DFCA System Description 107c

Dynamic Frequency and Channel Allocation (DFCA)

System Description

© Nokia Networks Nokia Customer Confidential

1 (93)

The information in this document is subject to change without notice and describes only the product defined in the introduction of this documentation. This document is intended for the use of Nokia Networks' customers only for the purposes of the agreement under which the document is submitted, and no part of it may be reproduced or transmitted in any form or means without the prior written permission of Nokia Networks. The document has been prepared to be used by professional and properly trained personnel, and the customer assumes full responsibility when using it. Nokia Networks welcomes customer comments aspart of the process of continuous development and improvement of the documentation. The information or statements given in this document concerning the suitability, capacity, or performance of the mentioned hardware or software products cannot be considered binding but shall be defined in the agreement made between Nokia Networks and the customer. However, Nokia Networks has made all reasonable efforts to ensure that the instructions contained in the document are adequate and free of material errors and omissions. Nokia Networks will, if necessary, explain issues which may not be covered by the document. Nokia Networks' liability for any errors in the document is limited to the documentary correction of errors. Nokia Networks WILL NOT BE RESPONSIBLE IN ANY EVENT FOR ERRORS IN THIS DOCUMENT OR FOR ANY DAMAGES, INCIDENTAL OR CONSEQUENTIAL (INCLUDING MONETARY LOSSES), that might arise from the use of this document or the information in it. This document and the product it describes are considered protected by copyright according to the applicable laws. NOKIA logo is a registered trademark of Nokia Corporation. Other product names mentioned in this document may be trademarks of their respective companies, and they are mentioned for identification purposes only. Copyright © Nokia Oyj 2004. All rights reserved.


2 (93)

Contents

1. SCOPE ................................................................................................................................................................................. 10

2. DFCA OVERVIEW ................................................................................................................................................................ 10 2.1 DFCA SYSTEM ARCHITECTURE................................................................................................................................... 10 2.2 HARDWARE REQUIREMENTS ....................................................................................................................................... 11

2.2.1 BSC ........................................................................................................................................................................ 11 2.2.2 BTS......................................................................................................................................................................... 11 2.2.3 LMU....................................................................................................................................................................... 11

2.3 SOFTWARE REQUIREMENTS......................................................................................................................................... 12 2.3.1 BSC ........................................................................................................................................................................ 12 2.3.2 BTS......................................................................................................................................................................... 12 2.3.3 NetAct .................................................................................................................................................................... 12 2.3.4 LMU....................................................................................................................................................................... 12

2.4 BSS SYNCHRONIZATION ............................................................................................................................................. 12 2.5 FREQUENCY BAND MANAGEMENT............................................................................................................................... 12 2.6 INTERFERENCE CONTROL PRINCIPLE ........................................................................................................................... 13

2.6.1 C/I estimation overview ......................................................................................................................................... 13 2.6.2 Interference control ............................................................................................................................................... 15

2.7 SUPPORTED FREQUENCY HOPPING MODES................................................................................................................... 16 2.8 SDCCH AND (E)GPRS CHANNELS............................................................................................................................. 17 2.9 SUPPORTED CONNECTION TYPES ................................................................................................................................. 17

3. DFCA OPERATING MODES ................................................................................................................................................. 17 3.1.1 DFCA hopping DFCA mode.................................................................................................................................. 17 3.1.2 Operation in case of a BSS synchronization failure .............................................................................................. 18 3.1.3 Operation in case of BSC-BSC connecion failure ................................................................................................. 19

4. DFCA CHANNEL SELECTION ALGORITHM ......................................................................................................................... 19 4.1 ESTIMATION OF C/I ..................................................................................................................................................... 20 4.2 INPUTS ........................................................................................................................................................................ 21

4.2.1 MS measurement reports ....................................................................................................................................... 21 4.2.2 Background Interference Matrix (BIM) tables ...................................................................................................... 22 4.2.3 DFCA Radio Resource Table................................................................................................................................. 23 4.2.4 DFCA adjacent channel lookup table.................................................................................................................... 24

4.3 DFCA C/I ESTIMATION PROCESS ................................................................................................................................ 24 4.3.1 STEP 1: BIM scaling ............................................................................................................................................. 25 4.3.2 STEP2: Incoming DL C/I estimation ..................................................................................................................... 25 4.3.3 STEP 3: Incoming UL C/I estimation .................................................................................................................... 26 4.3.4 STEP 4: Outgoing DL C/I estimation .................................................................................................................... 28 4.3.5 STEP 5: Outgoing UL C/I estimation .................................................................................................................... 29 4.3.6 HR specific considerations in C/I estimations ....................................................................................................... 30

4.4 CHANNEL RANKING AND SELECTION........................................................................................................................... 31 4.4.1 Channel search with the DFCA allocation method 0 (primary method) ............................................................... 32 4.4.2 Channel search with the DFCA allocation method 1 (load saving method).......................................................... 33 4.4.3 Rotation in channel allocation............................................................................................................................... 33 4.4.4 Overload protection............................................................................................................................................... 33 4.4.5 BTS selection within a segment.............................................................................................................................. 33 4.4.6 Soft Blocking C/I.................................................................................................................................................... 34 4.4.7 C/N check for cell access control........................................................................................................................... 34

4.5 TRAINING SEQUENCE SELECTION ................................................................................................................................ 36


3 (93)

4.6 CHANNEL MODE SELECTION (FR/HR)......................................................................................................................... 37 4.6.1 Overview of channel mode selection...................................................................................................................... 37 4.6.2 C/I based rate selection between HR and FR (Not available in BSS11)................................................................ 37

5. BIM UPDATE PROCESS ....................................................................................................................................................... 39 5.1 GENERAL .................................................................................................................................................................... 39 5.2 C/I STATISTICS COLLECTION ....................................................................................................................................... 40 5.3 PROCESSING OF THE C/I STATISTICS FOR BIM UPDATE............................................................................................... 41 5.4 UPDATING OF THE BIM TABLES.................................................................................................................................. 43

5.4.1 Case 1: Creating a new BIM entry ........................................................................................................................ 43 5.4.2 Case 2: Updating an existing BIM entry ............................................................................................................... 44

5.5 INTERFERENCE RELATION TERMINATION .................................................................................................................... 45 5.6 BCCH & BSIC CONFLICT MANAGEMENT................................................................................................................... 45

6. INTER BSC COMMUNICATION ........................................................................................................................................... 46 6.1 INTRODUCTION OF THE BSC – BSC INTERFACE ......................................................................................................... 46 6.2 ADDRESSING IN THE BSC - BSC INTERFACE .............................................................................................................. 47 6.3 SIGNALING TRAFFIC LOAD .......................................................................................................................................... 49

7. OPERATION AND MAINTENANCE...................................................................................................................................... 49 7.1 NOKIA NETACT NETWORK MANAGEMENT SYSTEM..................................................................................................... 49 7.2 SPECIAL CONSIDERATIONS RELATED TO BSS SYNCHRONIZATION .............................................................................. 50

7.2.1 Time slot offset for better FR SACCH performance .............................................................................................. 50 7.2.2 FN offset for optimized BSIC decoding performance ............................................................................................ 53

7.3 NETWORK PLANNING/CONFIGURATION REQUIREMENTS.............................................................................................. 55 7.3.1 BCCH & BSIC planning ........................................................................................................................................ 55 7.3.2 Frequency band reorganization............................................................................................................................. 55 7.3.3 Frame number and Time slot offsets...................................................................................................................... 56

7.4 MANAGING DFCA...................................................................................................................................................... 56 7.4.1 Definition of DFCA MA lists.................................................................................................................................. 56 7.4.2 Creation, activation, modification and deletion of DFCA MA lists ....................................................................... 57 7.4.3 Attaching/detaching DFCA MA lists to/from a BTS .............................................................................................. 58 7.4.4 Handling of BA lists with DFCA............................................................................................................................ 58 7.4.5 Defining DFCA TRXs of a BTS.............................................................................................................................. 59 7.4.6 Modifying DFCA mode of a BTS ........................................................................................................................... 59

7.5 AUTOMATIC CONFIGURATION CHANGES ..................................................................................................................... 62 7.5.1 Loss of synchronization ......................................................................................................................................... 62 7.5.2 Return of synchronization...................................................................................................................................... 62 7.5.3 Loss of inter BSC connection................................................................................................................................. 63 7.5.4 Return of inter BSC connection ............................................................................................................................. 63 7.5.5 TRX fault on a DFCA BTS..................................................................................................................................... 63

7.6 BCCH/BSIC CHANGE & CO-BCCH BLIND SPOT AVOIDANCE.................................................................................... 63 7.7 INTERACTIONS WITH OTHER FEATURES....................................................................................................................... 64

7.7.1 Intelligent Underlay Overlay (IUO) & Intelligent Frequency Hopping (IFH) ...................................................... 64 7.7.2 (E)GPRS ................................................................................................................................................................ 64 7.7.3 HSCSD................................................................................................................................................................... 64 7.7.4 Half Rate and AMR................................................................................................................................................ 64 7.7.5 Dynamic SDCCH................................................................................................................................................... 64 7.7.6 Intelligent Coverage Enhancement (ICE).............................................................................................................. 64 7.7.7 Enhanced Coverage by Frequency Hopping ......................................................................................................... 64 7.7.8 Extended Cell Radius............................................................................................................................................. 64 7.7.9 Common BCCH and Segments .............................................................................................................................. 65 7.7.10 Queuing ............................................................................................................................................................. 65 7.7.11 Directed Retry ................................................................................................................................................... 65


4 (93)

7.7.12 FACCH Call Setup ............................................................................................................................................ 65 7.7.13 Pre-emption ....................................................................................................................................................... 65 7.7.14 Power Control ................................................................................................................................................... 65 7.7.15 Power Optimization in Handover...................................................................................................................... 65 7.7.16 Interference Band Recommendation.................................................................................................................. 65 7.7.17 Dynamic Hot Spot.............................................................................................................................................. 65 7.7.18 AMH & DADL/B ............................................................................................................................................... 65 7.7.19 Antenna Hopping............................................................................................................................................... 65 7.7.20 Automated Planning Enhancements .................................................................................................................. 66 7.7.21 Enhanced Measurement Report......................................................................................................................... 66 7.7.22 FER Measurement ............................................................................................................................................. 66 7.7.23 Location Services .............................................................................................................................................. 66 7.7.24 Nokia Smart Radio Concept (IDD, 4UD, IRC) ................................................................................................. 66 7.7.25 BTS HW Issues .................................................................................................................................................. 66

8. DFCA PARAMETERS............................................................................................................................................................ 67 8.1 BSC LEVEL PARAMETERS ........................................................................................................................................... 67

8.1.1 Connection type specific C/I targets ...................................................................................................................... 67 8.1.2 C/I target UL offset ....................................................................................................................................... 67 8.1.3 Connection type specific soft blocking C/I limits ................................................................................................... 68 8.1.4 Connection type specific soft blocking C/N limits.................................................................................................. 68 8.1.5 Parameters related to BIM update......................................................................................................................... 69 8.1.6 Expected BSC-BSC interface delay........................................................................................................... 71 8.1.7 DFCA channel allocation method.......................................................................................................................... 71 8.1.8 Mobile allocation frequency list state.................................................................................................................... 72 8.1.9 Parameters related to the LAC to SPC mapping table .......................................................................................... 72

8.2 BTS LEVEL PARAMETERS ........................................................................................................................................... 73 8.2.1 DFCA mode.................................................................................................................................................... 73 8.2.2 DFCA MA list ids................................................................................................................................................... 73 8.2.3 DFCA unsynchronized mode MA list..................................................................................................................... 74 8.2.4 Forced HR mode related parameters ....................................................................................................... 74 8.2.5 UL/DL noise levels for C/N based access control ................................................................................. 75

8.3 TRX LEVEL PARAMETERS........................................................................................................................................... 76 8.3.1 DFCA indication.................................................................................................................................................... 76

9. DFCA PERFORMANCE COUNTERS AND MEASUREMENTS ............................................................................................... 76 9.1 HANDOVER MEASUREMENT ........................................................................................................................................ 76

9.1.1 HO ATTEMPT TO DFCA TRX.............................................................................................................................. 76 9.1.2 UNSUCC HO TO DFCA TRX ............................................................................................................................... 77 9.1.3 SUCC HO TO DFCA TRX..................................................................................................................................... 77

9.2 TRAFFIC MEASUREMENT ............................................................................................................................................. 77 9.2.1 TCH REL DUE TO BSC BSC CONFLICT CALL.................................................................................................. 78 9.2.2 TCH REL DUE TO BSC BSC CONFLICT TARGET............................................................................................. 78

9.3 RESOURCE AVAILABILITY MEASUREMENT .................................................................................................................. 78 9.3.1 TIME IN FORCED DFCA HR MODE .................................................................................................................. 78 9.3.2 TIME IN FORCED DFCA AMR HR MODE ......................................................................................................... 79 9.3.3 TIME IN FORCED DFCA HR AND AMR HR MODE.......................................................................................... 79

9.4 BSC-BSC MEASUREMENT .......................................................................................................................................... 79 9.4.1 BSC- BSC DELAY.................................................................................................................................................. 80 9.4.2 BSC - BSC DENOMINATOR 1.............................................................................................................................. 80 9.4.3 BSC- BSC PEAK DELAY....................................................................................................................................... 80

9.5 DFCA MEASUREMENT................................................................................................................................................ 81 9.5.1 C/I TARGET........................................................................................................................................................... 81


5 (93)

9.5.2 C/I TARGET UL OFFSET ..................................................................................................................................... 81 9.5.3 DFCA assignment incoming interference C/I counters ......................................................................................... 81 9.5.4 DFCA assignment outgoing interference C/I counters .......................................................................................... 82 9.5.5 SUCC DFCA ASS .................................................................................................................................................. 83 9.5.6 SUCC DFCA ASS HIGH LOAD ............................................................................................................................ 83 9.5.7 SOFT BLOCKED DFCA ASS DUE TO C/I........................................................................................................... 83 9.5.8 SOFT BLOCKED DFCA ASS DUE TO C/N ......................................................................................................... 84

9.6 DFCA ASSIGNMENT MEASUREMENT ......................................................................................................................... 84 9.6.1 DFCA ASSIGNMENTS .......................................................................................................................................... 84

9.7 BSC LEVEL CLEAR CODE (PM) MEASUREMENT........................................................................................................ 84 9.7.1 INTRA HO TO DFCA............................................................................................................................................ 84 9.7.2 INTER BSC DFCA ASSIGNMENT SUCC............................................................................................................. 85 9.7.3 INTER BSC DFCA ASSIGNMENT REJ ................................................................................................................ 85

10. DFCA PERFORMANCE SIMULATIONS............................................................................................................................ 85 10.1 GENERAL .................................................................................................................................................................... 85

10.1.1 Simulation tool .................................................................................................................................................. 85 10.1.2 Simulation settings............................................................................................................................................. 86 10.1.3 Capacity gain indicators ................................................................................................................................... 88

10.2 SIMULATION RESULTS................................................................................................................................................. 90 10.2.1 EFR performance .............................................................................................................................................. 90 10.2.2 AMR performance ............................................................................................................................................. 91 10.2.3 AMR in Narrowband environment (3.6 MHz) ................................................................................................... 91


6 (93)

ABBREVIATIONS

AMR Adaptive Multirate

BA BCCH Allocation

BB Base Band

BCCH Broadcast Control Channel

BCF Base Station Control Function

BCSU Base Station Controller Signalling Unit

BIM Background Interference Matrix

BSC Base Station Controller

BSIC Base Station Identification Code

BSS Base Station System

BTS Base Transceiver Station

C Carrier power level

C/I Carrier to Interference Ratio

CIR Carrier to Interference Ratio

CPU Central Processing Unit

CS Circuit Switched

DFCA Dynamic Frequency and Channel Allocation

DL Downlink

EFR Enhanced Full Rate

EGPRS Enhanced General Packet Radio Service

FH Frequency Hopping

FN Frame Number

FR Full Rate

GPRS General Packet Radio Service

GPS Global Positioning System

HR Half Rate

HSCSD High Speed Circuit Switched Data

HSN Hopping Sequence Number

HW Hardware

I Interference Level

IP Internet Protocol

LMU Location Measurement Unit

MA list Mobile Allocation List

MAIO Mobile Allocation Index Offset

MCMU Marker and Cellular Management Unit

MML Man to Machine Language

MR Measurement Report

MS Mobile Station


7 (93)

PC Power Control

PS Packet Switched

PWR Power

QoS Quality Of Service

RF Radio Frequency

RR Radio Resource

RRM Radio Resource Management

SACCH Slow Associated Control Channel

SDCCH Standalone Dedicated Control Channel

SEG Segment

SW Software

TCH Traffic Channel

TDMA Time Division Multiple Access

TRX Transmitter / Receiver, Transceiver

TSC Training Sequence Code

TS, TSL Timeslot


8 (93)

CONCEPTS

DFCA Layer

DFCA TRXs; those TRXs of a cell that apply DFCA channel allocation algorithm and utilise DFCA Hopping. DFCA layer is for circuit switched connections only. No conventional frequency planning but common MA list settings.

Regular Layer

Non-DFCA TRXs, regular TRXs; the BCCH TRX and possible other non-DFCA TRXs of a cell. These TRXs - and only these TRXs - may carry control channels and packet switched channels. Normal frequency planning in use.


9 (93)

1. SCOPE

This document is intended to provide a reasonable understanding of the DFCA feature including the DFCA principle, performance benefits, SW and HW requirements and deployment as well as operation of a DFCA network.

2. DFCA OVERVIEW

DFCA is a BSS radio resource management functionality that selects the radio channel individually for each connection from a dedicated channel pool based on C/I criteria. The different degrees of interference tolerance of different connection types such as EFR, AMR FR and AMR HR are taken into account in the channel selection process.

This chapter presents the basic principle of DFCA and the system requirements for DFCA operation.

2.1 DFCA system architecture

The main DFCA functionality is located in the BSC. The DFCA RRM algorithm in BSC controls the radio channel assignments in all DFCA TRXs in all BTSs controlled by the BSC.

The BTSs using DFCA must be synchronized to a global clock reference provided by the GPS satellite system. This is achieved by having a Location Measurement Unit (LMU) installed in every BTS site. The LMU incorporates a GPS satellite receiver and provides a common clock signal that is used by all BTSs in the site.

The DFCA RRM algorithm must have real time knowledge of the radio channel usage situation of all BTSs in the area. Near BSC area borders some significant interfering BTSs may be controlled by another BSC. Therefore the DFCA RRM algorithms in different BSCs must be able to exchange information between them. This information exchange is provided by the IP based BSC-BSC connection that allows DFCA signaling flow between BSCs running DFCA RRM algorithm.

Overview of the DFCA system architecture is presented in Figure 1.


10 (93)

BSC 1

DFCA

BSC 2

DFCA

IP network

LMU

Figure 1. DFCA system architecture

2.2 Hardware requirements

2.2.1 BSC

BSC2i or BSC3i is required.

In BSC2i the MCMU and BSCU units must have CP6MX CPU cards.

BSCs using DFCA and that have adjacent service areas must be connected to each other with BSC-BSC connection. This may require CPLAN-S panel in the BSC, LAN cabing, hubs/switches and other IP networking equipment.

2.2.2 BTS

UltraSite BTSs and MetroSite BTSs are supported. Wideband combiners are required because DFCA utilises RF hopping.

Since the DFCA hopping mode is not available inTalk Family BTSs the full DFCA performance gains cannot be achieved in a network with considerable amount of Talk Family BTSs. Similarly, DFCA performance is reduced if RTC combiners are used in some UltraSite BTSs leaving non-DFCA capable sites within DFCA area.

2nd generation BTS, InSite BTSs and PrimeSite BTSs are not supported.

2.2.3 LMU

DFCA requires BSS synchronization meaning that one LMU unit must be installed in every BTS site.


11 (93)

2.3 Software requirements

2.3.1 BSC

DFCA is implemented in BSC SW release S11 for trialling purpose only. Full DFCA functionality is in S11.5 release.

2.3.2 BTS

DFCA trial support is implemented in UltraSite CX4.0 SW release. Full DFCA support for UltraSite and MetroSite BTSs is available in SW version CX4.1.

2.3.3 NetAct

BSS 11.5 features including DFCA are supported in OSS4 release.

2.3.4 LMU

Requirement for LMU SW is version 4.1 or higher.

2.4 BSS synchronization

It is required that all BTSs using DFCA operate in BSS synchronization. BSS synchronization feature with necessary recovery enhancements is available in BSS11 release.

BSS synchronization is provided by a Location Measurement Unit (LMU) that is installed in the BTS site. LMU provides synchronized clock signal that is distributed to all BTS cabinets at a site. The LMU derives the synchronized clock signal from GPS time reference.

2.5 Frequency band management

DFCA requires a separate frequency band that is reserved for DFCA use only. These frequencies should not be used in any non-DFCA TRX within the area where DFCA is used. This ensures that DFCA is in full control of the usage of DFCA band frequencies and can therefore effectively control the interference. Usage of DFCA band frequencies for non-DFCA TRXs may cause some local DFCA performance degradation due to the uncontrolled interference.

A separate dedicated BCCH frequency band is recommended as this best ensures accurate neighbouring cell signal level measurements that are crucial for in the DFCA C/I determination process.

The DFCA frequency band is further divided into one or more MA lists. A DFCA MA list can contain 1-32 frequencies and a maximum of 32 different DFCA MA lists are allowed. The DFCA MA lists must be defined the same way in all BSCs within each continuous DFCA area. Since DFCA uses cyclic frequency hopping, it is recommendable to ensure that the DFCA MA lists do not contain consecutive frequencies in order to ensure better frequency diversity gain. Furthermore, any two DFCA MA lists containing adjacent frequencies are required to be of the same length. This is necessary to be able to control the possible adjacent channel interference situations.

More examples of frequency band management can be found in Section 7.4.1.


12 (93)

DFCA MA lists must be of equal length in order forthe algorithm to be able to take inter MA list adjacent

channel interference into account

Split into MA lists

29 34 39 44

30 35 40 45

31 36 41 46

32 37 42 47

33 38 43 48

DFCA MA list 1

DFCA MA list 2

DFCA MA list 3

DFCA MA list 4

DFCA MA list 5

29 31 33 35 37 39 41 43 45 47

30 32 34 36 38 40 42 44 46 48

DFCA MA list 1

DFCA MA list 2

Example 1:

Example 2:

BCCH band (16) Regular band (12) DFCA band (20)

optional

Figure 2. Example frequency band split

2.6 Interference control principle

2.6.1 C/I estimation overview

The DFCA channel selection decisions are based on C/I criteria. The DFCA RRM algorithm examines all available radio channels and estimates the C/I the user would be exposed to for each of them. The C/I estimation is based on several inputs that are listed in Figure 3.


13 (93)

DFCA RRM

Background Interference Matrix

MS measurement reports

UL/DL PC power reductions

Radio channel usage information

Non-real time information

Near real time information

Real time information

Radio channel selection

Figure 3. Information used by DFCA

The MS measurement reports and the real time radio channel usage information are the key inputs in the C/I estimation. The MS measurement report includes the serving cell signal level and the signal levels of the strongest neighbouring cells. The potential downlink C/I impacting the user if interfering co-channel is used in one of the reported neighbouring cells can be calculated as follows (units in decibels):

CHneigbourBCservingTCH RXLEVreductionPWRRXLEVCIRDLDL

−+= _ (Eq. 1)

The PWR_reduction term corresponds to the current downlink serving channel power reduction.

This C/I is realized only if there is an ongoing connection in the neighbouring cell using the same radio channel. Therefore, after calculating the potential C/Is for incoming interference coming from the neighbouring cells the DFCA RRM algorithm will use the real time radio channel usage information to determine the actual C/I.

The Background Interference Matrix (BIM) is used to derive a C/I estimate for the nearby cells that have not been included in the MS measurement report. The BIM is covered with more detail in section 4.2.2.

The power control power level reduction information tells the current power level reduction that has been applied to the interfering connection. This information is used to make the C/I estimate more accurate as it is otherwise based on the full power assumption that would frequently lead to too pessimistic C/I estimates if power control is used. The power level information is updated every couple of seconds so that this information is not available in absolute real time. However, this small delay has not been found to degrade the performance in any significant way.


14 (93)

2.6.2 Interference control

The BSS synchronization leads to controlled TDMA frame alignment in all BTSs. Generally, the time slots are co-incident so that time slot 0 occurs simultaneously in all base stations although it is recommendable to shift the TDMA frame alignment by one time slot in order to reduce the interference load on SACCH channels as described in Section 7.2.1.

DFCA uses cyclic frequency hopping. The cyclic frequency hopping combined with BSS synchronization and TDMA frame number control mean that the frequency usage of every connection is orderly and repetitive allowing DFCA to have accurate interference control despite the usage of frequency hopping. In cyclic frequency hopping mode the MA-list frequencies are used sequentially from the lowest to the highest frequency changing to the next MA list frequency for every new TDMA frame. After the highest frequency the hopping sequence returns to the lowest one. If two connections (A and B) using the same time slot use the same MA-list and MAIO and the used TDMA frame numbers satisfy the following equation (eq. 2) then the two connections (A and B) will always use the same radio channel at the same time thus being potential co-channel interferers to each other. This is illustrated in Figure 4.

lengthMAFNlengthMAFN BA _mod_mod = (eq. 2)

time

freq

time

freq

Fixed interference

relationsConnection

Interferingconnection

time

freq

time

freq

Fixed interference

relationsConnection

Interferingconnection

Figure 4. Priciple of cyclic FH

Differences in the TDMA frame number cause offsets in the cyclic frequency hopping sequence. Offset in the frame number means similar offset (modulo MA-list length) in the MAIO of the interfering connection as illustrated in Figure 5 where the red arrows indicate a co-channel relation. Therefore,


15 (93)

co-channel interference condition exists between connections A and B if the same DFCA MA-list is used for both connections (A and B) and the following equation holds true:

lenghtMAMAIOFNlenghtMAMAIOFN BBAA _mod)(_mod)( +=+ (Eq 3.)

50 52 54 56

50 52 54 56

MA list 1 FN=n

MAIO 0 1 2 3

52 54 56 50

52 54 56 50

MAIO 0 1 2 3

MA list 1 FN=n

MA list 1 FN=n+1

MA list 1 FN=n+1

50 52 54 56

54 56 50 52

MA list 1FN=n

MAIO 0 1 2 3

52 54 56 50

56 50 52 54

MAIO 0 1 2 3

MA list 1FN=n+2

MA list 1FN=n+1

MA list 1FN=n+3

Cyclic FH without frame number offsets Cyclic FH with frame number offsets

50 52 54 56

50 52 54 56

MA list 1 FN=n

MAIO 0 1 2 3

52 54 56 50

52 54 56 50

MAIO 0 1 2 3

MA list 1 FN=n

MA list 1 FN=n+1

MA list 1 FN=n+1

50 52 54 56

54 56 50 52

MA list 1FN=n

MAIO 0 1 2 3

52 54 56 50

56 50 52 54

MAIO 0 1 2 3

MA list 1FN=n+2

MA list 1FN=n+1

MA list 1FN=n+3

Figure 5. Interference relations with FN offsets

The deterministic nature of the cyclic frequency hopping when combined with BSS synchronization and TDMA frame number control means that the interference relations closely resemble the situation in a non-hopping network. The radio channel used by each connection at any moment can be calculated based on the used MA-list, MAIO and TDMA frame number. If the connections use the same MA-list and the combination of used MAIOs and frame numbers is such that co-channel interference occurs, it then occurs constantly all the time for every single burst. The same applies in case of MA-lists containing adjacent frequencies in which case continuous adjacent channel interference may occur, provided that the MA-lists containing adjacent frequencies have equal length. The interference relations are therefore constant and predictable regardless of the utilization of cyclic frequency hopping so that the interference sources can be determined and they stay unchanged from burst to burst much the same way as in a non-hopping network. DFCA uses these properties of cyclic hopping together with BSS synchronization to predict and control the interference.

2.7 Supported frequency hopping modes

DFCA feature introduces a new hopping mode 'DFCA Hopping' that utilises cyclic RF hopping on timeslots of DFCA TRXs. Differently from conventional RF hopping, DFCA hopping allows TSL specific selection of MA and MAIO for each connection.

Other TRXs than DFCA TRXs of a cell, so called regular layer, can be configured for conventional RF hopping or left non-hopping.


16 (93)

In addition, in case of lost BSS synchronisation, DFCA TRXs apply 'normal' RF hopping over pre-defined 'un-synchronised MA list'.

2.8 SDCCH and (E)GPRS channels

The DFCA feature does not support SDCCH or (E)GPRS channels on DFCA TRXs. Therefore all permanent SDCCH time slots must be configured to the BCCH TRX or any regular TRX. The same also applies with the dynamic SDCCH feature so that the system will only create dynamic SDCCH channels on BCCH or regular TRXs. Similarly the (E)GPRS territory is only allowed on BCCH and regular TRXs.

2.9 Supported connection types

The DFCA channel selection is based on the C/I criteria. The aim is to maximize the C/I and at the same time adapt to the differences in the interference tolerance of different user types. In order to achieve this, a target C/I is defined to each of the user classes. The user classes and the default target C/I values are listed in Table 1. Note that HSCSD connections are allocated at regular layer only. The channel selection criteria is described with more detail in Section 4.4.

Table 1. DFCA user classes

User class Default C/I Target

EFR, FR, 9.6 kbps data 14 dB

HR 14 dB

AMR FR 8 dB

AMR HR 12 dB

CS Data (14.4) 16 dB

3. DFCA OPERATING MODES

3.1.1 DFCA hopping DFCA mode

DFCA frequency hopping is a new frequency hopping mode supported by UltraSite and MetroSite base stations with wide band combining from CX4.1 software release onwards.

DFCA hopping is based on the basic principle of synthesized frequency hopping where the TRX unit changes the used frequency according to the given hopping sequence. The difference when compared to the RF hopping mode is that with DFCA hopping the TRX supports independent cyclic hopping sequences for each time slot that can be freely selected with each channel activation whereas with RF hopping the used hopping sequence is the same for all time slots of a TRX and it cannot be changed dynamically without BTS blocking.


17 (93)

With DFCA hopping the BSC can freely select the MA-list, MAIO and TSC for each TCH activation allowing the DFCA algorithm to choose the most suitable radio channel for each new connection or handover based on C/I criteria. This full channel selection freedom allows DFCA to achieve the best performance with DFCA hopping mode.

The DFCA hopping mode is applied only in the TRXs dedicated to DFCA use (DFCA TRXs). When DFCA hopping mode is used, the BCCH TRX will always be in non-hopping mode and the other regular TRXs can use either cyclic or random RF hopping or operate on a fixed frequency without frequency hopping. Since the DFCA algorithm takes care of selecting the most suitable frequency hopping parameters for each connection, there is no need to plan frequencies or FH parameters for the DFCA TRXs as each time slot in a DFCA TRX can use any MAIO and any of the DFCA MA-lists. However, the BCCH TRX needs to be assigned a frequency and the regular TRXs (if they exist) need also to be assigned a frequency or the frequency hopping parameters. The general configuration of DFCA hopping BTS is presented in Figure 6.

The DFCA hopping mode should be used whenever possible to ensure optimum DFCA performance.

BCCH

Regular TRX(optional)

DFCA TRX

DFCA TRX

DFCA TRX

DFCA TRX

BTS

No FH or RF FH

No FH

TSL level MA list, MAIO, TSC settings. HSN=0

SDCCH

SDCCH SDCCH

BCCH GPRS territory

GPRS territory

BCCH

Regular TRX(optional)

DFCA TRX

DFCA TRX

DFCA TRX

DFCA TRX

BTS

No FH or RF FHNo FH or RF FH

No FHNo FH

TSL level MA list, MAIO, TSC settings. HSN=0TSL level MA list, MAIO, TSC settings. HSN=0TSL level MA list, MAIO, TSC settings. HSN=0

SDCCHSDCCH

SDCCH SDCCH

BCCH GPRS territory

GPRS territoryGPRS territory

Figure 6. Example of BTS configuration in DFCA hopping mode

3.1.2 Operation in case of a BSS synchronization failure

A BTS can fall out of BSS level synchronization if the LMU unit looses the GPS signal for so long that the clock drift causes the time slots to become sufficiently dealigned. The time slot dealignment can cause the DFCA interference control (that is based on BSS synchronization) not to function properly anymore so that continuation using normal DFCA operation could lead to significant performance degradation. When this happens the BTS is configured to unsynchronized DFCA operation.

When BTS enters unsynchronized DFCA operation the DFCA TRXs are re-configured to operate in random frequency hopping mode. This configuration change requires momentary blocking of the BTS


18 (93)

that is performed automatically. Prior to the blocking the BSC applies forced handover procedure to pre-empt the BTS from traffic in order to avoid dropped calls. During the unsynchronized DFCA operation the DFCA algorithm is suspended and the conventional channel allocation procedure is used also in the DFCA TRXs.

In the unsynchronized DFCA mode the DFCA TRXs are configured to use a special operator definable DFCA unsynchronized mode MA-list.

The HSN to be used for in the DFCA TRXs is defined automatically using the formula presented below:

( ) 163mod_ += frequencyBCCHHSN (Eq 4)

This guarantees that a different HSN will be used in any co-located BTS that may still be following the same clock. Also this takes care of the HSN re-use distance corresponding to the BCCH reuse distance.

If the original DFCA hopping mode is DFCA or RF hopping then the MAIOs for the DFCA TRXs in the unsynchronized DFCA mode are also assigned automatically so that the MAIO offset is 0 and the MAIO step is 2. This means that the unsynchronized mode MA-list must have at least double the number of frequencies than there are DFCA TRXs in the cell.

The usage of random frequency hopping on DFCA TRXs causes local DFCA performance degradation as the interference affecting the connections in the DFCA TRXs of an unsynchronized cell is completely randomized. Also, if the unsynchronized mode DFCA MA-list consists of DFCA frequency band frequencies then the unsynchronized BTS will cause an uncontrolled random interference component to the DFCA connections in the nearby cells somewhat degrading the DFCA performance in those surrounding cells.

When GPS signal is re-acquired the BSC automatically performs re-configuration actions and return back to normal synchronized DFCA operation. These re-configuration actions involve cell pre-emption and a BCF reset.

3.1.3 Operation in case of BSC-BSC connecion failure

In case the BSC-BSC connection becomes unavailable the DFCA cannot function properly over the BSC area borders. In this case the DFCA BTSs controlled by another BSC cause cause uncontrollable interference that cannot be accounted for. Therefore, the DFCA BTSs located close to the BSC area border cannot function in DFCA mode anymore and they are automatically configured to operate in unsynchronized DFCA mode with random hopping. The consequences of this are identical to those explained in Section 3.1.2.

4. DFCA CHANNEL SELECTION ALGORITHM

DFCA channel selection algorithm is responsible for obtaining the C/I estimations and performing channel allocation decisions based on those estimations. The flowchart of the channel selection process is shown in Figure 7. This chapter describes the operation of the DFCA algorithm in detail.


19 (93)

Incoming BIM

Outgoing BIM

For each DFCA BTS

For each BSC

DFCA RR table

DFCA adj.-channel lookup table

Calculate BIM scaling factor

Pick TSL, MA and MAIO

Incoming DL&UL CIR estimation + TX pwr setting

Outgoing DL&UL CIR estimation

Determine CIR difference for this TSL, MA and MAIO

All checked

Choose the most suitable

Above soft blocking threshold

Determine TSC

Block

No

Yes

No

Yes

MS measurement report

For each connection

INPUTS

Incoming BIM

Outgoing BIM

For each DFCA BTS

For each BSC

DFCA RR table


DFCA RR tableDFCA RR table



Calculate BIM scaling factor

Pick TSL, MA and MAIO

Incoming DL&UL CIR estimation + TX pwr setting

Outgoing DL&UL CIR estimation

Determine CIR difference for this TSL, MA and MAIO

All checked

Choose the most suitable

Above soft blocking threshold

Determine TSC

Block

No

Yes

No

Yes

MS measurement report

For each connection

INPUTS

Figure 7. DFCA channel selection process

4.1 Estimation of C/I

The C/I estimation relies on the fact that the interference relations in a DFCA network are stable and predictable due to BSS synchronization and controlled usage of cyclic frequency hopping. The C/I estimation is done by combining information from several sources such as the MS measurement reports, background interference matrix, real time DFCA channel usage information and near real time UL/DL TX power reductions. These are introduced with more detail in the sections below.

Several C/I estimations are produced for each radio channel. The incoming C/I describes the interference coming from existing connections that would affect the new connection for which the channel assignment is being done. The new connection can also potentially cause interference to existing connections using the same or adjacent radio channel. This is examined by determining the outgoing C/I for every potentially affected existing connection. Furthermore, both downlink and uplink C/I are estimated separately for both outgoing and incoming interference as illustrated in Figure 8.


20 (93)

BTS 1

BTS 6

New connection

Existing connection

Incoming DL

Outgoing DL

Incoming UL

Outgoing ULBTS 1

BTS 6

New connection

Existing connection

Incoming DL

Outgoing DL

Incoming UL

Outgoing UL

Figure 8. DFCA C/I estimations

4.2 Inputs

This section describes the inputs used by the DFCA channel selection process.

4.2.1 MS measurement reports

Each MS on a dedicated channel sends measurement reports to the BSC every 0.48s. The measurement report contains the measured serving cell signal level and the measured signal levels for the identified neighboring cells as well as current link quality related measures. The neighboring cells are identified by the BCCH frequency and the BSIC. Illustration of the measurement report is shown in Figure 9.

• Serving cell RXLEV• Neighbouring cell 1: BCCH freq, BSIC, RXLEV• Neighbouring cell 2: BCCH freq, BSIC, RXLEV• Neighbouring cell 3: BCCH freq, BSIC, RXLEV• Neighbouring cell 4: BCCH freq, BSIC, RXLEV• Neighbouring cell 5: BCCH freq, BSIC, RXLEV• Neighbouring cell 6: BCCH freq, BSIC, RXLEV

… • Neighbouring cell n: BCCH freq, BSIC, RXLEV• (+current link quality information)

• Serving cell RXLEV• Neighbouring cell 1: BCCH freq, BSIC, RXLEV• Neighbouring cell 2: BCCH freq, BSIC, RXLEV• Neighbouring cell 3: BCCH freq, BSIC, RXLEV• Neighbouring cell 4: BCCH freq, BSIC, RXLEV• Neighbouring cell 5: BCCH freq, BSIC, RXLEV• Neighbouring cell 6: BCCH freq, BSIC, RXLEV

… • Neighbouring cell n: BCCH freq, BSIC, RXLEV• (+current link quality information)

MS measurement reportMS measurement report

Figure 9. MS downlink measurement report


21 (93)

The potential DL C/I towards each of the neighboring cells can be determined by calculating the measured difference between the serving cell signal level and the neighboring cell signal level also accounting for the possible serving cell TX power reduction due to power control as shown in Equation 5.

gcellneigbourinlservingcelDL RXLEVreductionPWRRXLEVCIR −+= _ (Eq.5)

This C/I represents the worst case C/I that would be realized if there was a full power transmission on the same radio channel that is used in the serving cell.

4.2.2 Background Interference Matrix (BIM) tables

For each DFCA cell the system generates two BIM tables: the incoming interference BIM table and the outgoing interference BIM table. The incoming interference BIM table lists the potentially interfering surrounding cells. Correspondingly, the outgoing interference BIM table lists the potentially interfered surrounding cells. The BIM tables are automatically created and updated by the BSC. The BIM table creation and update process is described in Section 5.

Example of incoming interference BIM table is shown in Table 2. Each interfering neighboring cell is identified by its BCCH frequency & BSIC and the level of interference is reflected in the C/I value listed for each interfering cell. The C/I value is defined as the C/I that is exceeded γ percent of the time based on long term MS measurements statistics gathered from all connections in the cell. The γ is a BSC level parameter “BIM confidence probability” with a default setting of 90%.

If the interfering cell is controlled by the same BSC then the segment id is in the range of 1- 660 (S11: 1-248). A higher segment id implies that the interfering cell is controlled by another BSC. The band offset field indicates the correction factor used for the C/I in the case of a remote neighbor controlled by another BSC when the interfering cell BCCH is on a frequency band that is not used at all in the interfered cell but there is also a secondary band in the interfering cell. The band offsets corresponds to the mean value of the nonBCCHLayerOffset parameters of the non-BCCH band BTSs. The missed update and status fields are explained in Section 5.

Table 2. Example of incoming interference BIM table

BCCH BSIC C/I (dB)

SEG Id

Missed update

Band offset

Status

102 54 10 12 0 0 ok 115 22 18 661 255 2 ok --- --- --- --- --- --- ---

The outgoing interference BIM table has simpler structure including only cell identification fields, the C/I describing the level of interference and information of the forced HR mode requests as seen in Table 3.

Table 3. Example of outgoing interference BIM table

BCCH BSIC C/I SEG Forced HR req* 74 30 13 dB 23 no


22 (93)

34 49 17 dB 667 AMR HR --- --- --- --- ---

* Forced HR req possible values: 'no', 'HR', 'AMR HR' and 'HR & AMR HR', see Section 4.6.2.

Note: In BSS11 the segment identification range of own BSC is 1-248.

The BIM tables can be displayed my MML command. (BIM display by MML is not supported in BSS11).

4.2.3 DFCA Radio Resource Table

The DFCA radio resource table contains the real time DFCA channel usage situation in each DFCA cell as illustrated in Table 4. Each ongoing connection on a DFCA TRX is listed in this table together with several parameters that are listed below:

• Time slot

• DFCA MA list & MAIO

• Sub channel (in case of HR)

• Training sequence code

• C/I target

• Soft blocking C/I limit

• UL power reduction (by power control)*

• DL power reduction (by power control) *

• Serving cell UL & DL RXLEVs*

• BCCH frequency, BSIC & C/I* for up to 20 neighboring cells (based on measurement reports received from this mobile)

• BIM scaling factor The entries marked with * are updated periodically during the connection.

The DFCA radio resource table is used internally in the C/I estimation process and it is not visible to the user.


23 (93)

Table 4. Example of DFCA radio resource table

TSC xtarget C/I

soft block C/IUL pwr reductionDL pwr reduction

Serv cell ul/dl RXLEV 11 BCCH BSIC C/I2 BCCH BSIC C/I3 BCCH BSIC C/I

…n BCCH BSIC C/IBIM scaling factor

HR 0TSC x

target C/Isoft block C/I

UL pwr reductionDL pwr reduction

Serv cell ul/dl RXLEV 11 BCCH BSIC C/I2 BCCH BSIC C/I3 BCCH BSIC C/I


HR 0 HR 1

0 1 2 3 4 5 6 7TSL

MA list 1, MAIO 0

MA list 1, MAIO 1

MA list 1, MAIO 2

MA list 2, MAIO 0

MA list n, MAIO n

…..

DFCA MA list, MAIO

(One table for eachDFCA BTS)

TSC xtarget C/I

soft block C/IUL pwr reductionDL pwr reduction

Serv cell ul/dl RXLEV1 BCCH BSIC C/I2 BCCH BSIC C/I3 BCCH BSIC C/I


HR 1

4.2.4 DFCA adjacent channel lookup table

Identification of co-channel situation is straight forward as one DFCA band frequency can only be used in one DFCA MA list. Therefore, a co-channel situation only occurs if two connections use the same MA list and MAIO in the same time slot. However, adjacent channel situation is bit more complicated to detect.

Adjacent frequencies may occur within a DFCA MA list and also between two different DFCA MA lists. During the creation or deletion of DFCA MA lists the BSC will automatically update an adjacent channel lookup table that indicates which DFCA MA-list & MAIO combinations can cause adjacent channel interference to each other. This table is used internally in the C/I estimation process and it is not visible to the user.

4.3 DFCA C/I estimation process

This section describes the DFCA channel selection process step by step. The flowchart of this process is shown in Figure 7.


24 (93)

4.3.1 STEP 1: BIM scaling

The C/I values in BIM tables are based on long-term statistics of all connections in the cell. This statistical nature means that the real C/I of one user towards an interfering cell may be significantly different depending on user’s actual location. For example, if the user is located close to the serving base station the high serving signal level usually implies that the C/Is tend to be much higher than the statistical C/I values provided in the BIM tables.

BIM scaling is used to correct the statistical C/I values of BIM tables to better correspond the user’s actual situation. The BIM scaling is based on comparing the measured C/I towards the neighboring cells included in the latest measurement report to the corresponding C/I values listed in the incoming interference BIM table. From this comparison the differences between the measured and the statistical C/Is can be determined. Finally, the BIM scaling factor corresponding to the average difference is calculated. This BIM scaling factor is then applied to all C/I values taken from the BIM table for C/I estimations for this connection.

BCCH, BSIC

MeasuredC/I

BIM C/I ∆ C/I

102, 23 8 dB 4 dB 4 dB

114, 52 12 dB 7 dB 5 dB

104, 12 12 dB 9 dB 3 dB

116, 32 15 dB 11 dB 4 dB

108, 12 19 dB 15 dB 4 dB

111, 43 20 dB 15 dB 5 dB

Average ∆ C/I (BIM scaling factor): 4 dB

Figure 10. Example of the BIM scaling factor calculation

4.3.2 STEP2: Incoming DL C/I estimation

In this step the interfering co- and adjacent channel connections in the neighboring cells are identified and the level of interference coming from each of these connections is estimated. The algorithm goes through the cells listed in the incoming interference BIM table and checks the DFCA RR table for ongoing co- and adjacent channel connections in these cells. When an interfering connection is found then the MS measurement report is examined to see if a directly measured C/I value can be extracted from the measurement report. If the interfering cell is not reported in the measurement report then the C/I value from the incoming interference BIM table is used and scaled by the BIM scaling factor that was determined before. Finally, the measured or the statistical C/I is adjusted by the actual DL tx power reduction of the interfering connection that is available in the DFCA RR table. The most significant DL incoming interference source is identified and if possible the initial DL tx power level is reduced so that the final C/I corresponds to the C/I target set to this connection type. The minimum C/I difference corresponding to the difference between the C/I target and the C/I caused by the dominant interference source is recorded. Typically, the initial DL tx power level setting makes this min C/I difference to be zero, but in high load situations the min C/I difference can even be negative indicating that the C/I target was not achieved.


25 (93)

Example of this process is presented in Figure 11.

Channel being checked: TSL 1, MA 2, MAIO 1

MS Measurement report

C/I target: 14 dB

Incoming BIM (100, 01)BCCH, BSIC C/I

102, 23 4 dB114, 52 7 dB104, 12 9 dB115, 30 10 dB101, 52 11 dB116, 32 11 dB103, 06 14 dB

111, 43 15 dB108, 12 15 dB

102, 36 18 dB

102, 23 4 dB114, 52 7 dB104, 12 9 dB115, 30 10 dB101, 52 11 dB116, 32 11 dB103, 06 14 dB

111, 43 15 dB108, 12 15 dB

102, 36 18 dB

Adj

-

-

CoCo

Co

Adj

Co

-

-

102, 23 6 dB104, 12 8 dB116, 32 10 dB103, 06 0 dB108, 12 0 dB102, 36 4 dB

BCCH, BSIC DL pwr reduction

DL tx power reductions from DFCA RR table

1. Determine if co- or adjacent channel is being used in the cells listed in incoming BIM by using the DFCA tables

2. Calculate the real C/I caused by each of the ongoing interfering connections.•For the directly measured neighbouring cells the measured C/I is used

•For the unmeasured ones the incoming BIM value is used scaled with the BIM scaling factor

•In case of adjacent channel interference +18 dB offset is used•The power reductions in the interfering connections are taken into account when the real C/I is determined

DFCA assignment in cell: 100, 01

14 dB

2 dB

7 dB0 dB

19 dB

8 dB

C/I diff after PC

4. The C/I difference after power reduction is calculated and the lowest C/I difference is stored

3. The most restrictive C/I is identified (18 dB in this case)The initial DL power reduction is determined so that as a result the connection will be operating in it's C/I target

DFCA RR table


102, 23 8 dBBCCH, BSIC C/I

114, 52 12 dB104, 12 12 dB116, 32 15 dB108, 12 19 dB111, 43 20 dB

Incom. DL 0Incom. UL Outgo. DL

Outgo. UL MIN

Incom. DL 0Incom. UL Outgo. DL

Outgo. UL MIN

MIN C/I diff

32 dB

20 dB

25 dB18 dB37 dB

26 dB

C/I real

32 dB

20 dB

25 dB18 dB37 dB

26 dB

C/I real

14 dB

14 dB

14 dB14 dB14 dB

14 dB

C/I target

18 dB - 14 dB = 4 dB

18 +4 +4 = 26

DL pwr reduction

BIM scaling factor+4 dB

Figure 11. Example of incoming DL C/I estimation and initial DL tx power setting

4.3.3 STEP 3: Incoming UL C/I estimation

The UL C/I estimation cannot be done as accurately as the DL C/I estimation as only DL measurements of the neighboring ceIl signal levels are available. As a consequence, the UL C/I estimation is based purely on the statistical C/I values available in the BIM tables. Although the BIM tables are also based on the DL measurements, a reasonable estimation of the UL interference situation can still be derived based on the fact that the propagation path of the incoming UL interference is the same as the propagation path of the outgoing DL interference as illustrated in Figure 12. Therefore, in order to estimate the incoming UL C/I, the outgoing interference BIM is used.


26 (93)

BTS 1

BTS 6

New connection

Existing connection

Incoming DL

Outgoing DL

Incoming UL

Outgoing UL

Same propagation path – opposite

direction

BTS 1

BTS 6

New connection

Existing connection

Incoming DL

Outgoing DL

Incoming UL

Outgoing UL

Same propagation path – opposite

direction

Figure 12. UL & DL interference propagation paths

The estimate can be improved by accounting for the new connection’s relative location in the cell that is reflected in the incoming interference BIM scaling factor. If the new user is located close to the cell site the BIM scaling factor will be a high positive number reflecting the fact that the serving signal path loss is low. If the new user is located at the cell edge then the BIM scaling factor may even be negative reflecting high serving signal path loss.

The final UL incoming C/I estimate is derived by taking the C/I values from the outgoing interference BIM and scaling them using the incoming interference BIM scaling factor that was determined before. Finally each UL interference contribution is adjusted by accounting for the possible UL tx power reductions that are available in the DFCA RR table. The most significant UL incoming interference source is then identified and if possible the initial UL tx power level for the new connection is reduced so that the final C/I corresponds to the C/I target set to this connection type. The minimum C/I difference corresponding to the difference between the C/I target and the C/I caused by the dominant interference source is recorded. Typically, the initial UL tx power level setting makes this min C/I difference to be zero, but in high load situations the min C/I difference can even be negative indicating that the C/I target was not achieved.


27 (93)

DFCA RR table


Channel being checked: TSL 1, MA 2, MAIO 1C/I target: 14 dB DFCA assignment in cell: 100, 01

DL Measurement report


114, 52 12 dB104, 12 12 dB116, 32 15 dB108, 12 19 dB111, 43 20 dB


1. Determine if co- or adjacent channel is being used in the cells listed in outgoing interference BIM by using the DFCA tables

BCCH, BSIC

Information from DFCA RR table

UL pwrreduction

Measured C/I for cell 100,01

102, 23 10 dB 11 dB103, 06 0 dB 19 dB105, 59 10 dB not measured

104, 12 2 dB not measured116, 32 0 dB 18 dB


Outgoing BIM (100, 01)BCCH, BSIC C/I


102, 23 8 dB103, 06 9 dB110, 02 10 dB105, 59 12 dB116, 32 13 dB104, 12 15 dB

108, 12 18 dB111, 43 17 dB

102, 36 20 dB

-

-

CoCo

Co

-

Co

Adj

Adj

Adj

2. Calculate the real C/I caused by each of the ongoing interfering connections.•The incoming UL C/I estimations are always based on the statistical C/I values presented in the outgoing interference BIM table and scaled with the BIM scaling factor. •In case of adjacent channel interference +18 dB offset is used•The power reductions in the interfering connections are taken into account when the real C/I is determined

14 dB

14 dB14 dB

14 dB

C/I target

14 dB

14 dB

14 dB

Incom. DL 0Incom. UL - 1Outgo. DL

Outgo. UL MIN

MIN C/I diff

13 dB

17 dB21 dB

34 dB

C/I real

40 dB

44 dB

40 dB

20 +4 +10 = 343. The most restrictive C/I is identified (13 dB in this case)The initial UL power reduction is determined so that as a result the connection will be operating in it's C/I target * No reduction possible

as C/I is below target

13 dB - 14 dB = 0 dB*

-1 dB

3 dB7 dB

20 dB

C/I di ff after PC

26 dB

30 dB

26 dB


UL pwr reduction

DFCA RR table





Channel being checked: TSL 1, MA 2, MAIO 1C/I target: 14 dB DFCA assignment in cell: 100, 01



114, 52 12 dB104, 12 12 dB116, 32 15 dB108, 12 19 dB111, 43 20 dB




BCCH, BSIC


UL pwrreduction







102, 23 8 dB103, 06 9 dB110, 02 10 dB105, 59 12 dB116, 32 13 dB104, 12 15 dB

108, 12 18 dB111, 43 17 dB

102, 36 20 dB

-

-

CoCo

Co

-

Co

Adj

Adj

Adj


BCCH, BSIC


UL pwrreduction







102, 23 8 dB103, 06 9 dB110, 02 10 dB105, 59 12 dB116, 32 13 dB104, 12 15 dB

108, 12 18 dB111, 43 17 dB

102, 36 20 dB


102, 23 8 dB103, 06 9 dB110, 02 10 dB105, 59 12 dB116, 32 13 dB104, 12 15 dB

108, 12 18 dB111, 43 17 dB

102, 36 20 dB

-

-

CoCo

Co

-

Co

Adj

Adj

Adj

2. Calculate the real C/I caused by each of the ongoing interfering connections.•The incoming UL C/I estimations are always based on the statistical C/I values presented in the outgoing interference BIM table and scaled with the BIM scaling factor. •In case of adjacent channel interference +18 dB offset is used•The power reductions in the interfering connections are taken into account when the real C/I is determined

14 dB

14 dB14 dB

14 dB

C/I target

14 dB

14 dB

14 dB

14 dB

14 dB14 dB

14 dB

C/I target

14 dB

14 dB

14 dB


Outgo. UL MIN

MIN C/I diff


Outgo. UL MIN


Outgo. UL MIN

MIN C/I diff

13 dB

17 dB21 dB

34 dB

C/I real

40 dB

44 dB

40 dB

13 dB

17 dB21 dB

34 dB

C/I real

40 dB

44 dB

40 dB

20 +4 +10 = 343. The most restrictive C/I is identified (13 dB in this case)The initial UL power reduction is determined so that as a result the connection will be operating in it's C/I target * No reduction possible

as C/I is below target

13 dB - 14 dB = 0 dB*

-1 dB

3 dB7 dB

20 dB

C/I di ff after PC

26 dB

30 dB

26 dB


UL pwr reduction

Figure 13. Example of incoming UL C/I estimation and initial UL tx power setting

4.3.4 STEP 4: Outgoing DL C/I estimation

In this step the co- and adjacent channel connections in the neighboring cells that may get DL interference from the new connection (=outgoing interference) are identified and the level of interference affecting each of these connections is estimated. The algorithm goes through the cells listed in the outgoing interference BIM table and checks the DFCA RR table for ongoing co- and adjacent channel connections in these cells. When an interfered connection is found then the DFCA RR table is examined to see if a directly measured C/I value is available. If the directly measured C/I is not available then the C/I value from the outgoing interference BIM table is used and scaled by the BIM scaling factor of the interfered connection that is listed in the DFCA RR table. Finally, the measured or the statistical C/I is adjusted by the actual DL tx power reduction of the new connection that was determined in step 2. The mostly affected DL outgoing interference victim is identified by calculating the minimum C/I difference for each interfered connection corresponding to the difference between the C/I target of each connection (that is available in the DFCA RR table) and the C/I caused by the new connection. The minimum C/I difference is then recorded.


28 (93)

Incom. DL 0Incom. UL - 1Outgo. DL - 2

Outgo. UL MIN

MIN C/I di ff

DFCA RR table





114, 52 12 dB104, 12 12 dB116, 32 15 dB108, 12 19 dB111, 43 20 dB


BCCH, BSIC


DL pwrreduction

Measured C/ I for cell 100,01

102, 23 6 dB +3 11 dB 14 dB103, 06 0 dB +6 15 dB 20 dB105, 59 8 dB +10 not measured 16 dB

104, 12 8 dB +7 not measured 14 dB116, 32 10 dB -2 18 dB 14 dB

102, 36 4 dB +3 not measured 14 dB108, 12 0 dB +4 20 dB 16 dB

BIM scaling factor C/I target

Outgoing BIM (100, 01)


102, 23 8 dB103, 06 9 dB110, 02 10 dB105, 59 12 dB116, 32 13 dB104, 12 15 dB

108, 12 18 dB111, 43 17 dB

102, 36 20 dB

-

-

CoCo

Co

-

Co

Adj

Adj

Adj

1. Interfered connections and interference type as determined in incoming UL interference check

DL pwr reduction = 4 dB** From Incoming DL check

2. Calculate the real C/I caused to each of the interfered ongoing connections.•If the C/I caused by cell 100,01 can be determined from measurement reports in DFCA RR table then the measured C/I is used

•For the unmeasured ones the outgoing BIM value is used scaled with the BIM scaling factor of the interfered connection

•In case of adjacent channel interference +18 dB offset is used

•The power reductions in the interfering connections as well as in the DL of the serving connection are taken into account when the real C/I is determined

19 dB

12 dB18 dB

23 dB

C/I real

27 dB

36 dB

42 dB

20 +3 +4 - 4 = 23

3. The calculated C/I is then compared to the C/I target and a C/I difference is calculated. The lowest C/I difference is stored

20 dB

14 dB14 dB

14 dB

C/I target

14 dB

16 dB

16 dB

-1 dB

-2 dB4 dB

9 dB

C/I diff

13 dB

20 dB

26 dB


Outgo. UL MIN

MIN C/I di ff


Outgo. UL MIN


Outgo. UL MIN

MIN C/I di ff







114, 52 12 dB104, 12 12 dB116, 32 15 dB108, 12 19 dB111, 43 20 dB


BCCH, BSIC


DL pwrreduction








102, 23 8 dB103, 06 9 dB110, 02 10 dB105, 59 12 dB116, 32 13 dB104, 12 15 dB

108, 12 18 dB111, 43 17 dB

102, 36 20 dB

-

-

CoCo

Co

-

Co

Adj

Adj

Adj


BCCH, BSIC


DL pwrreduction











102, 23 8 dB103, 06 9 dB110, 02 10 dB105, 59 12 dB116, 32 13 dB104, 12 15 dB

108, 12 18 dB111, 43 17 dB

102, 36 20 dB

-

-

CoCo

Co

-

Co

Adj

Adj

Adj




102, 23 8 dB103, 06 9 dB110, 02 10 dB105, 59 12 dB116, 32 13 dB104, 12 15 dB

108, 12 18 dB111, 43 17 dB

102, 36 20 dB

-

-

CoCo

Co

-

Co

Adj

Adj

Adj



102, 23 8 dB103, 06 9 dB110, 02 10 dB105, 59 12 dB116, 32 13 dB104, 12 15 dB

108, 12 18 dB111, 43 17 dB

102, 36 20 dB


102, 23 8 dB103, 06 9 dB110, 02 10 dB105, 59 12 dB116, 32 13 dB104, 12 15 dB

108, 12 18 dB111, 43 17 dB

102, 36 20 dB

-

-

CoCo

Co

-

Co

Adj

Adj

Adj


DL pwr reduction = 4 dB** From Incoming DL check

2. Calculate the real C/I caused to each of the interfered ongoing connections.•If the C/I caused by cell 100,01 can be determined from measurement reports in DFCA RR table then the measured C/I is used

•For the unmeasured ones the outgoing BIM value is used scaled with the BIM scaling factor of the interfered connection


•The power reductions in the interfering connections as well as in the DL of the serving connection are taken into account when the real C/I is determined

19 dB

12 dB18 dB

23 dB

C/I real

27 dB

36 dB

42 dB

19 dB

12 dB18 dB

23 dB

C/I real

27 dB

36 dB

42 dB

19 dB

12 dB18 dB

23 dB

C/I real

27 dB

36 dB

42 dB

20 +3 +4 - 4 = 23


20 dB

14 dB14 dB

14 dB

C/I target

14 dB

16 dB

16 dB

-1 dB

-2 dB4 dB

9 dB

C/I diff

13 dB

20 dB

26 dB


20 dB

14 dB14 dB

14 dB

C/I target

14 dB

16 dB

16 dB

-1 dB

-2 dB4 dB

9 dB

C/I diff

13 dB

20 dB

26 dB

Figure 14. Example of outgoing DL C/I estimation

4.3.5 STEP 5: Outgoing UL C/I estimation

In this step the UL interference that would be caused by the new connection to the ongoing connections in nearby cells is estimated. The outgoing UL C/I estimation exploits the fact that the interference propagation path of the outgoing UL interference is the same as the path of the incoming DL interference as illustrated in Figure 12. Therefore, the incoming interference BIM table is used.

The UL outgoing C/I estimate is derived by taking the C/I values from the incoming interference BIM and scaling them using the individual BIM scaling factor of each of the interfered connection as listed in the DFCA RR table. Finally the UL interference estimate for each of the potentially interfered connection is adjusted by accounting for the possible UL tx power reduction for the new connection that was determined in step 3. The mostly affected UL outgoing interference victim is identified by calculating the minimum C/I difference for each interfered connection corresponding to the difference between the C/I target of each connection (that is available in the DFCA RR table) and the C/I caused by the new connection. The minimum C/I difference is then recorded.


29 (93)


Outgo. UL - 5MIN - 5

MIN C/I di ff

DFCA RR table





114, 52 12 dB104, 12 12 dB116, 32 15 dB108, 12 19 dB111, 43 20 dB

UL pwr reduction = 0 dB*

2. Calculate the real C/I caused by each of the ongoing interfering connections.•The statistical C/I from the incoming BIM table is used and scaled with the connection specific BIM scaling factor that is listed in DFCA RR table


•The UL power reductions in the interfering connections as well as in the UL of the serving connection are taken into account when the real C/I is determined


* From incoming UL check

Adj

-

-

CoCo

Co

Adj

Co

-

-


102, 23 10 dB +3 14 dB104, 12 2 dB +7 14 dB116, 32 0 dB -2 14 dB103, 06 0 dB +6 20 dB108, 12 4 dB +4 16 dB102, 36 10 dB +3 14 dB

BCCH, BSIC


1. Interfered connections and interference type as determined in incoming DL interference check

UL pwrreduction


BCCH, BSIC C/I

102, 23 4 dB114, 52 7 dB104, 12 9 dB115, 30 10 dB101, 52 11 dB116, 32 11 dB103, 06 14 dB

111, 43 15 dB108, 12 15 dB

102, 36 18 dB

14 dB

14 dB

14 dB20 dB16 dB

14 dB

C/I target

15 dB

14 dB

9 dB20 dB41 dB

11 dB

C/I real

18 +3 +0 - 10 = 11

C/I diff

3. The calculated C/I is then compared to the C/I target and a C/I difference is calculated. The lowest C/I difference measures the quality of the evaluated DFCA assignment

1 dB

0 dB

-5 dB0 dB

25 dB

-3 dB

C/I difference for TSL 1, MA 2,

MAIO 1



MIN C/I di ff





MIN C/I di ff







114, 52 12 dB104, 12 12 dB116, 32 15 dB108, 12 19 dB111, 43 20 dB



114, 52 12 dB104, 12 12 dB116, 32 15 dB108, 12 19 dB111, 43 20 dB

UL pwr reduction = 0 dB*UL pwr reduction = 0 dB*

2. Calculate the real C/I caused by each of the ongoing interfering connections.•The statistical C/I from the incoming BIM table is used and scaled with the connection specific BIM scaling factor that is listed in DFCA RR table


•The UL power reductions in the interfering connections as well as in the UL of the serving connection are taken into account when the real C/I is determined


* From incoming UL check

Adj

-

-

CoCo

Co

Adj

Co

-

-



BCCH, BSIC



UL pwrreduction


BCCH, BSIC C/I

102, 23 4 dB114, 52 7 dB104, 12 9 dB115, 30 10 dB101, 52 11 dB116, 32 11 dB103, 06 14 dB

111, 43 15 dB108, 12 15 dB

102, 36 18 dB

Adj

-

-

CoCo

Co

Adj

Co

-

-

Adj

-

-

CoCo

Co

Adj

Co

-

-




BCCH, BSIC



UL pwrreduction


BCCH, BSIC C/I

102, 23 4 dB114, 52 7 dB104, 12 9 dB115, 30 10 dB101, 52 11 dB116, 32 11 dB103, 06 14 dB

111, 43 15 dB108, 12 15 dB

102, 36 18 dB

BCCH, BSIC C/I

102, 23 4 dB114, 52 7 dB104, 12 9 dB115, 30 10 dB101, 52 11 dB116, 32 11 dB103, 06 14 dB

111, 43 15 dB108, 12 15 dB

102, 36 18 dB

102, 23 4 dB114, 52 7 dB104, 12 9 dB115, 30 10 dB101, 52 11 dB116, 32 11 dB103, 06 14 dB

111, 43 15 dB108, 12 15 dB

102, 36 18 dB

14 dB

14 dB

14 dB20 dB16 dB

14 dB

C/I target

14 dB

14 dB

14 dB20 dB16 dB

14 dB

C/I target

15 dB

14 dB

9 dB20 dB41 dB

11 dB

C/I real

15 dB

14 dB

9 dB20 dB41 dB

11 dB

C/I real

18 +3 +0 - 10 = 11

C/I diff

3. The calculated C/I is then compared to the C/I target and a C/I difference is calculated. The lowest C/I difference measures the quality of the evaluated DFCA assignment

1 dB

0 dB

-5 dB0 dB

25 dB

-3 dB

C/I difference for TSL 1, MA 2,

MAIO 1

Figure 15. Example of outgoing UL C/I estimation

4.3.6 HR specific considerations in C/I estimations

Some special considerations have to be made if HR is used. These are presented below:

HR – HR

When a HR connection is interfering another HR connection the DFCA interference estimations and the channel search will function the normal way. In such a case the DFCA resource table will indicate the HR sub channel making it possible to distinguish between two HR connections sharing the same TSL.

HR – FR

When the new channel assignment is done for a HR connection and the interfering/interfered connection is using FR, the DFCA interference estimations and the channel search will function the normal way for the incoming interference. For outgoing interference some special considerations must be made: Since the new connection will be using HR, the resulting interference to any FR connection will be present in every other burst. This means that the interference impact is reduced. This is taken into account by adding 3dB to the estimated outgoing interference C/I when the interfered connection is FR.


30 (93)

FR – HR

When the new channel assignment is done for a FR connection and the interfering/interfered connection is using HR the interference will be split into two parts each affecting every other burst. In this case the channel search algorithm will calculate the C/I values separately for both of the interfering HR sub-channels. The resulting C/Is coming from the HR connections will be increased by 3dB. For outgoing interference estimations when FR is interfering a HR connection, the estimated C/I value is used as it is without any offsets.

4.4 Channel ranking and selection

In the channel search process the DFCA algorithm determines the interference situation that the new user would have in different available time slot, DFCA MA list and MAIO combinations. Also, the interference caused to the existing connections in the surrounding cells is also determined for the different time slot, DFCA MA list and MAIO combinations. For each combination the most restrictive min C/I difference is determined that is the lowest min C/I difference from the DL/UL incoming interference and DL/UL outgoing interference C/I estimations.

Basically DFCA channels can be divided into the two groups according to the minimum C/I of the channel: The channels on and above the target (min C/I difference >= 0) and below the target (min C/I difference < 0). Channels on target or above target can be both treated as good channels. In a handover there is also third group, the channels below the soft-blocking limit, channels in this group can be allocated for a handover but only if there is no available channels in any above group.

Example of DFCA channel ranking is shown in Table 5.

There are two DFCA channel ranking methods (0 and 1). The method 0 is the primary method providing the best performance. The method 1 can be used to reduce the BSC processing load and internally BSC will change to method 1 if the processing load with mode 0 reaches critical level. The DFCA channel ranking methods are explained with more detail in the sections below.


31 (93)

Table 5. DFCA channel ranking

TSL MA MAIO Min C/I Difference

Reduced power (UL+DL)

Preference order with DFCA alloc method 0

Preference order with DFCA alloc method 1

1 1 0 Idle = 63 10 1 1*

1 1 4 Idle = 63 10 1 1*

1 2 20 20 10 3 1*

2 5 11 0 8 5 1*

2 8 8 0 3 6 1*

3 5 7 0 10 4 1*

3 7 1 0 2 7 1*

4 5 30 -2 4 Soft blocked Soft blocked



* With method 1 the first channel in the channel search that equals or exceeds the C/I target level (C/I difference >= 0) is selected.

4.4.1 Channel search with the DFCA allocation method 0 (primary method)

DFCA C/I group order is above target, on target and below target. With this method the channel with the highest minimum C/I difference is the most preferred. So this method aims always to maximize the C/I. In order to find the channel with the highest C/I the min C/I difference has to be determined for all the available channels first. Therefore this method is computationally intensive. However, the channel search is stopped if a totally idle channel is found. The totally idle channel means a MA, MAIO and time slot combination, for which a co- or adjacent channel is not allocated in any of the cells found in the incoming or outgoing interference BIM tables. In that case it is unnecessary to continue the search because it is not possible to find better channel than totally idle channel.

With this method the power reduction is also taken into account when the MA, MAIO and time slot combination is selected. The sum of uplink and downlink power reductions defines the priority on the target C/I level. The MA, MAIO and time slot combination is on the target level if the C/I difference is zero. It can be assumed that the less power needed to achieve the C/I target level the better the channel causing less interference to the network. On the target C/I level the channel with the highest power reduction sum is selected.


32 (93)

This method is expected to yield the best network performance and it is assumed in all DFCA performance simulations.

4.4.2 Channel search with the DFCA allocation method 1 (load saving method)

DFCA C/I group order is on target, above target and below target. With this method the searching of the MA list can stopped when a totally idle channel or a channel on the target C/I level is found. This method is therefore computationally less intensive at least when network load is low.

The primary target on this group is a channel nearest the target level and so the channel with the lowest C/I difference. If there are not available any channel on target level or above target then the channel between the target level and soft blocking limit are checked.

4.4.3 Rotation in channel allocation

With both DFCA allocation methods the round robin method is used to circulate MAs, MAIOs and time slots so that the searching is always started from the MA, MAIO and time slot following the one which was allocated the previous time. This means that after allocation the allocated MA, MAIO and time slot are saved to BTS level and when a channel is next time searched all these indexes are increased by one to indicate the starting point. If all channels show equal quality, this method allows allocating of all channels with successive calls one by one using one phone (e.g. testing situations). Also MA, MAIO and time slot combination spreading is better, which helps examination of the DFCA assignment measurement.

4.4.4 Overload protection

When MCMU overload situation is detected the DFCA channel search is modified irrespective of the selected DFCA allocation method. In overload situation the channel search is stopped right away when a channel that indicates a positive or zero C/I difference is found.

4.4.5 BTS selection within a segment

In the segment environment the searching rules order is as following:

1. BTS load level according to the parameter BTSloadInSEG.

2. Channel type: In full rate channel allocation ranking order the permanent full rate channels are preferred to the dual rate channels according to minimum C/I difference as follows: 1. Full rate channel with idle MA and MAIO combination 2. Full rate channel on C/I diff >= 0 level 3. Dual rate channel with idle MA and MAIO combination 4. Dual rate channel on C/I diff >= 0 level 5. Full rate channel on C/I diff < 0 level 6. Dual rate channel on C/I diff < 0 level In half rate channel allocation ranking order the permanent half rate channels are preferred to the dual rate channels according to minimum C/I difference as follows:


33 (93)

1. Permanent half rate channel with idle MA and MAIO combination 2. Permanent half rate channel on C/I diff >= 0 level 3. Half occupied dual rate channel with idle MA and MAIO combination 4. Half occupied dual rate channel on C/I diff >= 0 level 5. Idle dual rate channel with idle MA and MAIO combination 6. Idle dual rate channel on C/I diff >= 0 level 7. Permanent half rate channel on C/I diff < 0 level 8. Half occupied dual rate channel on C/I diff < 0 level 9. Idle dual rate channel on C/I diff < 0 level

3. Actual minimum C/I difference: If the load of the BTS is below the BSC-level parameter load rate for channel search or the BTS-level parameter cell load for channel search the time slot is selected purely according to the selected DFCA allocation method (method 0 or 1).

4. And finally the total load of the BTS is taken into account.

4.4.6 Soft Blocking C/I

In addition to the C/I target, each connection type can be given a soft blocking C/I threshold. If in the C/I estimation phase any of the four C/I estimates produced for each radio channel candidate does not exceed the soft blocking limit of the interfered connection, then the radio channel candidate is deemed soft blocked.

If there are no acceptable assignment candidates (i.e. all the candidates breach the connection type specific soft blocking C/I limit), the assignment will be directed to a regular TRX if available. If there are no free TSLs on any regular TRXs of the cell, the call will be finally rejected (DFCA soft blocking situation). Directed retry could still be used and if successfull the directed retry will move the call to another cell in which case the call is not rejected.

In case of handovers soft blocking is not applied except if the handover is for DR or DADLB then the soft blocking is applied as in normal call setup.

4.4.7 C/N check for cell access control

The C/N check allows more flexible signal level based cell access control. The traditional cell access minimum signal level check cannot take into account the different properties of different connection types. For example, an AMR connection can tolerate several decibels lower signal levels than an EFR connection due to more robust channel coding. By using the connection type specific C/N soft blocking limits with C/N check these differences are taken into account. C/N check also ensures that the usage of half rate channels is not allowed to degrade the connection quality of users in low field strength situation.

In the beginning of the DFCA TCH channel assignment procedure the radio link signal level situation is checked by the BSC. The signal level check can lead to exclusion of some channel modes if the signal level is found to be inadequate for usage of these channel modes. In extreme case none of the available channel modes is acceptable leading to the soft blocking of the connection. The blocking is not applied in case of a handover.


34 (93)

The procedure is as follows:

The reported UL and DL RXLEVs are compared to the UL and DL noise levels as defined by operator. These parameters are listed in the parameter section. The C/N ratio is calculated as follows:

Call setups and intra cell handovers

The nonBCCHLayerOffset is subtracted only in case of a non-BCCH layer of a segment. For call setups and intra cell handovers the C/Ns are calculated as follows:

( )110__max +−−= levelnoiseDLerOffsetnonBCCHLayRXLEVNC

DLDL

(Eq.6)

( )110__max +−−= levelnoiseULerOffsetnonBCCHLayRXLEVNC

ULUL

(Eq.7)

Intra BSC inter Cell handover

In inter cell handover the uplink measurements are not available and therefore the C/N is based only for the downlink measurement. The nonBCCHLayerOffset is subtracted only in case of a non-BCCH layer of a segment. For inter cell handover the C/Ns are calculated as follows:

( )110__)max( +−−= levelnoiseDLerOffsetnonBCCHLaynRXLEVNC

DLDL

(Eq.8)

( )110__)max( +−−= levelnoiseULerOffsetnonBCCHLaynRXLEVNC

DLUL

(Eq.9)

Inter BSC handovers

In inter BSC handover the measurement results are not available, hence the C/N checks are not performed at all. All channel modes can be used.

The calculated C/N values are compared to the connection type specific C/N soft blocking limits. If the either of the calculated C/N ratios is smaller than the C/N soft blocking limit, then the channel mode use is prohibited due to insufficient C/N ratio. This way some of the channel modes may not be allowed to be used. This restriction overrides the preference for usage of HR channels in case of high cell load situation or forced HR mode as presented in Section 4.6.

If the either of the calculated C/N ratios (DL/UL) is below the soft blocking C/N limits of all channel modes, then the channel assignment is soft blocked from DFCA TRXs. However, soft blocking is not applied if the channel assignment is due to a handover.

For example:

Calculated C/N ratio = 10dB Soft blocking C/N (FR) = 12 dB Soft blocking C/N (HR) = 14 dB


35 (93)

Soft blocking C/N (AMR FR) = 7 dB Soft blocking C/N (AMR HR) = 12 dB In this case the low C/N ratio prohibits the usage of all other channel modes except the usage of the AMR FR channel. Therefore, the connection is attempted to be assigned to a AMR FR channel irrespective of the cell load or the forced HR mode. This is to ensure that users in marginal conditions are not forced to use a channel mode that cannot perform adequately in such conditions. If the AMR feature is not supported by the mobile or it is not activated in the network, then the connection establishment is soft blocked as both non-AMR FR and non-AMR HR channels require higher C/N than the calculated C/N ratio is.

4.5 Training sequence selection

The worst-case situation regarding the training sequence usage in synchronized network is a case where a significant co-channel or adjacent channel interference source uses the same training sequence code (TSC). This may cause the receiver to obtain incorrect channel estimate and therefore lead to link level performance degradation and poor AMR link adaptation performance. To avoid this the following procedure is used when a TSC is selected for a new connection on DFCA TRX.

If a suitable channel is found then the most suitable training sequence code is determined by searching through the interfering connections and finding the TSC that has been used with the highest C/I. This means that for the selected MA, MAIO and time slot combination minimum C/I differences are checked TSC by TSC and the one with the highest C/I difference is selected. Example of the TSC selection is shown in Table 6.

Table 6. Selection of the TSC

Co-TSC 0 1 2 3 4 5 6 7

Min C/I diff (incoming DL)

0 5 7 0 3 7 8 0

Min C/I diff (incoming UL)

1 2 8 2 6 5 9 0

Min C/I diff (outgoing DL)

3 5 12 3 8 4 11 2

Min C/I diff (outgoing UL)

4 7 15 5 3 5 13 3

Min C/I diff (all considered)

0 2 7 0 3 4 8 0


36 (93)

In this case the TSC 6 is chosen because the minimum C/I difference is the highest so that with TSC 6 the new connection will have the smallest chance for adverse effects from TSC autocorrelation.

4.6 Channel mode selection (FR/HR)

4.6.1 Overview of channel mode selection

DFCA has no impact on the basic channel mode adaptation mechanism that is based on the BTS load in terms of the number of available time slots. This will function normally with DFCA.

However, simulations have shown that DFCA achieves the best gains when HR channels are used all over the DFCA area as shown in Figure 16 below. This can lead to a situation where it is beneficial to use HR channels even if there are still sufficient HW resources available in the BTS. During low traffic hours it is still beneficial to prefer FR operation as FR channels can provide better subjective speech quality in these circumstances. In this situation the existing BTS load based method alone is not sufficient.

DFCA AMR FR and AMR HR performance Downlink FER(>4.2%)

0

0.5

1

1.5

2

2.5

3

15.00% 21.00% 27.00% 33.00% 39.00%

EFL (%)

outa

ge %

DFCA AMR FR DFCA AMR HR

Transition to HR operation

DFCA AMR FR and AMR HR performance Downlink FER(>4.2%)

0

0.5

1

1.5

2

2.5

3

15.00% 21.00% 27.00% 33.00% 39.00%

EFL (%)

outa

ge %

DFCA AMR FR DFCA AMR HR

Transition to HR operationTransition to HR operation

Figure 16. DFCA performance with AMR FR 5.9 and AMR HR 5.9

4.6.2 C/I based rate selection between HR and FR (Not available in BSS11)

In order to ensure utilization of HR in all high load situations, a C/I based method is used together with the BTS load based method. The C/I based method is based on averaging the estimated incoming DL C/Is of all DFCA assignments within each BTS over a pre-defined averaging period. The average incoming DL C/I provides a good benchmark of the load on the DFCA frequencies and can therefore be used to trigger a forced HR operation whenever the load becomes so high that the additional DFCA gains of HR operation are needed.

The incoming DL C/I figure of each successful DFCA channel assignment is added up by BSC in a BTS level temporary statistic. Only the C/Is that are calculated using DL measurement results are included to this C/I statistic. This excludes e.g. BSC external handovers from this statistic. The average BTS level incoming DL C/I is periodically determined from this statistic after which the statistic is reset and started from the beginning. The averaged C/I value represents the load on the DFCA frequencies until


37 (93)

the next C/I average is determined. A separate statistic and averaged C/I value is maintained for each frequency band (i.e. DFCA BTSs in 800 band use different C/I average than the DFCA BTSs in 1900 frequency band.)

The load thresholds for forced HR operation are defined by means of 'DFCA forced HR mode C/I threshold' and 'DFCA forced HR mode C/I threshold AMR' –parameters and the C/I averaging period is defined by the 'DFCA forced HR mode C/I averaging period' parameter. If the current incoming DL C/I average is below the above threshold for the corresponding channel type (non-AMR/AMR), then a HR channel will always be assigned to a speech connection that is placed on a DFCA TRX. Once the forced HR mode has been switched on, it will only be switched off after the BTS level incoming DL C/I average has improved at least to the level of the corresponding 'DFCA forced HR mode C/I threshold' plus the hysteresis defined by the parameter 'DFCA forced HR mode hysteresis'.

A DFCA cell can be in the following Forced HR modes:

1. No forced HR mode

2. Forced HR mode for non-AMR channels • A cell enters this mode when the latest average incoming DL C/I is below the 'DFCA forced HR

mode C/I threshold' and above the 'DFCA forced HR mode C/I threshold AMR'

3. Forced HR mode for AMR channels • A cell enters this mode when the latest average incoming DL C/I is below the 'DFCA forced HR

mode C/I threshold AMR' and above the 'DFCA forced HR mode C/I threshold'

4. Forced HR mode for non-AMR and AMR channels • A cell enters this mode when the latest average incoming DL C/I is below the 'DFCA forced HR

mode C/I threshold' and also below the 'DFCA forced HR mode C/I threshold AMR'

The HR performance benefits are only achieved if the interfering cells are also using HR channels. Therefore, every time a cell enters the forced HR mode due to the C/I criteria presented above, the forced HR mode will be also taken into use in all interfering cells listed in the incoming interference BIM table. Example of this is shown in Figure 17 where forced HR mode in cell A also triggers forced HR mode in the surrounding interfering cells. When a forced HR mode is requested to be used in a cell the requested mode is stored into the outgoing BIM table of that cell.

Similarly, when a cell exits Forced HR mode due to the C/I criteria the Forced HR mode request to the interfering cells is cancelled for all the cells that are listed in the incoming interference BIM table.

Thus, DFCA must be operating in Forced HR mode if:

1. The average incoming DL C/I over the last averaging period is below the corresponding "DFCA forced HR mode C/I threshold" or

2. The cell has been in Forced HR mode during the previous period due to a C/I reason. The average incoming DL C/I over the last averaging period is below the corresponding "DFCA forced HR mode C/I threshold" + "DFCA forced HR mode hysteresis" or at least one of the interfered cells listed in the outgoing interference BIM table has requested forced HR mode and this request has not been cancelled.


38 (93)

Cell level average CIR0

Forced HR mode ON threshold

Hysteresis

Forced HR mode OFF threshold

A A

Average CIR in cell A0 Cell level average CIR

Average CIR in cell ACell level average CIR0

Forced HR mode ON threshold

Hysteresis

Forced HR mode OFF threshold

A A

Average CIR in cell A0 Cell level average CIR

Average CIR in cell A

Figure 17. Cell A in normal mode and forced HR mode

Rate selection between full rate and half rate

BSC selects the rate between full rate and half rate at first according to the load-based method. And if the recommended rate is full rate the C/I method is evaluated. After evaluation the recommended rates could be different for the different bands. In the load based method the rate selection is on segment level and so always the same for all bands. Note that the C/N based cell access control presented in Section 4.4.7 can prohibit the use of HR for some connections regardless of the cell load based or C/I based channel rate selection.

In the channel allocation BSC prefers the non-forced band. In the first round of the search only the non-forced full rates are searched. On the second round BSC prefers the band that is not forced to half rate according to the BSC level C/I. For an example if there is a channel available on acceptable minimum C/Is on both bands the BSC selects the channel from the non forced band. If these two rounds don’t make the result then the full rate channels of the HR forced band are searched.

5. BIM UPDATE PROCESS

5.1 General

The BIM update process collects statistical C/I data from all the cells that are reported by the mobiles in each DFCA BTS. The BSC will collect this C/I statistics for the DFCA algorithm from the active TCH connections of a DFCA BTS as the signal levels are reported by the mobiles. This means from all the connections in the BTS and not only from those in DFCA TRXs. It should be noted that the C/I statistics within the BIM update procedure means data for the internal use of the DFCA algorithm. It is not referring to the counters that are part of the Performance Monitoring functionality of the BSC and the BIM tables themselves only indicate the potential C/I between two cells and not the actual C/I that has been realized.

DFCA is controlled on per BTS basis and the C/I statistics collection for the DFCA algorithm is made in a BTS where DFCA has been switched on (DFCA mode is set to other than OFF). The BIM tables that are formed based on the collected data are however maintained on cell level (segment level) between the interfered and the interfering cells. This has significance in segment environment where cells can have several BTSs (with features Multi BCF Control and Common BCCH Control).


39 (93)

A drawback with the cell level C/I definition is that it does not take into account the differences between segment's BTSs if they are of different base station site types in the case of Multi BCF Control. Different frequency bands in the Common BCCH Control do not pose a problem because resources on different frequency bands do not interfere with each other. Frequency band here refers to GSM800, GSM900, GSM1800 or GSM1900.

In a multi band cell the C/I information is collected only on the frequency band that corresponds to the BCCH frequency band of the neighboring cell that the MS reports. Thus, the C/I information is not necessarily collected in every DFCA BTS for a certain neighbor if the frequency bands do not match. The C/I relation defined between two cells on one frequency band is used also on the secondary frequency band between the cells if there are such resources in both cells.

A single band cell on a band that differs from that used for common BCCHs in the network is an exceptional case. In such a cell it is necessary to collect C/I data of a neighbor on a different band due to the fact that in the neighboring cell there may also be a secondary frequency band which can interfere with the resources of the serving cell.

The C/I information in multi band cells is collected only on the frequency band on which the interfering neighbor cell is detected (BCCH freq band of the neighbor). As DFCA mode is controlled BTS by BTS a configuration is possible where only the secondary frequency band of a segment is used for DFCA. As the C/I is determined based on the BCCH of an interfering common BCCH controlled cell it requires that the necessary DFCA measurements are collected also in the BCCH band of the interfered segment. There are no extra actions in the BSC software to implement this. Instead, the operator is required to set the BTSs of the BCCH band in 'Standby' DFCA mode in order to have the BIM update procedure working properly.

5.2 C/I statistics collection

The C/I is determined as a downlink signal level ratio towards each reported neighboring cell in a DFCA BTS as presented in Equation 6. A reported neighboring cell refers to any cell with any BCCH&BSIC combination that is included in one or more received measurement report. The C/I statistics are therefore collected and the BIM C/I values are determined to all reported cells (not only the defined adjacent cells*). If DL Power Control is being used in the serving BTS, the power reduction factor applied must be taken into account to determine the maximum potential level of the serving TCH.

*Note: Enhanced Measurement Reporting (EMR) feature may restrict the reporting of the cells that have not been defined as adjacent cells depending on the parameter settings for that feature. For more information refer to the EMR feature documentation.

CHneigbourBCservingTCH RXLEVreductionPWRRXLEVIC

−+= _ (Eq.10)

If the reported neighbor is controlled by the same BSC and it is not in the 'DFCA hopping' or 'Standby' DFCA mode then the C/I statistics for this neighbor is not collected. For the neighbors controlled by other BSCs the check on DFCA usage is made later in the process by a message transfer between DFCA algorithms of different BSCs.


40 (93)

In each DFCA cell BSC maintains a C/I statistics table where the measured C/I distribution is collected for every neighboring cell reported in the cell's DFCA BTSs. For each reported cell there are 44 counters. 43 of these are C/I counters that are ticked whenever a received measurement report indicates a C/I corresponding to the C/I value that the counter represents. The last counter shows the total number of times when any of the counters representing the neighbor cell has been ticked. An example of the C/I statistics counters is presented in the following Table 7.

The collecting of the C/I data is made in every DFCA BTS of a segment but the obtained data is saved only in one segment specific table, which sums up the results that are received for a particular neighbor cell in different BTSs.

Table 7. Example C/I statistics (C/I counters 7 dB – 29 dB omitted from this picture)

BCCH BSIC <-5 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 ... 30 31 32 33 34 35 >35 Σ102 54 1722 293 322 376 421 532 832 1443 2109 2890 3782 4390 5873 ... 3453 3209 2877 2477 2029 1567 5488 172629115 22 120 47 78 102 176 258 329 467 598 832 1290 1790 2598 ... 8902 8637 8276 7839 7483 7108 25489 145763... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...

C/I counters

5.3 Processing of the C/I statistics for BIM update

The BIM update is a periodically repeated process where segment's C/I statistics data is processed and the BIM tables are updated. The BIM update period is an operator definable parameter. This may range from 10 minutes to 24 hours and the BIM updates can also be frozen completely if needed.

Step 1: Filtering of insignificant cells

The BSC sums up the total amount of samples collected for each particular neighboring cell. If the total amount of reported samples is less than 1000 the cell is excluded from the BIM update process. This is to ensure the statistical reliability of the BIM.

The requirement of a thousand samples per neighbor is reasonable also with the shortest update period values since the samples are collected frequently of every active connection in a DFCA BTS. A connection that is collected information about can be also on a regular TRX. The samples collected in regular TRXs give similar information about neighboring cells as those in DFCA TRXs and improve the statistical reliability of the BIM.

Step 2: Compensation for unreported measurement results

In the second step the impact of unreported measurement results is minimized. Unreported measurement result means a situation where the RXLEV measurement result of the neighboring cell is not within the x strongest measurement results where the x is the maximum number of measurement results that can be packed into one measurement report.

In this step the BSC will determine the neighboring cell that has been reported most frequently (i.e. the cell that has the highest Σ This highest Σ is denoted as Σmax.

For each reported neighbor in the C/I statistics the difference

Σ−Σ=Σ maxdiff (Eq.11)


41 (93)

is calculated. This calculated difference Σdiff is then added to the >35 counter of the neighbor. It now appears that all the neighboring cells were reported in equal number of reports and in case of a neighboring cell RXLEV not actually being included in the report it is now assumed to indicate a C/I of over 35 dB.

Step 3: Determining the C/I corresponding to γ

In the third step the C/I corresponding to the BIM confidence probability is determined. The BIM confidence probability γ is a BSC level operator parameter indicating the level of confidence for the estimation used to build the BIM. The C/I is determined as follows:

1. The amount of C/I samples corresponding to γ is determined by the formula γn

( ) Σ⋅−= γγ 1n (Eq.12)

where � is the BIM confidence probability and � is the sum of all C/I samples in the C/I statistics for this neighboring cell after the Σdiff has been added to the >35 dB counter.

2. The counters in the C/I statistics table are summed up from left to right until the sum is equal to or higher than . The C/I corresponding to ��is the C/I that this counter represents. γn

Step 4: Band offset correction

The C/I information is always collected using the normal equation (Eq 6). For all interference relations where the interfered single band cell and the interfering multi band cell BCCH are on different frequency bands BSC makes a "correction" by taking into account the signal level difference between the bands in the interfering cell. For this the mean value of the nonBCCHLayerOffset parameters of the non-BCCH band BTSs is used as shown in the following equation.

meanfinal

erOffsetnonBCCHLayIC

IC

+= (Eq.13)

Downward rounding is used with the mean value for the offsets in order to avoid being too optimistic about the final result.


42 (93)

BCCH1800

1800BCCH900

Interference

Serving cellInterfering cell

CI

nonBCCHLayerOffset =RXLEVBCCH900 - RXLEV1800

C/Ifinal = C/I + nonBCCHLayerOffset

Figure 18. Band offset correction

5.4 Updating of the BIM tables

The statistical C/I values that have been determined after the processing described in the previous section are used to update the BIM tables. The BIM table updating process is slightly different in case the neighboring cell is already included in the BIM tables compared to the case where the neighboring cell is not listed in the BIM tables causing a new BIM table entry to be created. Both of these cases are described separately below.

5.4.1 Case 1: Creating a new BIM entry

In case of a new BIM table entry the interfering cell identification and the corresponding statistical C/I value are added to the incoming interference BIM of the interfered cell (= the cell in which the C/I statistics were collected). Correspondingly, the same statistical C/I value with the identification of the interfered cell are added to the outgoing interference BIM table of the interfering cell (= the cell that is causing the interference). So, the same statistical C/I value is always present in two BIM tables: Incoming interference and outgoing interference BIM tables of the interfered cell and the interfering cell respectively. Together these BIM table entries form a so called “interference relation” that identify the possibility of interference and it’s direction in both ends of the relation. The interference relation can be internal within one BSC or it can span between two BSCs if the interfered and the interfering cells are controlled by different BSCs and the BSCs are connected by the inter BSC interface. In either case the relation is automatically setup and maintained.

However, before a new interference relation is established the BSC will check if the interference is significant enough.


43 (93)

If the C/I value is less than or equal to the level indicated by the ‘Interference Threshold’ -parameter then new BIM entries are created and the interference relation is established. The missed update counter in the incoming BIM entry is set to 0.

If the C/I value is greater than the ‘Interference Threshold’ -parameter then no new BIM entries (incoming interference BIM/outgoing interference BIM) are created. This is done to avoid creating interference relations for cases where the expected level of interference is so low that it can be considered irrelevant from the connection quality point of view. This way unnecessary signaling and processing load in BSC can be avoided.

It should be noted that DFCA uses the Interference Threshold parameter that is valid in the BSC of the interfering cell. It is assumes that the threshold values in all the BSCs of the network are the same. It is however a responsibility of the operator to set values equal throughout the network.

If the connection to the remote BSC cannot be established the BSC removes the remote interfering cell from the incoming interference BIM table. If the statistical C/I for the unknown foreign cell was lower than the lowest defined soft blocking C/I threshold in the DFCA cell the interference is regarded as significant and BSC sets a new disturbance alarm 'Unknown potentially interfering cell for DFCA'.

5.4.2 Case 2: Updating an existing BIM entry

If an interference relation and therefore the associated BIM table entries already exist then the BSC performs the following actions:

1. If the old C/I value in the BIM is the same as the newly determined one then only the missed update counter for the neighbor BTS in the incoming interference BIM table is set to 0.

2. If the old and the new C/I value differ from each other the information of the two values is combined. This combining forms a statistical filtering process that controls the rate of change of the BIM value. The combining is controlled by an operator definable parameter ‘BIM Update Scaling Factor’ denoted as α. The parameter can have values between 0 and 1. It identifies the impact of the latest C/I distribution when it is combined with the long term distribution. The default value for α is 0.5.

The combination of the new and the old BIM value is done as follows:

αβ −= 1 (Eq.14)

oldnewIC

IC

IC

γγγ

βα ⎟⎠⎞

⎜⎝⎛⋅+⎟

⎠⎞

⎜⎝⎛⋅=⎟

⎠⎞

⎜⎝⎛

(Eq.15)

a) If the combined C/I value is greater than the level indicated by the Interference Threshold

parameter then the interference relation is terminated by removing the BIM table entry for the neighbor BTS from the incoming interference BIM table of the serving BTS. Also information of the serving BTS is removed from the outgoing interference BIM table of the interfering BTS.


44 (93)

b) If the combined C/I value is less than or equal to the level indicated by the Interference Threshold parameter the BIM entry for the neighbor BTS is updated in the incoming interference BIM table of the serving BTS. Also the C/I value for the serving BTS is updated in the outgoing interference BIM table of the interfering BTS.

For the updated incoming interference BIM table entries the Missed Update counter is set to 0.

5.5 Interference relation termination

In constantly changing environment and network topology it can happen that a cell that was previously found to be a relevant interference source may cease to be significant enough to be considered in DFCA C/I estimations. In this case such interference relation must be terminated. The termination of the interference relation can happen for the following reasons:

1. DFCA is turned off in all the BTSs of the interfering cell in which case the BSC will automatically terminate all interference relations for that cell.

2. If the BSC still receives sufficient number of measurement results (1000) where the interfering cell signal level is included and the determined statistical C/I is found to be above the level set by the Interference Threshold parameter then the interference relation is terminated immediately as described in Section 5.4.2.

3. If a previously found interfering cell does not appear in n successive BIM updates where n corresponds to the ‘BIM update guard time’ –parameter setting then the interference relation is terminated. It is recommended that the BIM update guard time parameter is set to sufficiently high value so that temporary faults and maintenance outages do not cause termination of the interference relation as a terminated relations must be re-established by the DFCA switch on process described in Section 7.2 before the DFCA operation can be resumed. For example, if the BIM update period is set to 1 hour and the BIM update guard time is set to 48 then the interference relations are preserved for 1 hour x 48 = 48 hours before they are deleted as a consequence of a cell being out of service. It is also possible to allow indefinite outages by setting the BIM update guard time –parameter to 63.

5.6 BCCH & BSIC conflict management

In some special circumstances an existing interference relation may become ambiguous or it may not be possible for the mobile terminals the measure the interfering cell signal level anymore. In these cases the interference relation is maintained by locking the incoming interference BIM table entry so that the particular interference relation termination is not allowed. This locking is done automatically when needed by the BSC by setting the missed update counter value to 255.

A locked BIM entry indicates that a BCCH conflict or a conflict between two (BCCH,BSIC) pairs exists. There are two different conflict cases:

1. The interfering cell is currently using the same BCCH frequency than the serving cell. Therefore, this cell cannot be measured by the mobiles and it will not appear in the DFCA C/I statistics. Without an explicit locking procedure this could eventually lead to the deletion of the BIM entry as


45 (93)

the 'Missed Update' counter reaches the set threshold. The locking procedure makes sure that a cell that was previously found to be interfering will be taken into account even after the BCCH frequency has been changed to be the same as in the serving cell. The BSC will identify the BCCH frequency conflict automatically and immediately after the BCCH frequency of a cell where DFCA is used is changed. Similarly, when the conflict is removed by BCCH frequency change the BSC will automatically remove the locking and resume the normal BIM updates.

2. Two of the interfering cells that are listed in the incoming interference BIM table are using the same (BCCH, BSIC) combination. Therefore, it is unknown which is the actual measured cell when a C/I is extracted from a DL measurement report with this (BCCH, BSIC) combination. Because of this ambiguity, the incoming interference BIM entries of both of these interfering BTSs with the conflicting (BCCH, BSIC) pair are locked by setting the ‘Missed Update’ counter to 255. While the incoming interference BIM entries are locked, the BIM updates for the entries are not performed and any measured C/I with the conflicting (BCCH, BSIC) combination is ignored (i.e. C/I estimations for the incoming interference from these conflicting BTSs are done based on the BIM only). The interference relations are kept active ensuring that the information flow regarding channel assignments and releases, power level and neighboring cell updates keep flowing in (in case the interfering cells are controlled by other BSCs). This situation can happen if operator changes the BCCH&BSIC combination of one interfering cell to be the same with another interfering cell. BSC will automatically detect this situation and take the described actions. Similarly when the conflict is removed the BSC will automatically restore the normal operation.

6. INTER BSC COMMUNICATION

6.1 Introduction of the BSC – BSC interface

The BSC-BSC interface is used for the transmission of BSC (BTS) measurements to neighboring BSCs, for example updating the DFCA Radio Resource table and the BIM tables. The BSC-BSC connection is required in order to exchange the information needed by the DFCA so that DFCA can operate as transparently as possible across BSC area borders.

This interface is physically located in the BCSU unit that has to have a CPU card for the Ethernet LAN connection. The following CPU cards are the minimum requirement (containing 2 Ethernet ports per CPU card):

• CP6MX for BSC2i

• CP710-A for BSC3i (installed as default with BSC3i)

The used protocol stack in this interface is presented in Figure 19. The use of M3UA & SCCP allows load sharing between BCSU units and switchovers are supported. If two links are configured it can handle switchovers so that at least one signaling link is working and application doesn't even notice the problem. The DFCA procedures will be added to be extension of RNSAP. The complete RNSAP is not implemented as part of the DFCA feature.


46 (93)

M3UA

IP

SCTP

ethernet

M3UA

RNSAP+

IP

SCTP

ethernet

SCCP SCCP

RNSAP+

BSC BSC

Figure 19. Protocol stacks for the BSC- BSC interface

6.2 Addressing in the BSC - BSC interface

Every BSC has been assigned an SCCP Signalling Point Code (SPC) during the creation of the A interface. This same Signalling Point Code is used also to identify and address a BSC through the BSC-BSC interface. Internally the DFCA functions in the BSC must be able to uniquely identify each DFCA cell in the area and the BSC (SPC) that is controlling each of the DFCA cells.

Finding out the BSC (SPC) controlling each of the DFCA cells is done automatically but with some help from the operator in the form of a LAC to SPC mapping table. The LAC to SPC mapping table helps to reduce the signaling and processing load in the BSC - BSC interface by giving the BSC a limited set of neighboring BSCs from where to search each DFCA cell that is not under it’s own control.

Example of the LAC to SPC mapping table is presented in Table 8. In this table each LAC is mapped to 1-6 different SPCs. The operator has to fill this table so that it contains the LACs used in the DFCA area and the BSCs (SPCs) that control one or more cells within each LAC.

Table 8. Example of the LAC to SPC mapping table

LAC to SPC mapping info

Index LAC SPC

10002 24 302 54 1202

1 10011

923 2032 254 0 0

2 10015

0


47 (93)

0 311 265 43 0 0

3 10020

0

… …

0 0 0 0 0

64 FFFEH

0

The DFCA related information is always concerning one particular DFCA cell that is identified by unique BCCH&BSIC combination. When a BSC needs to communicate DFCA related information to another BSC it will proceed in the following order:

1. BSC will find out if it has already identified the BSC (SPC) that is controlling the cell with the required BCCH&BSIC combination. - If the SPC for the BCCH&BSIC combination can be identified then the message is sent to the correct SPC - If the SPC for the BCCH&BSIC combination is not known the BSC will proceed to the next step

2. The BSC will check if the LAC in which the cell with the required BCCH&BSIC combination is located can be identified from the adjacent cell definitions. - If the LAC can be identified the BSC will then query all the BSCs (SPCs) listed for the LAC in the LAC to SPC mapping table for the required BCCH&BSIC. In normal case only one of the BSCs (SPCs) replies with a positive answer and thus the correct SPC for the required BCCH&BSIC combination is identified for messaging. - If the LAC cannot be identified (meaning that the required cell (BCCH&BSIC) is not defined as an adjacent cell) the BSC will then send a query to every known BSC (SPC) listed in the LAC to SPC mapping table. In normal case only one of the BSCs (SPCs) replies with a positive answer and thus the correct SPC for the required BCCH&BSIC combination is identified for messaging. - If the LAC can be identified but the LAC is not listed in the LAC to SPC mapping table, then a disturbance alarm: 'Unknown potentially interfering cell for DFCA' is raised - If the LAC can be identified and the LAC is listed in the LAC to SPC mapping table but with no SPCs defined for this LAC, then BSC assumes that there are no DFCA cells in this LAC and terminates the attempt to send the DFCA related message without raising alarms - If more than one SPC reply positively implying that they control a cell with a required BCCH&BSIC combination then a BCCH&BSIC conflict situation is identified and a 'DFCA neighbor cell conflict' –alarm is raised identifying the conflicting BCCH&BSIC combination and the SPCs. In this case the message sending is cancelled.


48 (93)

6.3 Signaling traffic load

The signaling traffic load on the BSC-BSC interface in a relatively high load situation has been estimated to be about 200kbit/s in each direction for one BSC. This estimate is based on the assumptions listed in Table 9. The signaling load is quite sensitive to the assumptions so the typical signaling rate will vary depending on the traffic load, network topology and radio propagation conditions in the area.

Table 9. Assumptions used to estimate the signalling traffic load

Nb Of Cells/BSC 100Nb Of Users/cell 22.4Max Reported Neighb 10Affected cells in own BSC 50%Nb Of Interfering BSCs 2

7. OPERATION AND MAINTENANCE

7.1 Nokia NetAct network management system

Nokia NetAct is a network and service management system, which provides centralised management functions for different network technologies and network domains. The system consists of functionality areas (Monitor, Reporter, Planner etc.), which provide management capabilities grouped together according to the most relevant operator processes. For BSS there is an extensive set of management functionality available: - monitoring - reporting - planning - configuring and optimising - system management - service quality management.

Nokia NetAct is interfacing BSS and will also collect data related to the DFCA feature, such as alarms and measurement counters. Reporting and monitoring functionality is possible to develop using the data collected. NetAct Radio Access Configurator will enable network wide management of DFCA parameter configuration providing support to feature rollout, activation and tuning.

Nokia NetAct support for DFCA feature is part of OSS4 release that will support BSS S11.5.


49 (93)

7.2 Special considerations related to BSS synchronization

7.2.1 Time slot offset for better FR SACCH performance

TDMA frame number offset and time slot offset are LMU parameters defined in BSC for each LMU. In order to support HR channels with DFCA, the used frame number offsets must be multiples of 26 as illustrated in Figure 20. The Frame Number Offset parameter [0…51] implements the requirement 'FN_offset = n x 26' so that it is actually 'n' that is set by the parameter in the BSC.

HR HR HR HRFN offset = 0x26 FN offset = 1x26 FN offset = 0 FN offset = 5

Sub channel Sub channel Sub channel Sub channel0 1 0 1 0 1 0 1

0 x 26 x 0 x 5 x1 x 27 x 1 x 6 x2 x 28 x 2 x 7 x3 x 29 x 3 x 8 x4 x 30 x 4 x 9 x5 x 31 x 5 x 10 x6 x 32 x 6 x 11 x7 x 33 x 7 x 12 SACCH8 x 34 x 8 x 13 x9 x 35 x 9 x 14 x10 x 36 x 10 x 15 x11 x 37 x 11 x 16 x12 SACCH 38 SACCH 12 SACCH 17 x13 x 39 x 13 x 18 x14 x 40 x 14 x 19 x15 x 41 x 15 x 20 x16 x 42 x 16 x 21 x17 x 43 x 17 x 22 x18 x 44 x 18 x 23 x19 x 45 x 19 x 24 x20 x 46 x 20 x 25 SACCH21 x 47 x 21 x 26 x22 x 48 x 22 x 27 x23 x 49 x 23 x 28 x24 x 50 x 24 x 29 x25 SACCH 51 SACCH 25 SACCH 30 x26 x 52 x 26 x 31 x27 x 53 x 27 x 32 x

FN FN FN FN

Cross sub channel interference

Inter sub channel interference

Controlled (FN_offset= n x 26) Uncontrolled (FN_offset≠ n x 26)

DFCA requires interference control => FN_offsets must be n x 26

(This is also the current LMU implementation)




FN FN FN FN




FN FN FN FN





Controlled (FN_offset= n x 26) Uncontrolled (FN_offset≠ n x 26)

DFCA requires interference control => FN_offsets must be n x 26

(This is also the current LMU implementation)

Figure 20. Illustration of the necessity of the n x 26 TDMA FN offsets when HR is used with DFCA

The FN offset = n x 26 –requirement leads to a situation where the SACCH burst are transmitted simultaneously through the network. As the SACCH bursts are always transmitted independently of DTX, the SACCH does not benefit from DTX gain at all. While in asynchronous network SACCH benefits from DTX as in most cases SACCH is interfered by TCHs that use DTX.

In case of AMR FR the C/I of the channel can be very low, as TCH is very robust. Low C/I condition together with the lack of DTX gain can cause very poor SACCH performance, which leads to lost measurement reports and PC commands and ultimately to degradation in TCH/FR performance.

Solution

AMR FR capacity degradation can be avoided by using the time slot offset –parameter. The parameter may have values 0 and 1. Using every second cell a different value of time slot offset the SACCH load decreases by 50% as illustrated in Figure 21. The FN for sending the FR SACCH depends on the time slot number. For even and odd time slots the frame number is different and when SACCHs for the evens are sent for the odds that frame is idle. For the timeslots 0,2,4,6 the SACCHs are sent when FN is 12+26 x n and for the 1,3,5,7 the SACCHs are sent when FN is 25 + 26 x n. If the network is


50 (93)

synchronized without applying time slot offsets then the SACCHs e.g. for time slot 0 are sent at the same time in whole network and in next time slot (time slot 1) nothing is sent.

CELL 1

2

3

4

SACCH oridle

FN 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

SACCH oridle

TSL 0 1 2 3 4 5 6 7

TSL 0 1 2 3 4 5 6 7

TSL 0 1 2 3 4 5 6 7

TSL 0 1 2 3 4 5 6 7

Time slot de-alignment

0 1 2 3 4 5 6 7

Time slot offset = 1,FN 12 and SACCHsendingstarts one time beforethe cells with the timeslot offset 0

Figure 21. 50 % less SACCH load for full rate with the FN de-alignment

The FN de-alignment for full rate causes exceptions in HR sub-channel alignment. As the time slot offset is 0 or 1, the de-alignment affects only 1 TCH of 8. If the time slot offset is set to 1 the TSLs are sent one time slot after the cell with the time slot value 0. In half rate every second frame is alternatively for the sub ch 0 or for the sub ch 1. TSLs 1-7 with TSoffset 0 and TSLs 0-6 with TSoffset


51 (93)

1 have the same frame number but the TSL 0 in TSoffset 0 and TSL 7 in TSoffset 1 have different frame numbers as illustrated in Figure 22.

TS offset0 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 01 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7

FN = 0 FN = 1 FN = 2

FN = 0 FN = 1 FN = 2

HR HRFN offset = 0x26 FN offset = 0x26

Sub channel Sub channel0 1 0 1

0 x 0 x1 x 1 x2 x 2 x3 x 3 x4 x 4 x5 x 5 x6 x 6 x7 x 7 x8 x 8 x9 x 9 x10 x 10 x11 x 11 x12 SACCH 12 SACCH13 x 13 x14 x 14 x15 x 15 x16 x 16 x17 x 17 x18 x 18 x19 x 19 x20 x 20 x21 x 21 x22 x 22 x23 x 23 x24 x 24 x25 SACCH 25 SACCH26 x 26 x27 x 27 x

FN FN

HR HRFN offset = 0x26 FN offset = 0x26

Sub channel Sub channel0 1 0 1

1 x 0 x2 x 1 x3 x 2 x4 x 3 x5 x 4 x6 x 5 x7 x 6 x8 x 7 x9 x 8 x10 x 9 x11 x 10 x12 SACCH 11 x13 x 12 SACCH14 x 13 x15 x 14 x16 x 15 x17 x 16 x18 x 17 x19 x 18 x20 x 19 x21 x 20 x22 x 21 x23 x 22 x24 x 23 x25 SACCH 24 x26 x 25 SACCH27 x 26 x28 x 27 x

FN FN

FN de-alignment

TS 0-6 (TSoffset = 0)TS 1-7 (TSoffset = 1)

TS 7 (TSoffset = 0)TS 0 (TSoffset = 1)

Figure 22. FN de-alignment causes exception in HR sub-channels

When the C/Is are calculated between de-aligned HR TSLs the C/I evaluation is done for both of the interfering TSLs’ half rate channels. If the TS offsets are the same then both sub-channels don’t need to be checked.

If half rate is in use in the BTS then the BSC prefers FR allocation to the de-alignned TSL. If TS offset is 0 the de-aligned TSL is TSL 0 and if the TS offset is 1 the de-aligned TSL is 7. FR channel is preferred according to the following rules:

• if the cell has two equal channels (de-aligned TSL and other) according to the load level in SEG, TSL type and minimum C/I then the de-aligned one is chosen. If the de-aligned TSL has worse minimum C/I or it is for some other reason not so attractive then the other TSL is chosen.


52 (93)

7.2.2 FN offset for optimized BSIC decoding performance

In order to identify a neighboring cell the mobile terminal has to decode the BSIC transmission by receiving the frequency correction burst and the synchronization burst during the idle frames. The frequency correction and the synchronization bursts are sent on the BCCH according to a pre-determined schedule based on the TDMA frame numbers as presented in Figure 23. and the idle frame occurs once every 26 TDMA frames for FR connections as shown in Figure 24.

• BCCH multi frame (MF51)BCCH0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50F S - - - - - - - - F S - - - - - - - - F S - - - - - - - - F S - - - - - - - - F S - - - - - - - - -

• Includes 5 frames for FCCH and 5 frames for SCH

Figure 23. Scheduling of the frequency correction and the synchronization bursts in the BCCH multi frame

• TCH multi frame (MF26)

• Includes 1 frame for SACCH and 1 idle frame

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25I S

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25S I

TS 0, 2, 4, 6

TS 1, 3, 5, 7

• TCH multi frame (MF26)

• Includes 1 frame for SACCH and 1 idle frame

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25I S

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25S I

TS 0, 2, 4, 6

TS 1, 3, 5, 7

Figure 24. Scheduling of the idle frame in TCH FR multiframe

Decoding of the BSIC requires a co-incidence of the idle frame and the frequency correction and synchronization burst transmissions. If the FN offsets are not used then the BCCH multi frames are co-incident in all the cells significantly decreasing the possibilities to decode the BSIC as generally the decoding is possible only every 10 TCH multi frames as illustrated in Figure 25 whereas with non co-incident BCCH multi frames the BSIC decoding is possible much more frequently.


53 (93)

BCCH0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50F S - - - - - - - - F S - - - - - - - - F S - - - - - - - - F S - - - - - - - - F S - - - - - - - - -

TCH/FR (TS 2,3,5,7)0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

I I25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

I I24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

I I23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

I I22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

I I21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

I I20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

I I19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

I I18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

I I17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

I I16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

I I15 16 17 18 19 20 21 22 23 25 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13

I I14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12

I I13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11

I12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10I I

11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9I I

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8I I

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7I I

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6I I

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5I I

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4I I

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3I I

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2I I

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1I I

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0I I

BCCH0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50F S - - - - - - - - F S - - - - - - - - F S - - - - - - - - F S - - - - - - - - F S - - - - - - - - -

TCH/FR (TS 2,3,5,7)0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

I I25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

I I24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

I I23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

I I22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

I I21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

I I20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

I I19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

I I18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

I I17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

I I16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

I I15 16 17 18 19 20 21 22 23 25 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13

I I14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12

I I13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11

I12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10I I

11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9I I

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8I I

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7I I

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6I I

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5I I

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4I I

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3I I

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2I I

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1I I

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 0I I

Figure 25. Illustration of BSIC decoding opportunities when FN offsets are not used

Similarly as the TS offset also the FN offset is LMU specific parameter. Thus all the synchronized BTSs in the same site in the same clock distribution chain always have the same FN offset. The planning of the FN offsets is therefore done on site basis. A simple optimized FN offset planning rule is presented in Figure 26. It should be noted that the BSIC decoding speed improvement may not be very critical and perfectly acceptable results may be achieved with even more simple planning rules (for example derive FN offset value from BCF ID) as long as the FN offset varies from site to site.

The fastest BSIC decoding times can be achieved when the network and mobile terminal support Enhanced Measurement Reporting (EMR) feature that allows the network to signal Real Time Differences (RTDs) of the neighboring cells to the mobiles.


54 (93)

+ 4+ 4

+ 4 + 4

+ 4+ 4+ 4

+ 4

+ 4+ 4

+ 1+ 1

+ 1

+ 1

+ 1

+ 1

+ 1

+ 1

+ 1

+ 1

17

16

20

2113

12

1814

MS

+ 4+ 4

+ 4 + 4

+ 4+ 4+ 4

+ 4

+ 4+ 4

+ 1+ 1

+ 1

+ 1

+ 1

+ 1

+ 1

+ 1

+ 1

+ 1

17

16

20

2113

12

1814

MS

Figure 26. A simple FN offset planning rule (FN offsets presented in red color)

7.3 Network planning/configuration requirements

7.3.1 BCCH & BSIC planning

The interfering cell identification in DFCA is based on unique BCCH frequecy and BSIC combination in each cell. It is therefore very important that the BCCH & BSIC planning ensures that the same BCCH & BSIC combination is not reused within adjacent BSCs. Normally there is no limitation in BSIC usage so that operator can utilize all 64 BSIC combinations in the network. Assuming 12 BCCH frequencies there are 64x12 = 768 unique BCCH & BSIC combinations allowing very large reuse distances. Reuse of a conflicting BCCH & BSIC combination too close may cause BCCH & BSIC conflicts explained in Section 5.6.

If a nearby cell has the same BCCH frequency as the serving cell then it cannot be measured by the mobiles causing a potential blind spot for DFCA as unmeasured cell cannot be considered as interfering cell for DFCA C/I estimation. Special care should therefore be taken to avoid severe co-BCCH situations. It should be noted that DFCA remembers a potential interference situation if a nearby cell has previously been measured and found to be potentially interfering but it has since become a blind spot as a consequence of a BCCH plan change. See Section 5.6 for further information.

7.3.2 Frequency band reorganization

DFCA requires a dedicated frequency band that is only used by DFCA. Also the accuracy of the BCCH level measurements directly impacts the C/I estimation accuracy and therefore DFCA performance. It is therefore recommended to use a dedicated BCCH frequency band ensuring that non-BCCH interference does not cause errors in the measured BCCH signal levels. Dedicated BCCH band also leads to shorter BA lists that makes it possible for the mobile to measure more level samples for each BCCH frequency and to decode the BSICs faster. All this is beneficial for DFCA C/I estimation accuracy.


55 (93)

7.3.3 Frame number and Time slot offsets

As described in Section 7.2 a proper setting of frame number and time slot offsets is important to ensure optimum performance in synchronized BSS environment. The FN offset is used to ensure optimum BSIC decoding speed and the TS offset is used to ensure sufficient SACCH performance.

7.4 Managing DFCA

7.4.1 Definition of DFCA MA lists

The system allows 64 different DFCA MA lists to be defined, of which 32 may be in used at a time. The reason for introducing two sets of 32 (max) frequencies is in making it easier to handle frequency band rearrangements: one set is active while another is modifiable without service impact. (Only one set of 32 freqs available in BSS11). The DFCA MA lists are defined in each BSC and MA list identification number ranges 661 … 692 and 693 … 724 (2 sets, max 32 frequences in each) are reserved for DFCA MA lists (256 – 287 in BSS11). There are a couple of rules that have to be followed when defining DFCA MA lists. These rules are listed below.

1. DFCA requires a dedicated frequency band

2. Any frequency can only be used in one DFCA MA list

3. If any two DFCA MA lists contain frequencies that are adjacent to each other then the two DFCA MA lists are considered to be adjacent. Adjacent DFCA MA lists must have the same length (i.e. contain identical number of frequencies).

4. DFCA MA lists in adjacent BSCs (BSCs whose service areas are adjacent to each other) must be defined the same way so that DFCA MA lists with identical id numbers are identical in both BSCs.

An example of DFCA MA lists definition is shown in Figure 27. In this example a DFCA frequency band has been defined to DFCA MA lists in three alternative ways.

In option 1 the DFCA band is simply divided into 2 DFCA MA lists each list containing every other frequency. Avoiding MA lists with consecutive frequencies is beneficial to maximize the frequency diversity gains of cyclic frequency hopping that is used with DFCA.

In option 2 the DFCA band is divided into 3 DFCA MA lists with every third frequency in each MA list. This may provide higher frequency diversity gains in environments that are not very dense. However, all three lists are adjacent to each other so that their lengths have to be identical. Because of this requirement, two frequencies of the example DFCA band cannot be used in this.

In option 3 the DFCA band is divided into 4 DFCA MA lists. Two of the DFCA MA lists are intentionally left short having only three frequencies each. These DFCA MA lists with different lengths are possible since there are no adjacent frequencies between the shorter and the longer lists. This is achieved by leaving frequency 35 unused. Also, frequency 48 is left unused as it cannot be used in any of the DFCA MA lists without braking the equal-length requirement of adjacent DFCA MA lists.


56 (93)

DFCA frequency band29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

MA 1: 29 31 33 35 37 39 41 43 45 47

MA 2: 30 32 34 36 38 40 42 44 46 48

MA 1: 29 32 35 38 41 44

MA 2: 30 33 36 39 42 45

MA 3: 31 34 37 40 43 46

unused: 47 48

MA 1: 29 31 33

MA 2: 30 32 34

MA 3: 36 38 40 42 44 46

MA 4: 37 39 41 43 45 47

unused: 35 48

Option 1:

Option 2:

Option 3:

Adjacent lists

Adjacent lists

Adjacent lists

Adjacent lists

Adjacent lists

DFCA frequency band29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

MA 1: 29 31 33 35 37 39 41 43 45 47

MA 2: 30 32 34 36 38 40 42 44 46 48

MA 1: 29 32 35 38 41 44

MA 2: 30 33 36 39 42 45

MA 3: 31 34 37 40 43 46

unused: 47 48

MA 1: 29 31 33

MA 2: 30 32 34

MA 3: 36 38 40 42 44 46

MA 4: 37 39 41 43 45 47

unused: 35 48

Option 1:

Option 2:

Option 3:

Adjacent lists

Adjacent lists

Adjacent lists

Adjacent lists

Adjacent lists

Figure 27. Example of DFCA MA list definitions

A special MA list is required for unsynchronised operation. It is recommended that the list contains all the DFCA layer frequencies and that the same list is attached into all the DFCA cells. The minimum length for the unsynchronised MA list is two times the number of DFCA TRXs in a cell.

7.4.2 Creation, activation, modification and deletion of DFCA MA lists

DFCA MA lists are created and treated separately from the normal MA lists used for conventional RF hopping. The identification number ranges reserved for the DFCA MA lists are 661 … 692 and 693 … 724 (2 sets, max 32 frequences in each) (256-287 in BSS11) allowing for a maximum of 64 different DFCA MA lists. Operator can create, delete, modify and output DFCA MA lists with the existing commands of the BCCH AND MOBILE ALLOCATION FREQUENCY LIST AND RA HANDLING command group (EB). The DFCA MA lists have specific state flag that can have two possible states: "out of use" and "in use". When a new DFCA MA list is created its initial state is always "out of use".

BSC performs several checks when a DFCA MA list state is changed to "in use" by the operator. When a DFCA MA list state is changed to "in use" –state the following checks ensuring correct DFCA operation are automatically done:

- there are frequencies on the DFCA MA list (the list is not empty),

- the frequencies on the DFCA MA list must not be found on any other existing and active (state = "in use") DFCA MA list,

- if there are adjacent frequencies on the modified DFCA MA list and an existing DFCA MA list then the amount of frequencies must be the same in each of them.


57 (93)

If all these requirements are not met the BSC will refuse the state change and return an error message.

A DFCA MA list can only be modified if its state is "out of use" and the list is not attached to any BTS. Changing the activity state of a DFCA MA list from "in use" to "out of use" requires that the list is not attached to any BTS. A DFCA MA list can be detached from a BTSs when the DFCA TRXs are locked (BSS11: BTS lock required). A DFCA MA list can be deleted if the following requirements are met:

- the activity state of the list is "out of use" or

- the activity state of the list is "in use" and the list is not attached to any BTS.

IMPORTANT: The DFCA MA lists with the same DFCA MA list id number and that are in the "in use" –state must be identical in all the BSCs controlling DFCA cells in the same continuous DFCA area. This is because in the BSC-BSC communication information of the actual DFCA MA lists is not exchanged. The DFCA MA lists are only identified by their id number that is required to be the same in all BSCs.

7.4.3 Attaching/detaching DFCA MA lists to/from a BTS

The operator can modify the DFCA MA list attachments of a BTS with command A in the 'Base Transceiver Station Handling in BSC' MML (EQ). The operator can modify the DFCA MA list attachments and the unsynchronized mode DFCA MA list attachment. The activity state of a DFCA MA list must be “in use” for the list to be attached.

A maximum of 32 DFCA MA lists can be attached to a BTS in the "DFCA hopping" mode for the synchronized operation.

In addition to actual DFCA hopping lists, also a separate MA list for the unsynchronized mode operation is needed. This list is a normal (id = 1 – 660, 1-255 in BSS11) MA list defined in the BSC and is used with the traditional channel allocation algorithm when the BTS in question has lost synchronization.

Modifying of the DFCA MA list attachments for the synchronized operation as well as for the DFCA unsynchronized mode MA list requires locking of the DFCA layer TRXs or locking the BTS (in BSS11 the whole BTS has to be locked).

7.4.4 Handling of BA lists with DFCA

In order to collect reliable C/I statistics and create complete BIM tables where all the relevant potentially interfering cells are listed, the mobiles should be able to measure the BCCH frequencies of all surrounding cells. Since the mobiles only measure the BCCH frequencies given in the BCCH allocation (BA) list, it is very important that the BA list contains all the relevant BCCH frequencies. Normally the BA list given to the mobiles in active state only contains the BCCH frequencies of the defined adjacent cells. This is sufficient for handover purposes, but if a nearby cell is not defined as adjacent cell then it may happen that the mobiles will never measure the signal level of such cell causing it to be invisible interference source that cannot be taken into account in the DFCA C/I estimation. This would degrade the DFCA performance.

To avoid DFCA performance degradation a BA list including all the BCCH frequencies used in the area should be created. This BA list should be attached as an idle state BA list to the DFCA cells and the active state BA list should be selected to be the same as the idle state BA list by using the EQB MML


58 (93)

command. This way the mobiles will monitor and measure all the possible BCCH frequencies minimizing the risk of some potential interfering cells not being measured and reported. The BSC will automatically remove the serving cell BCCH frequency from the list unless the Common BCCH functionality requires it.

7.4.5 Defining DFCA TRXs of a BTS

The operator can define a TRX to be used for DFCA by modifying the TRX level parameter DFCA indication. The parameter can be modified without locking of the object if the DFCA mode of the related BTS is OFF or STANDBY. If the BTS is in DFCA hopping mode then the procedure requires TRX locking.

The following requirements must be met on a DFCA TRX:

− There are only traffic channel timeslots in the TRX,

− (E)GPRS is not enabled for the TRX,

− The TRX is not a super reuse TRX,

− The TRX is not an extended range TRX,

− The antenna combiner (if any) is of wide-band type allowing RF hopping.

7.4.6 Modifying DFCA mode of a BTS

The BTS "DFCA mode" –parameter can be modified using the EQM MML command. The following picture illustrates the possible transitions between the different DFCA modes.


59 (93)

OFF

STANDBY

DFCA HOPPING

Figure 28 Possible transitions between DFCA modes

BSC prevents switching of DFCA mode from OFF directly DFCA Hopping.

The DFCA STANDBY mode is an interim mode that is required to be used before the actual DFCA operation in a cell is started by moving to DFCA hopping mode. In the STANDBY mode the DFCA C/I statistics collection with the BIM updates is started allowing the potentially interfering cells to be identified by the system. Also, the interference relations are established and the associated information updates from the interfering DFCA cells are started. All this makes it possible to start DFCA operation in the cell. However, for all this to happen at least one BIM update must be completed before the DFCA operation can be started. Therefore after changing the DFCA mode from OFF to STANDBY the operator must keep the BTS in STANDBY state for at least one BIM update period duration (for parameter, see section 8.1.5) before changing the DFCA mode to DFCA hopping. The BIM update period length is defined by the BSC level parameter.

7.4.6.1 Switching DFCA mode from OFF to STANDBY

The BTS DFCA mode can be switched from OFF to STANDBY “on the fly” without BTS locking as BTS configuration is not changed when the DFCA mode is changed. BTS stays in the STANDBY mode until the DFCA mode is manually switched to OFF or to DFCA hopping.

The switch to the STANDBY mode causes the following actions in BSC:

• BSC starts maintaining the incoming interference BIM table for the cell,


60 (93)

• BSC starts maintaining the outgoing interference BIM tables for other local DFCA cells included in the BIM updates

• BSC initializes and maintains inter BSC interference relations to DFCA cells under neighboring BSCs when the source of detected interference is a BSC external cell

• Interference relation initialization requests from neighboring BSCs can now be accepted for the cell leading to the maintenance of the outgoing interference BIM table for the cell.

The switch from OFF to STANDBY mode does not cause any actions in the BTS.

7.4.6.2 Switching DFCA mode from STANDBY to OFF

The BTS DFCA mode can be switched from STANDBY to OFF “on the fly” without BTS locking, as BTS configuration is not changed when the DFCA mode is changed.

When this mode change happens the BSC will delete the associated BIM tables and BIM entries and cancel all the DFCA interference relations relating to this BTS.

7.4.6.3 Switching DFCA mode from STANDBY to DFCA hopping

Before the DFCA mode can be changed from STANDBY to DFCA hopping the BTS must be in a locked state. After the DFCA mode has been changed the BTS must be unlocked for the changes and reconfigurations to take place.

When this change takes place the BSC checks that the following requirements are met:

• The BTS site type must be either UltraSite or MetroSite.

• The BTS synchronization status must be “LMU synchronized”. (Note: In BSS11 this check is not done).

This change causes the DFCA TRXs in the BTS to be reconfigured to the DFCA mode and the DFCA channel selection algorithm starts working for the channel assignments on the DFCA TRXs. The BSC also starts updating the DFCA Radio Resource Table both internally and externally to the neighboring BSCs if there are any foreign DFCA cells listed in the BIM tables of the cell in question.

7.4.6.4 Switching DFCA mode from DFCA hopping to STANDBY

Before the DFCA mode can be changed from DFCA hopping to STANDBY the BTS must be in a locked state. After the DFCA mode has been changed the BTS must be unlocked for the changes and reconfigurations to take place.

This change causes the DFCA TRXs in the BTS to be reconfigured to the normal mode and conventional channel selection algorithm starts working for all the channel assignments in the BTS. The BSC also stops updating the DFCA Radio Resource Table both internally and externally.

The BIM updating process and maintenance of the BIM tables continues as normal.


61 (93)

7.4.6.5 Switching DFCA mode from DFCA hopping to OFF

Before the DFCA mode can be changed from DFCA hopping to OFF the BTS must be in a locked state. After the DFCA mode has been changed the BTS must be unlocked for the changes and reconfigurations to take place.

This change causes the DFCA TRXs in the BTS to be reconfigured to the normal mode and conventional channel selection algorithm starts working for all the channel assignments in the BTS. The BSC also stops updating the DFCA Radio Resource Table both internally and externally. Also the BIM updating process and maintenance of the BIM tables is stopped and the BIM tables and BIM entries related to the cell in question are deleted.

7.5 Automatic configuration changes

DFCA operation requires that the network is working synchronized. To implement the synchronization the BCFs receive clock synchronization signal from the LMU unit. There can however be situations where the LMU clock synchronization is temporarily lost. In these situations the BSC automatically switches between synchronized and unsynchronized operation in the BTSs that are controlled by the BCF that has lost its synchronization. This means temporary changes in DFCA operation.

The synchronization of the network is monitored by the BSS11 feature 'Recovery for BSS and Site Synchronization' which also provides the synchronization status information needed for DFCA purposes. In DFCA the operation is regarded as synchronized when the synchronization status is “LMU synchronized” and as unsynchronized with all other synchronization status values.

Another thing that is required for the efficient DFCA operation in a BTS is that there are no detected problems in connections to the neighboring BSCs controlling DFCA cells that cause interference to the BTS. If such connection failures take place during DFCA operation it leads to the same recovery actions as when the synchronization of a BTS is lost.

7.5.1 Loss of synchronization

As the network synchronization is lost in a BTS the DFCA TRXs of the BTS are set in the random RF frequency hopping mode.

The configuration change to the random frequency hopping involves the unsynchronized mode DFCA MA list that is taken into use in the DFCA TRXs. The HSN to be used is defined by the formula (BCCH frequency MOD 63)+1. This guarantees that a different HSN will be used in co-located BTS that may still be following the same clock. Also this takes care of the HSN re-use distance corresponding to the BCCH reuse distance.

As BSC gets informed about the loss of synchronization it blocks the BTSs that are in the control of the BCF whose synchronization was lost. BSC changes the BTSs to unsynchronized random frequency hopping operation, performs the related reconfiguration actions, requests restart for the BTSs and opens them again for use.

7.5.2 Return of synchronization

When the synchronization status of a BTS in DFCA hopping mode changes back to synchronized, the normal DFCA mode of operation is restored for the DFCA TRXs.


62 (93)

7.5.3 Loss of inter BSC connection

Signaling Connection Control Part (SCCP) of the BSC-BSC interface informs the connected BSC if the connection to another BSC breaks down. On receiving of the information about the BSC-BSC connection failure the BSC determines first the foreign cells that are controlled by the BSC to which the connection was lost. The BSC checks the incoming interference BIM table of every local DFCA cell to see if any of the foreign cells is included in the BIM and if the C/I for any such foreign cell is lower than the highest defined soft blocking C/I threshold in the BSC. If this is the case the communication fault may cause excessive uncontrolled interference in the DFCA TRXs of the cell because the DFCA algorithm is operating without the up-to-date knowledge on some of its interfering cells.

If an inter BSC communication fault leads to uncontrolled interference in a DFCA BTS then the BSC sets an alarm 'DFCA use prevented due to inter BSC connection failure'. As it is very likely that there are several BTSs that are affected by the failure, it is possible to inquire a list of affected BTSs by an MML command (not in BSS11; BSS11 raises an alarm per BTS) . After raising the alarm the DFCA TRXs in the affected BTSs are configured to random frequency hopping mode quite similarly as is done when the synchronization is lost (see Section 7.5.1). Short inter BSC comminication breaks may be allowed without need for immediate actions. The allowed break duration can be adjusted by MML.

7.5.4 Return of inter BSC connection

When receiving the information on the return of inter BSC communication to a remote BSC BSC determines if there are cells where DFCA usage can be recovered due to the detected inter BSC communication status change. To return the DFCA operation in a DFCA BTS the BSC cancels the alarm 'DFCA use prevented due to inter BSC connection failure' and configures the BTSs automatically back to the DFCA operation.

7.5.5 TRX fault on a DFCA BTS

When a fault is found in a regular non-DFCA TRX, a DFCA TRX is configured to replace the faulty TRX. The faulty TRX leading to the recovery actions can be any of the regular layer TRXs including the BCCH TRX. The TRX DFCA indication parameter is also swapped during the recovery actions.

The swapped faulty TRX remains blocked with the original configuration of the DFCA TRX. As the TRX fault is cancelled the DFCA TRX is de-blocked and returned to operation without any further configuration actions.

There are no reconfiguration actions needed due to a fault of a DFCA TRX.

7.6 BCCH/BSIC change & Co-BCCH blind spot avoidance

As explained in Section 4.2.2 DFCA identifies the potentially interfering cells by the BCCH & BSIC combination that should be unique in the continuous DFCA area. However, if BCCH & BSIC conflicts happen the BSC can deal with them automatically as explained in Section 5.6 and inform the operator of the conflict by raising an alarm. During a conflict situation the system will cease the BIM updates for the conflicting BIM entries and the C/I estimations for interference concerning the conflicting cells will be done based on the BIM information only (i.e. the measurement reports are not utilized due to ambiguity in the cell identification).


63 (93)

Similarly, if a DFCA cell that was previously found to be a potential interference source or victim changes to use the same BCCH frequency as the cell in question then the BSC will automatically detect a co-BCCH conflict situation. In this situation the corresponding BIM entries will be locked to ensure that the interference relation is maintained and the DFCA C/I estimation concerning the co-BCCH cell will be done based on the BIM entries only (i.e. the measurement reports are not utilized as the co-BCCH cell cannot be reported).

Changes in the BCCH frequency plan during DFCA operation actually help to remove the possibly detrimental impacts of interference blind spots caused by co-BCCH frequency conflicts as after a BCCH plan change it is likely that a previously conflicting and therefore invisible interfering cells become visible to DFCA C/I estimation as the co-BCCH conflict is removed. Any new co-BCCH conflicts introduced by the BCCH plan change are less critical as DFCA already has determined the statistical interference level and stored in the BIM tables before the co-BCCH conflict happened.

7.7 Interactions with other features

7.7.1 Intelligent Underlay Overlay (IUO) & Intelligent Frequency Hopping (IFH)

DFCA can't be used in a cell/segment where IUO or IFH is activated, but there is no problem in having DFCA BTSs and IUO BTSs controlled by the same BSC.

7.7.2 (E)GPRS

The packet switched territory is only allowed in the BCCH and regular TRXs, not in DFCA TRXs. The same applies to PCCH/PBCCH.

7.7.3 HSCSD

Supported in regular layer, HSCSD not supported in DFCA TRXs.

7.7.4 Half Rate and AMR

Supported. Additionally DFCA introduces a concept of 'forced HR mode' as discussed in Section 4.6.2.

7.7.5 Dynamic SDCCH

SDCCH timeslots are not allowed in a DFCA TRX. This also applies to the dynamic SDCCH feature.

7.7.6 Intelligent Coverage Enhancement (ICE)

ICE is not supported in a DFCA cell/segment.

7.7.7 Enhanced Coverage by Frequency Hopping

Not supported in a DFCA cell/segment.

7.7.8 Extended Cell Radius

Not supported in DFCA cell/segment.


64 (93)

7.7.9 Common BCCH and Segments

Works normally with DFCA.

7.7.10 Queuing

Works normally with DFCA. In soft blocking case queuing is not used for DFCA resources, only for regular resources.

7.7.11 Directed Retry


7.7.12 FACCH Call Setup

FACCH call setup to a DFCA TRX is not possible. FACCH setup to a regular TRX works normally.

7.7.13 Pre-emption

Works normally in regular layer, not supported in DFCA TRXs.

7.7.14 Power Control

The BSC may make initial power reduction based on C/I criteria for the connection on the DFCA TRXs as described is Section 4.3. After initial power reduction the power control functions normally.

7.7.15 Power Optimization in Handover

Not supported for connections that are to be assigned to DFCA TRXs. BSC makes initial C/I based power control decisions for a DFCA connection. For a connection on a regular TRX power optimisation works normally.

7.7.16 Interference Band Recommendation

This feature is not used for TRXs in DFCA mode. For connections on regular TRXs the interference band recommendations are used normally. DFCA algorithm does not use UL idle channel measurements.

7.7.17 Dynamic Hot Spot

Not supported in a DFCA cell/segment.

7.7.18 AMH & DADL/B

For the load used in AMH and in DADL/B triggers, the regular and DFCA channels are seen as one resource group.

7.7.19 Antenna Hopping

Antenna Hopping (AH) cannot be used in DFCA TRXs. AH is possible on the regular layer provided that it is configured to be different BTS object of a cell/segment from DFCA layer.


65 (93)

7.7.20 Automated Planning Enhancements

If both DFCA and Automated Planning Enhancements are used in a cell/segment then the BSC uses the BA list definitions that the operator has made for the Automated Planning Enhancements feature. In order to ensure the proper functioning of DFCA the operator has to include all the BCCH frequencies in the used MeasBAList.

7.7.21 Enhanced Measurement Report

EMR is supported. Additionally to the defined adjacent cells, the DFCA functionality adds the BSICs of the potentially interfering cells in the BIM to the EMR BSIC list thus ensuring that each enhanced measurement report can contain measurements from as many neighboring cells as possible.

As explained in Section 7.4.4 a special BA list containing all the BCCH frequencies should be used in order to ensure that mobiles report all cells that are potentially interfering irrespecticely if they have been defined as adjacent cells for handover purposes or not. In case of EMR the reporting of cells with “invalid BSIC” (i.e. the cells that have not been defined as adjacent) should be allowed. As a consequence the number of reported cells in one EMR report may be reduced. The extent of this reduction depends on how many cells that have not been defined as adjacent are reported. Table 10 presents the total number of reported cell in one enhanced measurement report with different number of “invalid BSIC” cells (i.e. cells not defined as adjacent) and varying reporting bitmap size (corresponding toi the total number of defined adjacent cells).

Table 10. Total number of reported cells in one EMR Invalid BSIC cells reported

Bitmap size 0 1 2 3 4 5 632 12 10 8 6 4 x x24 13 11 9 7 5 x x16 15 13 11 9 7 5 x8 8 9 10 10 8 60 0 1 2 3 4 5 x

x

7.7.22 FER Measurement


7.7.23 Location Services

Work normally with DFCA.

7.7.24 Nokia Smart Radio Concept (IDD, 4UD, IRC)

Supported in regular layer. IDD and 4-way UL diversity are not supported in DFCA TRXs, IRC is supported in all TRXs that have EDGE capable HW.

7.7.25 BTS HW Issues

DFCA is available in UltraSite and in MetroSite BTSs, not in Talk Family of BTSs. (MetroSite support not yet in BSS11)


66 (93)

Wide Band Combiner is required for DFCA BTSs. RTC is not supported.

8. DFCA PARAMETERS

8.1 BSC level parameters

8.1.1 Connection type specific C/I targets

These parameters specify the target C/I for different connection types that are used in the DFCA C/I estimation and channel selection process described in Section 4.3.

Name Range Default MML Modification notes

C/I target FR 0 .. 63 dB 14 dB EEH, EEO online

Target C/I value for FR and EFR speech connections as well as for CS data connections of up to 9.6 kbit/s.

C/I target HR 0 .. 63 dB 14 dB EEH, EEO online Target C/I value for HR speech connections.

C/I target AMR FR

0 .. 63 dB 8 dB EEH, EEO online

Target C/I value for AMR FR speech connections.

C/I target AMR HR


Target C/I value for AMR HR speech connections.

C/I target 14.4 0 .. 63 dB 16 dB EEH, EEO online

Target C/I value for 14.4 kbit/s CS data connections.

8.1.2 C/I target UL offset


C/I target UL offset

-31... 31 dB

0 dB EEH, EEO online

This parameter defines an offset that is added to the C/I targets and soft blocking C/I limits of all connection types when uplink interference checks are performed. It can be


67 (93)

used to compensate for any differences between UL and DL link level performance.

8.1.3 Connection type specific soft blocking C/I limits

These parameters specify the soft blocking C/I limits for different connection types. The Soft blocking C/I limits are used in the channel selection process described in Section 4.4.6.


Soft blocking C/I FR


This parameter sets the minimum acceptable C/I value for FR and EFR speech connections as well as for CS data connections of up to 9.6 kbit/s.

Soft blocking C/I HR


This parameter sets the minimum acceptable C/I value for HR speech connections.

Soft blocking C/I AMR FR


This parameter sets the minimum acceptable C/I value for AMR speech connections.

Soft blocking C/I AMR HR


This parameter sets the minimum acceptable C/I value for AMR HR speech connections.

Soft blocking C/I 14.4


This parameter sets the minimum acceptable C/I value for 14.4 kbit/s CS data connections.


68 (93)

8.1.4 Connection type specific soft blocking C/N limits

These parameters specify the soft blocking C/N limits for different connection types. The Soft blocking C/N limits are used cell access control as described in Section 4.4.7.


Soft blocking C/N FR


This parameter sets the minimum acceptable carrier to noise ratio for FR and EFR speech connections as well as for CS data connections of up to 9.6 kbit/s.

Soft blocking C/N HR


This parameter sets the minimum acceptable carrier to noise ratio for HR speech connections.

Soft blocking C/N AMR FR


This parameter sets the minimum acceptable carrier to noise ratio for AMR speech connections.

Soft blocking C/N AMR HR


This parameter sets the minimum acceptable carrier to noise ratio for AMR HR speech connections.

Soft blocking C/N 14.4


This parameter sets the minimum acceptable carrier to noise ratio for 14.4 kbit/s CS data connections.

8.1.5 Parameters related to BIM update

These parameters govern the BIM update process described in Section 5.



69 (93)

BIM confidence probability

50% to 95% 90% EEH, EEO online

This parameter sets the level of confidence for the estimation used to build a background interference matrix. It gives the share of users experiencing a C/I equal to or greater than the C/I contained in the matrix. This parameter has the biggest impact when BIM scaling cannot be done due to lack if measured C/I values such as in case of inter BSC handover. In those cases the BIM confidence probability can have significant impact on the aggressiveness of initial power reductions and soft blocking. See section 5.3.

BIM interference threshold


This parameter indicates an upper limit for C/I values that are considered as relevant for the interference estimations. C/I values above the Interference threshold can be discarded during the BIM update procedure. See section 5.4.

BIM update scaling factor

0.0 to 1.0 (0.1 step size)

0.5 EEH, EEO online

The weighting factor for the C/I value defined during the latest BIM update period when this is combined with the long term C/I statistics. See section 5.4.2.


70 (93)

BIM update period

10 min to 60 min (10 min step

size); 2h to 6h (1h step size);

12h and 24 h, 0 turns off the BIM

update procedure

60 min EEH, EEO online

This parameter defines the period between successive background interference matrix updates. Thus, it is the length of the data collection period for a single BIM update. See section 5.1.

BIM update guard time.

0…63, value range for guard time is

0…62 BIM update periods,

value 63 indicates that a neighbor is not removed from

the BIM even if the neighbor is not included in any new BIM

update

10 EEH, EEO online

Defines the number of BIM update periods that the DFCA algorithm allows without removing a neighbor cell from the BIM table of a DFCA cell even if the neighbor is not included in the BIM updates made during that time. See section 5.5.

8.1.6 Expected BSC-BSC interface delay


Expected BSC-BSC interface delay*

0 .. 2 seconds, 10 ms steps

50 ms EEH, EEO online

This parameter is used for the channel assignment control to prevent simultaneous conflicting channel allocations in neighboring BSCs.

* Not available in BSS11.

8.1.7 DFCA channel allocation method


DFCA channel allocation method

0, primary target of a DFCA assignment is a MA, MAIO and tsl

0 EEH, EEO online The parameter defines if DCFA assignments are made primarily to


71 (93)

method combination with the highest C/I difference from the target level 1, primary target of a DFCA assignment is a MA, MAIO and tsl combination on the connection specific target C/I level.

channels having the connection specific C/I target level or to channels, which have the highest positive C/I difference from the target level.

See section 4.4.

8.1.8 Mobile allocation frequency list state


MA list state

0 = normal (non-DFCA), 1 = DFCA in use, 2 = DFCA out of use

'0 = non-DFCA' for MA id values 1 - 660 '2 = DFCA out of use' for MA id values 661-724

EBE, EBT

online.

Indicates if a mobile allocation frequency list is a normal or a DFCA MA list and further in the DFCA case if the list is ready for use or if it is only created but not yet employed by the DFCA algorithm.

See section 7.4.2.

(BSS11: MA id = 1-255, DFCA MA id = 256-287)

8.1.9 Parameters related to the LAC to SPC mapping table

These parameters are used to define the LAC to SPC mapping table that is used to identify the BSCs in the BSC-BSC interface as described in Section 6.2.



72 (93)

Mapping entry index

1…64 different LACs 1…6 different SPCs

The next free index

EEC, EEF, EED, EES

read only

Indication for a mapping relation between a location area code and a signaling point code in the LAC to BSC address mapping table.

Location area code 1…65533 undefined EEC, EEF online

A location area code that will be mapped with a signaling point code.

Signaling point code

0…0xFFFFFFFF undefined EEC, EEF online

A signaling point code that will be mapped with a location area code

8.2 BTS level parameters

8.2.1 DFCA mode


DFCA mode

0 = off 1 = standby 2 = DFCA hopping

0 = off EQO, EQM, EFL

Between 0 and 1 => online

To/from mode 2 => BTS locking

This parameter defines the operational mode of a DFCA BTS.

See section 7.4.6.

8.2.2 DFCA MA list ids


DFCA MA list ids

A list of values 661-724 - EQO, EQA BTS locking

IDs of the DFCA MA lists attached to a BTS for the synchronized operation. See section 7.4.3.

(BSS11: DFCA MA id =


73 (93)

256-287)

8.2.3 DFCA unsynchronized mode MA list


DFCA unsynchronized mode MA list

0 - 660 0 indicates that no MA list is attached

to the BTS

0 EQO, EQA BTS locking

The MA list to be used in the DFCA TRXs if a BTS in "DFCA hopping" mode loses synchronization. This is a reference to any existing normal (non-DFCA) MA list defined in BSC. See sections 7.5.1 and 7.5.3.

(BSS11: MA id = 1-255)

8.2.4 Forced HR mode related parameters

These parameters control the usage and behavior of the forced HR mode *). More information in section 4.6.2.


Forced HR mode C/I averaging period*

5 to 30 minutes with 5 min step size 15 min

EQO, EQM

online

The period for calculating an average C/I of the incoming DL C/I estimates used in DFCA channel assignments. This BTS specific average value is used in deciding between full rate and half rate in DFCA channel assignment.


74 (93)

Forced HR mode C/I threshold*

0 .. 63 dB 0 dB EQO, EQM online

If the BTS level average C/I is below the threshold defined by this parameter, then HR will be preferred for all the DFCA speech channel assignments of non-AMR requests

Forced AMR HR mode C/I threshold*


If the BTS level average C/I is below the threshold defined by this parameter, then HR will be preferred for all the DFCA speech channel assignments of AMR requests.

Forced HR mode hysteresis*


This parameter defines how many decibels the BTS level C/I average must be above the corresponding forced HR mode C/I threshold in order for the forced HR mode to be switched off.

* Not available in BSS11.

8.2.5 UL/DL noise levels for C/N based access control


DL noise level -120 ... –80 dB -117 dB EQO, EQM

online

This parameter is used in the C/N check where cell access control is applied based on channel type specific C/N soft blocking limit. See section 4.4.7.


75 (93)

UL noise level -120 ... –80 dB -117 dB EQO, EQM

online

This parameter is used in the C/N check where cell access control is applied based on channel type specific C/N soft blocking limit. See section 4.4.7.

8.3 TRX level parameters

8.3.1 DFCA indication


DFCA indication. F = non-DFCA TRX T = DFCA TRX F

ERC, ERM, ERO

If the DFCA mode of the BTS is "off" or "standby" the modification can be made online. If the DFCA mode of the BTSis "DFCA hopping" the modification requires TRX locking.

This parameter defines whether a TRX belongs to the regular or to the DFCA layer when the DFCA mode of the BTS is "DFCA hopping". See section 3.

9. DFCA PERFORMANCE COUNTERS AND MEASUREMENTS

9.1 Handover measurement

Note: Counters presented in this section are not implemented in BSS11.

New counters related to the new handover case where an intra cell handover is made from a regular TRX to a DFCA TRX based on regular layer TCH load.

9.1.1 HO ATTEMPT TO DFCA TRX

• Indicates the times BSC has started an intra cell handover from a regular layer TRX to DFCA TRXs of the cell based on load on regular TRXs.


76 (93)

No COUNTER NAME

XXXX HO ATTEMPT TO DFCA TRX EXPLANATION: Number of attempts to perform a TCH-TCH handover from a regular layer TRX to DFCA resources of the cell. UPDATED: Whenever a handover is initiated to move a call from a TCH on a regular TRX to a DFCA TRX based on the high TCH load of the regular area.

9.1.2 UNSUCC HO TO DFCA TRX

• Indicates the number of unsuccessful handover attempts from a regular layer TRX to DFCA TRXs of the cell when the handover has been initiated based on the high TCH load on regular TRXs.

No COUNTER NAME

XXXX UNSUCC HO TO DFCA TRX EXPLANATION: Number of unsuccessful attempts to perform a TCH-TCH handover from a regular layer TRX to DFCA resources of the cell. UPDATED: When a handover to move a call from a regular TRX based on load to a DFCA TRX fails.

9.1.3 SUCC HO TO DFCA TRX

• Indicates the number of successful handovers from a regular layer TRX to DFCA TRXs of the cell when the handover has been initiated based on the high TCH load on regular TRXs.

No COUNTER NAME

XXXX SUCC HO TO DFCA TRX EXPLANATION: Number of successful TCH-TCH handovers from a regular layer TRX to DFCA resources of the cell. UPDATED: When a handover to move a call from a regular TRX based on load to a DFCA TRX succeeds.

9.2 Traffic measurement


These counters are related to the mechanism that supervises DFCA channel assignments in a BSC and it’s neighboring BSCs looking for co-incident conflicting DFCA channel assignments that were made simultaneously by two DFCA algorithms in different BSCs. The channel assignment conflicts are


77 (93)

detected before the assignments are executed, so the impact of a conflict to the end user is just a small additional delay in the channel assignment.

9.2.1 TCH REL DUE TO BSC BSC CONFLICT CALL

No COUNTER NAME

01XXX TCH REL DUE TO BSC BSC CONFLICT CALL EXPLANATION: Number of TCHs releases due to BSC-BSC conflict. UPDATED: When a TCH is released.

9.2.2 TCH REL DUE TO BSC BSC CONFLICT TARGET

No COUNTER NAME

01XXX TCH REL DUE TO BSC BSC CONFLICT TARGET EXPLANATION: Number of TCHs releases due to BSC-BSC conflict. . UPDATED: When a TCH is released.

9.3 Resource availability measurement


These counters indicate the time a BTS has been in different forced HR modes.

9.3.1 TIME IN FORCED DFCA HR MODE

No COUNTER NAME

X2XXX TIME IN FORCED DFCA HR MODE EXPLANATION: The counter counts the time the BTS has been in the forced HR mode. UPDATED: The BSC updates the counter when the mode starts and ends. COUNTER TYPE: Time. UNIT: 10 ms DEPENDENCIES WITH OTHER COUNTERS: The counter X2XXX is updated together with this counter


78 (93)

9.3.2 TIME IN FORCED DFCA AMR HR MODE

No COUNTER NAME

X2XXX TIME IN FORCED DFCA AMR HR MODE EXPLANATION: The counter counts the time the BTS has been in the forced AMR HR mode. UPDATED: The BSC updates the counter when the mode starts and ends. COUNTER TYPE: Time. UNIT: 10 ms DEPENDENCIES WITH OTHER COUNTERS: The counter X2XXX is updated together with this counter

9.3.3 TIME IN FORCED DFCA HR AND AMR HR MODE

No COUNTER NAME

X2XXX TIME IN FORCED DFCA HR AND AMR HR MODE

EXPLANATION: The counter counts the time the BTS has been simultaneously in HR forced mode and in AMR HR forced mode. UPDATED: The BSC updates the counter when the mode starts and ends. COUNTER TYPE: Time. UNIT: 10 ms DEPENDENCIES WITH OTHER COUNTERS: The counter X2XXX is updated together with this counter

9.4 BSC-BSC measurement

This is a new measurement with object level SPC address.


79 (93)

9.4.1 BSC- BSC DELAY

No COUNTER NAME

102000 BSC- BSC DELAY EXPLANATION: Average delay between Radio Resource Managers of the BSCs UPDATED: The BSC measures and updates this counter in one-minute periods.

COUNTER TYPE: Average. UNIT: 10 ms DEPENDENCIES WITH OTHER COUNTERS: The counter 102001 is updated together with this counter

9.4.2 BSC - BSC DENOMINATOR 1

No COUNTER NAME

102001 BSC - BSC DENOMINATOR 1 EXPLANATION: Denominator for the 102000 counter UPDATED: DEPENDENCIES WITH OTHER COUNTERS: The counter 102000 is updated together with this counter

9.4.3 BSC- BSC PEAK DELAY

No COUNTER NAME

102002 BSC- BSC PEAK DELAY EXPLANATION: : Peak delay between Radio Resource Managers of the BSCs. UPDATED: The BSC measures and updates this counter in one minute period. COUNTER TYPE:Peak. UNIT: 10 ms


80 (93)

9.5 DFCA measurement

This is a new measurement with object level BTS_id + connection type (FR/EFR, HR, 14.4, AMR or AMR HR).

9.5.1 C/I TARGET

No COUNTER NAME

100000 SOFT BLOCKING C/I EXPLANATION: The minimum acceptable C/I value for the connection type (FR/EFR, HR, 14.4, AMR or AMR HR). UPDATED: The RRMPRB updates this at the end of measurement period. COUNTER TYPE: Variable UNIT: dB LIMITS: 0..63

9.5.2 C/I TARGET UL OFFSET

No COUNTER NAME

100001 C/I TARGET UL OFFSET EXPLANATION: The offset that is added to the C/I targets and soft blocking C/I limits of all connection types when uplink interference checks are performed. UPDATED: The RRMPRB updates this at the end of measurement period. COUNTER TYPE: Variable UNIT: LIMITS: -31..31

9.5.3 DFCA assignment incoming interference C/I counters

• These counters record the most restrictive incoming interference C/I for the assigned DFCA channels based on DFCA C/I estimate. The counters record the C/I relative to the C/I targets set for each connection type.


81 (93)

No COUNTER NAME

100002 DFCA C/I TG UL EXPLANATION: The number of DFCA assignments where the uplink C/I level for the connection equals the target C/I. UPDATED: The BSC updates this counter according to the incoming interference C/I

No COUNTER NAME

100003 DFCA C/I TG DL EXPLANATION: The number of DFCA assignments where the downlink C/I level for the connection equals the target C/I. UPDATED: The BSC updates this counter according to the incoming interference C/I

There are a total of 37 of these counters for each direction (UL and DL) ranging from <-15dB, -15dB, -14dB.....-1dB, Target, +1dB....+19dB, +20dB, >+20dB

9.5.4 DFCA assignment outgoing interference C/I counters

• These counters record the most restrictive outgoing interference C/I for the assigned DFCA channels based on DFCA C/I estimate. The counters record the C/I relative to the C/I targets set for each connection type.

No COUNTER NAME

100078 WORST INTERF C/I TG UL EXPLANATION: Number of DFCA assignments where uplink C/I value of the most restrictive connection equals the target C/I value. UPDATED: The BSC updates this counter according to the outgoing interference C/I


82 (93)

No COUNTER NAME

100079 WORST INTERF C/I TG DL EXPLANATION: Number of DFCA assignments where downlink C/I value of the most restrictive connection equals the target C/I value. UPDATED: The BSC updates this counter according to the outgoing interference C/I

There are a total of 37 of these counters for each direction (UL and DL) ranging from <-15dB, -15dB, -14dB.....-1dB, Target, +1dB....+19dB, +20dB, >+20dB

9.5.5 SUCC DFCA ASS

No COUNTER NAME

100154 SUCC DFCA ASS EXPLANATION: Number of successful DFCA assignments. UPDATED: The BSC updates this in TCH allocation when a DFCA TCH is allocated.

9.5.6 SUCC DFCA ASS HIGH LOAD

No COUNTER NAME

100155 SUCC DFCA ASS HIGH LOAD EXPLANATION: Number of successful DFCA assignments in high MCMU load situation. Due to high MCMU load the DFCA channel search is limited and therefore the most optimal DFCA channel cannot be necessarily allocated. UPDATED: The BSC updates this in TCH allocation when a DFCA TCH is allocated.

9.5.7 SOFT BLOCKED DFCA ASS DUE TO C/I

No COUNTER NAME

100156 SOFT BLOCKED DFCA ASS DUE TO C/I EXPLANATION: Number of soft blocked DFCA assignments due to C/I. UPDATED: Updated in a TCH allocation if a DFCA assignment fails due to C/I soft blocking threshold


83 (93)

9.5.8 SOFT BLOCKED DFCA ASS DUE TO C/N

No COUNTER NAME

100157 SOFT BLOCKED DFCA ASS DUE TO C/N EXPLANATION: Number of soft blocked DFCA assignments due to C/N. UPDATED: Updated in a TCH allocation if a DFCA assignment fails due to C/N soft blocking threshold

9.6 DFCA Assignment Measurement

This is a new measurement with object level BTS_id + MA_id + MAIO.

9.6.1 DFCA ASSIGNMENTS

No COUNTER NAME

101000 DFCA ASSIGNMENTS EXPLANATION: Number of DFCA assignments using the specified combination of MA list and MAIO. UPDATED: After each successful DFCA assignment

9.7 BSC Level Clear Code (PM) Measurement


A new success counter in the BSC Level Clear Code (PM) Measurement is introduced due to the new intra cell handover type.

9.7.1 INTRA HO TO DFCA

• Indicates the number of successful handovers from the regular frequency layer to the DFCA layer of the same cell when the handover has been initiated based on the high TCH load of the regular layer.


84 (93)

No COUNTER NAME

XXXX INTRA HO TO DFCA EXPLANATION: Number of completed and successful TCH-TCH handovers from a regular TRX to a DFCA TRX of the cell. UPDATED: When a call has been successfully transferred from a regular TRX to a DFCA TRX based on the high TCH load of the regular area. DEPENDENCIES WITH OTHER COUNTERS: Handover Measurement counter SUCC HO TO DFCA TRX is updated along with this counter.

9.7.2 INTER BSC DFCA ASSIGNMENT SUCC

No COUNTER NAME

051XXX INTER BSC DFCA ASSIGNMENT SUCC EXPLANATION: Number of successful DFCA assignments in an inter-BSC communication procedure. UPDATED: When the timer set up after a BSC BSC conflict control process has expired. The assignment has been completed successfully and no conflict between BSCs have been generated.

9.7.3 INTER BSC DFCA ASSIGNMENT REJ

No COUNTER NAME

051XXX INTER BSC DFCA ASSIGNMENT REJ EXPLANATION: Number of rejected DFCA assignments after a BSC BSC conflict control communication procedure. UPDATED: Each time a DFCA assignment process is finally aborted due to a coincident channel assignment processes

10. DFCA PERFORMANCE SIMULATIONS

10.1 General

10.1.1 Simulation tool

All the simulations presented here have been executed with a dynamic network-level GSM/(E)GPRS simulator called SMART, developed at Nokia Research Center. This simulator models the system behavior of the network in a highly detailed level, and contains most of the features existing in real


85 (93)

networks: call control, handovers, power control, DTX, etc. The link level modeling is based on previous results from a link-level simulator, and so the step from physical level (C/I) to network level is performed by means of mapping tables, one for every kind of codec and/or modulation scheme.

The simulator contains several cells that can be loaded with a desired amount of users, which will be moving around the network, trying to carry out voice conversations (or data transfer in case of GPRS). SMART will establish calls, handle the radio resources, take care of ongoing connections, and provide all the statistics needed to evaluate the performance.

10.1.2 Simulation settings

Real mobile networks are placed in an irregular way, with antennas at different heights and different angles; with obstacles, hotspots, different mobility users, etc.

We have tried to emulate a realistic environment, taking into account some of these effects. There are, of course, lot things that have not been modeled, but the interesting part is to study the behavior of a simulated GSM system that is no longer ideal.

The network layout corresponds to a real GSM network in Brisbane, Australia. Following pictures show the environment and cell locations (left), and the cell dominance map from the simulator (right):

Figure 29. Simulated network

With the information retrieved from the real operational network, we have modeled the traffic distribution in the network. This distribution was very uneven, as usual in real life, and there were cells with very small number of TRXs, and some others carrying huge amount of traffic.

For our simulation purposes, a maximum network load was established, which guaranteed that blocking was under 4% in the more loaded cells (only 4 cells were above 2% blocking rate). At this point, the average network load was 40%, and the average blocking rate 0.14%. The average number of hopping TRXs per cell was 3.92.

Two different MAIO plans were done, one for a 12 frequency band test, and another one for a 9 frequencies test. These plans were done manually, using a 1/1 hopping scheme, and trying to avoid as


86 (93)

much as possible intrasite interference. Obviously, these plans are not perfect, as happened in the ideal network, and this fact will make DFCA gains to be improved.

The following tables describe the general settings used in the simulations.

Scenario

Antenna type Sectorized antennas, 65º 3dB beam width

Antenna high 15 m

Mobile speed 3 km/h

FH ON (1/1 reuse)

PC ON

DTX ON (ul & dl), speech activity factor 0.5

Uplink Diversity ON

Propagation

Model Macrocell environment

Frequency Band GSM 900

Slow fading std 6 dB

SF corr. Distance 50 m

SF corr. Coef. 0.367879

Fast fading model Jakes900 file (narrowband channels)

Frequency Hopping Settings

Correlation model Correlation within a frequency, uncorrelated between frequencies

Reference case Random hopping, 1 MA-list (reuse 1/1)

DFCA case Cyclic hopping, 1 MA-list

Power control Settings

PC_Algorithm RXLEV-RXQUAL based

levqualBasedL_RXLEV 20 (-90 dBm)

levqualBasedU_RXLEV 62 (-48 dBm)

levqualBasedL_RXQUAL 3 (EFR, AMR-HR), 5(AMR-FR)

levqualBasedU_RXQUAL 2 (EFR, AMR-HR), 4(AMR-FR)

Handover Settings

IntracellHO check period 6 (SACCH)


87 (93)

SimpleHOMargin (dB) 3

SimpleHOPeriod 10 (SACCH)

MS services settings

BCCH preference BCCH blocked

Used speech codecs EFR, AMR FR 5.9 and AMR HR 5.9

Call length 120 s (mean), 1 s (minimum)

10.1.3 Capacity gain indicators

Network performance will be presented by means of Frame Erasure Rate (FER) vs. Effective Frequency Load (EFL) curves, in which a reference case will be compared to the new DFCA behavior in order to get the capacity gain figures.

FER

FER samples have been collected for every user in an active call, in periods of 2 seconds (up to 1% FER granularity). According to real life statistics, a FER limit of 4%1 has been chosen, so that FER samples higher than 4% will be considered as bad quality samples.

To simplify the analysis of the network performance, dropping calls has been disabled. This way, the only indicator for bad quality in the network are the FER samples collected.

EFL

EFL (Effective Frequency Load) is a measure that relates traffic information with the effective use of the available resources (timeslots and frequency spectrum):

ºº effR

HWLcellTRXn

freqsnHWLEFL =×= , (Eq.16)

where HWL stands for Hardware Load (%) and Reff is the effective reuse of the network.

Looking at the above formula, it can be seen that there are two ways to achieve a certain EFL point in the network:

1. Fixed Bandwidth: increase the hardware load admitting more users in the network.

2. Variable Bandwidth: the effecting reuse is reduced by cutting down the available spectrum (number of frequencies).


88 (93)

The second method is not recommended, because reducing the number of frequencies has secondary effects such as frequency diversity losses or adjacent channel interference increase because of non-optimal MAIO planning. These effects are only present when the number of frequencies is quite reduced (below 12). Although the major part of the simulations have been run applying the first method, in some cases the second method was used to speed up simulations time.

In any case, the hardware load limit has been established to be so that gets up to a 2% blocking probability. When this blocking rate is exceeded, a higher number of TRXs have to be introduced, so the MAIO planning may change as well between different EFL points.

BCCH layer has been setup to carry no traffic, so we are considering only hopping layer statistics. However, when printing the results the whole BW used for the simulation is considered; assuming that BCCH reuse is 4/12 the number of frequencies used for the TCH layer can be calculated (ie, 5 MHz spectrum means 12 frequencies for TCH layer).

Gain Figures

The results for downlink and uplink performance are presented independently, because usually their performance is different. When calculating the capacity gain, theoretically the most restrictive link should be selected in every case; in practice we won't look too much at the uplink performance2, because it is known that in real systems uplink is getting profit of much more advanced processing than downlink (diversity, IRC, …).

Capacity gain figures at a given quality percentage outage can be directly obtained from the FER vs EFL curves:

) %() %() %(

outagexEFLoutagexEFLoutagexEFL

GainREF

REFNEW −= (Eq.17)

Since the hardware load in the equation 16 is directly proportional to the carried traffic, the EFL gain percentage shows how much more traffic can be carried with same quality.

In the following picture an example is presented for a 2% bad quality users outage:


89 (93)

2 In any case, downlink is most of the times the limiting link

Example FER vs EFL CurveDownlink FER(>4.2%)

00.5

11.5

22.5

33.5

44.5

7.00% 10.00% 13.00% 16.00% 19.00% 22.00%

EFL (%)

% u

nsat

isfie

d us

ers

Curve 1 Curve 2

Gain = (18-16)/16 = 12.5 %

Example FER vs EFL CurveUplink FER(>4.2%)

00.5

11.5

22.5

33.5

44.5

7.00% 10.00% 13.00% 16.00% 19.00% 22.00%

EFL (%)

% u

nsat

isfie

d us

ers

Curve 1 Curve 2

Figure 30. Example of capacity gain calculation

10.2 Simulation results

10.2.1 EFR performance

For EFR performance tests the 12 frequencies hopping plan was considered, which seems to be the limit when using acceptable load figures.

Irregular ScenarioGSM Case, 5 MHz

Downlink FER(>4.2%)

02468

10121416

4.00% 6.00% 8.00% 10.00% 12.00% 14.00% 16.00%

EFL (%)

% u

nsat

isfie

d us

ers

DFCA REF

Irregular ScenarioGSM Case, 5 MHzUplink FER(>4.2%)

02468

10121416

4.00% 6.00% 8.00% 10.00% 12.00% 14.00% 16.00%

EFL (%)

% u

nsat

isfie

d us

ers

DFCA REF

Figure 31. GSM-FR speech performance in irregular network

DFCA is able to provide significant capacity gains by improving the network performance in both DL and UL. The gains with EFR are summarized below.

Outage Gain

2% 78%

4% 66%


90 (93)

10.2.2 AMR performance

Following simulations show AMR FR & HR performance. The frequency band used was 1.8 MHz (9 frequencies); this small band was necessary to get to the EFL points where the quality starts to degrade, keeping the Blocked Calls Rate (BCR) of the network into the allowed limits.

AMR Case, 4.2 MHzDownlink FER(>4.2%)

00.5

11.5

22.5

33.5

4

4.00% 8.00% 12.00% 16.00% 20.00% 24.00%

EFL (%)

% u

nsat

isfie

d us

ers

REF AMR-FR DFCA AMR-FR DFCA AMR-HR

AMR Case, 4.2 MHzUplink FER(>4.2%)

00.5

11.5

22.5

33.5

4

4.00% 8.00% 12.00% 16.00% 20.00% 24.00%

EFL (%)%

uns

atis

fied

user

sREF AMR-FR DFCA AMR-FR DFCA AMR-HR

Figure 32. AMR speech performance in irregular network

The behavior of the network is similar as in the previous case and DFCA is able to provide clear performance improvements. It should be noted that DFCA with AMR HR clearly outperforms the case with DFCA and AMR FR. The explanation to this is that the usage of HR channels effectively doubles the number of unique radio channels available to DFCA channel selection algorithm. This provides the DFCA algorithm much higher flexibility for the channel selection that ultimately makes it easier for DFCA to find good quality radio channels.

Gains in Downlink

Outage Gain (AMR-FR) Gain (ARM-HR)

1% 24% 55%

2% 25% -

10.2.3 AMR in Narrowband environment (3.6 MHz)

Next study shows the performance of this network when narrowband considerations are taken into account. A correlated frequency model was used to run the simulations, using a very reduced set of frequencies (6 frequencies for TCH).

Only AMR service was tested, provided that this would probably be the most used service in this environment.

In order to cope with such low number of frequencies, the network setup had to be changed, reducing the number of TRXs per cell to a maximum of 3. The average number of TRXs went down from previous


91 (93)

3.92 to 2.42. The MAIO plan had to be changed as well, trying to optimise in every site. The following scheme was used to setup the frequency parameters:

Site scheme Cell Number (in the site) Frequency parameters

1st cell (3 trx) Maio = 0, Step = 2, hsn = X

2nd cell (3 trx) Maio = 1, Step = 2, hsn = X

3-3-3

3rd cell (3 trx) Maio = 0, Step = 2, hsn = Y


2nd cell (3-2 trx) Maio = 1, Step = 2, hsn = X

3-3-2 / 3-2-2

3rd cell (2 trx) Maio = 0, Step = 2, hsn = Y

1st cell (3-2 trx) Maio = 0, Step = 2, hsn = X


3-1-1 / 2-1-1

3rd cell (1 trx) Maio = 5, Step = 2, hsn = X



2-2-2

3rd cell (2 trx) Maio = 2, Step = 3, hsn = X

The pictures below show the results of the simulations. DFCA algorithm was run both with AMR full rate and half rate codecs, based on previous studies it was known that the algorithm performs better with HR traffic.

Spectrum BW = 3.6 MHzPerformance of Hopping Layer

Downlink FER(>4.2%)

00.5

11.5

22.5

33.5

5.00% 8.00% 11.00% 14.00% 17.00%

EFL (%)

% u

nsat

isfie

d us

ers

REF AMR DFCA AMR HR DFCA AMR FR

Spectrum BW = 3.6 MHzPerformance of Hopping Layer

Uplink FER(>4.2%)

00.5

11.5

22.5

33.5

5.00% 8.00% 11.00% 14.00% 17.00%

EFL (%)

% u

nsat

isfie

d us

ers

REF AMR DFCA AMR HR DFCA AMR FR

Figure 33. AMR Performance with narrow band (3.6 MHz)

As we can see in the graphs, DFCA performance is gaining as much as 26% in downlink direction (@2% outage). It seems that the DFCA gains are significantly reduced in a narrow band environment.

Gains in Downlink

Outage Gain (AMR-FR) Gain (AMR-HR)

1% 9% 17%

2% 19% 26%


92 (93)

The causes for this degradation of performance in the narrowband case are the following:

- Reduced number of frequencies gives lower frequency diversity gain, which is the only source of gain for DFCA FH (there is no interference diversity).

- Having only 6 frequencies is reducing considerably the choice for the channel allocation algorithm.


93 (93)

DFCA System Description 107c

Documents

nokia networks liability

nokia logo

copyright nokia oyj

mentioned hardware

customer comments aspart

dfca system architecture

free of material errors

dfca overview