Optimization Methodology Using GSM Forte

Optimization Methodology Using GSM Fort

Schema Confidential & Proprietary May 2008 Page 1 of 17


Copyright Notice Due to a policy of continuous product development and refinement, Schema reserves the right to alter the

specifications and descriptions outlined in this publication without giving prior notice of any kind. In addition, no part of this publication, taken as a whole or separately, shall be deemed to be part of any contract for

equipment or services.

Schema retains the sole proprietary rights to all information contained in this document. No part of this publication may be reproduced, stored in a retrieval system, transmitted in any form or by any means,

including but not limited to: electronic, mechanical, magnetic, photocopy, recording, or otherwise, in use now or in the future, without prior written consent from Schema.

Copyright 2008 Schema


Table of Contents 1 Introduction.............................................................................................. 3 2 Project Workflow for Single Band Market.................................................. 4 3 Project Workflow for Dual Band Markets................................................... 5 4 MS Recording Creation .............................................................................. 6

4.1 Pre-project Neighbor Optimization.....................................................................6 4.2 BSIC Check ...................................................................................................6 4.3 Recording Implementation ...............................................................................6

4.3.1 Recording Length..................................................................................6 4.3.2 Shadow-Breaking..................................................................................7 4.3.3 Conclusion ...........................................................................................9

5 Neighbor Optimization ............................................................................ 10 6 Frequency Planning................................................................................. 11

6.1 Available Frequency Planning Strategy.............................................................11 6.1.1 Theoretical Background of Frequency Hopping ........................................11 6.1.2 Frequency Hopping .............................................................................11

6.2 Hopping Method ...........................................................................................12 6.3 BCCH Planning for SFH ..................................................................................13 6.4 Taking Advantages of Fort Planning Capabilities ..............................................14

7 Parameter Optimization .......................................................................... 16 7.1 Dual Band Parameter Optimization ..................................................................16 7.2 Feature Activation.........................................................................................16 7.3 Hopping ......................................................................................................16

8 Project Planning...................................................................................... 17



1 Introduction

This document presents the methodology for optimizing any GSM network (single band or dual band markets GSM900/DCS1800 or GSM850/PCS1900) using Schemas GSM Forts optimization tool.

The complete optimization cycle includes:

Preliminary analysis Set-up Recording Neighbor Optimization Frequency Planning Parameter Optimization (traffic management between bands, feature activation) This document offers a general approach to the optimization process, and is not meant to replace the existing training document used for each vendor (Seven Basic Steps to Optimize an Ericsson Market).



2 Project Workflow for Single Band Market

INITIAL RECORDING (OPTIONNAL)

INITIAL RETUNE (OPTIONNAL)

FINAL RECORDING

FINAL RETUNE

NEIGHBOR OPTIMIZATION

PRE-PROJECT KPIs COLLECTION

NEIGHBOR DELETION/ BSIC CHECK

TRAFFIC PARAMETERS FINE TUNING



3 Project Workflow for Dual Band Markets

INITIAL GSM 900 RECORDING

INITIAL GSM 900 RETUNE

FINAL GSM 900 RECORDING

FINAL RETUNE (BOTH BANDS)

NEIGHBOR OPTIMIZATION

DUAL BAND TRAFFIC PARAMETER OPTIMIZATION

PRE-PROJECT KPIs COLLECTION

NEIGHBOR DELETION/ BSIC CHECK

DUAL BAND TRAFFIC PARAMETERS FINE TUNING

DCS 1800 RECORDING (optional)

FINAL DCS1800 RECORDING

INITIAL DCS 1800 RETUNE (optional)



4 MS Recording Creation

The MS recording process must be performed carefully to minimize the length of the project, especially with regard to Siemens, Alcatel, Motorola, Huawei, or Nortel, where the BA list cannot be modified by the user, but only through the creation of fake neighbors.

4.1 Pre-project Neighbor Optimization

Before starting the MS recording, an analysis of the neighbor list should be performed using Fort in order to identify all cells with more than 25 neighbors.

Based on this analysis, a list should be created with neighbors to be deleted, using only Handover Statistics, (as a model is not yet available) based on the following workflow:

1. Identify all cells in the optimization set that have more than 25 neighbors.

2. Rank each neighbor relation by descending number of handover attempts.

3. Delete 6 of the last 12 neighbor relation by targeting inter-band HO relations first (in case of dual band market)

4.2 BSIC Check

Before beginning recording, the Fort BCCH-BSIC Reuse report should be checked to ensure that the BSIC plan does not include any close BCCH/BSIC reuses.

Any reuse below 5 or 10 km (depending on the willingness of the customer to make changes) should be corrected, and a new BSIC implemented, to avoid decoding problems during the modeling phase.

4.3 Recording Implementation

4.3.1 Recording Length

Single Band Markets

The length of recordings varies, based on the number of BCCHs to measure, and the vendor.



Markets using free BCCH/TCH planning (with or without Base band hopping) may require several recording sessions per cell when more than 30 BCCHs are used.

Dual Band Markets

The length of recordings varies for each band, based on the number of BCCHs to measure and on the vendor.

Since the DCS1800 (PCS1900) layer is normally the capacity layer with more available spectrum, it uses more potential BCCH. On average, it takes twice as long to record the DCS layer as it does to record the GSM layer.

In theory, cross band measurements are only needed:

If a network has a mixed configuration of Common BCCH or CBCCH (e.g., BCCH on band 850/900 and TCH on band 1900/1800 within the same sector) and Multiple BCCH or MBCCH (e.g., two co-sited sectors, one with 850/900 BCCH and one with 1900/1800 BCCH); all sectors should consider measuring interferences from both bands (mandatory).

If cross-band HO optimization is required: The C/I between sectors belonging to different bands must be available for the user to optimize the handovers between bands. Therefore, measurement recording on both bands should be activated. If this not possible (as in siemens), Fort can estimate cross-band measurements in the model by using intra-band measurements, and applying an attenuation value based on the difference in propagation between the bands.

The Fort Cross Band Measurements feature is available for dual band networks. While it is possible and recommended to measure both bands in a dual band network, when more than one band is measured, the measurement recording time increases. Therefore, this feature should only be used in one of the circumstances described above.

Note than cross-band measurements are not available in Siemens, so the only way to evaluate cross band impact is by using the Fort Cross Band Measurements feature.

4.3.2 Shadow-Breaking

Recordings should provide the most accurate model possible. Because of shadowing (blind spots), they may be useful to perform an initial retune and collect two sets of recordings (before and after the initial retune). For example, two sectors using the same BCCH will be considered shadowed. The interference between the two will only



be estimated except if measurements can be collected before and after the BCCH have been changed.

Shadow breaking may be required when the number of BCCHs used is limited (below 20) In most cases, an initial retune is recommended for the GSM layer.

Check the following for shadowing:

The BCCH reuse report before a project shows the number of close BCCH reuses below 5 km may indicate poor BCCH planning and high shadowing

A low correlation with an under-predicted model may indicate a high level of shadowing (but the first recording session must be finished before it is possible to see the shadowing).

If an initial BCCH retune is performed for shadow-breaking, the Reveal shadowing option must be used in the initial plan. That option should not be used for the second and final plan.

Note: the initial BCCH retune goal, also called mini-BCCH retune, is not to improve the network statistics, but simply to reveal shadowed impacts (and eventually clean up the BSIC plan). It is recommended to that Fort change all the BCCHs in the retune area. Otherwise, the Minimum change option can be used to allow only a pre-determined number of changes.

If no initial BCCH reuse is performed, the user may or may not use that option in the Frequency plan.



4.3.3 Conclusion

An initial retune is recommended for markets having less than 20 BCCHs. MS statistics should be recorded before and after the retune.

In dual band markets, the capacity layer (DCS 1800) does not always require a shadow-breaking retune. Since twice the recording time is required, only one retune (one set of recordings) should be implemented for that layer. The recordings should be made after dual-band parameter optimization is implemented, if possible.

If additional time is available, inter-band (cross-band) measurements could be recorded.



5 Neighbor Optimization

Neighbor optimization should be prepared with a shadow-free model, if possible, and implemented before the final retune.

The main goal of the neighbor optimization is to decrease the constraints on the BCCH plan, not to improve network statistics (HO Success Rate), since most customers already have a reasonable HO list.

Fort uses neighbor relations as logical constraints when planning BCCH to avoid using co-channel BCCHs between neighbors, which is usually not permitted in the OSS. A large number of intra-band neighbors, creates many logical constraints (not physical constraints), which are NOT based on measured interference from MS.

The Handover optimizer should be run on the layer with the most frequency planning constraints. Minimizing the number of neighbor relations improves the quality of Forts BCCH FP output.

Optimizing inter-band HO for dual band markets reduces the length of the recording process, but does not necessarily improve performance.

If cross-band MS cannot be recorded because of the extra length associated with it, Fort should not be used to optimize inter-band HO relations, since those relations are only based on HO Statistics (not measured interference).

The DCS1800 to 1800 Handover relations can be optimized along with GSM900 to 900, but this will usually not result in significant improvement in the performance or quality of the Fort 1800 BCCH plan because of the high number of available BCCH frequencies.



6 Frequency Planning

Frequency plans can be prepared separately for both bands, according to the Fort methodology.

6.1 Available Frequency Planning Strategy

Several frequency planning strategies are available to operators, which involve BCCH vs. TCH planning, hopping vs. non-hopping strategy, and base band hopping vs. synthesized hopping.

Activating frequency hopping will maximize improvement to the network.

6.1.1 Theoretical Background of Frequency Hopping

GSM networks evolve with the goals of providing better quality of service and more system capacity. Frequency hopping helps to achieve these goals.

Cellular systems are limited by interference. Multiple co-channel interference, though controlled, is normal, and determines the limits of the service area. The higher the interference level, the harder it is to reuse available frequencies within the smallest area. Since quality of service depends on the carrier/interference ratio (C/I) more than on the signal/noise ratio, the system can tolerate the trade-off between quality and capacity.

Higher levels of capacity and quality are needed to support fast network growth. All possible techniques should be used to progressively enhance radio and network performance. GSM has some powerful mechanism intended to reduce the effect of interference through frequency hopping, discontinuous transmission (DTX) and power control.

6.1.2 Frequency Hopping

Frequency hopping can be classified as either Base Band hopping (BB) or as Synthesized Frequency Hopping (SFH). SFH uses only one transmitter for all bursts in a specific connection, while base band hopping uses as many transmitters as frequencies in the hopping sequence. Hopping can be cyclic or random, but in random hopping, a HSN other than 0 should be chosen.

SFH involves changing channel frequency in every transmitted burst (217 hopes per second) thus providing frequency diversity and interference averaging. This randomizes the risk of interference and improves channel behavior (for selective fading).



The following factors affect SFH performance:

1. Number of Hopping Frequencies A higher number of hopping frequencies improves system performance by increasing frequency diversity. It is not helpful to use more than eight hopping frequencies since the GSM interleaving period consists of eight bursts.

2. Hopping frequencies separation A larger frequency separation between hopping frequencies, improves system performance as the effects of propagation become more uncorrelated. Frequency spacing directly affects Fast Fading. A separation of three to five channels between hoppers provides the maximum gain.

3. System load Since a low system load results in lower interference probability in each hopping frequency, this directly affects SFH performance.

The choice of frequency hopping strategy depends on each networks optimization level and available spectrum.

6.2 Hopping Method

Frequency hopping should be activated before using any frequency planning methodology.

The choice of one method over another (Base band vs. SFH) depends on the available spectrum, the system load (number of TRXs and EFL), and the maturity of the market (either fast growing or mature with very few new site activations). The hopping strategy influences BCCH planning strategy.

1. Base band hopping does not require separating the BCCH and the TCH bands and therefore allow free BCCH/TCH planning (BCCH and TCH will use all available frequencies). In addition, all traffic TS on the BCCH TRX will be hopping. But it has two main disadvantages:

There is no gain, or a limited gain, of cells with a low number of TRXs (2 or fewer).

Extensive planning is required, since a new frequency must be planned for every new TRX in the system (TRX addition or new cells)

2. The standard SFH (SFH1:1 or SFH1:3), involves splitting a spectrum into two separate groups of frequencies for BCCH and TCH -- two blocks or a staggered allocation (1 BCCH, 1 TCH). The disadvantages are:



A specific number of frequencies must be reserved for BCCH planning only (between 12 and 21), potentially affecting the quality of the BCCH plan.

The traffic TS on the BCCH TRX will not hop. Without frequency planning, the random collisions between cells cannot be

controlled.

The main advantage of SFH is that it does not requires extensive planning, so that TRX or new sites can easily be added to the network, which is especially good for fast-growing markets. In addition, SFH brings the quality gain of hopping even to cells with low number of TRXs, since the number of hoppers is much greater than the number of TRXs.

3. SFH Ad Hoc is a different version of SFH that allows planning a specific MAL (with a different frequency and length) for each sector, instead of a fixed MAL used by all sectors (SFH 1:1). MAL length is based on the number of TRXs in each cell. Normally, a minimum of three or four hoppers is used for each cell with a MAL length equal to the number of TRXs+1.

SFH Ad Hoc planning reintroduces frequency planning within SFH, and keeps SFH quality gain even for cells with few TRXs. Random collisions are avoided through the Interference matrix that Fort creates.

6.3 BCCH Planning for SFH

Two dedicated groups of frequencies for BCCH and TCH should be used with SFH (especially with fixed MAL), but this does not necessarily mean that there will be two frequency blocks.

A staggered BCCH vs. TCH plan can be used, interleaving some blocks of BCCH and TCH within the spectrum:

Example 1: available spectrum 1-24, channels 1-2 will be TCH, 3-4 BCCH Example 2: available spectrum 1-24, channels 1will be TCH, 2 BCCH



The main advantages of using a staggered plan are:

Capacity: Along with SFH1:1 or SFH1:3, the staggered plan allows use of all available MAIO, so that more TRXs can be planned.

Quality: The staggered plan increases frequency diversity more than using two blocks, especially when limited frequencies are available.

6.4 Taking Advantages of Fort Planning Capabilities

Fort offers a highly accurate Interference Matrix, based on real traffic distribution within the network.

To obtain the best possible quality vs. capacity ratio, a non-random frequency planning strategy should be used, such as:

Example 1 Example 21 TCH 1 TCH2 TCH 2 BCCH3 BCCH 3 TCH4 BCCH 4 BCCH5 TCH 5 TCH6 TCH 6 BCCH7 BCCH 7 TCH8 BCCH 8 BCCH9 TCH 9 TCH

10 TCH 10 BCCH11 BCCH 11 TCH12 BCCH 12 BCCH13 TCH 13 TCH14 TCH 14 BCCH15 BCCH 15 TCH16 BCCH 16 BCCH17 TCH 17 TCH18 TCH 18 BCCH19 BCCH 19 TCH20 BCCH 20 BCCH21 TCH 21 TCH22 TCH 22 BCCH23 BCCH 23 TCH24 BCCH 24 BCCH



Base band hopping for markets with many TRXs/ cells (3 or more TRX/cells), including free BCCH/ TCH planning.

SFH Ad Hoc planning, preferably with dedicated BCCH and TCH channels (blocked or staggered to increase frequency diversity) for markets with limited spectrum or few TRX/cell (2 TRX/cells average).

Both methods have advantages and disadvantages, but also significantly increase network performance and capacity.



7 Parameter Optimization

7.1 Dual Band Parameter Optimization

This topic is covered in a separate document that deals only with Dual band markets.

7.2 Feature Activation

GSM Fort easily checks the activation status of the major GSM features, including Intra cell Handover, Power control, DTX, and TCH allocation strategy.

When the environment is exported to Schema format, it can be checked, regardless of vendor, using the Sector and Channel Group.txt files.

The Sector.txt file includes the Intra-cell Handover activation status and the TCH allocation strategy used in the network.

Intra-cell Handover should be activated for Non-hopping or SFH (since it is irrelevant when base band is used).

The TCH allocation priority should be set to TCH first in the case of SFH (which is irrelevant when base band is used) since hopping gain provides better quality on the TCH layer.

The Channel group.txt file includes the DTX and Power control activation status:

DTX Uplink should always be activated (downlink may provide some benefit, but also results in voice quality distortions).

Power control should be activated on both links (if only 1 TRX is present, PC DL cannot be activated).

7.3 Hopping

Frequency hopping should always be activated (either BB or SFH) and all sectors in the optimization set and guard zone should use it.



8 Project Planning

D

a

y

1

D

a

y

2

D

a

y

3

D

a

y

4

D

a

y

5

D

a

y

6

D

a

y

7

D

a

y

8

D

a

y

9

D

a

y

1

0

D

a

y

1

1

D

a

y

1

2

D

a

y

1

3

D

a

y

1

4

D

a

y

1

5

D

a

y

1

6

D

a

y

1

7

D

a

y

1

8

D

a

y

1

9

D

a

y

2

0

D

a

y

2

1

D

a

y

2

2

D

a

y

2

3

D

a

y

2

4

D

a

y

2

5

D

a

y

2

6

D

a

y

2

7

D

a

y

2

8

D

a

y

2

9

D

a

y

3

0

D

a

y

3

1

D

a

y

3

2

D

a

y

3

3

Optimization Activities Project Preparation and Logistic Kickoff meeting NBR Clean-up MS Statistics Collection (900 ) (*) N1 N2 N3 Network Modeling/Planning Dual-band Parameters Optimization Frequency plan implementation MS Statistics Collection (1800 cross band) (**) N1 N2 MS Statistics Collection (900 and 1800) (***) N1 N2 N3 N4 N5 Network Modeling/Analysis Neighbor list implementation Frequency Plan Implementation Fine tuning of fP and parameters BB Hopping Activation Performance collection/analysis Performance Benchmarking Before Performance Benchmarking After Final Report Preparation (*) N1 to N3 for 900 (N1 to N5 for 1800 optional) (**) N1 and N2 for 1800 measuring 900 BCCH (optional)

Optimization Methodology Using GSM Forte

Documents

project neighbor optimization

gsm fort copyright notice

project workflow

bcch planning

recording implementation

recording length

hopping method

gsm fort table of contents