU2 of
XPUM 1 62.828 94 71 5 1
hit around 90%. the XPUM in
XPUM 2 60.902 92 63 -3 1 subrack 1 and
Solution: CPU0/CPU1/CPU
XPUT 0 49.355 80 64 -2 0
Adjust the homing of sites to balance the XPU load. XPUT 1 50.15 78 69 3 0 2/CPU3 of the
XPUT.
Adjust the homing of SS7 signaling links to balance the XPUT 2 51.186 80 73 7 0
Select sites according
XPU load.
to the following criterion:
Add XPUs. Reduce 10% CPU
usage by adjusting
2 XPUT 3 50.723 83 67 1 0 27 407 every 100 Erl traffic.
The call setup success rate fell in Office XXX of Country XXX. Large number of assignment
failures due to equipment failures occur on the A interface.
Analysis:
2. Check the cable connection of the interface and whether any crossed pair connections exist.
Conclusion:
According to the detailed analysis, no equipment fault alarm occurred. No crossed pair connection
existed. Circuits of all interfaces were sufficient. Generally DSPs were sufficient. However, DPUa
boards configured for different TC subracks were uneven. Some DSPs were seriously overloaded.
As inter-subrack DSP sharing was not supported, assignment failures occurred when DSP
resources were unavailable for calls assigned to the second TC subrack.
Solution:
Large number of assignment failures occurred occasionally on the BSC of an office during busy hours. During idle
hours, however, the BSC worked properly. During busy hours, the traffic increased drastically on occasion.
Analysis:
Check whether the A interface, Ater interface, and DSPs are overloaded.
Conclusion:
The switching capacity of the inter-subrack TNU is as follows: Each TNU cable consists of 4096 64 kbit/s links.
Among the 4096 64 kbit/s links, eight are used for system monitoring and 4088 are used for service switching,
which are generally sufficient for use. According to the detailed analysis, the service volume of the BSC was large.
CS services used lots of inter-subrack resources. PS cells were assigned disorderly. Many PS cells of subrack 2
were assigned to DPUs of subrack 0, which also used inter-subrack switching resources. Many services adopt high
coding rates, which used much more inter-subrack resources. When the switching volume of CS and PS services
in subrack 2 exceeded 8176 (2 x 4088), no inter-subrack resources were available for subsequent calls and
assignment failures occurred.
Solution:
Symptom:
To adapt to the fast development of PS services and the fast growth of traffic, the customer added PDCHs. A large
quantity of users still complained about the low service rate and poor experience. As a result, the customer thought
that Huawei equipment had performance defects.
Analysis:
Check usage of Um interfaces, transmission resources, interface boards, and DSPs for PS services.
Conclusion:
According to the comprehensive analysis, Um interfaces were sufficient, network interference was at a normal level,
the coding rate of PS services was properly set, and no PDCH congestion occurred. Transmission resources on
the Abis interface, however, were seriously congested. During busy hours, the congestion of Abis TDM timeslots
exceeded 15%, which seriously affected the service rate of PDCHs. Most PDCHs carried services at a low coding
rate due to insufficient Abis resources.
Solution:
Expand Abis transmission resources and increase the number of idle timeslots for the Abis interface.
Capacity
X Measures for Capacity Planning and Optimization
Principles for Capacity Evaluation and Optimization
How to Use the Capacity Evaluation Tool
Standard Operating Procedure and Deliverables
Click the evaluation item link in the resource capacity evaluation report generated by the OMStar.
Detailed analysis results of each item are displayed. In the resource capacity planning part, the
OMStar automatically generates the corresponding capacity expansion and adjustment solution. It
provides guidance for the planning and optimization of network resource capacities.
Capacity
X Measures for Capacity Planning and Optimization
Principles for Capacity Evaluation and Optimization
How to Use the Capacity Evaluation Tool
Standard Operating Procedure and Deliverables
resource.
Operation 3 Analyze data and generate an evaluation report using the OMStar.
Operation 4 Check whether the capacity is restricted. If yes, generate the list of restricted cells. If
the problem is caused by restricted capacity, go to operation 5. Otherwise, submit the problem to
Operation 6 Implement the optimization solution and evaluate the optimization effect. If the
congestion is mitigated or eliminated, promote the related service or sell related products.
Coverage
Principles
Troubleshooting Process
Cases
The terrain is flat, with densely distributed buildings. Lots of high-rises Such areas are in the center of 1. Densely populated
(over 30 floors) exist in the area, whose average height exceeds 50 m. super large cities, with a small areas, constituting hot
Dense urban area and densely distributed spots
The spacing between buildings is small and irregular. Most streets in
area (CBD) buildings. Many business 2. High requirements on
the area are narrow. The population is very large and the majority of
people are in buildings within certain periods. A activities of the regions take the PS service rate
place here. Users frequently use 3. Key areas for PS
The terrain is flat, with densely distributed buildings. The average mobile communications and service development
Dense urban
height of buildings ranges from 25 m to 30 m. The spacing between have high requirements on the 4. High requirement on
area (residential
buildings is small and irregular, averaging 10 m to 20 m. Many streets quality of PS services. service quality
area)
(not arteries) in the areas are narrow with a dense population.
Such areas are in the center of cities. The height of buildings is Such areas feature prosperous 1. Densely populated
High-density industries and businesses, all areas, with a high traffic
moderate, averaging about 20 m. The average spacing between
common urban necessary transport facilities and volume
buildings is the same as the height of buildings. There is open and
area infrastructure, and many arteries, 2. PS services with a
green space within the area. B
with a high urbanization level, moderate rate
Such areas are on the edge of cities. Buildings are low and scattered, dense populations, fast 3. High requirement on
Low-density
with an average height of 15 m – 20 m. The average spacing between economic development, and high service quality
common urban
buildings exceeds the average height of buildings. Streets in the areas income per capita.
area
are wide. There are parks and large green space.
Such areas are on the border between cities and rural areas. Such areas feature well 1. Small service volume
Buildings are scattered, with an average height of 10 m and an developed businesses and quite 2. Only low-rate PS
Suburban area a high level of urbanization, with services provided
average building interval of 30 m – 50 m. Streets in the area are wide,
with lots of plants or open space. C many enterprises, convenient
transportation, and middle-level
In such areas, buildings are very scattered. Most buildings are economic development and
Rural area bungalows, with an average height of about 5 m. There are large income per capita.
areas of open space, farmlands, plants, and roads.
Straight road Straight roads and highways. Two types of areas are included: 1. The traffic volume is
1. Main roads and railways very small.
Sea/lake Sea or lake surface, characterized by vast open space and no D 2. Less developed rural and 2. The purpose of site
surface obstacles within the line of sight. mountainous areas deployment is to provide
network coverage.
Dense wood Select this clutter model for special areas, such as forests and parks with vast forests.
Water surface Select this clutter model for special areas, such as lakes and rivers.
No received
signals
The uplink loss is bigger than the The downlink loss is bigger than
downlink loss. the uplink loss. Uplink access
is impossible.
The coverage scope is determined by the party whose signal strength is poorer.
Coverage quality
C I
Signal Network
strength interference
Transmit power Path loss Handover timeliness Internal interference External interference
Possible root causes of coverage problems are as follows: equipment, parameter settings, handover,
networking solution, and interference.
Coverage
Principles
Troubleshooting Process
Cases
17:00
21:00
17:00
21:00
13:00
17:00
21:00
13:00
21:00
13:00
17:00
21:00
13:00
17:00
21:00
13:00
17:00
13:00
17:00
21:00
13:00
17:00
21:00
13:00
17:00
1:00
5:00
9:00
1:00
5:00
9:00
1:00
5:00
9:00
1:00
5:00
9:00
1:00
5:00
9:00
1:00
5:00
9:00
1:00
5:00
9:00
1:00
5:00
9:00
11/07/201111/08/2011 11/08/2012 11/08/2013 11/08/2014 11/08/2015 11/08/2016 11/08/2017 11/08/2018
Detailed procedure:
For details, see the Detection and Troubleshooting of GSM RF Tunnel Faults: Principle + Tools + Cases + Deliverables.
List of parameter
(1) MML configuration file OMStar and PFMStudio Adjust
check results. This
(2) Intelligent parameter inappropriate
list contains cell
baselines, user-defined GSM Parameters Check: Check parameter
parameters that do
parameter baselines, or Rules + Tools + Cases + settings and
not meet baselines
parameter settings before the Deliverables.ppt generate the
and corresponding
swap (in swap scenarios) List of coverage parameters (focus) summary case.
causes.
Detailed procedure
For details, see the GSM Parameters Check: Check Rules + Tools + Cases + Deliverables.
(1) Find out cells whose neighboring cells are missing. (1) GIS software, such as Nastar, Mapinfo,
When neighboring cells are missing, MSs cannot be (1) List of missing and Google Earth
handed over to the neighboring cells. As a result, MSs neighboring cells of (2) Signaling analysis tools, such as
(1) Add missing neighboring
suffer quality deterioration caused by the low level in the problematic cells PFMStudio
cells according to their
serving cell. (1) Engineering (3) DT analysis tools, such as the Probe
priorities.
(2) Find out cells with poor coverage due to no handovers parameters, MML TEMS
(2) Adjust and optimize
or delayed handovers. When the handover algorithm or configuration file, and e- (4) Parameter query and browsing tools,
(2) List of parameters of parameters that cause
parameters are inappropriately configured, handovers will map such as PFMStudio, OMStar, and LMT
problematic cells handover delay.
be delayed. As a result, MSs suffer quality deterioration (2) Top cell signaling Note: A prototype tool is being developed
(3) Locate "island" cells
caused by the low level in the serving cell. tracing for automatically analyzing missing
based on engineering
(3) DT data neighboring cells, coverage overlaps, poor
(3) Find out "island" cells caused by coverage overlaps. parameters and adjust the
(3) List of "island" cells (4) MR tracing data coverage, and unbalanced uplink
Cells become islands when coverage overlaps occur. In downtilt or add neighboring
caused by coverage /downlink traffic using MR tracing data.
this case, handovers cannot be triggered and the quality cells.
overlaps and list of TA Currently, the tool is undergoing the Beta
deteriorates due to the low level. As a result, call drops
samples test. It is scheduled to be launched in early
occur.
2013.
Implementation condition: optional. Implemented when operations 1 to 4 fail to resolve the problem or areas
affected by abnormal coverage need to be accurately located after the swap or engineering adjustment.
Purpose: Accurately locate areas affected by abnormal coverage by service geographic observation. Perform DT or
dial-up test if necessary. Confirm the impact scope and symptom of the problem.
Principles: Accurately locate areas affected by abnormal coverage by using the MR-based service geographic
observation functions of the Nastar and service geographic observation functions of the DT software.
Data sources, tools, and outputs:
If poor coverage is detected in the DT after the swap or engineering adjustment, we usually use the
following service geographic observation methods:
(1) Direct analysis method: If the coverage changes significantly in continuous areas, areas and serving
cells where the network coverage becomes poor can be located directly from the geographic distribution
of levels.
Formulate
the RF
adjustment
solution
based on
simulation
results.
Coverage
Principles
Troubleshooting Process
Cases
Adjustment result:
Percentage of MRs Involving Poor Percentage of MRs Involving Percentage of
Percentage of
Service Quality Due to Uplink Poor Service Quality Due to Unbalanced UL
Cell ID Unbalanced UL and DL
Poor Coverage (Before Uplink Poor Coverage (After and DL Paths
Paths (Poor DL Quality)
Adjustment) Adjustment) (Poor UL Quality)
142 10% 2% 50% 5%
80 15% 1% 45% 3%
153 5% 0% 35% 7%
48 9% 0% 40% 8%
Adjustment result: The uplink and downlink become balanced (see the traffic statistics on July 1).
Action1:
Perform the next
hardware/channel/ant No exception is detected. N/A
operation.
enna check
Action2: parameter Perform the next
No exception is detected. N/A
check operation.
(1) Some neighboring cells are Perform the next
N/A
missing. operation.
Perform the next
(2) No handover is delayed. N/A
Action3: handover operation.
exception check and (3) Analyze island cells due to
coverage overlap coverage overlap. Locate the The serving range of this cell
Adjust the downtilt and The call drop rate
check missing neighboring cells is too large. As a result, calls
azimuth of the top cell and becomes normal
whose high level exceeds the cannot be handed over
limit its coverage along after the
TA based on the uplink and quickly to nearby neighboring
the roads and streets. adjustment.
downlink level information in the cells, and call drops occur.
MRs involving poor coverage.
Adjustment result: The call drop rate in the top cell becomes abnormal.
Comparative analysis of
Skip this operation if the area is located.
coverage & traffic
Action1:
hardware/channel/antenna No exception is detected. Perform the next operation.
check
Action2: parameter check No exception is detected. Perform the next operation.
Action3: handover exception
check and coverage overlap No exception is detected. Perform the next operation.
check
Compare level distribution information in the
The percentage of low levels
DT results before and after the network swap:
Action4: geographic increases after the network swap
The percentage of low levels increases 0.5%. Perform the next operation.
coverage analysis and continuous areas affected by
In terms of geographic distribution, the level
poor coverage exist.
deteriorates significantly.
According to the site survey, the area is As the problem is caused by the
Action5: RF and coverage Add sites to make up
mountainous and coverage holes exist at high terrain, antenna adjusting is not a
solution check coverage holes.
altitudes. solution.
Interference
Principles
Troubleshooting Process
Cases
Affected cell 2
Maximum
interference
direction 2 Affected cell 1
Interference
Principles
Troubleshooting Process
Cases
When the following conditions are met, it is deemed that the problems are
caused by interference:
1. Troubleshooting trigger conditions are met.
2. The interference band threshold and measurement algorithm are set correctly.
Procedure
See the Detection and Troubleshooting of GSM RF Tunnel Faults: Principle + Tools + Cases + Deliverables.
Procedure
See the GSM Parameters Check: Check Rules + Tools + Cases + Deliverables.
When PS DTX is enabled, power transmission stops when PDCHs are idle or not occupied by TBFs, to In China
PS downlink
reduce network interference, improve the performance of CS and PS services, improve the network quality, GBSS9.0 BTS3000V300R009C00SPC050 Downlink
DTX phase 2
and reduce the power consumption of the BTS. GBSS9.1 BTS3000V300R009C01SPC033
GBSS13.0 BTS3000 V300R013C00SPC011
MBTS:
SRAN3.0 BTS3900 V100R002C00SPC340
SRAN5.0 BTS3900 V100R003C00SPC360
SRAN6.0 BTS3900 V100R004C00SPC121
The BSC determines PS power control after preprocessing the downlink measurement quality. If power
adjustment is required, the BSC calculates the power attenuation value by using the PS power control
PS downlink BSC: R013 and later
algorithm and delivers the value to the BTS. The BTS adjusts the transmit power on the current radio block Downlink
power control BTS: R013 and later
based on the power attenuation value. In this way, the transmit power of the BTS is decreased and the
network quality is improved.
PS uplink power The MS adjusts the uplink output power based on the downlink receive level and decreases the
R8+ Uplink
control transmit power, reducing interference in other channels.
Interference
Influence Factor
Source
When GSM channels are in service, non-BCCH
CS DTX (percentage of power
TRXs do not require full power transmission.
CS traffic transmission timeslots)
Different interference reduction measures
volume CS power control
should be developed for different services Percentage of TCHHs
during the calculation of total network Number of PS DTX (percentage of dummy
interference to correctly quantify interference. activated blocks)
channels for
PS power control
PS services
Assume that cell A is the serving cell (Server) and cell B is the interfering cell
(Interference).
Co-frequency correlation factor: indicates the overlapping coverage of two cells. If
both the capacity (numbers of TRXs of Server and Interference) and frequency are the Network
same, the higher overlapping coverage indicates stronger interference between the cells. structure
Number of MRs involving cell A that may be index
affected by the intra-frequency cell B
formulas
Number of MRs involving cell A adopted
Number of MRs involving cell A that may be affected by the intra- by China
frequency cell B =∑Number of MRs in each C/I interval x Corresponding Mobile
interference probability
Overlapping coverage =
For details, see the GBSS R13 Frequency Planning Solution and the section about
frequency replanning in this document.
Evaluation method
Optimization solution
Increase the TCHH percentage and the capacity (voice quality deteriorates).
Enable the VAMOS function (depending on the percentage of MSs supporting the VAMOS function).
Enable load sharing, for example, migrate traffic between the GSM and 3G networks or between the GSM and
4G networks.
Replace antennas, for example, use antennas that perform better in terms of immunity to interference.
Purpose: Check whether downlink external interference exists. If downlink external interference exists,
identify the interference type.
Principle: If internal CCI or ACI does not exist, and the percentage of MRs involving high level (greater
than –90 dBm) and low quality (RXQUAL_6 and RXQUAL_7) is greater than 5%, downlink external
interference may exist. The source of downlink external interference is determined only by frequency
scanning using the frequency scanner, not the HQI.
Data sources and tools: frequency scanner (such as Tektronix YBT250 and Agilent 9340 A) and yagi
antenna
Interference
Interfering Interfering Interfering Structure direction
cell service frequency interference analysis
location location location location
ACI
CS ACI
PS ARFCN on Number of
Interference Interference ARFCN on Average
Total Interference CCI ARFCN ACI ARFCN CCI ARFCN the Non- Distance Coverage MRs with the
Poor-Quality Interferen Interference Interference Traffic on Traffic on the Non- Average Azimuth Level of
Server CI Interference % Interference Traffic on on the on the on the Non- BCCH Between Overlap High-Level
Signals % ce CI BCCH BSIC the BCCH the Non- BCCH TA Relationship Affected
Traffic the Non- BCCH TRX BCCH TRX BCCH TRX TRX Sites Type Neighboring
TRX BCCH TRX TRX Signals
BCCH TRX (Interferen Cell
(Server)
ce)
49199 6.72% 0.58% 49178 17 61 1058.52 203.59 92.80 762.12 [17] [13, 17] 1.14 4328.97 no Opposite 500 -77.40
Neighboring
49198 5.15% 0.09% 48138 3 62 62.88 61.02 0.05 0.80 [4] 1.56 1580.77 no cell including 350 -79.71
the Server
49198 0.15% 0.09% 32752 5 60 0 0 0 0 [4] [16] [15] 3.10 1671.89 no Parallel 10 -79.71
Interference
Principles
Troubleshooting Process
Cases
Step 1 Confirm that the high call drop rate is caused by interference.
After the swap, the downlink HQI in 900 MHz cells deteriorated, which led to high call drop rates.
Required Action Analysis Data Analysis and Data Analysis Next Issue Closure Result
Symptom Conclusion Operation
Action1: RF channel check No exception Normal N/A N/A
The CDMA antenna was about 1.7 meters from the GSM antenna. In the figure above, the CDMA antenna is on the
left, facing the GSM antenna. The CDMA antenna was about 0.5 meters lower than the GSM antenna. When the
CDMA BTS and the GSM BTS were enabled, strong GSM uplink interference was caused by spurious CDMA signals
because no filter was configured on the two sides.
Step 3 Perform issue closure and export the summary report.
Configure the CDMA stop filter on the GSM side. Then perform issue closure and export the summary
report.
Common dual-band Not recommended. Adopted The negative PBGT is a It is complex to set the negative Mobistar and OL in
network — 900 MHz cells when required by the customer neighboring cell-level PBGT. Belgium
2 and 1800 MHz cells at the and for network swapping parameter and can be
same layer adjusted flexibly.
Enhanced dual-band Not recommended. 1. Resource and hardware 1. Heavy maintenance workload FOA site in Iran
network sharing 2. Small capacity (compared
2. Tight frequency reuse with CoBCCH)
3
3. Traffic balancing
4. Compatibility of PS services
Concentric circle (CoBCCH) See the section about the 1. Capacity improvement: One 1. CoBCCH is demanding on VDF in South Africa
CoBCCH in this document. BCCH is shared and TCHs are application scenarios. If MTN in Nigeria
Not recommended. Adopted added. CoBCCH is enabled in a Warid in Pakistan
when required by the customer.2. Network quality scenario that does not meet the
improvement: The number of requirements, network
inter-cell handovers and the performance will be worse than
number of BCCH TRXs are that of a dual-band network.
decreased. The interference 2. It is difficult to optimize
4
becomes weaker. 1800 MHz CoBCCH cells because
BCCH frequency interference parameter configurations vary
is weaker. with scenarios.
3. Convenience for
maintenance: The number of
cells and neighboring cells is
decreased. The maintenance
workload is reduced.
Perform this Operation when the configuration is correct but the network
Check the consistency between networking
1 performance cannot meet customer requirements.
policies and customer requirements
Identify customer requirements and export the current networking policies
and effects.
Mandatory
Check the consistency between
2 Check whether networking parameters map or comply with the baselines
configurations and networking policies
and whether configurations are correct.
Generally, you are advised to clarify networking policies rather than adjust them. If networking policies need to
be adjusted, learn about customer requirements and provide networking policy deliverables. After confirmation,
jointly replan networking policies with the customer.
Output: Modify and then check parameter baselines based on networking policy rules in operation 1 and
export the networking parameter check result.
Huawei Parameter
N Parameter Name Mapping Relationship
Name
The layer of 1800 MHz cells is one level lower
Layer of The Cell /
than that of 900 MHz cells.
Inter-layer HO
/ Mapping failure. It is set to 30.
Threshold
Adjacent Cell
HO_LEVEL_UMBRELL
Inter-layer HO Huawei=[HO_LEVEL_UMBRELLA] – 30 + 64
A
Hysteresis
Edge HO UL RX_LEV HO_T_L_UL_RX_LEVE
Huawei=HO_T_L_UL_RX_LEVEL
Threshold L
Edge HO DL RX_LEV HO_T_L_DL_RX_LEVE
Huawei=HO_T_L_DL_RX_LEVEL
Threshold L
PBGT HO
HO_MARGIN_PBGT Huawei=[HO_MARGIN_PBGT] + 64
Threshold
Inter-cell HO
HO_MARGIN_LEV Huawei=HO_MARGIN_LEV
Hysteresis
Huawei=HO_MARGIN_LEV–
BQ HO Margin HO_MARGIN_QUAL
HO_MARGIN_QUAL + 64
Notes: (1) Parameters of 900 MHz cells and 1800 MHz cells are set following the preceding table.
(2) Other parameters are set based on parameter review comments.
Cases
References
The following problems may occur if 900 MHz and 1800 MHz networks
cover different areas after Co-BCCH is activated:
The overlaid subcell fails to absorb traffic to the maximum, leading to congestion in the underlaid
subcell.
KPI deterioration resulting from blind assignment and handover becomes severer with a larger ISD.
MSs failing to measure the level of the overlaid subcell Overlayer: DCS1800 lead to an increase in the call drop rate
and a decrease in the handover
-75
success rate and assignment success
rate.
KPI deterioration resulting from blind
-95
assignment and handover becomes
severer if a the coverage difference
Short and medium- Long distance
long distance (larger ISD) increases between the
overlaid subcell and the underlaid
subcell.
Handover command
RxLev, RxQual…..
DCS1800
Overlay
? As good as 900
MHz?
After Co-BCCH reconstruction or swap, if KPIs are deteriorated and fail to meet the
expectation, check each factor listed above as described in following sections.
Cases
References
Before swap or
2: Version verification Swap or reconstruction None To verify the current version, functions not supported by the version and impacts on KPIs.
reconstruction
To check whether the antenna azimuth and downtilt of the 900 MHz and 1800 MHz
During the
3: Antenna check Reconstruction None networks to be reconstructed are consistent and adjust the azimuth and downtilt (if
reconstruction
necessary) two weeks before commencement of reconstruction.
To configure the TRX output power of the Co-BCCH cell: For the underlaid subcell, keep
During swap or
4: Power configuration Swap or reconstruction None the TRX output power unchanged. For the overlaid subcell, configure the TRXs to make
reconstruction
the output power close to the TRX output power of the underlaid subcell to the maximum.
During swap or To configure channels and attributes according to applicable swap/reconstruction
5: Channel configuration Swap or reconstruction None
reconstruction requirements.
During swap or To check parameters in strict accordance with parameter mapping rules (from dual-band
6: Parameter check Swap or reconstruction None
reconstruction network to Co-BCCH network or from a vendor's network to Huawei network).
1. After swap: To analyze whether the total traffic volume and traffic distribution in the
8: Analysis on voice and overlaid and underlaid subcells after Co-BCCH swap are consistent with the original
data services (including network, which decides directly whether KPIs are worsened compared with those of
After swap or
the ratio and changes) of Swap or reconstruction None the original network.
reconstruction
the overlaid and 2. After swap/reconstruction: To check the PS traffic volume and number of activated
underlaid subcells PDCHs in addition to the CS traffic volume and traffic proportions of the overlaid and
underlaid subcells in consideration of interference of PDCHs in CS services.
1. To check whether the power before and after swap is matched with each other by
comparing the level of the overlaid and underlaid subcells before and after swap.
Before and after 2. To verify whether the triggering time for handover from the underlaid subcell to
9: Drive test Swap None
swap overlaid subcell and from overlaid subcell to underlaid subcell is consistent with that
of the original network and whether the coverage of the overlaid subcell is the same
as that of the original 1800 MHz network.
Co-BCCH 1. TRX power of the Co-BCCH overlaid subcell and underlaid subcell is inherited from the original
reconstruction
1800 MHz and 900 MHz networks, respectively. In addition, the power of each TRX in the same
frequency band must be equal. If not, adjust the TRX power.
2. If the ISD exceeds 5 km, the coverage discrepancy between the overlaid and the underlaid
subcells is great. The coverage discrepancy can be decreased by enabling MCPA and
configuring the maximum power for TRXs in the 1800 MHz frequency band.
Co-BCCH swap 1. Data for power match and actually measured data need to be provided. To ensure accurate
power match, the BTSs to be measured must be representative and the number is determined
depending on the number of cells to be swapped.
2. For complete power match requirements, see GSM BTS Power Match.
3. In case of incorrect data of the original network or other problems relating to power match, drive
tests for some BTSs are required both before and after swap. For drive test requirements and
methods before and after Co-BCCH swap, see section "Required Action 9: Drive Test" in this
document.
Power here refers to the actual output power, which is calculated using the following formula: Actual
output power = Value of power type – Value of power class + Fine-tuning value.
Note: Data services are not distributed to the overlaid subcell in versions earlier than R014. The methods for adjusting the
traffic of both the overlaid and underlaid subcells will be supplemented after being verified in R014.
HUAWEI TECHNOLOGIES CO., LTD. Page 172
Required Action 9: Drive Test
Triggering condition of the action: mandatory before and after Co-BCCH swap
A Co-BCCH cell is under the coverage of both the 900 MHz and 1800 MHz frequency bands. Therefore,
drive tests must be performed with caution as described in Drive Test Methods and Precautions issued
by the R&D department before and after Co-BCCH swap.
Test Item Purpose Method Remarks
Cases
References
Cause analysis
The number of PDCHs in the underlaid subcell is increased by 9% after swap, leading to
greater interference in the underlaid subcell.
Conclusions
In the current version (earlier than R013C00SPH556), PDCHs cannot be initially allocated to the
overlaid subcell, leading to great difference in PS traffic distribution between the live network and the
original network. Because PDCHs which are in the overlaid subcell before swap are allocated to the
underlaid subcell after swap, the PDCHs in the underlaid subcell are increased by 9% after swap,
leading to greater difference in the underlaid subcell.
Reduce Reduce
DL PS PC
PDCHs PDCHs
Before Activated
on NG3 on NG2
SWAP
Easter
Holiday
With PDCHs being limited, the call drop rate is significantly reduced, almost close to that before swap.
R014 versions support PDCH allocation optimization and therefore reduce the number of PDCHs and improve the
PDCH bearing efficiency.
900 M 900 M
Layer 1
The 1800 MHz cell has a higher priority and a wider logical
area (by means of CRO configuration) than the 900 MHz
cell; therefore, MSs are more likely to reside in the
neighboring 1800 MHz cell.
Conclusions
Analysis by performing required actions fails to identify any factor that affects traffic distribution.
According to previous theoretical analysis, scattered reconstruction results in failure of traffic
distribution match with the original network, which can be rectified by optimizing configurations of
parameters.
Note: For other performance optimization measures required after reconstruction from a dual-band
network to a Co-BCCH network, see section 3.5 in GPSS V100R003C00 Technical Guide for the
Solution on Reconstruction from a Dual-Band Network to a Co-BCCH Network.
Cases
References
Location Area
Overview
Paging Principles and Relevant Parameters
Calculation of Paging Capacity on the Um Interface
LAC Boundary Dividing Principles
LAC Planning and Optimization Procedure
Cases
The LAC size of areas may vary with traffic characteristics. Usually, sizes of different LAs
are suggested as follows: LA in dense urban areas < LA in common urban areas < LA in
suburban areas < LA in rural areas
When determining the LAC boundary, you must consider geographic features and the
distribution of MSs to reduce the LA update frequency. Usually, the area with fewer users or
handover possibility is selected as the boundary. Avoid selecting an area with a dense
population as the boundary, such as roads and shopping malls.
Avoid selecting an area where high-end users are located as the LA boundary, such as
group users and VIP users.
BTSs that use the multicarrier system must be located in the same LA.
Areas where BTSs of the same LAC are located must be continuous. Avoid and reduce the
possibility of BTS overlapping in adjacent LAs.
04
04 Evaluate the cell-level paging overload
02 Calculate the safe paging (evaluate CCCH expansion) and the necessity
Estimate the number of delivered
Circle the LAC scope Calculate the safe paging capacity
capacity
of LAC or RAC splitting 02 of the new LAC, the traffic volume 04 paging messages of the LAC and
supported by the new LAC, and the calculate the traffic volume and
number of TRXs after the number of TRXs after the
06 05 implementation of the MOCN. implementation of the MOCN.
CCCH expansion Y Is the number of BCHs that need to be expanded in the cell
Calculate the traffic volume smaller than the number of BCHs that the customer allows to
03 05 Calculate the designed and optimization expand?
allowed per LA and the Evaluate the risk of paging
traffic volume and TRXs in
number of TRXs allowed per overload after the implementation
an LAC
LA N 05 of the MOCN (evaluate the
reasonableness of the LAC
07 Need to split or adjust the new LAC or RAC (the scope).
customer circles the rough scope of the new LAC
or RAC based on the geological position)
Evaluate the reasonableness of the LAC
06 scope and provide the evaluation result Is the LAC so large as to cause Y
(compare the traffic volume and TRX 06
Calculate the safe paging the paging overload? Does the
09 Estimate the number of delivered
number allowed per LA with those of the 08 capacity of the new LAC or paging messages of the new LAC
LAC scope need to be
circled area.) RAC or RAC in the live network adjusted?
N
End
Is the LAC so large as to Evaluate the reasonableness of LAC or RAC
cause paging overload? Does
Y 10 splitting (compare the estimated paging
07 the LAC scope need to be capacity of the live network with the safe
adjusted? paging capability)
N
The U-Net Delivery 2.0 can automatically
11 Is the new LAC or RAC too large Y
End and does it needs to be adjusted support the three scenarios. The tool is
or re-split?
under development. Please keep
N
yourself updated on the release
End
information.
Verification of LAC combination for YYY2 site during The paging capacity of the Um interface under the existing
MOCN reconstruction in Country XXX2 parameter configurations and traffic model was about 110,000. The
The MOCN involves two operators, TELENOR and TELE2. The LAC was divided as follows before maximum total paging capacity of LACs involved in the combination
and after reconstitution of the MOCN:
was 48,000, which was far smaller than the paging capacity.
Therefore, LAC combination will not incur the risk of paging
overload.
Application scenarios:
Neighboring cell optimization after network adjustment: The neighboring cell relationships change
because of the addition, swap and capacity expansion of the cells.
Periodically neighboring cell optimization on the entire network: For network adjustment because of
changes to the network scale and subscriber distribution, you need to check and optimize neighboring
cells periodically on the entire network.
Neighboring cell optimization in certain areas: Centrally perform neighboring cell optimization for the cell
clusters or VIP areas with low KPI performance.
After suggestions on neighboring cell optimization are exported using Nastar, you can refer to additional
information (for example, distance and the number of MRs for neighboring cells) exported for further
verification.
Verify the following items when collecting MR data to ensure reliable neighboring cell optimization:
Parameter settings
Absolute level threshold and Relative level threshold are correctly set.
BTSMESRPTPREPROC is changed to BSC_Preprocessing.
Neighbor Cell Measurement Optimization Allowed is disabled (if V 8.1 is used).
Polling scripts of the BA2 table are delivered.
Measurement objects
All to-be-optimized cells are added to the measurement object list.
Polling measurement is conducted on all BCCH frequency.
Verify that the time of data collection is reasonable. (It is recommended that performance data for seven
consecutive days be collected.)
Network changes occurring during data collection
Check whether any to-be-optimized cells and frozen cells are activated or deactivated during data collection,
or any new BTS accesses the network during data collection.
Check whether any BCCH frequencies or BSICs are adjusted during data collection. (The adjustment affects
accuracy of the mapping of neighboring cell names and the final sequence of neighboring cells.)
Theoretically, to-be-optimized areas and frozen areas must be blocked from networks during data collection to
ensure accuracy and consistency of collected information. Operations that do not affect MR data collection,
however, can be performed. If any of the preceding network changes occurs during data collection, check the
final optimization results manually.
In MOCN scenarios, especially where 64 frequencies are supported, the trimming of the BA2
table is difficult. KPIs may decrease when excessive neighboring cells are deleted; if the
automatic filtering of the BA2 table is disabled and automatic level measurement of
neighboring cells is enabled, frequencies measured by users of operator A may, in extreme
situations, belong to operator B, causing an increase in the measured call drop rate. You are
advised to negotiate this scenario with the operator before data collection.
Before performing neighboring cell optimization based on MR data, conduct the following verification:
Verify consistency between engineering parameters and configurations.
Synchronize engineering parameters with configurations and verify the consistency between them
to check the reasons. First, check whether some cell information is available in the configuration
table but are unavailable in the engineering parameters. Collect the inconsistent cell information.
Then, check whether some cell information is available in the engineering parameters but are
unavailable in the configuration table. Collect the inconsistent cell information. Finally,
synchronize engineering parameters with the configuration table. Nastar has this function but
OMStar has not.
Verify accuracy of performance data.
Verify that collected data is accurate, all to-be-optimized cells have reported MR data, and each
traffic packet is in normal size.
Verify that neighboring cell optimization results are valid based on the following information:
possible neighboring cells, distance, number of MRs, topology structure diagram, and accurate
black-lists and white-lists. (Black-lists include neighbor relationship that cannot be added, and
white-lists include neighbor relationship that cannot be deleted. Both black-list and white-list
information must include the source and destination BSC names, source and destination cell
names, and the reason why the cell is black-listed or white-listed.) Pay special attention to the
neighbor relationship in special scenarios, such as indoor, high-speed railway, trestle, tunnel, and
large building BTSs. Update black-lists and white-lists in special scenarios to accumulate
experience.
Collect handover measurement data under GSM cell –GSM cell for one week and check the data as follows:
1. If the sum of H370c: CELLCELL_OUTCELL_HO_REQ and H380: CELLCELL_INCELL_HO_REQ is
smaller than 10, the related neighboring cell may be redundant.
2. If the ratio of H375B: CELLCELL_OUT_CELL_HO_FAIL_RECONN_FAIL_TIMES to H370c:
CELLCELL_OUTCELL_HO_REQ is greater than 10%, the related neighboring cell may be redundant.
3. Determine the redundant neighboring cell based on geographical position. If the neighboring cell is far
from the serving cell, or the neighboring cell is near the serving cell but the azimuths of the two cells are
opposite to each other, you need to disable the redundant neighboring cell. Otherwise, you need to check
whether the handover parameters are set properly.
If the software for optimizing neighboring cells is not installed onsite, use the preceding algorithm and
geographical display software to optimize redundant neighboring cells.
Collect handover measurement data under GSM cell –GSM cell for one week and check the data as follows:
1. If the sum of H370c: CELLCELL_OUTCELL_HO_REQ and H380: CELLCELL_INCELL_HO_REQ is
smaller than 10, the related neighboring cell may be redundant.
2. If the ratio of H375B: CELLCELL_OUT_CELL_HO_FAIL_RECONN_FAIL_TIMES to H370c:
CELLCELL_OUTCELL_HO_REQ is greater than 10%, the related neighboring cell may be redundant.
3. Determine the redundant neighboring cell based on geographical position. If the neighboring cell is far
from the serving cell, or the neighboring cell is near the serving cell but the azimuths of the two cells are
opposite to each other, you need to disable the redundant neighboring cell. Otherwise, you need to check
whether the handover parameters are set properly.
If the software for optimizing neighboring cells is not installed onsite, use the preceding algorithm and
geographical display software to optimize redundant neighboring cells.
The results of drive tests show that call drop occurs because the level decreases. The neighboring cell RxLev
is higher than the serving cell RxLev by more than 10 dB (if the MS use TCH carrier, the serving cell RxLev is after
a Power Control. However, because before Call-drop the serving cell RxLev has been relatively poor, indicating
that the MS has to the edge of the cell, the power control level is not too large, and can be approximated by
comparing serving the RxLev and the neighboring cell the RxLev.), but handover does not occur. The serving cell
SJGH0244_2 does not configure SJGH0354_0 as its neighboring cell. Call drop occurs due to the missing
neighboring cell.
After neighboring cell optimization results are delivered, perform the following steps
to verify that optimization results for all neighboring cells are delivered:
1. Collect the up-to-date configuration data, and check whether each neighboring
cell optimization result is configured.
2. If any neighboring cell optimization results are not delivered, deliver
optimization results of the neighboring cells again, and conduct verification
again until optimization results of all neighboring cells are delivered.
For the list of Required Actions for neighboring cell optimization, see the
template of Analysis Report of Neighboring Cell Optimization:
Traditional Mode
The traditional mode relies on massive drive tests, emulations, and engineer experience,
requires a large cost, and takes a long time.
Current Mode
A network interference model is set up according to a large amount of accurate MR data reported
through MSs. The analysis results of automatic frequency planning are provided through the AFP
software platform. Compared with the traditional mode, the current mode has the following
advantages:
Free: The required data is obtained in equipment rooms.
Convenient: After the MR data for about a week is obtained, a network interference model is
set up in two hours. The software searches for appropriate frequencies according to the
interference model.
Accurate: Call measurement records collected from the entire network are used to simulate
the true user behaviors and experiences, while in the traditional mode an estimation is
made according to the information from DTs.
The necessity of each step in the process will be described in detail subsequently.
5. Optimizing the RF
Note: The frequency replanning only adjusts frequencies based on the existing network
architecture. It cannot solve severe problems in the structure, parameters, and
equipment of the network and external interference. To achieve the optimum effects, you
need to prepare for the frequency replanning.
Any mistake may affect the frequency replanning result. Therefore, make
sufficient and subtle preparations according to the requirements.
Verifying SVIP and VIP users can be specified in engineering parameters. In frequency replanning, VIP cells should be given the highest
6
VIP cells priority.
1. Supported frequency bands: For example, whether the GSM900 supports the EGSM and the bandwidth supported by
Collecting the GSM1800.
information about 2. Cavity restriction: The differences between frequencies in a single cavity combiner must be greater than or equals to
7
hardware 600 KHz.
restrictions 3. Multi-carrier module information to be collected: module type, module version, number of TRXs in each module, and
the module where BCCH TRXs are.
Notes:
(1) The tool for frequency replanning determines cell names according to the BCCH_BSIC in the MR data. If
many cells that share BCCHs and BSICs and near each other exist on the existing network, the
interference source matching may be incorrect and the planning results are affected.
(2) The tool for BSIC planning is used to replan BSICs of the whole network and maximize the distances
between cells that share BCCHs and BSICs.
(3) The old Nastar platform is used for partial adjustment. The Nastar selects pairs of cells (two cells in a pair
share BCCHs and BSICs and the distance between the two cells is less than 5 km). The BSICs need to be
optimized manually. The Nastar is applicable to the situation where few such pair of cells exists.
For detailed evaluation methods, see the Qualitative Evaluation Scheme for the Gain of a
GBSS Auto Frequency Plan.
HUAWEI TECHNOLOGIES CO., LTD. Huawei Confidential Page 239
Case 1
Determining Principles of Frequency Replanning
Determine the frequency use principles of each TRX before frequency replanning. The principles have great impact on
planning results.
Taking sites in City X in Country X as an example: GSM900 indoor and outdoor sites use different frequency bands.
The indoor sites use frequencies 1, 4, 7, 10, 13, and 16.
Solution 1: TRXs of indoor cells with more than six frequencies use outdoor frequencies without authorization.
Solution 2:
1. TRXs of indoor cells with more than six frequencies use outdoor frequencies without authorization.
2. In a pair of cells that have handover relationship and the total number of TRXs of which is greater than 6, the
BCCH TRX and two TCH TRXs use indoor frequencies; other TRXs use outdoor frequencies without authorization.
Co-channel Interference Traffic of the Entire Network Adjacent-channel Interference Traffic of the Entire Network
Note: The co-channel and adjacent-channel interference traffic of the entire network reflects the actual traffic interference.
The co-channel interference traffic is greatly reduced in solution 2 as compared with that in solution 1. This is because solution 1 cannot
distribute enough indoor frequencies to two indoor cells with many TRXs in a building. As a result, the two cells share frequencies and
greatly interfere with each other.
Register measurements of
neighboring cell levels on the
M2000
Use the "Tool for modifying the BA2 table" to output MML commands for BA2 table modifications. The tools are applicable to
V8.1 and V9.0.
The upgrade of the BA2 table has a great impact on an interference matrix; therefore, you need to check whether all
frequencies are measured.
1. Frequency replanning uses the indexes S363 to S370 under Neighbor Cell Level
Measurement per Cell. It is recommended that you select the indexes under the whole
measurement set.
2. When enabling the performance measurement of neighboring cell levels, check the CPU
usage of the BSC XPUM. Stop the operation, if the average CPU usage of XPUM CPUs
0 to 3 reaches 70%. Enable flow control if the average CPU usage reaches 80%.
November 13 4858.68 323.01 99.75% 0.58% 0.00% 99.76% 0.33% 0.00% 98.56% 99.98% 100.00%
November 14 5004.63 305.63 99.74% 0.63% 0.01% 99.76% 0.33% 0.00% 98.48% 99.99% 100.00%
November 15 5160.03 283.65 99.65% 0.65% 0.02% 99.70% 0.36% 0.02% 98.27% 99.97% 99.98%
November 16 5051.13 298.19 99.71% 0.62% 0.01% 99.74% 0.37% 0.01% 98.30% 99.98% 99.99%
November 17 5109.93 298.9 99.67% 0.62% 0.01% 99.71% 0.38% 0.02% 98.26% 99.97% 99.98%
November 18 5107.35 290.29 99.68% 0.64% 0.01% 99.72% 0.38% 0.01% 98.21% 99.97% 99.98%
November 19 5017.66 304.14 99.68% 0.58% 0.01% 99.74% 0.37% 0.03% 98.38% 99.95% 99.96%
Average value 5044.2 300.54 99.70% 0.62% 0.01% 99.73% 0.36% 0.01% 98.35% 99.97% 99.98%
Morning on
3534.24 91.32 99.08% 1.35% 0.00% 99.12% 0.57% 0.03% 96.72% 99.96% 99.97%
November 20
It is found that the civil service examination was held on November 20. Exam centers in universities started jammers, which greatly
affected the counters of the entire network. It is found that 5 cells with the poorest counters, 34 in total, are in or near universities.
The accuracy of collected data is vital for frequency replanning; therefore, check
carefully during collection of frequency replanning data.
1. Before collecting data for frequency replanning, field engineers need to verify the BSC data
configuration to ensure that all MR collection switches for the BSC are enabled, preventing
information loss.
Deliverable: verification results of the mapping relationship between the cell list and the
frequency replanning cell list after all MR collection switches are enabled
2. During data collection, ensure that all frequency replanning cells report MR data properly.
Deliverable: verification results of mapping relationship between the cell list and the
frequency replanning cell list when BSCs performing frequency replanning report MR data
during peak hours
According to the protocol, BCCH frequency levels in the BA2 table are measured during a call and the
information about six neighboring cells with the strongest levels are reported. During the measurement and
reporting, the factors affecting the accuracy of an interference matrix are as follows:
1. A serving cell shares a BCCH frequency with its neighboring cells. Usually, the serving cell level is stronger
than the levels of neighboring cells. In this case, neighboring cell signals cannot be reported.
2. A serving cell shares a BCCH frequency with two neighboring cells. Two signals of the neighboring cells
interfere with each other and cannot be reported.
3. A serving cell shares a BCCH frequency with two neighboring cells. The weak signal is covered by the strong
one. In this case, the interference traffic of the neighboring cell with the weak signal is less than the true value.
4. The BCCH frequency is modified during data collection. This leads to mis-match of neighboring cell names.
5. MR data cannot be reported due to cells with large interference.
6. A serving cell is out of service. The interference data of the serving cell does not exist in the interference
matrix. In addition, other cells cannot detect signals of the serving cell.
7. In RF adjustment and parameter modification, the coverage scope of cells are changed. In this case, the
interference traffic in the matrix is inaccurate.
The distance between the two sites is 1,022 m. Senior Citizen Association in Jinniu District 2 (serving cell) shares a BCCH frequency with Adult
Education College of Southwest Jiaotong University 3. The interference traffic for an hour is 644, ranking 32nd. After the frequency replanning,
the BCCH frequencies of the two cells are different. In this case, the interference traffic for an hour is 17,467, ranking fourth.
QQ Business
Hotel 1
Grand Southwest Building
Material Market 2
Xuefu Tea
House 1
VCDB303: Sweet Osmanthus Garden 3 has GSM1800 neighboring cells that are in the same
site and direction, but there are not handover requests.
Interference cell
Interference cell
Frequency 3
Factors affecting the effects of frequency replanning: interference matrix Interference cell
and search algorithm
The following table describes an interference matrix. The matrix describes the interference relationship
between the serving cell and interference cells. High traffic of co-channel interference or adjacent-channel
interference indicates a close relationship between two cells. If two cells are assigned with a same frequency
or adjacent frequencies, high interference occurs and user call experience is adversely affected.
According to the interference matrix, two cells with high traffic of co-channel or adjacent-channel interference
cannot be assigned with a same frequency or adjacent frequencies. This is because high interference traffic of
the entire network indicates high interference.
Serving Cell Name Neighboring Cell Name Co-channel Interference Traffic Adjacent-channel Interference
Traffic
QQ Business Hotel 1 QQ Business Hotel 2
The traffic of interference of neighboring cells with a serving cell indicates the following information:
1. Level strength of the neighboring cells
2. The cross traffic of the serving and neighboring cells
Frequency resources are insufficient. Therefore, you can assign the same frequency to two cells with
small interference traffic but not two cells with high interference traffic.
During frequency planning for newly deployed BTSs or when a large number of newly
deployed BTSs access the network during frequency replanning, the U-Net is required
for generating interference matrix based on coverage prediction.
Data sources
Electronic map and engineering parameters
Principles
Based on the difference of prediction signal level of serving cells and prediction signal level of neighboring
cells, we calculate the co-frequency or adjacent-frequency interference traffic volume for the cell-cell.
Interference matrix based on coverage prediction
If all BTSs or cells on the network are newly deployed, import the interference matrix generated according
to coverage prediction into the AFP tool, and then perform frequency replanning calculation.
If certain BTSs are newly deployed, compare, normalize, or combine various interference matrices
provided by the AFP and then add the interference relationship of newly deployed BTSs to the
interference matrix generated based on the MR. Then, use the combined interference matrix to perform
frequency replanning calculation.
The precision of the interference matrix based on coverage prediction depends on the precision of electric
maps, engineering parameters, simulation performed on the U-Net.
The outputted frequency scheme needs to be imported to the old Nastar platform for manual check. Generally, you are not
recommended to greatly adjust the frequency scheme output by the AFP if the interference matrix is accurate. For two cells
whose azimuths are opposite to each other, conduct the following analysis according to the interference matrix:
(1) If the interference traffic of the two cells is low, you do not need to adjust the frequencies.
(2) If the interference matrix is incorrect, you need to analyze which step is abnormal.
(3) If the tool assigns same frequencies to the two cells in consideration of interference traffic on the whole network, you
need to adjust the frequencies manually by using the old Nastar platform.
Total interference traffic of the entire network: If the total interference traffic reduces after
frequency replanning, it indicates that the frequency replanning is feasible. If the interference
matrix is accurate, this assessment method is most effective.
Number of co-channel and adjacent-channel collisions: Compare the numbers within a range
that is 1.5 times greater than the cell distance before and after frequency replanning. If the
number reduces after frequency replanning, it indicates that the frequency replanning is feasible.
This assessment method is based on geographical locations and therefore cannot reflect the
actual condition of the existing network.
Emulation: Compare the percentages of different ranges of C/I before and after frequency
replanning. If the percentage of C/I values that are greater than 9 increases, it indicates that the
frequency replanning is feasible. Closely related to electric maps, engineering parameters, and
parameter configurations, the emulation is hard to control.
Co-channel handover proportion: Count how many times two cells that have handover
relationship share frequencies or the frequencies of the two cells are adjacent before and after
frequency replanning. If the times reduce after frequency replanning, it indicates that the
frequency replanning is feasible.
Co-frequency 208 38 45 1 15 5 1 3
TCH 26571 26252 15019 14574 4950 4305 2908 2271 935 551 785 386
Uplink adjacent-
10105 10462 6334 7024 2191 2141 1284 1111 428 282 359 199
frequency
After the planning, the co-frequency interference value (Int Co) and adjacent-frequency value (Int
Adj) decrease at different extent. The decrease in the interference value helps to improve the
network quality. The co-frequency and adjacent-frequency interference traffic volumes on the entire
network reflect the actual traffic interference status on the live network. After the optimization, the co-
frequency traffic decreases by 69% and the adjacent-frequency traffic volume decreases by 12.6%.
This method is mainly used for selecting a proper solution among multiple sets of frequency options.
After obtaining the MR data, use the Nastar to generate an interference matrix, import the
interference matrix into the AFP tool, and then perform frequency replanning calculation.
According to the comparison of the interference traffic volume, co-frequency and adjacent-
frequency handover, and CI simulation evaluation, the new frequency solution can bring gains.
After the frequency scheme is implemented on the network, you need to wait for one or two
days until the scheme is stable. In the waiting, you need to pay attention to top poor cells.
Process the cells where the ratio of the sum of the third to fifth values of Interference area
in measurement report (TRX) to the sum of the first to fifth values and the ratio of the sixth
and seventh values of Receiving quality allocation in measurement report (TRX) to the
sum of the first to seventh values are abnormal as follows:
1. Pay attention to the proportion of the frequency values in the "0 to 3"group and
the "0 to 5"group of the MR. RecvQualOrig. TRX.
2. Pay attention to the proportion of the frequency values in the "3 to 5" group of
the MR. Iterf. TRX.
3. Pay attention to the KPIs required by customers.
4. Use the counters in the following template to evaluate the frequency replanning
result. Add or delete counters and change the calculation formulas according to
onsite situations and customers' requirements.
Note: Frequency replanning is designed to reduce the frequency interference and improve the speech
quality of current networks. It can also reduce call drops and handover failures caused by poor network
quality due to interference. Frequency replanning, however, cannot reduce call drops and handover failures
caused by coverage problems or other reasons.
Increase After
Counter Name May 21 to May 27 May 30 to June 5
Frequency Replanning
UL HQI (0~5) 99.08% 99.14% 0.06%
DL HQI(0~5) 99.29% 99.43% 0.14%
Proportion of interference
93.83% 95.05% 1.23%
bands from 1 to 3
UL HQI (0~3) 96.74% 97.21% 0.48%
DL HQI (0~3) 96.07% 97.03% 0.96%
According to the preceding table, the UL HQI (0–3) increases by 0.48%, the DL
HQI (0–3) increases by 0.96%, and the interference bands from 1 to 3 increases by
1.23%. Other KPIs increase steadily. For details, see the Frequency Replanning
Evaluation Performed by an Operator at a Provincial Capital of China.
Background: After Huawei performs a non-Like-to-Like swap for an operator 1 in Europe, the call drop rate in certain clusters
fails to meet the operator's requirement, and Huawei faces penalty by the operator.
Troubleshooting:
1. Analysis on the causes of the cluster-level call drops:
Check through DT whether intranet interference causes the call drops.
2. Analysis on the causes for BTSs with the highest call drop rates:
(1) It is found that the high call drop rates in these BTSs are due to poor quality
(2) Through frequency scanning, it is found that the poor quality results from intranet interference
Through frequency scanning, it is found that the poor quality results from intranet interference. It is predicted that the call drop
rate can be optimized after frequency replanning.
The output is as follows:
After the BCCH frequency replanning is performed in about 40 cells in a cluster,
counters recover (see the following figure).
Note
1. Frequency replanning is a paid service project. In principle, the frequency
replanning cannot be provided for free to improve network quality during the
migration; otherwise, the subsequent professional service sales are affected.
Unless in special projects, the frequency replanning center is entitled to reject
the application for offering free frequency replanning.
2. Before the frequency replanning, a specific analysis conclusion must be given:
The unreasonable intranet frequency plan causes the poor network
performance.
0.7 80000
Perform frequency replanning to the BSC. Then the counters recover (see the right figure). 0.6
70000
60000
0.5
Note: 0.4
50000
40000
0.3
30000
1. Formulate the frequency planning policy with caution. Identify risks as many as 0.2 20000
0.1
possible before the frequency replanning and continue to optimize the project after the 0
10000
0
cannot be provided freely to improve network quality during the migration, or the
following professional service sales are affected.
Emulates the frequency schemes before and after the frequency replanning and
3 U_Net
compares the values of C/I.
Counts the number of co-channel cell-cell in handover relationships. Compares the
Tool for calculating co-
4 advantages and disadvantages of the frequency schemes before and after the
channel handover rate
frequency replanning.
Tool for counting the
Counts the number of co-channel or adjacent-channel cell-cell in the frequency
number of collisions of
5 scheme base on engineering parameters. Compares the advantages and
intra-frequencies or
disadvantages of the frequency schemes before and after the frequency replanning.
adjacent frequencies
1. Optimizes the BSIC.
6 Stand-alone Nastar
2. Checks whether the frequency scheme is proper after the frequency replanning.
Subtopic Link
1. Capacity Contents
2. Coverage Contents
3. Interference Contents
4. CoBCCH Contents
5. Dual-Band Networking Policies Contents
6. Location Area Contents
7. Neighboring Cell Optimization Contents
8. Frequency Replanning Contents