FP7 SOCRATES Hans van den Berg (TNO, NL) Ljupco Jorguseski (TNO, - - PowerPoint PPT Presentation

FP7 SOCRATES

FP7 SOCRATES

Hans van den Berg (TNO, NL) Ljupco Jorguseski (TNO, NL) Neil Scully (Vodafone, UK)

Informa SON Conference, London, 30 Nov – 1 Dec 2010

WWW.FP7-SOCRATES.EU

2

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

• 1. SOCRATES Overview (15 min)

a) Background, approach, use cases b) Main results

• 2. SON Concepts in SOCRATES (15 min)

Q&A (5 min)

• 3. SON Simulation Demo’s (30 min)

Q&A (10 min)

Outline

2/17

Hans Neil Ljupco

WWW.FP7-SOCRATES.EU

3

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

SOCRATES

– Self-Optimisation and self-ConfiguRATion in wirelEss networkS

Project period

– 3-year duration: From 01/01/2008 until 31/12/2010

Effort

– Number of person months:

378

– Total project costs:

€ 4,980,433

Consortium

FP7 SOCRATES

WWW.FP7-SOCRATES.EU

4

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Self-organisation in wireless networks

– Self-configuration

– e.g. ‘plug-and-play’ of new base

stations – Self-optimisation

– measurements, processing,

parameter adjustment, …

– continuous loop

– Self-healing

– failure detection – automatic minimisation of

coverage/capacity loss

Focus on 3GPP LTE (E-UTRAN) Evolutionary approach

Measurements

(Gathering and processing)

Self-

• ptimisation

Setting parameters Self- healing Self- configuration

continuous loop triggered by incidental events

Measurements

(Gathering and processing)

Self-

• ptimisation

Setting parameters Self- healing Self- configuration

continuous loop triggered by incidental events

FP7 SOCRATES: KEY ISSUES

WWW.FP7-SOCRATES.EU

5

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

SELF-ORGANISATION: EXPECTED GAINS (1/2)

OPEX reductions …

– Primary objective! – Network operations increasingly complex – Less human involvement in

– Network planning/optimisation – Performance monitoring, drive testing – Troubleshooting

Co ver age GS M Co ver age ED GE Co ver age UM TS

WWW.FP7-SOCRATES.EU

6

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

SELF-ORGANISATION: EXPECTED GAINS (1/2)

… and CAPEX reductions …

– Efficiency improvement delayed capacity expansions

– smart eNodeBs may however be more expensive

… and performance enhancements

– Enhanced service availability (robustness, resilience), QoS

IMPACT OF 'SELF-HEALING'

20 40 60 80 100

TIME LOCAL SERVICE QUALITY SITE FAILURE WITH SELF-HEALING: QUICK RECOVERY TO TOLERABLE LEVEL WITHOUT SELF-HEALING: DRAMATIC DROP TO INTOLERABLE LEVEL LOCAL SERVICE QUALITY

WWW.FP7-SOCRATES.EU

7

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Main objectives of SOCRATES

– Development of concepts, methods and algorithms for self-

• rganisation

– e.g. handover parameters, antenna parameters, admission control

parameters, … – Specification of architectural requirements – measurements, interfaces, protocols – Assessment of the operational impact – e.g. radio network planning, capacity management, …

SOCRATES: OBJECTIVES

WWW.FP7-SOCRATES.EU

8

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

FP7 SOCRATES

Overview

FP7 SOCRATES

– Background – Approach – Use cases – Results

WWW.FP7-SOCRATES.EU

9

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

SOCRATES: APPROACH (1/2)

9

Defining use cases, assessment criteria, reference scenarios and the framework Development phase Developing solutions for individual use cases Integration phase Integrating use cases into an overall solution Requirements phase

OVERALL OVERALL OVERALL OVERALL

SOA

GoS/QoS ROBUSTNESS OPEX CAPEX GoS/QoS ROBUSTNESS OPEX CAPEX

SOB

GoS/QoS ROBUSTNESS OPEX CAPEX

SOC

GoS/QoS ROBUSTNESS OPEX CAPEX

SOD

OVERALL OVERALL OVERALL OVERALL

SOA

GoS/QoS ROBUSTNESS OPEX CAPEX GoS/QoS ROBUSTNESS OPEX CAPEX

SOB

GoS/QoS ROBUSTNESS OPEX CAPEX GoS/QoS ROBUSTNESS OPEX CAPEX ROBUSTNESS OPEX CAPEX

SOB

GoS/QoS ROBUSTNESS OPEX CAPEX

SOC

GoS/QoS ROBUSTNESS OPEX CAPEX GoS/QoS ROBUSTNESS OPEX CAPEX

SOC

GoS/QoS ROBUSTNESS OPEX CAPEX

SOD

GoS/QoS ROBUSTNESS OPEX CAPEX GoS/QoS ROBUSTNESS OPEX CAPEX

SOD

• Three phases

WWW.FP7-SOCRATES.EU

10

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

SOCRATES: APPROACH (1/2)

Evolutionary approach towards self-optimisation

– Take current architecture as starting point

– Works quite well, when parameters are properly tuned …

– ‘Make’ existing functionalities ‘self-optimising’

– E.g. RRM mechanisms, … – Determine actual need for self-optimisation by sensitivity analysis – Algorithms for ‘automatic’ adaptation of parameters

– Required architectural modifications impact on standardisation

– Measurements, interfaces, signaling, …

Many ‘use cases’ defined and investigated

– Stand-alone functionalities – Interacting functionalities

– coordination needed!

WWW.FP7-SOCRATES.EU

11

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

SOCRATES USE CASES (stand-alone functionalities)

11

Self-optimisation Handover parameter optimisation Self-configuration Automatic generation of initial params Self-healing Cell outage management SON enabler X-map estimation Load balancing Packet scheduling optimisation Admission control optimisation Interference coordination HeNB handover optimisation HeNB interference and cov. optimisation

WWW.FP7-SOCRATES.EU

12

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

SOCRATES USE CASES (integrated functionalities)

12

Self-optimisation

Admission Control and Handover Optimisation Load balancing and Handover Optimisation Macro & Home eNodeB Handover Optimisation Packet Scheduling and Interference Coordination Combinations of use cases where there is either a control

parameter conflict or an observability dependency

WWW.FP7-SOCRATES.EU

13

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

FP7 SOCRATES

Overview

FP7 SOCRATES

– Background – Approach – Use cases – Results

WWW.FP7-SOCRATES.EU

14

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

MAIN PROJECT RESULTS (1/2)

Framework for development of self-organisation methods

– high-level guidelines, requirements, …

Assessment criteria and methodologies

– OPEX, CAPEX, performance gains – reference scenarios

Self-organisation methods and algorithms for selected use cases

– stand-alone – integrated

• therwise

CASE WITH SELF-HEALING CASE WITHOUT SELF-HEALING

• therwise

CASE WITH SELF-HEALING CASE WITHOUT SELF-HEALING

WWW.FP7-SOCRATES.EU

15

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

MAIN PROJECT RESULTS (2/2)

SON coordination framework

– alignment of interacting SON functions

Measurements, architecture & interfaces for SON

– based on use case studies, SON coordinator

Demonstrations

– visualisation through simulation – validation

Implications of SON (ongoing work)

– network planning & operations

AUTOGNOSTICS POLICY ALIGNMENT

• Incl. Guard function

NETWORK SUBSYSTEM (e.g., eNB, neighbours, OSS, NMS) OPERATOR SON FUNCTIONS

• bjectives

SON System SON FUNCTIONS SON FUNCTIONS SON FUNCTION

• bjectives &

Communication with peers Note SON system will often be in the eNB, but could be in any node

AUTOGNOSTICS POLICY ALIGNMENT

• Incl. Guard function

NETWORK SUBSYSTEM (e.g., eNB, neighbours, OSS, NMS) OPERATOR SON FUNCTIONS

• bjectives

SON System SON FUNCTIONS SON FUNCTIONS SON FUNCTION

• bjectives &

Communication with peers Note SON system will often be in the eNB, but could be in any node

WWW.FP7-SOCRATES.EU

16

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

SOCRATES – DISSEMINATION OF PROJECT RESULTS

Project deliverables, publications, presentations, … Organisation of workshops

– at ICT Mobile Summit 2009, Santander, June 2009 – joint SOCRATES / COST2100 workshops

– Braunschweig, February 2009 – Athens, February 2010

Demonstration stand at Future Networks & Mobile Summit 2010

– Florence, June 2010

Interactions with NGMN Liaisons to 3GPP

– several contributions delivered and planned

http://www.fp7-socrates.eu

WWW.FP7-SOCRATES.EU

17

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

SOCRATES – FINAL WORKSHOP

WWW.FP7-SOCRATES.EU

18

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

• 1. SOCRATES Overview (15 min)
• 2. SON Concepts in SOCRATES (15 min)

a) Assessment Criteria b) SON Coordination

Q&A (5 min)

• 3. SON Simulation Demo’s (30 min)

a) Home eNodeB b) Handover c) Load Balancing or Cell Outage Compensation

Q&A (10 min)

Outline

18/17

WWW.FP7-SOCRATES.EU

19

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Outline

Assessment criteria for self-organising networks

– Introduction – Metrics – Performance / Coverage / Capacity – OPEX – Benchmarking approach

Coordination of SON use cases

– Overview – Policy Function – Alignment Function – Guard Function

19

WWW.FP7-SOCRATES.EU

20

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Introduction

Why define assessment criteria?

– Evaluation of SON algorithms – Comparing multiple algorithms: Which algorithm is best? – Fine-tuning of an algorithm to optimise its performance – Assessment of the gains that can be achieved using SON – By comparison with manual network operation – Essential part of development of algorithms – But also useful for SON trials!

WWW.FP7-SOCRATES.EU

21

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Metrics: Performance – Coverage – Capacity Performance (GoS/QoS) GoS (Grade of Service) Call blocking ratio, call dropping ratio, … Coverage Service coverage, data rate coverage, … Capacity Maximum supportable traffic load, spectrum efficiency, … QoS (Quality of Service) Packet delay, packet loss ratio, transfer time, throughput, MOS, fairness, …

21

‘Conventional’ metrics, but important for SON

WWW.FP7-SOCRATES.EU

22

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Metrics: OPEX

OPEX reduction is often quoted as an important SON gain

– Important to be able to quantify the impact

In the SOCRATES approach:

– OPEX without SON is determined by summing together all components that

contribute to OPEX

– OPEX with SON is determined by assessing impact on various components – Difference is then assessed

WWW.FP7-SOCRATES.EU

23

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Determining OPEX savings – Methodology

OPEX without SON

• 1. Determine effort for an individual task

– Task is defined as optimising or adjusting a parameter or parameter set

• 2. Determine OPEX per task, per network, per year

– Cost per task (Euro) = Effort per task (days) × Cost per day (Euro) – OPEX per task / year =

(Cost per task) × (#Changes per network) × (#Changes per year)

• 3. Determine total OPEX per year

– OPEX / year = SUMall-tasks(OPEX per task / year)

OPEX with SON: use the same method as for ‘without SON’, but assess impact on

different components

– In some cases OPEX may be reduced to zero, but definitely not always

Can the SON algorithm completely autonomously:

– Obtain the required inputs? – Analyse the available data? – Determine new parameter settings?

WWW.FP7-SOCRATES.EU

24

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Benchmarking: Overall assessment

• Approaches to determine overall assessment:

–

Single target metric, with constraints on the other metrics

–

Combining different metrics using a utility function (with possible constraints on additional metrics)

• Absolute gains:

OVERALL OVERALL OVERALL OVERALL

SOA

GoS/QoS ROBUSTNESS OPEX CAPEX GoS/QoS ROBUSTNESS OPEX CAPEX

SOB

GoS/QoS ROBUSTNESS OPEX CAPEX

SOC

GoS/QoS ROBUSTNESS OPEX CAPEX

SOD

OVERALL OVERALL OVERALL OVERALL

SOA

GoS/QoS ROBUSTNESS OPEX CAPEX GoS/QoS ROBUSTNESS OPEX CAPEX

SOB

GoS/QoS ROBUSTNESS OPEX CAPEX GoS/QoS ROBUSTNESS OPEX CAPEX ROBUSTNESS OPEX CAPEX

SOB

GoS/QoS ROBUSTNESS OPEX CAPEX

SOC

GoS/QoS ROBUSTNESS OPEX CAPEX GoS/QoS ROBUSTNESS OPEX CAPEX

SOC

GoS/QoS ROBUSTNESS OPEX CAPEX

SOD

GoS/QoS ROBUSTNESS OPEX CAPEX GoS/QoS ROBUSTNESS OPEX CAPEX

SOD

Operator invests very little in manual optimisation Operator invests a lot in manual optimisation Gain is mainly in network quality Gain is mainly OPEX reduction

WWW.FP7-SOCRATES.EU

25

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Why SON Coordination?

The research and development work on SON has up to now concentrated on

stand-alone functions

With an increasing number of SON functions, the likeliness of conflicting

goals and targets of the individual SON functions also increases

The goal of SON coordination is to avoid or resolve potential conflicts, and

provide a single interface towards the operator regarding policies

SON Function A SON Function B Metric A Metric B P1 P2 P3 P4 SON Function A SON Function B Metric A Metric B P1 P2 P3 P4 SON Function A SON Function B Metric A Metric B P1 P2 P3 P4 SON Function A SON Function B Metric A Metric B P1 P2 P3 P4

Control Parameter Conflict Observability Dependency

WWW.FP7-SOCRATES.EU

26

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

SON Coordinator Framework - Overview

THE OPERATOR NETWORK SUBSYSTEM Policy Function

SON System

Guard Function Alignment Function Communi- cation with Peers Auto- gnostics Function SON Functions THE OPERATOR NETWORK SUBSYSTEM Policy Function

SON System

Guard Function Alignment Function Communi- cation with Peers Auto- gnostics Function SON Functions

WWW.FP7-SOCRATES.EU

27

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Policy Function

Operator Policy Function

SON Func. A Policy

SON-specific policies – these change according to cell type High-level performance

• bjectives, e.g.,

accessibility, retainability

Guard Align- ment Auto- gnostics

SON Func. B Policy SON Func. C Policy

Operator Policy Function

SON Func. A Policy SON Func. A Policy

SON-specific policies – these change according to cell type High-level performance

• bjectives, e.g.,

accessibility, retainability

Guard Align- ment Auto- gnostics

SON Func. B Policy SON Func. B Policy SON Func. C Policy SON Func. C Policy

WWW.FP7-SOCRATES.EU

28

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Alignment Function

Arbitration

Short-term driven, conflict resolution

Network Subsystem

Trigger

SON Function Activation

Cause analysis, management of SON functions

Policy Function Alignment Function Guard Function

Requests Feedback, Notifications, Settings, Start/Stop Arbitration, Activation & Cause Analysis Settings Confirmed Parameter Changes

Arbitration

Short-term driven, conflict resolution

Network Subsystem

Trigger

SON Function Activation

Cause analysis, management of SON functions

Policy Function Alignment Function Guard Function

Requests Feedback, Notifications, Settings, Start/Stop Arbitration, Activation & Cause Analysis Settings Confirmed Parameter Changes

WWW.FP7-SOCRATES.EU

29

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Guard Function

Guard Function

Oscillation Detection Absolute Performance Surveillance

Data request definitions

Align- ment Function Auto- gnostics Function Policy Function

Performance, Fault, Config. Data Triggers Settings for the detection of extreme network behaviour

Guard Function

Oscillation Detection Absolute Performance Surveillance

Data request definitions

Align- ment Function Auto- gnostics Function Policy Function

Performance, Fault, Config. Data Triggers Settings for the detection of extreme network behaviour

WWW.FP7-SOCRATES.EU

30

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Questions

30/17

WWW.FP7-SOCRATES.EU

31

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

SOCRATES Simulation Demo’s

WWW.FP7-SOCRATES.EU

32

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

• 1. SON Simulation Demo’s

a) Handover b) Load Balancing or Cell Outage Compensation c) Self-optimisation of Home eNodeBs

Q&A

WWW.FP7-SOCRATES.EU

33

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Handover self optimisation

Possible handover problems:

– HPIHOF: Handover failures (too early, too late) – HPIHPP: Ping-pong handovers – HPICD: Call dropping

HP = w1HPIHOF + w2HPIHPP + w3HPICD

Objective

– Minimization of HP via automatic adjustment of

hysteresis and time to trigger (TTT)

Ping-pong

time

HO Drop

X

33/17

WWW.FP7-SOCRATES.EU

34

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Handover self optimisation approach

Off-line HO system behaviour

– System level simulations with different

hysteresis and time-to-trigger

– Define operating region

34/17

Derive on-line (self-) optimisation

– Start within the operating region – Change HYS and/or TTT and measure

handover performance (HP)

– If HP improved further change in same

‘direction’. Otherwise change in other ‘direction’.

• !

WWW.FP7-SOCRATES.EU

35

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Handover demo (Thomas Janssen, TU Braunschweig)

35/17

WWW.FP7-SOCRATES.EU

36

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Load balancing self optimisation

Possible load problems:

– Users concentrated in one cell i.e. unequal load

distribution causing cell overload

– Users cannot be served with the required quality

due to lack of resources Objective

– Load transfer from overloaded cell towards

neighbour cells via automatic updates of HO

• ffset

– Maximise number of satisfied users

"# $#

• %&'

"# $# "# $# ((

• )

)

WWW.FP7-SOCRATES.EU

37

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Load balancing self optimisation approach

Estimate DL and UL load at target cell after

load balancing

– Based on DL and UL SINR estimation – UEs do not change location during load

balancing Decide which and how many UEs from

• verloaded cell should be transferred

Decide which neighbour cells will accept the

• verload traffic

Adjust HO offsets and send HO commands

WWW.FP7-SOCRATES.EU

38

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Load balancing demo (Szymon Stefanski, NSN)

WWW.FP7-SOCRATES.EU

39

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Cell outage compensation self optimisation

39/17

Cell outage problems:

– Lost of coverage – Unsatisfactory users and lost of revenue

Objective

– Coverage (partially) restored by neighbour cells

via automatic updates of e.g. antenna tilt, or uplink open loop power control operating point P0

– Maximise number of served users under the

constraint that UL and DL quality target at neighbour cells are above thresholds

WWW.FP7-SOCRATES.EU

40

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Cell outage compensation self optimisation approach

QUL QDL Quality target checkL

Measure UL and DL quality at neighbours Quality target check

– QUL <> QT,UL – QDL <> QT,DL

Adjust control parameters accordingly

– Decrease/increase DL power/tilt – Decrease/increase P0

WWW.FP7-SOCRATES.EU

41

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Cell Outage Compensation Demo (Mehdi Amirijoo, Ericsson)

WWW.FP7-SOCRATES.EU

42

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

HeNB handover optimisation: Introduction and scenario

Handover optimisation in scenario with

• pen access Home eNodeBs (HeNBs)

HeNBs are placed indoors, but provide

coverage to outdoor users

The objective is to determine how to

effectively use open access HeNBs

– Reliability: Minimise dropped calls due

to bad SINR conditions at handover

– User experience: Maximise the

throughput that the user is experiencing Optimisation for HeNB scenario may be

different than for macro-only scenario

Figure shows best server based purely on signal strength

X (m) Y (m)

• 200
• 100

100 200 300 400 500 600 700

• 500
• 400
• 300
• 200
• 100

100 200 300 400 500

3x3 layout of HeNBs 3 sector macro site

WWW.FP7-SOCRATES.EU

43

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Handover optimisation algorithm

Statistics (HPI = Handover

Performance Indicator) used as input to the algorithm are:

– Dropped call ratio (CDR) – Ping-pong handover ratio (PPHR)

Metrics are calculated separately for

each type of handover

– ‘X’ can be M (= Macro) or H (=Home

eNodeB)

Start values:

– PPHRThreshold = 0.1 – CDRThreshold = 0.1 – Hysteresis = 5 dB – TTT = 320 ms (fixed) PPHRX < PPHRThresholdX CDRX > CDRThresholdX Decrease CIOX by 1 dB PPHRX > PPHRThresholdX CDRX < CDRThresholdX Increase CIOX by 1 dB PPHRX < PPHRThresholdX CDRX < CDRThresholdX Decrease HPI thresholds by 0.02 PPHRX > PPHRThresholdX CDRX > CDRThresholdX Increase HPI thresholds by 0.02

Algorithm

WWW.FP7-SOCRATES.EU

44

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Macro & HeNB – Simulation results – Call drop ratio (CDR)

SON activated – start values: Hyst=5 dB, TTT=320 ms No SON – param. values: Hyst=3 dB, TTT=320 ms

Ping-pong handover ratio increases for

macro-macro handover

Ping-pong handover ratio is negligible for all

• ther handovers

100 200 300 400 500 600 10 20 30 40 50 60 70 80 90 Time (s) Call drop ratio (%) Macro-macro CDR - No SON Macro-macro CDR - SON activated HeNB-HeNB CDR - No SON HeNB-HeNB CDR - SON activated

Even with SON, HeNB handover performance is still an issue (for the considered scenario)

WWW.FP7-SOCRATES.EU

45

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Questions

45/17

WWW.FP7-SOCRATES.EU

46

Hans van den Berg (TNO), Ljupco Jorguseski (TNO), Neil Scully (Vodafone)

Contact Details

Hans van der Berg (TNO)

– j.l.vandenberg@tno.nl – +31 88 8667031

Neil Scully (Vodafone)

– Neil.Scully@vodafone.com – +44 7919 994699

Ljupco Jorguseski (TNO)

– ljupco.jorguseski@tno.nl – +31 88 8667154

Thomas Jansen (TU Braunschweig)

– Jansen@ifn.ing.tu-bs.de

Szymon Stefanski (NSN)

– szymon.stefanski@nsn.com

Mehdi Amirijoo (Ericsson)

– mehdi.amirijoo@ericsson.com

46/17