Ericsson Spectrum Re-Farming

Ericsson Internal SERVICE DELIVERY INSTRUCTION Prepared (also subject responsible if other)

No.

SLU/PD Jose Antonio Gimenez

370/102 60-FAY111 10

Approved

Checked

Date

Rev

EAB/GB/PB EAB/GB/PB Stevan Filipovic

EBCSFIL

15/5/2011

E

Reference

Spectrum Re-Farming Technical Guideline

1 (89)


2 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

Contents 1

Introduction Introduction ............................ .......................................... ............................ ............................ ............................ ...................... ........ 4

2

Description Description .......................... ......................................... ............................. ............................ ............................ ........................ .......... 4

3

Background Background ............................ .......................................... ............................ ............................ ............................ ...................... ........ 4

4

The Service and Expected Results....................................................... 4

5

Service Delivery process ...................................................................... 5

6

Feasibility Feasibility Study .......................... ......................................... ............................. ............................ ............................ ................7 6.1 General Network Audit ............................................................... 8 Several points have to be checked It consists in analyzing the overall Radio Network by checking the following points : .................... .......... ............. ... 8 6.2 WRAN/LTE Audit ........................... ......................................... ............................ ............................ ................ .. 8 6.3 GRAN Audit – Frequency Plans evaluation ............................... 9 6.3.1 xAFP project creation ................................................................. 9 6.3.2 Nbr Audit ............................. ........................................... ............................ ............................ ......................... ...........10 6.3.3 Traffic offloads of floads and TRX dimensioning impact ..................... .......... ............... .... 10 6.3.4 Frequency Plan strategies ....................................................... 13 6.3.5 Scenario Comparison Compari son and Degradation Analys is .................... .......... ............ 19 6.3.6 Study analysis with frequency plan comparison ...................... 23

7

Pre Refarming Refarming actions ............................. ........................................... ............................ ............................ ................ .. 24 7.1 Features introduction ............................................................... 25 7.1.1 Prerequisite for FLP ................................................................. 25 7.1.2 Strongly Recommended .................... .......... ..................... ..................... ..................... .................. ....... 25 7.1.3 Recommended Recommended ............................. ........................................... ............................ ............................ ................ .. 27 7.2 RF Optimization Optimization ........................... ......................................... ............................ ............................ ................ .. 28 7.2.1 General process ..................... ........... ..................... ..................... ..................... ..................... .................. ........ 29 7.2.2 Targets ........................... ......................................... ............................ ............................. ............................. ................ 32 7.2.3 Parameters to Optimize ........................................................... 32 7.2.4 Optimization guidelines ............................................................ 32 7.3 RRM Optimization Optimization ............................ .......................................... ............................ .......................... ............34 7.3.1 2G Traffic Balancing ................................................................. 35 7.3.2 3G/LTE Optimization process .................................................. 41

8

Refarming Refarming Realization Realization ............................. ........................................... ............................ ............................ ................ ..42 8.1 Frequency planning strategy definition .................................... 42 8.2 Measurement data input consistency check ............................ 43 8.2.1 BAR Files/ICDM Files/ICDM ........................... ......................................... ............................ ............................ ................ ..43 8.3 Frequency plan methodology ................................................... 44 8.3.1 FLP planning planning ........................... ......................................... ............................ ............................ ..................... .......44 8.3.2 Ad Hoc planning ..................... ........... ..................... ..................... ..................... ..................... .................. ........ 48 8.4 Ad-Hoc frequency plan creation c reation with xAFP .................... .......... .................... .......... 51 8.4.1 General Process ...................................................................... 51 8.4.2 Targets ........................... ......................................... ............................ ............................. ............................. ................ 55 8.4.3 Optimization guidelines ............................................................ 55 8.5 Implementation Implementation............................. ........................................... ............................ ............................ ................ .. 59


3 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

8.6

Reference

Neighbors Neighbors Update ............................ .......................................... ............................ .......................... ............59

9

Post Refarming actions ....................................................................... 60 9.1 Performance Evaluation .................... .......... ..................... ..................... ..................... .................. ....... 60 9.2 RRM Optimization Optimization ............................ .......................................... ............................ .......................... ............61 9.2.1 2G Optimization Optimization ........................... ......................................... ............................ ............................ ................ .. 61 9.2.2 3G/LTE Optimization .................... ......... ..................... ..................... ..................... .................... ............. ... 71

10

Acceptance and Conclusion ............................................................... 71

11

Roles and Competence Competence .......................... ......................................... ............................. ............................ ................ .. 71 11.1 Delivery Delivery manual ........................... ......................................... ............................ ............................ ................ .. 71

12

APPENDIX A. Co-Existence between GSM and WCDMA in the Same Frequency Band ........................... .......................................... ............................. ............................ ................ .. 72 12.1 Interference and site scenarios ................................................ 73 12.2 WCDMA DL capacity loss due to GSM .................................... 75 12.3 WCDMA UL capacity loss due to GSM .................................... 75 12.4 GSM UL capacity loss due to WCDMA .................................... 75 12.5 GSM DL capacity loss due to WCDMA .................................... 76 12.6 Conclusion Conclusion ............................ .......................................... ............................ ............................ ....................... ......... 77

13

APPENDIX B. Impact of BTS Hardware on xAFP settings ............... 77 13.1 Hybrid combiner ........................... ......................................... ............................ ............................ ................ .. 78 13.1.1 Description Description ............................ .......................................... ............................ ............................ ....................... ......... 78 13.1.2 GSM frequency restrictions implications .................................. 78 13.2 Cavity combiner ........................... ......................................... ............................ ............................ ................ .. 79 13.2.1 Description Description ............................ .......................................... ............................ ............................ ....................... ......... 79 13.2.2 GSM frequency restrictions implications .................................. 80 13.2.3 Hopping capability implication .................................................. 81 13.3 Antenna combining (also called cal led air combining) .................... ......... ............... .... 81 13.3.1 Description Description ............................ .......................................... ............................ ............................ ....................... ......... 81 13.4 Wide band Amplifier combining .................... .......... ..................... ..................... ................. ....... 82 13.4.1 Description Description ............................ .......................................... ............................ ............................ ....................... ......... 82 13.5 Repeaters Repeaters............................ .......................................... ............................ ............................ ......................... ........... 85 13.5.1 Synchronization equipment ..................... .......... ..................... .................... ..................... ............. .. 85 13.5.2 Legacy equipment restriction ................................................... 85

14

APPENDIX B. Ericsson Radio Equipment solutions ........................ 86 14.1 New Ericsson filter in the node B ................... ..................... .......... ................ ..... 86 14.2 Antenna solutions ..................... ........... ..................... ..................... .................... ..................... ................. ...... 87

15

References References........................... .......................................... ............................. ............................ ............................ ...................... ........ 88



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

1

4 (89)

No.

Reference

Introduction This document is aimed to be used as reference information for the engineers who will deliver the GSM Spectrum Refarming Service.

2

Description The technical guideline comprises relevant technical information and important details onhow to deliver the Spectrum Refarming service.

3

Background The radio frequency spectrum is a scarce and expensive resource so it should be used in the most efficient way. Usually, a GSM mobile operator starts facing frequency spectrum issues when one of the following situations happens: •

•

•

The regulator/government decides to reduce the available spectrum for the operator (i.e. Frequency Load 1 increases); Opportunity to deploy a new radio technology in a spectrumalready occupied by another technology (for example, deploy WCDMA at 900 MHzwhen there is already a GSM900 in place); Increasing traffic in parallel with quality degradation (i.e. Frequency Load is becoming too high for the Frequency Planning techniques currently used).

For all these scenarios a Spectrum Refarming service will be needed. Ericsson Spectrum Refarming for GSM Networks is a service that makes possible an existing GSM Network to operate with a tighter spectrum with minimum disturbances for performance, quality and capacity of the network.

4

The Service and Expected Results The solution offered by Ericsson for a GSM Spectrum Refarming Serviceconsists in reducing, to a certain limit, the spectrum needed for a certain capacity without compromising the overall performance.

1

Frequency Load is the recommended way of calculating the number of TRXs per cell. See reference [1] for its definition and calculation.


5 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

This goal is reached by means of optimized frequency planning techniques in combination with advanced OSS-based optimization tools plus Ericsson features used for improving capacity, reducing interference, and distributing traffic between bands and technologies. Refarming services are nowadays especially requested by operators when they want to introduce new radio technology in their GSM limited spectrum. I.e. in recent years, WCDMA requirements for operation in the 900 MHz band have been standardized in 3GPP paving the way for the Refarming of GSM900 spectrum to WCDMA 900 deployment. We can also expect that LTE requirements for operation in the 900 MHz band may become true in the near future. WCDMA900 combines the benefits of WCDMA and better propagation of lower frequency waves. WCDMA900 offers CAPEX gains (i.e. less number of NodeBs) in rural areas and better in building penetration in urban areas compared to WCDMA2100. In this casethe expected result of the Spectrum Refarming Service by the operator is to free up enough frequency band to deploy a WCDMA or LTE technology in its current GSM band and still have the GSM network performing as close as possible to how it was before the spectrum reduction.

5

Service Delivery process


6 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Figure 1

Reference

Spectrum Refarming for GSM Networks Delivery Process

After a feasibility study, when an advanced advance d analysis using simulations simulati ons base on current OSS KPIs is performed to find best FP strategy, some actions may be required before carving the spectrum. For that purpose, Ericsson provides automatic processes based on advanced tools that minimize the cost of those operations. Once the new frequency plan is implemented, it is recommended to perform a new RRM optimization to make sure the rest of network settings are aligned with the new scenario. Finally, performance evaluation for final service acceptance will be carried out taking advantage of the current OSS data interfaces set up in place for the tools used during the process. The delivery process of the service is divided in seven sub-processes. •

Feasibility Study

•

Pre Refarming actions

•

Frequency Re-Farm Plan elaboration

•

Implementation


7 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

•

Post Refarming actions

•

Performance Monitor and evaluation

•

Acceptance and Conclusion

Reference

Below a more detailed workflow is presented to complement high level solution description, indicating some of the actions that will be later described in detailed

Figure 2. Workflow - Ericsson Spectrum Refarming solution

6

Feasibility Study During the strategic advice phase a detailed analysis of the customer’s network is performed. In addition to this all requirements from the service are taken into consideration during a feasibility study. This study will generate a Feasibility Assessment and Scenario Analysis Report stating if the service is feasible, all limitations detected (software/hardware), expected results and the necessary actions to be taken during the service. These are the main steps to complete this phase:



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

6.1

8 (89)

No.

Reference

General Network Audit Several points have to be checked It consists in analyzing the overall Radio Network by checking the following points: •

Design Issues: both coverage and capacity;

•

Hardware Limitations (combiners/filters/antenna system);

•

BSS Release and installed features;

KPIs

•

•

6.2

Other network technology/layer which might absorb traffic from the re-farmed layer.

WRAN/LTE Audit For the other new/existing technologies it is important to assess the impact that the refarming may cause and the feasibility of the use of the carved spectrum for those technologies. As for the WCDMA system, regarding the introduction of a carrier in the re-farmed band, the first task is to assess the coexistence of GSM and WCDMA which will be addressed in [Co-Existence Issues between GSM and WCDMA in the Same Frequency Band] Band]. It is also necessary to assess the possibility to offload the GSM network in case the spectrum reduction does notallow the network to maintain sufficient quality with the current traffic levels2. Several drivers will also be potentially analyzed: o

Increase of WCDMA traffic (add capacity)

o

Introduction of a new service (e.g. MBMS)

o

o

2

Improve WCDMA coverage and increase traffic served by WCDMA Improve PS users and HSPA performance

This is in case the main reason for GSM spectrum re-farming is a regulatory reason. In general, this is always needed in order to manage the transition phase before an addition WCDMA 900 layer is introduced.



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

6.3

9 (89)

No.

Reference

GRAN Audit – Frequency Plans evaluation Complete GSM audit, including the evaluation of several FP alternatives will be performed at this step. For this purpose, xAFP toll will be used, following the next procedure xAFP Project Creation Physical + Network data ñam

Run Measurement Campaign

Traffic, IM, HOs, QoS Data

Neighbor Audit

Yes Offloading Coverage vs. Ca acit la ers? Apply selected Offloading methodology (inc. TRXs forecast) No

Best Frequency Strategy Scenario Identification

Top Worst Cells Analysis Baseband Hopping /No

Synthesized Frequency

Hopping

Hopping Hopping Type?

MAL length Reduction /

Extra TRXs deletion

Hopping Type Change

Final Strategies /Actions for Frequency Refarming Planning

6.3.1

xAFP project creation See section 8.4.1f section 8.4.1f or or detailed instructions



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

6.3.2

10 (89)

No.

Reference

Nbr Audit See section 8.6 section 8.6on on details how to perform NBR audit with xAFP

6.3.3

Traffic offloads and TRX dimensioning impact The main purpose of this section, is to identify the maximum amount of combined traffic in GSM coverage layer (900/850 Band) to be moved based on RxLev statistics to GSM capacity layer (1800/1900 Band). This step is dependent on the strategy adopted for the refarming, i.e. offloading traffic between GSM Coverage vs. Capacity layers may not take place in every refarming scenario. In the case there is need to analyze beforehand the expected performance when offloading traffic between GSM layers, Ericsson is proposing two methodologies in order to estimate the amount of traffic & TRXs expansion need. The decision on which methodology to use will be a compromise between accuracy of the estimation and effort required for the analysis.

6.3.3.1

Methodology 1: TRX Utilization Global methodology: •

•

•

•

Adapt the current traffic share between Coverage/Capacity layers to a new TRX configuration assumed after the refarming takes place. Identify the amount of extra GSM-capacity layer TRXs needed to support the offloaded traffic from GSM-coverage layer. The traffic that cannot be carried by GSM layers will be designated as the amount of traffic that must be migrated to the 3G network or lost. Create a traffic profile with the proposed offload in order to be used in different scenarios to be evaluated.

Dimensioning Process: •

•

Determine initial and post-refarming TRX configurations for every layer per site (i.e. moving from 2/2/2 in every layer to 1/1/1 coverage and 3/3/3 - capacity) Set an expected HR usage for every layer.


11 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

•

Reference

Computation of needed TRXs in Capacity Layer to support offloaded traffic will be based on Blocking > 2% (Erlang-B formula).

Dimensioning Process Outputs •

•

•

6.3.3.2

Traffic Profile. This traffic profile will contained the result of the offload between GSM layers, and it will be used during scenario evaluations. Number of TRXs to be removed in Coverage-Layer due to estimated traffic moved to Capacity-Layer. Number of TRXs to be added in Capacity-Layer due to traffic offloads from Coverage-Layer.

Methodology 2: RxLev based + Capacity layer IM scaling Global methodology: •

•

•

Move the High Signal Level traffic from GSM-Coverage layer to GSM-Capacity layer. Identify the amount of extra GSM-capacity layer TRXs needed to support the offloaded traffic from GSM-coverage layer. The traffic that cannot be carried by GSM layers will be designated as the amount of traffic that must be migrated to the 3G network or lost. Modify the interference matrix to model the offload effect.

•

•

Create a traffic profile with the proposed offload in order to be used in different scenarios to be evaluated.

Dimensioning Process: •

The traffic that is stronger than the Low Signal Strength cutoff threshold should be encouraged to move to the GSM-capacity layer. o

Cuttoff threshold = -90 dBm Each GSM-Coverage layer cell will be assumed to have:

•

o

1 SDCCH


12 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

•

•

•

Reference

o

1 EDGE timeslot allocated on BCCH carrier

o

1 BCCH timeslot

o

5 timeslots for voice on BCCH carrier

All GSM-Coverage layer cells will have 2TRXs (1 BCCH and 1 TCH) if traffic is greater than 5 E, otherwise the cells will have 1 TRX (BCCH only). The Half Rate usage will be assumed to be 100% for both GSM layers. Computation of needed TRXs in Capacity Layer to support offloaded traffic will be based on Blocking > 2% (Erlang-B formula), considering the number of timeslots for data traffic. In any case, the total number of TRXs will not exceed 4.

Dimensioning Process Outputs •

•

•

Traffic Profile. This traffic profile will contained the result of the offload between GSM layers, and it will be used during scenario evaluations. Number of TRXs to be removed in Coverage-Layer due to High Signal Strength traffic moved to Capacity-Layer. Number of TRXs to be added in Capacity-Layer due to traffic offloads from Coverage-Layer.

Interference Matrix offload effect modeling The IM which has to be used to simulate scenarios where the offload effect is under study must be processed in order to follow the traffic movement in the proper way. The offload effect will have impact only in GSM-Capacity layer IM. The IM C/I mean will be decreased depending on the offloaded traffic from GSM-Coverage layer to GSM-Capacity layer. The calculations must be performed as described as follows: •

•

PCO (Probability co-channel interference) = % C/I < 12 dB = NORMDIST (12, C/I mean, C/I std, 1). Offload = [E (before) – E (after)]/E(after)  Defined for Coverage-Layer offload.


13 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

•

•

•

Reference

Factor Alpha = [E(Coverage-Layer, before) – E(Coverage-Layer, after)]/E(Capacity-Layer, before) PCO(Capacity-Layer, before) = #Bad Erlangs / E(CapacityLayer, before) PCO(Capacity-Layer, after) = [#Bad Erlangs+HSE(CoverageLayer)]/[E(Capacity-Layer, before)+HSE(Coverage-Layer)] = (PCO(Capacity-Layer, before)+Factor Alpha)/(Factor Alpha+1) Adjusted Mean = 12 – STD*Norm(PCO,0,1)

•

A conclusion, it has to be highlighted that the main target of this process is to obtain a traffic offload which is going to relax the interference conditions in order to obtain the minimum degradation as possible after frequency refarming. The outputs provided by this process are: •

Traffic profile

•

Scaled IM

•

TRXs forecast

Attached a presentation containing c ontaining details about Traffic Traf fic offload process and xAFP interaction:

Traffic Offload and TRXs Study.pptx

6.3.4

Frequency Plan strategies See section 8.4 on details how to generate a frequency plan following a particular strategy The objective of this step is to generate the FPs of the potentially optimum strategies for the refarming scenario. In general, the following guidelines are used to select different strategies according to the particular conditions of the refarming: •

Amount of available spectrum after refarming:


14 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

o

High (>10MHz) -

Split-Band is usually the best performing plan in these scenarios.

-

The MAL will be automatically generated thanks to xAFP AdHoc optimizer, following the settings which were set up for the optimization. The user will define:

•

Range of frequencies to be used by BCCH layer

•

Range of frequencies to be used by TCH layer

•

o

Reference

MAL constraints: MAL length, MAL planning level (assign the MAL per sector or per site).

Medium (5-10MHz) -

Semi-AdHoc: Semi-AdHoc: Launch several scenarios to determine the number of required BCCH channels to keep acceptable performance. All the available channels will be used by TCH layer.

•

•

•

Depending on the amount of channels for BCCH in baseline configuration, the user will select different semi-adhoc scenarios changing the number of BCCH channels, testing the same number of channels than in baseline, lower values and higher values. The TCH layer will be planned by using the AdHoc optimizer, the MAL will be automatically generated following the specified constraints, as explained in previous scenarios. In this type of scenario it is possible to have cochannel reuses between layers, it is recommended to load power control statistics to help the optimizer to model/solve the inter-layer interference.


15 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

-

Split-Band: Split-Band: better performance as much spectrum is available. No co-channel reuse between layers, good approach when the power control is not enabled in the network.

•

•

-

As in the previous scenario, the user must create different blocks of channels in order to look for the amount of BCCHs which allows reaching an acceptable performance. The MALs will be automatically generated by AdHoc optimizer, as explained before.

Staggered: Staggered: as split-band but BCCH and TCH channels are selected alternatively

•

•

-

Reference

This type of scenarios avoids the co-channel between BCCH and TCH although the adjacency interference is clearly increased due to channel distribution. The MALs will be automatically generated by AdHoc optimizer, as in previous scenarios.

Staggered, TCH AdHoc. This AdHoc. This scenario introduces the possibility of using the entire available spectrum by TCH layer. The advantage of this type of scenarios is that there is no adjacency between BCCHBCCH.

•

•

The co-channel between TCH and BCCH is possible in this scenario, the power control statistics could help to solve the channel reuse in the proper way. The MALs have to be generated automatically using the AdHoc optimizer.


16 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

-

Reference

Full Ad-Hoc. Ad-Hoc. Important strategy when the baseline scenario was using a big amount of channels for both layers. It is important to activate/import Power Control statistics in this type of scenarios.

The AdHoc optimizer has to manage the cochannel and adjacency interference between layers.

•

The MALs will be automatically generated by the AdHoc optimizer, like in previous scenarios.

•

o

Small (<=5MHz) -

Synchronized 1/1 or 1/3 clusters. clusters. BCCH channels are selected. No need of TCH planning, MAIO and HSN will avoid the collisions. All cells will have the same reuse. The 1/3 strategy has better performance than 1/1 (as there is some gain due to frequency diversity).

•

•

The MALs will be pre-defined by user, using the total amount of channels to be used by TCH layer. In the case of running a 1/3 synchronized cluster, the “hopping layer optimizer”, which is one optimization engine available in xAFP tool, will assign the MALsper sector in the optimal way; meanwhile the HSN and MAIO optimizers will avoid the TCH collisions. •

v


17 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

Other factors:

•

o

o

o

Amount of microcells existing. If it is possible the microcells might use different spectrum than the macros. If the spectrum is reduced, then the conditions have to be relaxed and micro/macro will share the spectrum. HW limitation (combiners, minimum separation required, etc., see Appendix A&B) Environment (dense urban, urban, rural). In dense urban/urban areas where the interference level is potentially high, the impact of the frequency carving is higher. For these urban scenarios, the actions to be done should be more drastic: •

•

o

Reducing the MAL length as maximum as possible in case that the hopping type is Synthesized Frequency Hopping. Performing drastic TRXs reductions (being more aggressive during the TRX/traffic study which is detailed in section 6.3.3).

Changes on Hopping Type. For those scenarios where the hopping type could be changed, it is recommended to run a scenario with a different hopping strategy than in “baseline” frequency plan, in order to verify if the change could provide some gain. This action could be particularly important in those cases where after the TRXs adjustment and the traffic offload cannot be juan faggressive due to network net work restrictions (it is not a dual band network or the Capacity-Layer cannot manage a high percentage of offloaded traffic from Coverage-Layer). When the final available spectrum is reduced, changing the hopping type from Synthesized Frequency Hopping (SFH) to Base Band Hopping (BBH) could decrease the level of interference by losing some frequency diversity gain which is provided by the SFH. By other hand, changing the hopping type from SFH to BBH will generate a robust network against fading effects.


18 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

SFH highlights:

•

o

o

o

Frequency diversity. Having a hopping list whose length is higher than the number of TRXs provides gain and protection against fading. For frequency carving, it is recommended to reduce the MAL length as maximum as possible in urban areas, keeping more channels for rural sectors, where the fading effects could affect the performance. Robust systems when the TRX has a failure. When one of the TRXs of a certain cell is failing, another one may assume the calls and avoid the complete failure of the cell. BBH Highlights

•

o

o

o

Good approach when the spectrum is reduced, since the number of channels in air-interface is the same than the number of TRXs. It allows including the BCCH frequency in the hopping list of channels (base band hopping on BCCH). When a TRX is down, the cell fails, it allows to quickly identify hardware failures.

Changing the hopping type in a network is something which must be carefully planned, since the hopping type change implies extra operations during the frequency plan implementation. It is a timeconsuming operation which limits the number of changes that could be done during the implementation of the optimized solution. Assuming that the final decision is changing the hopping type of the network, the planned are must be resized and planned in clusters/stages in order to perform the implementation in the proper way. o

Traffic allocation method

Changing the traffic allocation method may help to introduce some gain in the optimized solution. After the carve off of the frequencies the optimization effort may be focused in one of the layers (BCCH of TCH). If the traffic is allocated in the layer which is suffering lower levels of interference, the global quality could be improved.


19 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

Combining the TRX adjustment and traffic reallocation study which is described in section 6.3.3, the channel reuse per layer of the optimized solution should be evaluated in order to look for the better traffic allocation according the given solution:

6.3.5

-

Good BCCH plan. When the optimized BCCH layer is predicting a good performance, the traffic allocation method might be changed to: BCCH First

-

Good TCH/MAL plan. When the optimized TCH layer is predicting a good performance, the traffic allocation method might be changed to: TCH First

-

BBH scenarios. In this type of networks, the traffic allocation method is always “random”, since the hopping is performed on the TRXs fixed frequencies instead of using a MAL. One possibility in this case could be forcing the system to be: Base Band Hopping on BCCH , or not, if the BCCH optimized plan is good, it is advisable to use the BCCH frequency as part of the hopping list.

Scenario Comparison and Degradation Analysis Objective of this step is to rank the different FPs according to the simulation results provided by the xAFP QoS module. This comparison should be performed in several phases in order to reach the optimal solution: 1.

Simulate different frequency strategies looking for the best approach.


20 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

The criterion that must be followed to select the frequency strategy is looking for the case which is closer to Baseline predicted performance. Some other checks must be considered, like the neighbor reuses and the percentage of interfered HOs as showed in the table below.

Baseline Neighbor reuses Co-Channel

Adj-Channel

2nd Nbr Cochannel

Full AdHoc

Semi AdHoc

Reuses

# HOs

% HOs

Reuses

# HOs

% HOs

Reuses

# HOs

% HOs

BCCH

BCCH

81

12738

0.02%

0

0

0.00%

0

0

0.00%

BCCH

TCH

141

81309

0.10%

19

1237

0.00%

27

1575

0.00%

TCH

BCCH

139

90287

0.11%

7

91

0.00%

6

20

0.00%

TCH

TCH

1742

1111731

1.40%

279

4612

0.01%

260

3471

0.00%

BCCH

BCCH

1463

1028087

1.29%

1554

918288

1.16%

1193

232398

0.29%

BCCH

TCH

1124

813477

1.02%

4264

4581451 4581451

5.77%

3944

4684416 4684416

5.90%

TCH

BCCH

1006

673980

0.85%

4247

4624848 4624848

5.82%

3892

4814026 4814026

6.06%

TCH

TCH

9240

12993761

16.36%

11482

16224992

20.43%

12185

17034436

21.45%

BCCH

BCCH

17163

15458500

19.47%

12954

9012360

11.35%

14754

10969600

13.81%

BCCH

TCH

18416

--

--

67691

--

--

59870

--

--

TCH

TCH

102515

--

--

121687

--

--

128800

--

--

HO attempts

Baseline

Total

79410100

With Average FHgain < 20 dB

1119610

Full AdHoc

Semi AdHoc

79411900 1.4%

96147

79411900 0.1%

76022

0.1%


21 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

BCCH-BCCH-CO (1st order Nbr) BCCH-BCCH-CO (2nd order Nbr) INTERFERED by 2nd order Nbr BCCH-BCCHCO BCCH-BCCH ADJ

Reference

12738

0.0%

0

0.0%

0

0.0%

15458500

19.5%

9012360

11.3%

10969600

13.8%

16897500

21.3%

9127360

11.5%

11325700

14.3%

1028090

1.3%

918288

1.2%

232398

0.3%

These scenarios should be generated taking in consideration all the suggestions described in section 6.3.4. 2.

The traffic profile, the load study and the scaled IM which is the result of all the proposed studios performed on section 6.3.3, must be considered to run again the simulations. Selecting only the best scenarios from step one, the simulations will be re-launched with the modifications in order to analyze the impact:

3.

Once the best plan is identified, use the Cost per frequency reports in order to identify worst contributors from interferenceand frequency limitation point of view. Combining these reports with TRX Usage report provided by xAFP it could be possible to remove some extra of the top-worst performing sectors thus improving the situation prior refarming. These are the xAFP tables to be considered: •

•

AdHoc Detail Cost  Identify sectors with the higher neighbor and IM costs AdHoc Cost per Frequency  Identify the most interfered frequencies/TRXs.


22 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

•

Reference

C/I contribution (Boomer)  Identify sectors with high interfered traffic rank.

The analysis of these three tables might provide a proposal of TRXs deletion or MAL size reduction (in case of SFH scenarios). Another action over the top-worst sectors could be extra-neighbor optimization (delete neighbors in order to relax the optimization constraints). The table below shows an example about the analysis performed with xAFP tool:

Cell Name SECTOR1 SECTOR2 SECTOR3 SECTOR4 SECTOR5 SECTOR6 SECTOR7

NUMBER TRX BCCH + 5 TCH BCCH + 3 TCH BCCH + 3 TCH BCCH + 4 TCH BCCH + 2 TCH BCCH + 7 TCH BCCH + 2 TCH

Freq cost

Most interfered Channel (TRX)

Interfered Traffic Rank

Average Distance Boomer (Km)

13.97

19.95

2

199.13

1.45

27.22

19.69

8.67

18

197.98

2.81

24.51

23.26

3.44

19

256.45

2.92

24.37

9.62

17.16

13

171.18

11.85

22.88

21.86

4.36

60

204.59

9.13

22.43

17.73

6.37

21

300.39

12.28

20.31

16.61

7.94

51

114.11

9.37

Total Cost = AdHoc cost+IM cost

C/I Cost (IM)

34.76

Action

Reduce #TRXs Reduce #TRXs Reduce #TRXs Reduce #TRXs Reduce #TRXs Reduce #TRXs Reduce #TRXs

The performance analysis of the best scenario should be accomplished again, taking actions on top worst performing sectors, in order to verify if these actions could finally improve the previous results.


23 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

Deliverable: Best FP proposal based on QoS simulation results 6.3.6

Study analysis with frequency plan comparison This section tries to describe the main parameters to decide which scenario may be chosen during the scenario evaluation process described in section 6.3.5. Quality indicators which must be compared:

o

•

•

•

•

o

% of Traffic with COMBINED Bad RxQual / Bad FER (Coverage + Interference and Interference Only). % of Traffic with BCCH Bad RxQual / Bad FER (Coverage + Interference and Interference Only). % of Traffic with TCH Bad RxQual/ Bad FER (Coverage + Interference and Interference Only). % of Traffic with Bad Signaling (Coverage + Interference and Interference Only). Neighbor reuse indicators which must be compared:



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

•

•

•

•

Reference

Number of co-channel reuses with BCCH layer and % of interfered HOs due to these reuses. Number of co-channel reuses with TCH layer and % of interfered HOs due to these reuses Number of adj-channel reuses with BCCH layer and % of interfered HOs due to these reuses. Number of adj-channel reuses with TCH layer and % of interfered HOs due to these reuses. Important considerations related to network refarming:

o

•

•

7

24 (89)

No.

Due to the carving of frequencies it is general expected that the quality indicators of the simulated scenarios will degrade the Baseline situation. The target of the comparison is to minimize the degradation of xAFP quality indicators and even improve (slightly) the simulated Baseline performance, if it is possible. Due to the spectrum reduction action, the number of reuses of simulated scenarios will be higher (in general) than in Baseline situation. The target of this analysis is to move the frequency reuses to those sectors where the impact of the reuse is lower (i.e. increasing the number of adjacencies in order to improve the co-channels or reducing the % of interfered HOs).

Pre Refarming actions Before the refarming takes place, and once the feasibility study is available, there are some actions that may need to take place prior refarming -

Activation of some features needed in order to boost capacity (AMR, FR/HR allocation), interference management (Power Control), traffic steering (Hierarchical Cell Structure)

-

RF Optimization: improving RF design may be a requisite if some clear issues are encountered (i.e. boomers) and/or for better traffic absorption between GSM bands.



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

-

7.1

25 (89)

No.

Reference

RRM Optimization: specific cell-level parameters can be optimized in order to 1) Better cope with the expected traffic shift when refarming is done, 2) General improvement of network performance

Features introduction For the implementation of a successful Refarming it is important to reduce the effects of interference to as low as possible. This is done by activating and tuning advanced Ericsson BSS features and by implementing a FLP frequency plan. Below, these features are listed with their role in the network briefly described.

7.1.1

Prerequisite for FLP

Features

Feature Identity

User description description

Frequency Hopping

FAJ 122 288

217/1553-HSC 103 12

Flexible MAIO Management

FAJ 122 870

221/1553-HSC 103 12

Dynamic BTS Power Control

FAJ 122 910

249/1553-HSC 103 12/7

Dynamic MS Power Control

FAJ 122 260

250/1553-HSC 103 12/7

DTX Uplink

FAJ 122 256

73/1553-HSC 103 12/3

DTX Downlink

FAJ 122 287

73/1553-HSC 103 12/3

Synthesizer frequency hopping is a prerequisite for FLP. To get the most out of frequency hopping the cells should be synchronized, at least within a site (Transceiver Group – TG), and MAIO Management should be utilized. DTX and Power control are necessary to reduce the interference levels present in FLP networks, which are greater when compared to BB-hopping networks 7.1.2

Strongly Recommended In a scenario of higher frequency load there are features which are strongly recommended in order to improve the network performance under increased interference situation.These features are listed and described below:


26 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Features

Feature Identity

Reference

User description

Idle Channel Measurement

FAJ 122 909

274/1553-HSC 103 12/7

Interference Rejection Combining

FAJ 122 083

252/1553-HSC 103 12/7

Intra-Cell Handover

FAJ 122 290

260/1553-HSC 103 12/7

AMR FR

FAJ 121 055

248/1553-HSC 103 12/7

AMR Power Control

FAJ 121 353

AMR Radio Link Timers

FAJ 121 826

Handover Power Boost Enhanced Handover Success Rate(06B)

FAJ 122 429

Enhanced AMR Coverage (07B)

Idle Channel Measurement is continuously measuring the interference level on unused idle channels, and uses this information at allocation to set up new calls on the best available channels. It is very useful on high loaded networks. Interference Rejection Combining is a feature that with its algorithm, filters out the interference by comparing signals from two antennas on the uplink.Mobiles suffering from high interference benefit the most. The feature AMR offers an enhanced speech quality for AMR capable mobiles in a network. The enhanced speech quality also provides better coverage at the edges of the cell, thus making it possible to increase the coverage area. AMR power control is used to minimize the interference i nterference in the radio network by controlling the output power for terminals using the AMR speech codec separately. The result is the possibility to increase the radio network capacity as well as improve speech quality by applying more aggressive power control settings, made possible by superior AMR robustness. AMR Radio Link Timers introduces separate radio link timers for calls using AMR. This makes possible to set separate values of RLINK timers for AMR and non-AMR connections.



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

7.1.3

27 (89)

No.

Reference

Recommended More features can also help increasing capacity by relieving congestion and better distributing the traffic load:

Features Cell Load Sharing AMR HR

Feature Identity FAJ 122 911

259/1553-HSC 103 12/7

FAJ 121 358

248/1553-HSC 103 12/8

Dynamic Half Rate Allocation

FAJ 122 582

Dynamic FR/HR Adaptation

FAJ 121 361

Automatic FLP Reduced Power Level after Handover

FAJ 123 162

Tight BCCH Frequency Reuse

User description

FAJ 123 164 FAJ 121 813

253/1553-HSC 103 12/7

There are no drawbacks on using Cell Load Sharing (CLS) in a FLP 1/1 network. Thisis because CLS will only work within the hysteresis area and the connections that are closest to the neighboring cells are moved first. Thus, it is recommended to use CLS to the full extent. It is recommended to Set RHYST to RHYST to 100%. Automatic FLP provides self-optimization of FLP cell parameters by analyzing daily cell to cell interference recording allowing more traffic in existing cells and improves network performance.When used in synchronized radio networks provides ~20% of capacity increase. Reduced Power Level after Handover – it will reduce the power level on the new channel after a handover using measurement data from serving channel and target cell in both uplink and downlink for all BTSs and MSs allowing up to 20% mote traffic per cell due to up to 20% power reduction in the network. Tight BCCH Frequency Reuse enables the operator to apply significantly tighter frequency reuse of BCCH frequencies and thereby increasing traffic capacity in the network. The below features can make less frequenciesto be used for the BCCH plan and increase the number of hopping frequencies:

Features

Feature Identity

User description

Multi-Band Cell

FAJ 122 085

224/1553-HSC 103 12/7

Overlaid/Underlaid Subcell

FAJ 122 430

239/1553-HSC 103 12/11


28 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

The Multi Band Cell functionality enables the possibility to have more than one system type in a cell, e.g. GSM900 and GSM1800. GSM1800. One BCCH instead of two BCCHs per sector is used. This means that less frequencyis used for the BCCH plan in favorof the hopping CHGRs. i.e. if the BCCH plans are donewith 14 BCCH reuse for each system type,the multi-band cells will lead to 14additional hopping frequencies on the other band (i.e. 1800). The trunking gain will also be higher for packet data services. Another benefit is that the BA list will be shorter meaning more accurate measurements forhandover decisions. The number of needed neighbor relations will also be smaller. It is important to notice that it is mandatory that the antennas of both bands in each sectors follow the same orientation (azimuth), otherwise MBC introduction will cause a severe degradation. The OL/UL feature can increase the number of hopping frequencies when it is used to tighten the BCCH reuse. Another feature can be used us ed only under certain condition. conditi on.

Features Assignment to another cell

Feature Identity FAJ 122 286

User description 245/1553-HSC 103 12/7 12/7

The assignment to worse cell at congestion should be used with care in a 1/1 situation since the connection might suffer very low C/I if set up on the wrong cell. The benefit of being able to set up a connection in a weaker cell has to be compared to the disadvantage of increased interference. The mobile will handover back as soon as there is an idle TCH in the dominant server. This can happen within a few seconds which means that the interference may only be a short occurrence. The first choice to resolve congestion therefore is to install more TRXs. During the preparation phase, features review shall highlight which features must be used prior to project start and which features will benefit the performance. If the customer does not own some of them this could be an add-on sales opportunity. Both the Basic and Optional features parameters have to be tuned and optimized according to the performance monitoring which will start immediately after the frequency plan implementation and features introduction. The final scope of this activity is the project acceptance according to the signed KPIs agreement.

7.2 RF Optimization The target of this action will be to improve RF footprint for the expected traffic balancing from the carved band to the other existing layers (i.e. DCS, UMTS, LTE).


29 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

For that purpose Ericsson solution provides an automatic process based on xACP tool, which uses basically the same OSS input data as xAFP tool plus some extra RF information (antennas parameters). Therefore the tool will be configured in order to accomplish the aforementioned targets. Note: RF Optimization is an optional action, and the final execution must be clarified as a result of the feasibility study phase

7.2.1

General process Below the main steps to complete a regular RF optimization campaign are summarized. Each step is linked to more detailed instructions on how to use the tool for that particular action

1.

Creation of xACP project (link to “1. xACP_Project_Creation”)

a.

Inputs needed i.

Topology data

ii.

Geographical data (clutter, terrain)

iii.

Antenna Models



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

b.

Reference

Define polygons to identify area to be optimized i.

Manually created

ii.

Imported from other tools (MapInfo etc)

c.

OSS data collection i.

ii.

2G: same measurements as described for xAFP process ( see section XXX) to create IM 1.

RxLev/RxQual

2.

Frequency Plan

3.

Timing Advance

4.

Traffic per sector

3G ( optional - CM/PM files for last 2 weeks before optimization)

d.

Propagation data generation per band

e.

Traffic Maps generation i.

Per band

ii.

Combined

f. 2.

30 (89)

No.

OSS scaling process

Preparation of xACP project for optimization targets (link to “2.xACP Setting Opt Parameters”) a.

Baseline performance assessment

b.

Per sector settings i.

Ranges for each parameter to be optimized for each sector

ii.

Cost of a particular change for each sector

c.

Global settings i.

Service levels

ii.

Clutter thresholds (coverage, quality)



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

3.

Setting optimizer targets i.

ii. b.

Weights for each criteria (coverage, quality, traffic quality, dominance) Multiband optimization for spectrum re-farming Special features

i.

Force DT based on overshooting statistics

ii.

Forced DT based on drop call rate statistics

c.

5.

Reference

Executing RF optimization (link to “3.xACP Executing Optimizer”) a.

4.

31 (89)

No.

Running the optimizer

Analysis of results (link to “4.xACP Analysis of Results”) a.

Main OLAP reports

b.

Main GIS reports

c.

Analysis using Delta Reports

Update OSS measurements after implementation a.

OSS data collection

b.

Report generation: main incidences

6.

Executing Neighbor Optimization (3G) with collected OSS data (OPTIONAL – for 3G side only) (link to “5.xACP NBR Optimization”, “6.xACP SC Optimization”)

7.

Delivery of recommended changes: the following list of logical parameters can be provided in XML/MML formats for direct implementation on OSS (OPTIONAL – for 3G side only)

8.

a.

CPICH power

b.

NBR list

c.

SCs

d.

RET (Ericsson only)

Final performance report



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

7.2.2

32 (89)

No.

Reference

Targets 

Facilitate the traffic balancing between the carved GSM band (i.e 900MHz) and the other existing GSM bands (i.e., DCS 1800MHz).



General improvement of RF design in order to further squeeze the performance once spectrum refarming is complete

7.2.3

Parameters to Optimize Below is showing general recommendation, this can change upon customer’s requirements/constraints:

7.2.4



Tilts (Electrical and/or mechanical). Relative range -2,4 degrees



Azimuths. Relative range -20,20 degrees Optimization guidelines

-

Specific Spectrum Refarming recommendations:



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

-

33 (89)

No.

Reference



The “target band” (i.e. the band where the traffic is intended to be shifted) should be flavored to absorb more traffic, therefore the final tilt distribution should be according to this premise – target band should be more uptilted as a general rule. Of course this has to be verified with the current baseline tilt configuration.



Check the traffic distribution per sector after committing the optimization changes into the tool, which will give an estimation of how much traffic per sector has been shifted from one band to another.

General RF improvement recommendations What are the main RF issues of this network?

 



Coverage (% of samples in Rxlev distribution below -95). If this is bigger than 10%. This network can be coverage limited. Quality •

•

bad RxQual vs. bad RxLev (xAFP report)

Is xACP identifying those problems in the xACP project built?

 

Global % of RxLev < -95 % should match with OSS stats cell level and area level: •

•





Share of % of Drops due to bad quality or low signal level

Cells with bad RxLev should have coverage holes xACP Coverage evaluation for -95dBm should be similar to OSS values.

% of bad dominance (bad quality) is significant is area is having significant interference?

Am I having boomers ( based on TA or IM) that should be clearly corrected: 

Boomers: Cells that are clearly over propagating based on TA or IM and are dropping a lot of calls? Those cases should be downtilted by using Force DT feature For Common/Single BCCH deployments:

 

If tilt difference between layers is following an specific rule (f.e : 1800t = 900t -2)


34 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

Optimize just 900 and apply the same rule for

•

1800. 

If tilt difference is not following any rule: 1st : 900 is optimized

•

•

2nd : 1800 is optimized based on optimized 900 tilts following these rules: o

o

1800t – 900 t should be less than 4 degrees Optimized tilt difference (1800t – 900t) should be equal or bigger than baseline ( meeting a) rule and assuming that 1800t< 900t)

7.3 RRM Optimization The target of this action will be selectively, in a cell basis, moving the traffic from the layer/band which has been carved to other layers (DCS, UMTS, LTE). According to the existing exist ing layers, the optimization optimiz ation will be focused on the following (see flowchar tFigure 2. Workflow - Ericsson Spectrum Refarming solution) solution): •

•

2G Traffic Balancing: if the scenario has extra GSM band where the traffic will be moved (i.e. DCS) then 2G Traffic Balancing optimization will be carried. 3G/LTE optimization: if the scenario is only single GSM band and UMTS/LTE, then the RRM optimization task will be carried in UMTS/LTE network.

For that purpose Ericsson solution provides an automatic process based on xParameters tool, which retrieves necessary OSS input data and KPIs in order to meet the required targets. Note: 3G/LTE RRM Optimization is an optional action, and the final execution must be clarified as a result of the feasibility study phase& Acceptance agreement Below it is depicted the process


35 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

7.3.1 2G Traffic Balancing 2G Traffic Balance optimization module embedded in xParameters toolis resulting from two integrated algorithms that work together

1. Capacity: Capacity: this optimization module aims at maximizing the capacity of the installed equipment, while maintaining network quality levels. Traffic is moved within a band and across bands to avoid congestion and/or equalize the utilization of the network.


36 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

2. HO Optimization algorithm: algorithm: achieving quality improvements, even in mature 2G networks by improving the quality of handovers. Overall HO performance will be improved, in particular is expected that quality KPIs like DCR will benefit from this optimization process. Before a refarming activity, if the network is dual band, it is desirable to shift traffic from the band to be refarmed, typically the coverage layer (GSM 850/900 Mhz) towards the capacity layer (GSM 1800/1900 Mhz)

This traffic shift makes sense only if current or expected quality in the capacity layer is better than in the refarmed layer, otherwise it may produce degradation in the overall quality of the network.This can be analyzed in overall terms during the Feasibility Study phase. However, xParameters optimization cell-based capabilities is able to balance traffic among cells of the different layers even if the capacity layer is more loaded than the refarmed layer (i.e., overall traffic share per band will not change much after optimization campaign, but locally there should be significant changes that will result in better performance). 7.3.1.1

Standard approach The standard usage of xParameters allows, as described in previous section, perform a capacity optimization and HO Optimization. When talking about capacity optimization, it refers to the ability of xParameters to push traffic from overloaded cells to less loaded cells while keeping the quality under control.


37 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

The mechanisms of traffic shifting are related to RRM parameters, more specifically to Locating/Handover control algorithms: -

-

Intra layer traffic balanced through Power Budget or Best Cell HO margins / offsets Inter layer traffic balanced through layer signal level thresholds

xParameters allows to define Min/Max/Default values for all optimized parameters. •

•

Min/Max values will define the boundaries of each parameter during the iterative optimization process. Default value will define the stable value towards which the parameter should tend in absence or congestion of the source and target cells under analysis.

For example: If cell A is very loaded and cell B is not loaded, the Handover Margin A->B will be reduced, so as a lower value of signal level lev el is required to handover towards B. B will absorb part of A’s traffic. The iterative process of xParameters algorithm will try to decrease this HO margin A->B as long as the quality is maintained and the delta of load A-B is still positive and the Max value of the parameter is not reached This could mean, for instance in Ericsson infrastructure, to increase offset from 0 to 3 dB. Default value for offset would be 0, meaning that if the imbalance of load between A y B disappears (for instance a TRX is added to A, or overload in A was caused by temporal causes) then xParameters will tend to set offset back to 0, the default (this mechanism is called “Step Back”) 7.3.1.2

Traffic Balancing between GSM bands The following description covers how to configure general settings of xParameters tool in order to move traffic from 900 to 1800 band; there are typically two use cases behind this scenario:



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

•

•

38 (89)

No.

Reference

Spectrum Refarming process of 900MHz band:the tool will be configured to offload traffic from 900 Mhz towards 1800 Mhz. Shifting traffic to a cleaner band (lower FreqLoad in 1800) resulting in an overall improvement of quality indicators.

In order to achieve this behavior, the followingactionscan be applied independently or combined (see section XXX) 1. Enabling the feature “push traffic to high band” (pushTrafficToHighBand = true, setting in internal algorithm settings) 2. Setting a lower Default/Minimum level access threshold to the High band in the algorithm GUI settings. 3. Setting different congestion/utilization low/high percentiles to 900Mhz and 1800Mhz Sectors a.

Lower Low & High Percentiles for Congestion/Utilization in 900 Mhz

b.

Higher Low & High Percentiles for Congestion/Utilization in 1800 Mhz

4. In general, if the baseline 900Mhz layer is more loaded in than 1800Mhz, the algorithm will pushby default a portion of traffic towards 1800Mhz.

Notes: -

The above strategies effectiveness will depend a lot on the baseline configuration and capacity of the 1800 Mhz to cope with more traffic.

-

All of the strategies are focused on the “Capacity” part of the optimization algorithm.

-

A close look up of the performance is required when we perform xParameters optimizations, also good analysis skills and comprehension of technology and RRM algorithms are required to properly address any problem and changes in the strategies that might be required to properly drive the activity in the right direction.

-

The HO optimization feature will always work in the background unless explicitly disabled (disabling it internally, covWeight=0 and/or disabling changes in the involved parameters step/step back=0)



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

7.3.1.3

39 (89)

No.

Reference

Summary of Inter-Layer and Intra-Layer Handover parameters optimized by xParameters Capacity algorithm The optimization of the following parameters is Active for all vendors by default.

POWER BUDGET HOs

LAYER MANAGEMENT

SUBCELL MANAGEMENT

Control Power Budget Hos (dB)

(rxlev / dBm)

(rxlev / dBm)

ALCATEL

HO_MARGIN (adce)

L_RXLEV_CPT_HO (cell)

RXLEV_DL_ZONE (cell) ZONE_HO_HYST_DL (cell)

ERICSSON

OFFSET (adce) HIHYST (adce)

LAYERTHR (cell)

LOL/DTCB (cell)

HUAWEI

PBGT_HO_THRSH (adce)

Inter-layer_HO_Threshold (cell) Inter-layer_HO_Hysteresis (cell)

RX-LEV_Threshold (cell) RXLEV_Hysteresis (cell)

NOKIA

PMRG (adce)

AUCL (adce)

LAR (out->in) (cell) LER (in->out)) (cell)

SIEMENS

HOM (adce)

DOESN'T EXIST, SIMULATED USING RXLEVMIN (adce)

HORXLVDLI (out->in) (cell) HORXLVDLO (in->out) (cell)

ZTE

HO_MARGIN_PBGT (adce)

MaxLossThs (adce)

PathLossMin (adce)



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

7.3.1.4

40 (89)

No.

Reference

Summary of Inter-Layer and Intra-Layer Traffic Management parameters optimized by xParameters Handover Optimization algorithm

The optimization of these parameters is Active in all vendors by default. Controled by algorithm internal setting (covWeight=1)

MINIMUM HO ACCESS LEVEL (rxlev / dBm) RXLEV_MIN_n (cell)

ALCATEL

MSRXMIN / BSRXMIN (cell)

ERICSSON

Min_DL_Level_on_Candidate_Cell (cell) Min_Access_Level_Offset (adce) HUAWEI NOKIA SIEMENS ZTE

SL (adce) RXLEVMIN (adce) RxLevMin (adce)

LEVEL HOs

ALCATEL ERICSSON

HUAWEI NOKIA SIEMENS ZTE

Trigger Threshold (rxlev / dBm)

Margin (dB)

Filters (SACCH)

L_RXLEV_DL_HL_RXLEV_UL_H (cell)

HO_MARGIN_LEV (adce)

A_LEV_HO (cell)

HYSTSEP (cell)

OFFSET (adce) LOHIST (adce)

SSLENSD SSEVALSD (cell)

Edge_HO_DL_RX_LEV_Thrsh Edge_HO_UL_RX_LEV_Thrsh (cell)

Inter-cell_HO_hysteresis (adce)

Filter_Length_for_TCH_Level Filter_Length_for_Ncell_RX_LEV (cell)

LDR/LUR (cell)

LMRG (adce)

NOT OPTIMIZED

HOLTHLVDL/UL (cell)

LEVHOM (adce)

HOAVLEV.aLevHo (cell)

HoDlLevThs HoUlLevThs (cell)

HoMarginRxLev (adce)

HoDlLevWindow (cell)


41 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

The optimization of the following parameters is Inactive in all vendors by default. Controled by algorithm internal setting (intWeight=0)

QUALITY HOs

ALCATEL ERICSSON

Trigger Threshold (rxlev /dBm)

Margin (dB)

Filters (SACCH)

NOT OPTIMIZED

HO_MARGIN_QUAL (adce)

NOT OPTIMIZED

QLIMUL/QLIMDL (cell)

BQOFFSET (adce)

QLENSD&QEVALSD (cell)

NOT OPTIMIZED

BQ_HO_Margin (adce)

NOT OPTIMIZED

NOT OPTIMIZED

QMRG (adce)

NOT OPTIMIZED

HOLTHLQUDL/UL (cell)

QUALLEVHOM (adce)

HOAVLEV.a_Qual_Ho (cell)

NOT OPTIMIZED

HoMarginRxQual (adce)

NOT OPTIMIZED


7.3.2 3G/LTE Optimization process

This document is focused on the actions to improve KPIs of GSM layer when a refarming process takes place. Therefore, even in some cases it can be recommended to perform optimization campaign of other technologies involved in the refarming process, cell-level Parameter optimization of other technologies over 2G layer is out of the scope. Check for other references in product catalog The figure below illustrates the process of 3G technology RRM optimization using Ericsson xParameters tool. Technical guidelines regarding 3G Generic Optimization will be published soon in product catalog.



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

8

42 (89)

No.

Reference

Refarming Realization

8.1 Frequency planning strategy definition The main idea behind the Frequency Refarming Service is to provide a more efficient usage of the frequency spectrum and most of this efficiency is achieved through a more optimized frequency plan. The commonly used frequency planning methods are MRP (Multiple Reuse Pattern), FLP (Frequency Load Planning) and AdHoc. MRP utilizes base band frequency hopping while FLP utilizes synthesized frequency hopping. MRP doesn’t work well when coping with the contradictionbetween high capacity and scarce frequency spectrum. Therefore, at the end two main alternatives for the frequency plan methodology are envisaged:


43 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

1. Automatic High Frequency Load Planning (AFLP), mainly 1/1 approach. The major benefits of FLP can be sorted into two areas: •

•

Minimized frequency planning and flexible network expansion support, high frequency utilization. Practically all operators will benefit from the reduced effort spent on TCH frequency planning. The capacity gain by using FLP is specifically useful for operators with a narrow frequency band, and sometimes maybe the only reasonable alternative. Limitation:as Limitation:as there is no gain to hop over less than four frequencies, the minimum required frequency spectrum for FLP is 3MHz(= 11 BCCH frequencies + 4 TCH frequencies). Frequency Load: when planning using FLP methodology, it is very important to understand the frequency load of the scenario to plan. For a detailed description of the concept please refer to the APPENDIX C.

2. AdHoc plan by using xAFP tool, adjusting the final plan to the particular constraints (final bandwidth available, TCH-BCCH allocation limitations, etc). The main benefits of this approach are: •

•

Networks with not homogeneous traffic distribution may not reach optimum performance by adopting 1/1 strategy Planning shared MALs between co-site sectors that way taking advantage of longer MALs while avoiding frequency reuse between mutually highly interfering sites.

The decision of the frequency planning method to use will be made once conclusions from the feasibility study report are available.

8.2

Measurement data input consistency check

8.2.1

BAR Files/ICDM If a BAR File or ICDM is used as input information for the planning tool it is important to note that, due to some particular situations in the network, some interfering cell relations in these files might be omitted. If this situation happens, the input information for the tool will not be reliable and will compromise the frequency plan quality.


44 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

When cells close to each other have a co-channel, BCCH or TCH, the mobiles in those cells will not be able to report the perceived co-channel signal strength from the other cell. Let’s say cell A coverage is overlapping cell B coverage and both are using the same channel X. The mobiles in cell A will be affected by interference in channel X from cell B but they will not report it. This missing information will be understood by the planning tool as if cell A and cell B are not interfering with each other, which is not the case, and in this situation the tool might assign the same channel for cell A and cell B. Hence, in order to have a reliable BAR File or ICDM, it is necessary to perform a preliminary verification in the network before starting the recordings of ICDM or extracting the BAR files. All close co-channels should be retuned in order to avoid that the BAR File/ICDM are not reliable. Besides the close co-channel issue, there is also another situation that is better to be avoided if possible — the the repeated BCCH/BSIC pair, which may also turn into erroneous information in the tool. The mobiles use BCCH/BSIC to report measurements and this is how the information is recorded. The planning tools use an algorithm that identifies BCCH/BSIC pairs in the recordings and assigns then a cell name. If the algorithm has more than one cell name related to a given BCCH/BSIC pair, it will try to figure which cell name each BCCH/BSIC pair is referring to in the recordings through estimations like azimuth, path loss and signal strength. If by any chance the information used for these estimations is not precise enough or even incorrect, there is a chance that the algorithm will inform an erroneous interfering cell name for the tool which will also compromise the frequency plan quality. To avoid such a situation, it is recommended to check BCCH/BSIC pairs after extraction of BAR/ICDM files and correct the information, if needed.

8.3

Frequency plan methodology This section presents the guidelines to perform a frequency plan using the two aforementioned main methodologies: FLP and AdHoc planning

8.3.1

FLP planning When reducing GSM spectrum in most of the cases there will be left only a narrow band for GSM and due to this fact if by any reason FLP cannot be used the limitations due to the reduced spectrum imposed to the network capacity and performance will be more evident.


45 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

Usage of the FLP planning method can result in major benefits for an operator. The benefits can be sorted into two major areas: Minimized frequency planning with flexible network expansion support and high frequency utilization. The capacity gain originated by FLP is specifically useful for operators with a narrow frequency band and sometimes maybe the only reasonable alternative. Almost all operators will benefit from the reduced effort spent on TCH frequency planning. FLP 1/1 reuse takes Frequency planning one step further due to its capacity advantages in narrow band networks. For instance, FLP 1/1 yields ~20% more capacity than MRP or FLP 1/3 in a 7.5 MHz network. And the smaller the frequency band gets the greater the gain with FLP 1/1 will be. The main reason for this gain is the larger number of frequencies per cell that is given by FLP 1/1. When the frequency band is greater than 7.5 MHz, the capacity advantage of FLP 1/1 becomes smaller and FLP 1/3 will be a good choice. From a GSM point of view, the implementation of a frequency plan in FLP modalitycannot start without the fulfillment of some constrains: 1. FLP is based on synthesizer frequency hopping and cannot be used with filter combiners. If filter combiners are used in the networkand the operator do not want to swap them out, FLP cannot be implemented. It will be necessary to use BB hopping and a different planning strategy needs to be used. 2. If FLP will be implemented, features like Frequency Hopping, Flexible MAIO Management, Dynamic BTS/MS Power Control and DTX Uplink/ Downlink are prerequisites. 3. As there is no gain to hop over less than 4 frequencies, the minimum required frequency spectrum for FLP is 3 MHz(= 11 BCCH frequencies 3 + 4 TCH frequencies). If this prerequisite is not fulfilled Spectrum Refarming will not be possible.

3

It is suggested to plan the BCCH frequencies, to values near to a 12 reuse, even if there are plenty of frequencies.The reason is that if the traffic is expected to increase a lot in the near future, it might be a good idea to plan the BCCH frequency plan a little tighter in favor of hopping TCH. The network plan can then cope with additional traffic, and TRXs can be added without any need to change the frequency plan.



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

8.3.1.1

46 (89)

No.

Reference

Considerations on Frequency Load & FLP The frequency load is defined by the percentage of time on which a specific hopping frequency is on air. During a traffic increase and/or spectrum reduction the frequency load will increase. It has been verified that SQI and TCH Dropped Call Rate due to bad quality are directly related to frequency load, i.e. if the frequency load increases it will also degrade SQI and increase TCH Dropped Call Rate due to bad quality. One of the conditions that will determine whether the Refarming project can be doneor not is how much of the original spectrum needs to be reduced. The limiting factor is how much Frequency Load is acceptable. An example of the calculation of how much spectrum can be reduced is given below: Suppose an operator has A MHz GSM band. If the operator wants to reduce B MHz in the same bandthen the available GSM spectrum is (A-B) MHz. Let’s assume that in order to keep good quality and network performance the Frequency Load should not be greater than 12%. The maximum Frequency Load in a FLP 1/1 network can be very different from network to network due to different radio network topology and cell planning. Important factors that can determine the achievable load to be carried with reasonable quality are: •

Low / high sites (confined coverage)

•

Antenna direction and tilt

•

How the BCCH BCCH/TCH plan)

frequencies

•

Microcells between FLP hopping cells in the same area

•

Dedicated or shared TCH band (with e.g. microcells)

•

•

•

are

chosen

(staggered/blocked

The relation between outdoor and indoor traffic (especially traffic in high rise buildings can be critical) Available spectrum (as the hopping gain depends on it) Quality standards for the operator (accessibility, retainability, service quality)

•

Radio network features used

•

Speech codec used


47 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

Thus, the acceptable Frequency Load is very operator/network dependant. The frequency load measure is defined as:

FRQLoad

=

Erlang cell 8 * (# FRQcell )

In order to calculate the load we select a cluster of 10~15 sites in the customer’s network that have the highest traffic to calculate the value of Erlang cell .For instance, if Erlang cell = 16.63 and FRQLoad = = 12%, then according to Frequency Load definition the calculation is shown below:

FRQLoad FRQLoad =

Erlang Cell

8 * (# FRQ

≤

)

0.12 ⇒ # FRQ

Cell

≥ Cell

Erlang Cell

8 * (0 (0.12)

=

16.63 8 * (0 (0.12)

= 1 7 .3 2

That is to say, the minimum number of hopping TCH frequency is 18.Assuming there is 12 BCCH frequenciesreuse. Then, 12 BCCH + 18 TCH =30 frequencies, 30*0.2 = 6MHz. If (A-B) >= 6MHz, the frequency Refarming can be done. However, here we have several assumptions: 1.

12 reuse BCCH frequencies

2. Frequency Load is not greater greater than 12% 3.

No HR traffic

4.

No Multi-band Cell

Once one or some of these assumptions are changed, the result will be changed accordingly. If (A-B) < 6 MHz, it doesn’t mean the frequency Refarming can’t be made. An additional effort is needed. This is done by means of decreasing the value of

Erlang cell by traffic steering to other existent cells or deployment of new cells. Choosingamethod to decrease the value of Erlang cell will depend on the real network traffic.


48 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

[1] Furthermore, some features can help to improve the network interference situation. Besides features, several important parameters also need to be set to a suitable value, for further information refer to the theFLP Guideline which contains several features and parameter values indications.

8.3.2

Ad Hoc planning

8.3.2.1

Introduction By analyzing different frequency planning strategies with xAFP tool, it is possible to choose the best methodology of carving, while ensuring least cost and minimizing customer impact. In this module it is detailed how to create an xAFP project, focusing on its usage for refarming purposes. First, a description of the steps to follow when building a basic xAFP project is given. This includes the importation of configuration data, as well as collected statistics data on which a frequency plan is based. Then, an explanation on how to run frequency plans according to different strategies is given. This part will help to understand the different utilities and settings that allow the analysis of different carving options, traffic and TRX dimensioning impact, as well as quality degradation due to the spectrum carving.


49 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

Inputs

Physical Data

Network Data

Measurements Settings

BTS hardware constraints Network Sector extra settings

Control Room

Entrence

XAFP

Optimization cost settings

Exit

Frequency Allocation strategy

Frequency plan

Neighbor audit

Rotation management

Outputs

8.3.2.2

Overview

IM correlation


50 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

The first necessary information to build a project consists of relevant physical data and network configuration for each of the sectors in the region of interest, plus the sectors surrounding it. Once this information is collected, imported and has proved to be consistent (steps 1 and 2 in the figure), it is possible to proceed with network statistics collection. Traffic, handovers, mobile measurement reports (MMR) and RxQual/RxLev data are the main inputs on which xAFP functions are based. In order to collect these, a period of measurement collection is needed, which can from two to four weeks. This measurement collection has to be carefully prepared beforehand in the case of MMR and rxQual/rxLev statistics (step 4 in the figure). Traffic and handover statistics of the measurement campaign period can be extracted from the operator databases. Based on the MMR, an interference matrix (IM) is generated, containing interference relationships between the sector pairs (step 5 in the figure). xAFP makes intensive use of IM information in order to plan the frequencies, so it has to be made sure that this IM is correctly built and that it reasonably represents the environment interferences. In order to validate and enhance the IM, traffic and rxQual/rxLev statistics can be used (steps 6 and 7 in the figure). Finally, the optimization settings can be prepared, spectrum ranges and allocation strategies indicated, and the optimizer can be launched (step 8 in the figure). Different scenarios, varying settings and frequency strategies, can be simulated and compared in order to analyze which methodology is the best for the specific requirements of the case.



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

8.4

51 (89)

No.

Reference

Ad-Hoc frequency plan creation with xAFP This section covers the whole process to create an Ad-Hoc FP using xAFP tool, from project creation, collection of OSS measurements and frequency plan generation & simulation.

8.4.1

General Process The steps to generate an Ad-Hoc plan using xAFP are described below. Each step is linked to more detailed instructions 2. Creation of xAFP project (link to “1. xAFP_Installation and General Tips”, “2. xAFP_ProjectSetUp”) o

o

Required Inputs 

Topology (physical) data



Network dump



Frequency constraints

Outputs: 

xAFP project



Some configurations will be created in the project: •

•

Baseline Configuration: all on-air sites with current FP – this configuration will be used for the consistency checks, BSIC plan, MMR import and initial IM creation Optimized Configuration: copy of baseline configuration + final IM, final optimization settings, additional TRUs, new sites added during the preparation phase

3. Consistency check of imported data (link to “2. xAFP_ProjectSetUp”) o

Database cross validation (on-air site list)

4. IM creation (link to “3. xAFP_IM”) o

Required inputs:


52 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E



10-15 working days of MMRrecorded during the busy period(s).



RxQual/RxLev statistics (collected during MMR campaign)



Traffic per cell (average of RxQual measured days/period)



Power Control Statistics (optional, although very helpful if these statistics are available).



Handover Attempts (corresponding to the measured days campaign). BCCH-BSIC mapping on a daily basis



o

o

o

o

Reference

Outputs: 

Interference Matrix



Simulation Traffic Profile

Recommended actions before measurements campaign: 

Neighbor Audit (Deletion). It is required to get enough gaps in BA lists in order to perform the BCCH scanning.



BSIC Audit Plan. Retune BSICs with “close” reuse (<15 miles).

Scan List generator: add needed BCCHs to BA list to complete measurements according to vendor’s specifics Import mobile measurement reports on a daily basis with the proper BCCH-BSIC map.

o

Import traffic / Rx Statistics /Power Control statistics /HO

o

Generate/Update the IM using xAFP

5. IM validation (link to “3. xAFP_IM”) o

Required inputs (to set the QoS simulation) 

Number of traffic voice slots in BCCH per sector.


53 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E



o

Traffic allocation method (TCH first, BCCH first, random).



DTX/Power control Enabled/Disabled



Set RxLev threshold corresponding to RxQual-5



Half Rate Users Enabled/Disabled

Output 

o

Reference

IM validation. The predicted RxQual considering all the inputs must be correlated with the measured RxQual.

Important: If the IM does not pass the quality requirements, the next stages of the process cannot be achieved. 

Actions: Increase the MMR/data collection. Check the consistency of all the data.

6. Import HO statistics o

Collect daily stats for last 2-4 weeksfor ISBH

o

Audit imported data – is up to date with NBR list

o

Generate a new HO profile for optimization (up to date)

7. Import Traffic statistics o

Collect daily stats for last 2 weeks for ISBH

o

Generate a new traffic profile for optimization

8. Neighbor List Optimization (optional at this phase) 9. Frequency Plan Optimization (link to “4. xAFP_Optimization”) o

o

Adjust optimization sector settings 

Permitted spectrum



Minimum/Maximum channel distance



MAL length constraints (if RF hopping is enabled)

Import network/sector constraints (forbidden parameters)


54 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

o

o

o

Reference

Import Additional Interference Constraints: 

AICs



IM Overrides

Network Forecast 

Include new site integrations in the optimization



TRXs forecast (remove/add TRXs if required)

Run the optimizer:BCCH, TCH, BSIC, HSN, MAIO, MAL, Synch. Clustering

10. Fine Tune using UQB module (link to “6. xAFP_FineTune_UQB”) o

DL PC adjustments

o

MAL length optimization

o

FER based MAIO list length optimization

11. Review plan (link to “5. xAFP_Reports_FinalPlanReview”) o

o

Identify potential issues from xAFP reports 

Run consistency checks (per sector/relationship)



Complete Channel Distribution



Neighbor Reuse



Channel Reuse



Co-BCCH-BSIC conflicts report

The attached task flow could be executed in order to perform all the standard checks over optimized configuration.

12. Export final plan for implementation



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

8.4.2

55 (89)

No.

Reference

Targets General Targets: Provide a new frequency plan that is able to meet the following requirements: 

Minimize the potential degradation in the new scenario after reducing the available spectrum. Adapt all the network restrictions to the final available spectrum



Traffic Offload and TRXs dimensioning study: 

Offload traffic from GSM Coverage-Layer to GSM CapacityLayer without increasing the blocking and reducing the TCH availability.



Look for the optimal TRXs forecast for both bands, in order to properly manage the traffic movement between bands, avoiding traffic lost as much as possible.



Reduce the number of TRXs in GSM Coverage-Layer in order to reduce the carving of frequencies impact.

Frequency Strategy Evaluation: 

Look for the optimal BCCH-TCH distribution in order to reduce the impact of the degradation as much as possible.

Top Worst Analysis: 

8.4.3

Identify those sectors with higher limitations (due to RF effects and due to frequency reuses), and run aggressive actions on this set of sectors in order to relax the optimization conditions and get a better solution.

Optimization guidelines Optimization Recommendations after Complete Feasibility



Study: •

•

Perform a general network audit. It is advisable to test which extra-features could help the xAFP optimizer work, such as the Power Control activation, DTX mode and Half Rate/AMR traffic coding activation. Optimization Settings to pay attention in this scenario


56 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

o

o

o

•



Reference

Co&Adj Interference management. UQB module which allows reaching good general QoS stats thanks to co-channel reduction in balance with a certain and controlled adjacency reuse increase. AdHoc costs adapted to each simulated scenario. Depending on the frequency strategy (relaxed conditions for scenarios which aggressive spectrum carving). 

Reduce / or set equal zero, those costs related to neighbor adjacencies. That type of reuses will be controlled by IM costs.



Use HOs weights to model the reuses between neighbors.



Control the 2nd order neighbor costs. Setting too much high costs, could imply sub-optimal solutions. The recommendation could be set up a value below 5.

Review Micro/Macro cells reuses and interference influence between layers. Due to spectrum reduction it could be that microcells have to share the available frequencies with macros. It is advisable to protect the microcells frequency reuse using different optimization settings or putting attention on the IM of these type of cells.

Review the performance and the suffered interference due to frequency carving for VIP sectors. An extra protection can be given in xAFP tool to these sectors by increasing the sector cost (by default it is equal 1).

Optimization Recommendations when there is no possibility of running the Feasibility Study (time/resources limitation): •

•

Run a neighbor audit. Having a proper neighbor list will facilitate the optimizer to find out an optimal frequency plan with high restrictions due to frequencies carve off. If the hopping type is Base Band Hopping: TRXs Deletion in order to gain some room for optimization


57 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

o

o

•

•

•

•

Reference

The frequency carving could be more aggressive in the coverage layer, and part of the traffic could be managed by the capacity layer. Delete TRXs if the traffic may be carried by 3G/capacity layers. Study of the top-worst performing sectors using xAFP tables/data. The optimization efforts and the carrier reduction should be focused on these sectors. 

Boomer analysis. Analysis of those sectors which are causing interference to a high number of victims and a high percentage of traffic.



AdHoc cost top contributors. Analysis of those sectors with higher interference/adhoc costs.



Most interfered frequencies. Analysis of the frequency reuse and the suffered interference per channel/TRX.

If the hopping type is Synthesized Hopping: Perform MAL length reduction in order to reduce the interference. Use UQB module in order to look for the optimal MAL length (running MAL length optimizer). Perform a general network audit. It is advisable to test which extra-features could help the xAFP optimizer work, such as the Power Control activation, DTX mode and Half Rate/AMR traffic coding activation. Run several scenarios with different amount of TRXs and frequency strategies, in order to look for the solution with lower performance degradation. Optimization Settings to pay attention in this scenario o

o

Co&Adj Interference management. UQB module which allows reaching good general QoS stats thanks to co-channel reduction in balance with a certain and controlled adjacency reuse increase. AdHoc costs adapted to each simulated scenario. Depending on the frequency strategy (relaxed conditions for scenarios which aggressive spectrum carving).


58 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

o

o

Reference



Reduce / or set equal zero, those costs related to neighbor adjacencies. That type of reuses will be controlled by IM costs.



Use HOs weights to model the reuses between neighbors.



Control the 2nd order neighbor costs. Setting too much high costs, could imply sub-optimal solutions. The recommendation could be set up a value below 5.

Review Micro/Macro cells reuses and interference influence between layers. Due to spectrum reduction it could be that microcells have to share the available frequencies with macros. It is advisable to protect the microcells frequency reuse using different optimization settings or putting attention on the IM of these type of cells. Review the performance and the suffered interference due to frequency carving for VIP sectors. An extra protection can be given in xAFP tool to these sectors by increasing the sector cost (by default it is equal 1).

General frequency plan recommendations

 •

•

•

•

•

Keep the consistency of the project/data along all the activity duration. Check the reliability of the collected data. Reactivate the measurement campaign if needed. Secure the reliability of the Interference Matrix. It is one of the most important keys of the frequency planning using the xAFP tool. Maintain the project synchronized with the network conditions. Different clusters could be optimized/implemented in different time slots, and the contour conditions of ongoing projects should be updated in consequence. Keep the congruence between 2G and 3G/4G 3G/4G networks.



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

•

•

8.5

59 (89)

No.

Reference

Always run different scenarios in order to reach the best solution with regard to the network which has to be optimized. Consider future network changes: such as TRXs forecast and new site integrations.

Implementation The implementation of the new plan should be carefully scheduled in order to reduce the disturbance in the network as much as possible. Maintenance window with low traffic is the preferred choice. Some operators even prefer that the changes take place on weekends. As a tool for that CAN is proposed, which whic h is an optional OSS-RC function f unction module or Winfiol to implement the frequency plan. Nevertheless, for both above mentioned cases, a CNAi script should be created and checked (for CNA implementation) and a MML script, for each BSC should be created and checked (in case of Winfiol implementation). Experiences have shown that implementation via CNA uses less human resources to implement in the case of a large network, since it can manage to execute the processes in parallel. However, if the MML script option is considered, the CP Side-switch method is strongly advised. It has to be assured that all the lines related to neighbors updates and external cells definition are included on the MML scripts. The Side-switch method consists in putting the running BSC side operating separately from the standby side.The MML script should be loadedin the standby area of the BSC while keeping the original plan running on the other. After the load is finished, invert the side that is running with the one that is on standby. With this strategy, you have the original plan as a fallback in the new standby side, in case it is necessary, while the plan loading time is the time required to switch the standby side of the BSC by the one that is running.

8.6

Neighbors Update Neighbor list optimization is one of the key tasks for operators in order to guarantee proper network performance This item also includes the update and sanity check of BA list from all impacted cells.


60 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

The xAFP tool used for AdHoc frequency plan tool performs this task in an automated way and through the utilization of different criteria that remove or add neighbors from the baseline neighbor list imported from OSS. Neighbor lists should be updated in the following milestones during the service delivery: After new frequency plan implementation

•

•

After new RF design implementation (Pre Refarming actions, see section 7.2) section 7.2)

More detailed instructions on how to perform Neighbor Audits using xAFP tool will be published shortly in service product catalog, and an overview of the process will be included in further version of this document.

9

Post Refarming actions The first action to be taken after the Refarming will consist of a performance analysis in order to evaluate the situation just after the carving & new frequency plan implementation have been completed.

9.1

Performance Evaluation Right after the implementation of the Refarming plan, the performance monitoringshallstart immediately. In case a serious problem which jeopardizes the network performance can’t be resolved before the traffic busy hour of next day, a Fall-Back needs to take place. Then the issues raised must be analyzed and the spectrum Refarming plan reviewed in order to correct the detected problems and then rescheduled to be implemented. Otherwise, the recommend process to monitor performance would be: •

•

Analysis of main KPIs as defined in the acceptance agreement Analysis of second order KPIs related to the main KPIs in order to understand further impact on the performance caused by the Refarming & post ref-arming actions

If further information is required due performance degradation, troubleshooting actions are required. For that purpose, a top-down process should be followed as described next:


61 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

•

•

Reference

Troubleshooting based on KPIs/counters: for that purpose Ericsson provides automatic reporting& diagnosis capabilities by using xNethealing tool Troubleshooting based on Trace Data: for that purpose Ericsson provides automatic reporting capabilities from geolocated sector level raster reports down to call trace & event-based reports by using xGeomanager tool

As a result of t his performance analysis, Ericsson will make a proposal on certain actions required after the Refarming in order to meet the agreed Acceptance criteria.

9.2

RRM Optimization The following actions are presented in order to mitigate the effects that can cause performance degradation or eventually could help to improve performance up to the established levels of acceptance criteria. There are two main actions envisaged: second round of traffic balancing in 2G layer after the spectrum refarming and optimization campaign in other existing/new technologies (3G/LTE).

9.2.1

2G Optimization

9.2.1.1

General Process In this section a brief description of the xParameters 2G Optimization process is depicted in the form of flow diagram and list of points to complete successfully xParameters iteration. This would be a daily or weekly process, depending on the operational schema chosen (typically daily iterations for fast optimization campaign, weekly iterations for “maintenance” job on the operator once the fast optimization campaign has taken place already). For more details on each of the steps and meaning of some of the blocks, the reader is invited to check xParameters and ODG training material and general guidelines. A general overview of the t he process is depicted in next diagram. 

There are 2 sequential processes which are the ODG and xParameters setups


62 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E



Reference

Then the cyclical xParameters process which involves, ODG execution to collect OSS data (Configuration and performance), xParameters algorithm execution, checks, implementation of changes etc.

Figure 3: xParameters Process

Next diagram shows a higher level of detail on the process and flow diagram and checks the xParameters user must perform during an optimization campaign.


No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Figure 4: xParameters Detailed Process

Reference

63 (89)


64 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

1.Optimi Data Gateway (ODG) set-up 1.

Pre-process tasks (preprocess.py script) – Optional

2.

Collect / Adapt Data

3.

Process input files

4.

Dump Output csv files and MySQL load to operdb

5.

Post-process tasks (postprocess.py script) - Optional

2.Consistency check of ODG results (database errors correction) 3. Run xParameters algorithm (i.e. daily) 1.

Create (only first iteration)/Open xParameters Project

2.

Select sectors to be Optimized

3.

Assign Settings Templates

4.

Define Measurement Period

5.

Define and Run Optimization Task

4. Check algorithm outputs and network performance - Corrective actions if needed •

Execution Log

•

Changes Report

•

OLAP - Execution Report

•

OLAP/GIS - Performance Reports/Layers

•

xNetHealing or other PM monitoring tools

5. Send cell-level parameters changes scripts to customer for implementation 6. Check implementation - Corrective actions if needed •

Check Implementation Logs

•

View Discrepancies with reports


65 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

Optionally, some traffic steering actions by activation of RRM features can be performed. In the case of Ericsson these are the recommended features: •

Cell Load Sharing

•

Cell Re-selection and Handover GSM-UMTS-LTE

•

Assignment to Other Cell

•

Hierarchical Cell Structures

Operational considerations when running xParameters

Figure 5: Operational Schema



The optimization tasks can be configured and scheduled with very high flexibility. Optimization runs incrementally, with a minimum period between optimizations of 1 day. The tool takes the performance and configuration inputs that are going to be used for the optimization (previous day performance, previous week, previous N working days, etc) and the optimization algorithms produce a proposal for parameter changes.



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

9.2.1.2

66 (89)

No.

Reference



The user can receive automatically a report of those changes and implementation files (MML, XML, CSV…) via email or directly by using the tool reporting and graphical representation features.



The implementation of changes can be also automated in the network without intervention of the optimization engineer (available only in some vendors and under specific connectivity/permissions environment)



Typically a fast optimization campaign will consist of daily iterations of the algorithm, while in a “maintenance mode” we can schedule weekly optimizations. Targets

Once the Refarming has taken place, as mentioned in previous section several actions can be applied to reduce the frequency load of the re-farmed layer and obtain a better spectrum usage. It was mentioned that a more aggressive strategy of MultiBand cell and HCS features could be applied so as the re-farmed band carries less traffic. The utilization of xParameters in a post-refarming environment should be focused in the traffic shift towards the cleanest layer, as in the pre-refarming stage. However at this stage the operator might be switching off TRXs on the refarmed layer, for example 900MHz and expanding the capacity of the capacity with new TRXs on 1800Mhz. Other activities like site integrations, new neighbors plan, and the inherent traffic pattern and quality changes after the refarming would make advisable a new xParameters campaign, as the pre-refarming parameterization achieved by the tool would be probably suboptimal now. In this environment xParameters continuous optimization objective will be that the new capacity is properly utilized and that as much traffic as possible is pushed towards the capacity band with no negative impact in the performance. xParameters 2G can help to optimize cell and handover parameters related to multiband and HCS features so as the strategy applied is enforced with intelligent optimization algorithms. It is very important to mention that xParameters 2G does not “deploy an strategy” for traffic management, it optimizes parameters on top of current network active configuration and features.



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

9.2.1.3

67 (89)

No.

Reference

Parameters to Optimize Here in next tables, same as those in sections 7.3.3 and 7.3.4, list of all parameters involved in xParameters 2G Traffic Balance algorithm, related to capacity and HO optimization features. The most relevant in terms of traffic distribution (capacity algorithm) are included in the first table.

POWER BUDGET HOs

LAYER MANAGEMENT

SUBCELL MANAGEMENT

Control Power Budget Hos (dB)

(rxlev / dBm)

(rxlev / dBm)

ALCATEL

HO_MARGIN (adce)

L_RXLEV_CPT_HO (cell)

RXLEV_DL_ZONE (cell) ZONE_HO_HYST_DL (cell)

ERICSSON

OFFSET (adce) HIHYST (adce)

LAYERTHR (cell)

LOL/DTCB (cell)

HUAWEI

PBGT_HO_THRSH (adce)

Inter-layer_HO_Threshold (cell) Inter-layer_HO_Hysteresis (cell)

RX-LEV_Threshold (cell) RXLEV_Hysteresis (cell)

NOKIA

PMRG (adce)

AUCL (adce)

LAR (out->in) (cell) LER (in->out)) (cell)

SIEMENS

HOM (adce)

DOESN'T EXIST, SIMULATED USING RXLEVMIN (adce)

HORXLVDLI (out->in) (cell) HORXLVDLO (in->out) (cell)

ZTE

HO_MARGIN_PBGT (adce)

MaxLossThs (adce)

PathLossMin (adce)

Figure 6:Inter-Layer and Intra-Layer Handover parameters optimized by xParameters Capacity algorithm MINIMUM HO ACCESS LEVEL (rxlev / dBm)

ALCATEL ERICSSON

RXLEV_MIN_n (cell) MSRXMIN / BSRXMIN (cell)

Min_DL_Level_on_Candidate_Cell (cell) Min_Access_Level_Offset (adce) HUAWEI NOKIA SIEMENS ZTE

SL (adce) RXLEVMIN (adce) RxLevMin (adce)

LEVEL HOs Trigger Threshold (rxlev / dBm)

Margin (dB)

Filters (SACCH)


68 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

ALCATEL ERICSSON


Reference

L_RXLEV_DL_HL_RXLEV_UL_H (cell)

HO_MARGIN_LEV (adce)

A_LEV_HO (cell)

HYSTSEP (cell)

OFFSET (adce) LOHIST (adce)

SSLENSD SSEVALSD (cell)

Edge_HO_DL_RX_LEV_Thrsh Edge_HO_UL_RX_LEV_Thrsh (cell)

Inter-cell_HO_hysteresis (adce)

Filter_Length_for_TCH_Level Filter_Length_for_Ncell_RX_LEV (cell)

LDR/LUR (cell)

LMRG (adce)

NOT OPTIMIZED

HOLTHLVDL/UL (cell)

LEVHOM (adce)

HOAVLEV.aLevHo (cell)

HoDlLevThs HoUlLevThs (cell)

HoMarginRxLev (adce)

HoDlLevWindow (cell)

Figure 7: Summary of Handover parameters optimized by xParameters Handover Optimization algorithm

9.2.1.4

Optimization Guidelines The global configuration and guidelines of a feature is what we call strategy and it has to be defined properly before starting xParameters optimization. The finetune of a subset of parameters of the feature is what we call xParameters optimization. The following steps show the usual approach to start an optimization campaign: •

•

An optimization area is selected: The target area of the optimization. This area can be built graphically selecting sectors, sites or complete BSCs. Also a list of sectors can be imported. An optimization algorithm is chosen and configured according to the Optimization objectives. The Optimization is scheduled.

•

o

o

For example, for weekly optimization:The optimization task will run every Monday at 2.a.m. A Measurement Period is configured: The decisions of the algorithms will be based on the KPIs taken from this measurement period. Example, last 5 working days using the Busy Period (1h, 2h, 3h…) of each cell, considering busy period CS traffic or PS traffic, combined CS+PS traffic, etc…


69 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

•

Reference

Continuously during the optimization campaign the parameter changes are submitted and performance is monitored in order to decide on algorithm settings changes, strategy changes, implementation errors etc.

A typical plan of f ast optimization campaign, campaign , with daily or bi-daily iterations would be:

•

•

Week 1:CM/PM 1:CM/PM sample collection, compatibility verification, OSS access request Week 2: 2: If previous tasks are successful, ODG installation, baseline KPIs collection, kick-off meeting

•

Week 3: 3: Daily iterations

•


•


•

Week 6: 6: •

•

•

Optimized KPIs collection if the KPIs in week 5 are not already stable Parameters reversion if needed

Week 7: Report 7: Report generation

Normally 10 to 15 iterations are required to see a convergence of the parameters and performance of the network. Examples of traffic management “strategies” vs. xParameters optimization: optimization: •

HCS feature (Ericsson) Parameter “LAYER” is assigned per cell to assign priorities in the hierarchy. Typical case in a dual band network with 1800Mhz being the preferred band for carrying traffic is that LAYER(1800Mhz)

70 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

What is more priority and what is less priority is the “strategy” applied to HCS, and xParameters just work on that scenario. •

Umbrella HO (Nokia) Parameters EUM (enable umbrella handover) and AUCL determine if this kind of handovers is active in a cell (EUM) and which is the minimum level threshold required in the target cell (AUCL) to perform it. xParmeters does not activate EUM , it just optimizes AUCL values if umbrella handovers are enabled in a given handover relation. EUM=Y and AUCL<-47dBm A->B PMRG=63 and AUCL=-47dBm B->A If xParameters founds this configuration, it will understand that this is an umbrella controlled pair of neighbors and then umbrella level threshold AUCL will be optimized to move traffic between A and B. If the configuration found is different the controlling parameter for this pair of neighbors would be PMRG or even an unsupported configuration might be found and xParameters would discard the pair of neighbors from optimization.

•

Half Rate Another typical case of feature that xParameters does not optimize. The parameterization of Half Rate, AMR etc is not part of xParameters algorithms. Indirectly in a congested network where xParameters is able to reduce congestion, a reduction of Half Rate usage might also be achieved, at least locally.

There is an exception to this traffic management strategies, Siemens 2G. xParameters will not work fine in the traffic management between bands unless an “initial deployment” of inter-band traffic management is implemented. This is because in Siemens BSS there is not a parameter in terms of “level threshold” that controls the balance of traffic between cells of different “layers” (like layerthr in Ericsson, AUCL in Nokia etc). That behavior can be emulated with power budget handovers by setting negative handover margin (HOM) between low and high priority cells and (900 towards 1800 for example) and limiting the access through minimum signal level for handovers (RXLEVMIN)


71 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

With this respect, xParameters provides and “initial deployment” feature which produces the script of changes to deploy this “strategy”. More details in xParameters helpset, Siemens Traffic Balance 2G.

9.2.2

3G/LTE Optimization According to the result resu lt of the performance analysis, anal ysis, some optimizat ion actions can be adopted for other new or existing technologies that are envisaged to absorb traffic coming from the carved technology (GSM). For further instructions on this topic, please refer to services product catalog where generic optimization services for 3G/LTE can be found. http://prodcat.ericsson.se/frontend/product.action?code=fgc%20101%200109

10

Acceptance and Conclusion If the pre-defined amount of spectrum has been re-farmed as agreed and KPIs are in accordance with the KPI Agreement, a Conclusion Report will be done. The acceptance of the Conclusion Report will constitute the acceptance of the Spectrum Refarming Service.

11

Roles and Competence The functional skills for Spectrum Refarming are Radio Network Design and Radio Network Performance Improvement. The functional roles and their competence requirements are listed below.

11.1

Delivery manual

Framework Functional role: N&TC Project Manager

Doc number: SR 30675210

Functional role: N&TC Solution Consultant


Functional role: N&TC Network Consultant


Tools Drive Test

TI, MCOM, MapInfo

Frequency Plan

xAFP

RF Optimization

xACP

RRM Optimization (cell level parameters

xParameters


72 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

optimization) Parameter Management

Nexplorer, xNetmanager

Consistency Checking

Nexplorer, CAN, xNethealing

Performance Monitor Command shell

xNetHealing Winfiol, CNA

RSG Access OSS Access

F-Secure, RSG Connect Citrix Client, Xmanager

Measurements data file parser

File Converter

Report Template

Frequency Re-farming report

Reference Projects List FLP Trials Retuning_Claro_Nicaragua_PA1 T-Mobile_JAX MTA_spectrum reduction phase 2 CIP_Spectrum_Reduction_Project_T-M CIP_Spectrum_Reduct ion_Project_T-Mobile obile NJ Celcom_WCDMA900_Trial_Project_Fr Celcom_WCDMA900_T rial_Project_Frequency_Refarming_Methodology equency_Refarming_Methodology Vodafone Xelerate Refarming Refarming Process. Doc Num:OPER/MUANZANum:OPER/MUANZA-08:006254 08:006254 Optimi Reference #1 Optimi Reference #2

System Guidelines & Product Information GSM Product Information & RF Guidelines [CPI Store] Store] [GASK GASK]] [Ericsson Product Catalog] Catalog] GSM RN Guidelines GSM BSC Dimensioning Handbook Tools N&TC Tools Portal Radio Network Services Tools Process PROPS-C

12

APPENDIX A. Co-Existence between GSM and WCDMA in the Same Frequency Band Given an operator that will deploy WCDMA within its current limited GSM spectrum the issues can be summarized as:



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

•

•

Reference

Refarming of many GSM carriers makes the GSM frequency replanning “difficult” but creates “few” inter-system interference issues (case a below) Refarming of few GSM carriers makes the GSM frequency re-planning “easy” but creates “severe” inter-system interference issues (case b below)

Figure 8

12.1

73 (89)

No.

Two Refarming scenarios

Interference and site scenarios Due to the imperfectness of the transmitter and/or receiver we may list some interference scenarios on how GSM and WCDMA interfere with each other.


74 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Figure 9

Reference

What and where the potential problems are

As Fig. 2 is illus trating there are four important interference cases4 •

GSM downlink interfering with the WCDMA downlink

•

WCDMA downlink interfering with the GSM downlink

•

GSM uplink interfering with the WCDMA uplink

•

WCDMA uplink interfering with the GSM uplink

In addition there are two site scenarios to consider: •

Coordinated sites i.e. the WCDMA and GSM antennas are co-located

•

Un-coordinated sites i.e. there is no site sharing

Finally there are essentially two common ways of replacing GSM carriers with a WCDMA carrier namely

4

•

WCDMA carrier(s) sandwiched in-between GSM carriers

•

WCDMA carrier(s) at the GSM bandwidth allocation edge

RBS-RBS and terminal-terminal interference is not an issue due to the large duplex distance



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

12.2

75 (89)

No.

Reference

WCDMA DL capacity loss due to GSM The WCDMA DL capacity loss is controlled by the WCDMA terminal channel selectivity requiring at least at 2.8 MHz separation. It is therefore difficult to make a prediction about performance if the carrier separation is decreased. However, regardless of the terminal performance, at a channel separation of 2.2-2.3 MHz the channel leakage increases dramatically and would make it very difficult indeed to operate with this kind of channel separation. However, if the GSM channel power is sufficiently controlled and the traffic load is small it is possible to operate at a tolerable impact on the DL capacity. One way of achieving this is to make sure that the GSM channels that overlap the WCDMA carrier (have spacing smaller than 2.6 MHz) are used in an low traffic sub-cell layer and aggressive BTS power control is used (and hence the impact on the DL WCDMA capacity also minimized).

12.3

WCDMA UL capacity loss due to GSM The WCDMA UL capacity loss is assumed to be controlled by the GSM terminal channel leakage. The GSM channel leakage behaves acceptable until 2.2-2.3 MHz carrier spacing, below which it becomes very difficult to operate. Note that GSM terminals have a limited dynamic range for power control and at some small path loss they simply don’t down regulate anymore. This implies that a single GSM terminal can cause severe WCDMA UL noise rise and corresponding severe degradation in coverage. The remedy here is to make sure that the load on overlapping carriers (any carrier with a channel separation to the WCDMA carrier lower then say 2.4 MHz) must be very low indeed. Another remedy is to avoid using these GSM carriers c arriers close to the base bas e station.

12.4

GSM UL capacity loss due to WCDMA The GSM UL performance is controlled by the WCDMA terminal channel leakage which is insignificant for a 2.8 MHz carrier separation. From the specification data the critical point comes below 2.5-2.6 MHz separation where the channel leakage suddenly increases.


76 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

The GSM UL performance should degrade at channel separation below 2.5MHz, howevergiven that WCDMA terminals have a much larger dynamic range in their power control it is a much less impacting effect then the one expected in WCDMA UL loss and the GSM UL performance on channels overlapping the WCDMA carrier is not significantly affected.

12.5

GSM DL capacity loss due to WCDMA The GSM DL outage is insignificant for a 2.8 MHz carrier separation. Assuming that the WCDMA base statio n controls the GSM DL performance at smaller channel separations a critical point appears to be at channel spacing around 2.5-2.6 MHz. Going below that seems to be very difficult. However, as data such as this is illustrating (Fig. 5 below) 99% of the energy is within 4.1 MHz and hence it is obviously possible to operate at lower channel spacing without disturbing any adjacent GSM carrier – at least if Ericsson RBSs are used.



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Figure 10

12.6

77 (89)

No.

Reference

Spectrum emission from one RBS3202 individual

Conclusion The preferred scenario is to use coordinated GSM and WCDMA sites and the WCDMA carrier sandwiched in-between GSM carriers. The closest/overlapping GSM carriers should be TCH only, have the smallest traffic load possible and aggressive power control. This setup allows the use of a carrier spacing as low as 2.5MHz with low performance degradation both on WCDMA and GSM.

13

APPENDIX B. Impact of BTS Hardware on xAFP settings We will see in this paragraph the interaction between sector or site hardware technology or configuration and their impact on some xAFP settings. Hardware has implication on, min and max channel spacing, hopping capabilities.


78 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

13.1

Hybrid combiner

13.1.1

Description

Reference

A hybrid combiner is a passive electronic circuit with two inputs and two outputs. The two inputs are isolated from one another. This means that there are no frequency restrictions on what can be fed as inputs. A hybrid combiner can be used in two different ways: •

•

It can be used for combining two identical signals: In that case we will have 3dB theoretical gain on one of the outputs and almost no energy on the other output. In reality combining two identical signals is not done perfectly and the typical gain is around 2.5dB. Ericsson can combine both TRXs of a DTRU to gain 2.5dB of output power. It can be used for combining two signals that are not identical: In that case both outputs are showing signal but only one of them is used. As both outputs are transmitting energy each input signal on the used output has a 3dB theoretical loss. In reality hybrid combiners introduce a slightly higher loss. The loss is typically of 3.5dB. Note that the unused output does transmit energy and needs a load that can dissipate it. This introduces power limitations when using hybrids to combine signals.

LOGISTIC

/

M

PLAN IMPLEMENTATION

13.1.2

GSM frequency restrictions implications


79 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

When hybrids are used to combine cell’s carriers, its frequency restrictions are only related to GSM spectrum capabilities. One can’t use Co channel carriers within a cell and shouldn’t use it within a site. One should not use adjacent channels within a cell but could within a site if the network has to carry a lot of traffic per carriers (high fractional load). •

•

Hopping capability implication: A cell has no hopping type restriction imposed by the usage of hybrid combiners. Only the TRU capabilities will tell if a type of hopping can’t be used. Some very old BTS do not support SFH. Repeater’s special case: It is important to take into account if a cell is donor to a repeater using cavities. Some repeaters amplify the entire band in one go; others amplify each carrier separately using cavities to recombine them after amplifications. Cavity combined repeaters have the same frequency constraints that a cavity combined TRXs of a BTS. A donor cell will have to take the constraints of the repeater if they are more restrictive that its own.

13.2

Cavity combiner

13.2.1

Description A cavity combiner c ombiner is a passive electronic circuit with one input per carrier and one output transmitting the sum of all carriers. To guaranty a good isolation between all transmitters a high Q pass band filter is used for each one of them. This filter usually uses tunable cavities to keep the flexibility to adapt to any carrier frequency. Those cavities can be remotely tuned to allow frequency retunes. The typical insertion loss applied to each carrier by the pass band filter and circuitry linking all of the filters together is around 1dB. This insertion loss goes up to a few dBs depending on filter quality. It is typically possible to combine up to16 carriers and still have insertion losses that are only of a few dBs. Cavity combiners are able to handle high level of input energy as their real limitation is if the electrical potential differences inside the cavities becoming so high that arcing begins to occur. It is the dielectric characteristics of the insulator used in the cavities that are the limiting factor.


80 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

DEL

POST

CHECKS

13.2.2

GSM frequency restrictions implications

High Q cavity pass band filter are like any filters sensitive to the load it sees at its output. If you change the load, the filter’s frequency response will change. If too much energy is radiated back from the spectrum side lobe of surrounding carriers into the pass band filter, within its bandwidth, it will see it as a VSWR or dis-adaptation of its impedance. Therefore this will change the characteristics of the filter. This can create inter-modulation or other adverse effect. As a result it is mandatory to respect channel spacing between carriers as each carrier radiates power outside its band (See 3GPP TS 45.005 for more detail or GSM spectrum graph in paragraph 3.1.3 ). • On a sector level: Typically GSM cavity combined cells will always need a minimum of 600 Khz between carriers (f1, f2, f3, f4; f1 and f4 are the useable channels). For 1800 or 1900 it is recommended to go up to 800 Khz of separation. In practice since cavity’s quality has improved the rule of 600 Khz is used for all GSM bands. On a site basis: The required separation should depend on how • isolated sectors are from one another. In practice adjacent channels are avoided for all GSM bands. Isolation between transmitting antennas of 2 sectors should be greater than 30dB.



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

13.2.3

81 (89)

No.

Reference

Hopping capability implication

A BTS using cavity combiners will not be able to use SFH. SF H requires changing almost instantaneously the carrier’s frequency. The minimum time to retune a cavity combiner is around 2 seconds. Therefore tunable cavity combiners are not versatile enough to support SFH.

13.3

Antenna combining (also called air combining)

13.3.1

Description It is too often forgotten. Why not transmit sub groups of carriers using several antennas. The only conditions are to have, a sufficient isolation between antennas and, enough antennas and feeders. With 4 antennas (or two cross polarized antennas) we could transmit 4 carriers with no combiner insertion losses or 8 carriers using one level of hybrid combiner keeping possible to use SFH with only 3.5 dB insertion losses. The required isolation between antennas is typically the isolation found in cross polarized antenna specifications between polarizations (>30dB).

o

GSM frequency restrictions implications: none

o

Hopping capability implication: none


82 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

This type of combing is mentioned as a BTS hardware constraint for its implications towards ERP and site or sector antenna mounting constraints. •

•

A cell using air combining might not have any insertion losses due to combining or only up to 3.5 dB carrying 4 TRXs using hybrids. This is a usual configuration using one cross polarized antenna. As a result when checking the ERP information given by the customer is it wise to keep this in mind to avoid adding too much losses trying to correct incorrect information. Radiating cables or indoor solutions might have several layers of hybrid combiners as cross or multiple antennas are not used. Regarding sector configuration, when more than one physical antenna is being used, it is not unusual to have them not pointing in the same direction or not sharing the same mechanical tilt. This can damage the cell’s quality as different converge patterns can occur depending on the antenna carrying the call. Only one of them has the BCCH. It is the BCCH coverage that will trigger incoming hand over. If the TRX has much less good coverage than the BCCH a mobile might hand off into a dangerous situation.

13.4

Wide band Amplifier combining

13.4.1

Description This equipment is also called MCPA multi carrier power amplification. The idea is to sum all carriers before amplification. Two drastically different approaches exist to achieve it.

a) Sum all input signal at low power level before amplification.


83 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

i. Combine digitally all signals before converting it to analog and amplifying it.

The amplification stage must be wide band and high power. The limitation to such a loss free transmission is its linearity combined with high power. Any nonlinearity will create inter-modulation between carriers or even in the case of high modulation schemes like 64 QAM, sensitive to phase and amplitude, it would degrade the transmitted signal itself. To build extra wide band amplifiers capable


84 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

of transmitting the whole GSM bandwidth, for one of its spectrums (850, 900, 1800, 1900), is possible but would require highly inefficient class A power stage on top of using all known tricks to make a none linear amplifier linear. Among those tricks you have feed forward or digitally pre distorted signal (DPD) in a mirror image way of how the amplifier will distort the signal. If an amplifier is inefficient it means that it will transform energy into heat as only a small part of the power is transmitted. This technology is limited by: • •

High utility costs: Very high power consumption. Higher installation cost, a lot of heat to dissipate means bigger cooling systems.

The larger the bandwidth, the higher the complexity and power consumption and price are. Avoiding class A output power stage is key if we want to make this combining sustainable. With today’s technology, it is common to build an efficient 20 to 25 mega hertz bandwidth amplifiers with adequate linearity and low consumption. GSM frequency restrictions implications: One must be very careful when setting the maximum channel separation an amplification system can withstand. The lower and upper carrier must have its entire spectrum included within the amplifier’s bandwidth. A 20 mega hertz bandwidth amplifier can’t have 20 000/200 of maximum channel separation; the upper and lower channels, if used, would have their own spectrum cut by half. In reality the maximum channel separation is (20 000/200)-1. Depending on the technology used, frequency restrictions can be either given as: a defined frequency range or a maximum channel separation.

•

Hopping capability implication: None

•

Repeaters special case:

It is important to take into account if a cell is donor to a repeater using band selective amplification. Current band selective repeaters have a bandwidth up to 25Meg. Linearity is the limiting factor as it was for base stations.



370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

13.5

85 (89)

No.

Reference

Repeaters If a cell is donor to a repeater, it is the most restrictive constraint between the BTS and the repeater that will dictate the cell’s frequency and hopping constraints. As repeaters use the same type of technologies to recombine all carriers than BTSs all previous paragraph of this section apply to them as well.

13.5.1

Synchronization equipment BTSs usually take their clock reference from the BSC link. TRUs of a same cabinet can be considered as always synchronized. Synchronization of two cabinets sharing the same technical room is not guaranteed. They need to be close enough from one another to share the same clock time reference and be properly configured to allow correct synchronization. Un-synchronized cells within a site are usually found for high traffic density configurations. The need of having several cabinets, for providing a sufficient number of TRUs, might force to install them too far apart from one another or not share the same transmission link, or even not share the same BSC. The only way to guaranty a perfect synchronization within a group of cells that are not necessarily collocated is by using an external GPS synchronization. Hopping MAL can be shared between sectors, or more generally between TRUS of different cabinets, only if they are synchronized; collisions are handled by MAIO and HSN planning. This is what is used in 1x1 hopping reuse configurations. If cells of a site are not synchronized, even if they share the same HSN, it is not guaranteed that MAIO planning will do much to prevent collisions. There are two main reasons for this:

a) To prevent collision between sectors we need a perfect time alignment between the clocks of each sector. If there is a phase shift between them the hopping frequency changes might not occur at the exact same time in each sector and therefore collisions will happen. b) The frequency channel sequence is indexed, among other parameters, using the frame number of the frame being sent by the BTS. If two cells are not synchronized they will drift from one another and not share the same frame number anymore. Therefore the hopping frequency channel sequence will differ even if HSNs are the same. Defining hopping and hand over between two sectors of a given site as if they were synchronized, when in reality they are not, will always result in a low hand over successful rate and high drop call rate. 13.5.2

Legacy equipment restriction

13.5.2.1 Legacy antenna


86 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

An antenna may be adapted for f or PGSM but not for GSMR G SMR or EGSM. This might occur when an historical operator has to accept a new channel distribution. Since extra channels of GSMR and EGSM are not too far away frequency wise from PGSM spectrum, the VSWR increases is not too severe. A cell will suffer at least coverage losses when using channels not within the antenna’s band. Problem of such situations is that sometimes operators did not keep track of the exact equipment they had installed. 13.5.2.2 Legacy TRUs or BTS restriction

Some legacy equipment had a limitation such as: a) Not Synthesized frequency hopping capable (Usually BBH is still an option). Those restrictions should be provided by the operator. b) The maximum number of frequencies a MAL could contain is more restrictive than 3GPP limitations. c) GSMR or EGSM frequencies are out of range for some PGSM TRUs. As a result using the extra set of channels on those cells becomes impossible. Even though for the time being it does not have any impact on XAFP setup parameters, it is good to recall that not all TRUs are 8PSK ready. As a result EGPRS is not possible on those equipments.

14

APPENDIX B. Ericsson Radio Equipment solutions

14.1

New Ericsson filter in the node B Ericsson has developed a new ASIC filter in the node B receiver that is adjustable to 3.8, 4.0, 4.2, 4.4 and 4.8 MHz. With this filter in place extensive simulations have shown that a channel spacing as low as 2.2 MHz (between WCDMA and GSM) is feasible based on typical performance and real characteristics of Ericsson RBS and Ericsson (EMP) UE platforms (both GSM and WCDMA). That is, since terminals perform better it is possible to re-farm fewer GSM carriers (to replace with a single WCDMA carrier) then was possible initially and thereby makes the GSM frequency re-planning easier. The system impact of reduced channel spacing to 2.2 MHz for a coordinated scenario is as follows •

•

GSM UL/DL is barely affected by WCDMA WCDMA suffers slight capacity degradation in DL for rel-99 and slight throughput loss for HSDPA (up to ~5% in worst case)


87 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

•

Reference

Since several new GSM carriers are made available, we would benefit from a net gain in number of users in such networks.

In an uncoordinated scenario the recommendation is to maintain at least a 2.6 MHz carrier separation since •

•

•

14.2

GSM DL affected from capacity loss especially when using tight frequency reuse i.e. 1/1 reuse WCDMA suffers from capacity and throughput losses in both UL and DL. No benefit of running WCDMA on 4.2 MHz un-coordinated due to the high capacity loss in both GSM and WCDMA/HSPA.

Antenna solutions There are essentially three different antenna solutions, illustrated in Figs. 6 - 7 below.

ο

ο

Figure 11 Antenna solutions n 1 and n 2 for the coordinated co-existence scenario


88 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

ο

Figure 12 Antenna solution n 3 for the coordinated co-existence scenario ο

The main advantage with separate antennas (solution n 1) is the flexibility i.e. the possibility to use separately optimized tilts (and azimuths) for the two technologies. The disadvantages are that in many cases there is either no space available for extra antennas or there are restrictions on the number of antennas the operator can use at a given site. Note that separate antenna tilts are available with a double X-pol antenna, a bulky antenna but a leaner solution than two separate antennas. ο

Solutionn 2 saves antennas on site and may therefore be much more feasible than solution n 1. ο

ο

Solutionn 3 is a neat solution but has the disadvantage of making a larger set of the GSM frequencies unavailable (extra guard band needed), creating an even higher demand on the planning of the GSM frequency allocation after re-farming. There are also other disadvantages such as inflexibility in WCDMA frequency allocation.

15

References [2] Functional Role Description: Radio Network Design Consultant. 0207-FBD 101 051 [3] Network Consulting Certification, Certification,Competence description & certification criteria [4] ECC Report 82: Compatibility Study For UMTS Operating Within The GSM 900 And GSM 1800 Frequency Bands


89 (89)

No.


370/102 60-FAY111 10

Approved

Checked

Date

Rev


EBCSFIL

15/5/2011

E

Reference

[5] FLP Guideline [6] T-Mobile_JAX MTA_spectrum reduction phase 2 [7] Celcom_WCDMA900_Trial_Project_Frequency_Refarming_Methodology [8] WCDMA deployment in 850/900 MHz band through better GSM spectrum efficiency

Ericsson Spectrum Re-Farming

Recommend Documents