Cell Outage Management in LTE Networks M. Amirijoo (Ericsson), L. - - PDF document

FP7 ICT-SOCRATES

Cell Outage Management in LTE Networks

• M. Amirijoo (Ericsson),
• L. Jorguseski (TNO ICT),
• T. Kürner (TU Braunschweig),
• R. Litjens (TNO ICT),
• M. Neuland (TU Braunschweig),
• L. C. Schmelz (Nokia Siemens Networks),
• U. Türke (Atesio)

Abstract Cell outage management is a functionality aiming at automatically detecting and mitigating outages that

• ccur in radio networks due to unexpected failures. The FP7 EU-funded project SOCRATES (Self-

Optimisation and self-ConfiguRATion in wirelEss networkS) is heavily engaged in developing a framework and solutions for cell outage management, where appropriate measurements, control parameters, policies, assessment methodologies, and algorithms are being developed. Our vision is that radio networks autonomously detect an outage based on measurements, from e.g., user equipment and base stations, and alter the configuration of radio base stations such that coverage and service quality requirements during an outage, as specified by the operator, are satisfied. In this presentation we give an introduction in the area of cell outage management and present an overview of the results achieved so far as well as the next steps that will be pursued.

1

WWW.FP7-SOCRATES.EU

2/20 2/20

Introduction Components in cell outage management Operator policy Measurement Control parameters Scenarios Assessment criteria Simulator Future work

Outline

WWW.FP7-SOCRATES.EU

3/20 3/20

The first release of the 3GPP Long Term Evolution (LTE) standard has

been finalized

Operators require significant reduction of manual network management for

LTE

Introduction of self-organisation functionalities in LTE

– Reduces manual network management – Enhances network performance

One aspect that benefits from self-organization is cell outage management

(COM), which consists of:

– Cell outage detection – Cell outage compensation

Introduction

The standardisation body 3rd Generation Partnership Project (3GPP) has finalised the first release (Release 8) of the UMTS successor named as Evolved UTRAN (E-UTRAN), commonly known as Long Term Evolution (LTE). Wireless network operators today spend considerable manual effort in planning, configuring, optimising, and maintaining their wireless access networks. These efforts consume a substantial part of their

• perational expenditure (OPEX). Consequently, an important E-UTRAN requirement from the operators’

side is a significant reduction of the manual effort in the deployment, configuration and optimisation phases for this future wireless access system. One way for achieving this requirement is the introduction of self-organisation functionalities into the E-

• UTRAN. Self-organisation functionalities not only reduce the manual effort involved in network

management, but they also enhance the performance of the wireless network. One aspect that benefits from self-organization is the management of cell/site outages, which can be divided into two parts, namely, the detection of an outage and the compensation of the detected outage.

3

WWW.FP7-SOCRATES.EU

4/20 4/20

Reasons for outages, e.g.:

– hardware and software failures, – external failures such as power supply or network connectivity

Outages

– may not be detected for hours or even days – may require manual analysis and unplanned site visits

Outage detection function must timely inform the operator about the

• ccurrence and the cause of an outage

Automatic compensation actions are triggered to alleviate performance

degradation

Introduction

There are multiple reasons for a cell outage, e.g., hardware and software failures (radio board failure, channel processing implementation error etc), external failures such as power supply or network connectivity failures, or even erroneous configuration. While some cell outage cases are detected by Operations Support System (OSS) functions through performance counters and/or alarms, some may not be detected for hours or even days. It is often through long term performance analysis and subscriber complaints that these outages are detected. Currently, discovery and identification of some errors involves sometimes manual analysis and may require unplanned site visits, which makes cell outage detection a costly task. It is the task of the cell outage detection function to timely inform the operator about the occurrence of an outage and the cause of the outage. In the event of a cell outage detection, appropriate compensation methods are triggered to alleviate the degraded performance due to the resulting coverage gap and loss in throughput by appropriately adjusting radio parameters in surrounding sites. In general, human involvement shall only be triggered when absolutely necessary, e.g., when manual repairs are needed.

4

WWW.FP7-SOCRATES.EU

5/20 5/20

Components of Cell Outage Management

• 2000 -1500 -1000
• 500

500 1000 1500 2000 2500

• 2000
• 1500
• 1000
• 500

500 1000 1500 2000

• 170
• 160
• 150
• 140
• 130
• 120
• 110
• 100
• 90
• 80
• 70

Measurements Detection Compensation Operator policy: Coverage, QoS Control parameters

• Cov. map

estimation Simulation tools Scenarios Assessment criteria

Various measurements are gathered from the user equipments (UEs) and the base stations (called eNodeBs in LTE). The measurements are then fed into the cell outage detection function, which detects whether at the current time an outage has occurred and triggers the cell outage compensation function to take appropriate actions. The goal of cell outage compensation is to minimise the network performance degradation when a cell is in outage. This is done by automatic adjustment of network parameters in order to meet the operator’s performance requirements based on coverage and other quality indicators, e.g., throughput. It is realized that performance requirements may not be achieved during an outage and, as such, the goal of the compensation function is to meet performance requirements to the largest possible extent. Cell outage compensation algorithms may for example alter the antenna tilt and azimuth, or the cell transmit power, in order to cover the area that is in outage. Altering the radio parameter of the neighbouring cells means that some of the UEs served by neighbouring cells may be affected. For example, if the coverage of a neighbouring cell increases then this implies that more UEs will be served and, consequently, the UE throughput may decrease. This should be taken into account and an appropriate balance between coverage and other quality indicators (e.g. throughput) should be achieved. This balance is indicated by means of an operator policy that governs the actions taken by the cell outage compensation function. In order to monitor and evaluate the actions taken by the cell outage compensation algorithm, there is a need to estimate the coverage around the vicinity of the outage area. This is provided by the coverage map estimation function, which continuously monitors the network coverage by means of measurements and possibly prediction data. There are also other activities being pursued along this work, namely, the development of simulation tools, scenarios that capture different outage cases, and methods and criteria for assessing the impact and performance of cell outage detection and compensation algorithms.

5

WWW.FP7-SOCRATES.EU

6/20 6/20

In a cell outage situation an operator may still target at the ideal goal of

– achieving the best coverage possible, – providing the highest accessibility, and – delivering the best possible quality in the cell outage area and all surrounding

cells. All these goals cannot be fulfilled at the same time

– the targets have to be weighted and/or ranked in order to provide quantitative

input to an optimisation procedure. The optimisation goal itself may vary depending on the operator’s policy

– coverage orientation – quality orientation – capacity orientation

Operator Policy

Every operator has its own policy for the network performance. For cell outage management the operator wants to implement its policy into the network. For this, optimisation goals and a cost function have to be defined. In a cell outage situation an operator may still target at the ideal goal of

• achieving the best coverage possible,
• providing the highest accessibility, and
• delivering the best possible quality in the cell outage area and all surrounding cells.

In most cases, all these goals cannot be fulfilled at the same time. As a consequence the targets have to be weighted and/or ranked in order to provide quantitative input to an optimisation procedure. The optimisation goal itself may vary depending on the operator’s policy. For example, one operator wants to maximize its income also during the cell outage, whereas another operator may have also the long-term user satisfaction in mind. In the first case an operator may target a high capacity sacrificing coverage in a large area yielding a low perceived coverage, which is often worse for the reputation compared to a higher blocking rate. Hence, the policy definition should be modular/flexible enough to capture different operator strategies, e.g., coverage oriented operator vs. capacity oriented operator.

6

WWW.FP7-SOCRATES.EU

7/20 7/20

Three groups of cells have to be distinguished:

– the cell in outage ( ) – all cells that are actively tuned by the COM algorithm ( ) – all cells that are not actively tuned but influenced by the

network changes at all other cells ( ) Cells may have different optimisation goals and different

priorities

– focus on coverage ( , ) – focus on accessibility ( , ) – focus on quality ( , )

The cost function has to take all these different foci into

account

Operator Policy

When defining the cost function it has to be decided which cells have to be taken into account. Therefore, three groups of cells have to be distinguished: the cell in outage, all cells that are actively tuned by the COM algorithms, and all cells that are not actively tuned by the COM algorithms but influenced by the network changes at all other cells. This is shown in the upper figure. The cell in the centre is the cell in

• utage. The first ring are all cells that are actively tuned by the COM algorithms. The second ring (outer

ring) are all cells that are not actively tuned but are influenced by the network changes at all other cells. Optimisation goals may vary for different cells. For example, in cells covering large areas coverage should be kept high, whereas in high-capacity cells located in the same region the focus will be more on accessibility and/or quality. As a consequence the optimisation goals have to be defined on a cell-basis. This concept is shown in the lower figure in which the different layouts of the cells represent different

• foci. In addition to the different optimisation goals cells may have different importance, too, so that

different priorities are assigned to the cells. The cost function has to take all important effects, such as coverage, accessibility, and quality, into

• account. Normalised values are used in the cost function because it is easier to weight the normalised and

dimensionless values of the different criteria against each other. The reference value to which the criteria are normalized, and the normalisation procedure have still to be defined. Accessibility and quality can be assessed by key performance indicators (KPI) retrieved from counters available in the network. The coverage can be estimated based on measurement reports. Therefore, geographic coordinates of the mobile position or at least an estimation of these coordinates have to be assigned to the obtained measurement reports. The main challenge here is how to accurately estimate the position of the UE. It has to be investigated to which degree coverage can be estimated based on measurement reports.

7

WWW.FP7-SOCRATES.EU

8/20 8/20

Measurements

Continuous and event-triggered measurements required for cell

• utage management

– Affected eNodeB: Counters and KPIs, Timers, Alarms, Reports

Measurement Sources and Types:

– Neighbour eNodeB: Counters and KPIs, Timers, Alarms, Reports – User Equipment: Measurements, Reports – Access Gateway: Alarms, Timers, Counters – OSS System: Statistics

eNodeB1 eNodeB2 eNodeB3 eNodeB4 aGW UE1 UE2

The following example nodes may be used when collecting the measurements. eNodeB1 denotes the eNodeB that encounters a cell/sector outage or a complete node outage (marked by a flash). UE1 is served by eNodeB1 (shown by solid line). UE1 furthermore receives signal from neighbouring eNodeBs 2 and 3 (shown by dotted line) and reports measurements to eNodeB1. eNodeB2 – eNodeB4 are neighbouring eNodeBs of eNodeB1. The X2 interface between affected eNodeB1 and neighbouring eNodeBs is shown by a dashed line. UE2 is served by eNodeB2. Furthermore, UE2 receives signal from eNodeB1 (shown by dotted line) and reports measurements to its serving eNodeB2. Access Gateway (aGW) where affected eNodeB1 is connected to via S1 interface is shown by a dashed line.

8

WWW.FP7-SOCRATES.EU

9/20 9/20

Measurements - Details

Measurements from UE:

– Reference Signal Received Power (RSRP) / Channel Quality Identifier (CQI) – Failure reports, e.g. after failed handover – Historic measurements

Measurements form aGW / OSS System:

– Blocked / dropped calls counters – Alarms from eNodeBs / aGW

Measurements from eNodeB:

– Cell capacity – Radio Link Failure (RLF) counter – Internal / External Handover failure rate – Relative Load Indicator – Resource status reporting – Cause information element – Interference

The following measurements may be useful when detecting and compensating an outage. It is to be further investigated to see which of the measurements that are needed by the cell outage management functions.

• Cell Capacity: zero capacity or sudden decrease of capacity may indicate an outage
• Radio Link Failure (RLF) counter: sudden increase may indicate a cell outage
• Internal handover failure rate: high failure rate of internal handovers (between sectors / cells of one

eNodeB) may indicate the outage of a sector / cell

• Incoming / outgoing handover failure rate: high inter-eNodeB handover failure rate towards dedicated

eNodeB may indicate outage of a sector of this eNodeB or the whole eNodeB

• Relative load indicator: message is exchanged between neighbouring eNodeBs via X2 interface to

transfer load and interference coordination information. The indication of sudden interference level decrease reported through load indicator messages may indicate a cell outage at the sending eNodeB

• Resource status reporting: an eNodeB may request resource status from neighbouring eNodeBs via X2
• interface. Different measurements can be requested, such as cell load or prioritised bit rates, e.g., speech

which is for further study (FFS), in standardisation. Suddenly changing values of measurements may indicate cell outages, and the resource request could also be used to confirm potential outages that have been indicated by other measurements

• Cause Information Element: multiple X2 messages include the (mandatory) information element (IE)

"Cause", e.g. Handover Request, Handover Preparation Failure, Handover Cancel, Error Indication, X2 Setup Failure, eNodeB Configuration Update Failure, and Resource Status Failure. Relevant cause values for cell outage management are: “Cell not Available”, “Transport resource unavailable”, “Hardware Failure”. This information exchanged via X2 interface can be used for cell outage detection and cause analysis

• Interference measurements: sudden changes (reduction) in interference measurements taken by

eNodeB or UE may indicate outage of neighbour eNodeB; this is still a topic FFS

• Reference Signal Received Power (RSRP): sudden decrease in received power level of serving or

neighbour cells may indicate a cell outage. RSRP includes measurements of serving and neighbouring cells but is reported only to the serving eNodeB

• Channel Quality Indicator (CQI): the CQI is a measurement about the radio channel quality of the serving

eNodeB which is taken by the UE, and reported only to the serving eNodeB. A sudden decrease of the CQI level may indicate an upcoming cell outage of the serving eNodeB. On the other hand, a suddenly increasing CQI may indicate the outage of a neighbouring eNodeB which may cause a decrease in channel interference

• Failure reports: failure reports are generated by the UE after connection or handover failures and sent to

eNodeB / OAM system for cause analysis. These reports can also be used for cell outage detection or cause analysis

• Historic measurements: the UE provides measurements such as RSRP, CQI and others to the serving

eNodeB during connection – in case of a connection failure to the serving eNodeB (e.g. due to a cell

• utage) the latest measurements are usually lost. These historic could be provided to the new serving

eNodeB after cell re-selection for failure cause analysis (FFS in standardisation)

• Blocked / dropped calls: the number of blocked or dropped calls is counted by the eNodeB and by

default sent to the OAM system – a sudden increase may indicate cell outage

• Alarms from eNodeB / aGW: eNodeB and aGW send alarms to the OAM system in case of connection

failures, handover failures etc. These alarms could be used for cell outage detection and cause analysis

9

WWW.FP7-SOCRATES.EU

10/20 10/20

Physical channel parameters

– Power settings (e.g. “downlink reference signal”) – Control of the coverage/capacity trade-off (increase/decrease cell sizes)

Antenna parameters

– Tilt, azimuth (only if adjustable remotely, e.g., remote electrical tilt – RET) – Control of the coverage/capacity trade-off – Multi antenna techniques – E.g., Steer beam pattern to coverage hole

Home-eNodeB related parameters (FFS)

– Improve coverage by temporarily carrying macro layer traffic by Home-eNodeBs – Reduce interference caused by Home-eNodeBs to improve coverage

• f macro layer

Control Parameters

The power allocated to physical channels dictates the cell size. One the one hand, by increasing the physical channel power the coverage area of a cell can be increased (in order to compensate for outage). On the other hand, by lowering the cell power the cell area is reduced and as a consequence load and interference caused by the cell can also be reduced. Modern antenna design allows influencing the antenna pattern and the orientation of the main lobe by electrical means (e.g., remote electrical tilt, beam forming) . This is the basis for an on-line control of the antenna characteristics and thus provides the option to modify the characteristics according to the needs

• f self-organisation. In case of a cell outage, the footprint of the surrounding cells change as one or more

cells "disappear". The changes in the footprints are coincidental. As part of cell outage management, these changes may be controlled as to optimize network coverage and load balancing among the remaining cells. The consideration of Home-eNodeB for outage compensation is for further study, because the implementation details of Home-eNodeBs are not yet completely fixed. Two aspects are currently foreseen as possible options. First, if there are sufficient Home-eNodeBs in the outage area, these may take over some macro-layer traffic, once a macro-level sector/eNodeB has failed. The feasibility of this approach depends on whether it is possible to open Home-eNodeBs for access by UEs. Second aspect, is that Home-eNodeB users – if operating in the same spectrum - cause UL interference to the macro-layer. During an outage, resulting in a high distance between UE and a functioning eNodeB, the coverage might be significantly reduced by the UEs connected to Home-eNodeBs. A possible approach is to turn Home eNodeBs off whenever there is an outage or at least decrease the transmission power of UE served by Home eNodeBs.

10

WWW.FP7-SOCRATES.EU

11/20 11/20

Scenarios

The developed cell outage management algorithms are assessed for different

scenarios, comprising

– Outage scenarios: type and extent of failure – Evaluation scenarios: considered network, traffic and environment scenarios – COM mechanism: detection (COD) vs compensation (COC)

Key outage scenarios

– Sleeping site/sector due to physical channel failure – COD:

methods depend on failing channel

– COC:

similar methods as below

– Whole site/sector failure – COD:

no foreseen need of algorithm

– COC:

compensation using control parameters

– Transport network link failure – COD:

no foreseen need of algorithm

– COC:

X2 failure may affect HO performance and the effectiveness of ICIC

• The objective is to formulate a number of scenarios that are considered in the development and

assessment of cell outage management methods. These scenario descriptions comprise three key elements (i) the considered outage scenario, e.g. sleeping sites/cells, failing sites/cells or failing transport network links; (ii) assumptions regarding network, traffic and environment aspects for which useful deployment of cell outage management solutions is anticipated; and (iii) whether cell outage detection and/or cell outage compensation methods are/should be developed. A distinction is made between four categories of outage scenarios: (i) physical channel failures causing sleeping sites/sectors; (ii) whole site/sector failure (where failure is reported to OSS); (iii) transport network link failures; and (iv) other types of malfunctions, e.g. reduced performance due to hardware aging. In category (i) different subcases can be distinguished offering different potential for (primarily) cell

• utage detection solutions, depending on the actual physical channel that fails, e.g. the synchronisation

channel, reference signal, broadcast information, paging channel or user plane transmission. The type of control/signalling channel that fails influence e.g. whether network/mobile-originating calls or handovers can be supported, which in turn affects the possibilities for developing cell outage detection methods. In category (iii), a distinction is made regarding the type of transport network link that fails, e.g. in case all S1 links to an eNodeB fail, the site is entirely unfunctional (as in category (ii)), whereas if an X2 link fails, handovers (HOs) may endure additional delays as they will be handled via S1 links, and inter-cell interference coordination (ICIC) algorithms could no longer function, which affects capacity and/or coverage.

11

WWW.FP7-SOCRATES.EU

12/20

Scenarios

Evaluation scenarios

Impact of eNodeB density and load More COC potential in a dense capacity-driven network layout Impact of service type More COC potential in an area with predominantly low-bandwidth service, e.g. VoIP telephony Impact of outage location More COC potential if a cell/site

• utage occurs at the inner part
• f an LTE island

Also study impact of:

user mobility spatial traffic distribution propagation aspects UE terminal class The following evaluation scenarios have been formulated for the development of cell outage compensation methods. Furthermore, the priority of the evaluation scenarios is given.

• Impact of eNodeB density and traffic load: in a sparse, coverage-driven network layout, little potential

exists for compensating outage-induced coverage/capacity loss. In a dense, capacity-driven network layout, this potential is higher, particularly when traffic loads are low. Priority: high.

• Impact of service type: the distinct elasticity and quality of service requirements of different services

affect the compensation potential. For instance, compensation actions may be able to alleviate local

• utage effects to handle only low bandwidth services. Priority: high.
• Impact of outage location: if cell outages occur at the edge of an ‘LTE island’ fewer neighbours exist to

help compensate. For outages in the core of such an ‘LTE island’, the compensation potential is larger. Priority: medium.

• Impact of user mobility: if mobility is low (high), few (many) users spend a relatively long (short) time in

an outage area. The perceived outage impact depends on the delay-tolerance and elasticity of the service. Priority: medium.

• Impact of spatial traffic distribution: if traffic is concentrated around sites, such traffic is typically

relatively far away from neighbouring sites and hence the compensation potential is limited. If traffic is concentrated ‘in between’ sites, the potential is larger. Priority: medium.

• Impact of propagation aspects: the higher the path loss exponent (higher frequency bands, more

urbanised environments), the more difficult it may be to cover such users with relatively remote compensation cells in case of an outage. Furthermore, a higher shadowing variation (also primarily depending on the propagation environment) may allow more distant cells to serve calls in an outage area. Priority: low.

• Impact of UE terminal class: the higher a UE’s maximum uplink transmit power (class-dependent) the

lower the need for outage compensation, since it may still be able to attach to a more distant cell even without compensation measures. The potential for cell outage compensation is higher, since retuned surrounding cells may be more able to serve such a user. Priority: low.

12

WWW.FP7-SOCRATES.EU

13/20

Assessment Criteria – Cell Outage Detection

Detection delay = Tdetect - Tfail Detection probability = Ndetect/Nfail False detection probability = Nfalse/(Nfalse + Ndetect) Other: signalling overhead, processing overhead, severity indication accuracy F Detection Delay D D True Detection F Missed Detection False Detection Tfail Tdetect time F Detection Delay D D True Detection F Missed Detection False Detection Tfail Tdetect time F D Failure occurrence Failure detection Failure duration F D Failure occurrence Failure detection Failure duration

A so-called failure duration interval is associated with each failure, which starts with the occurrence of a failure and ends with the elimination of the failure (e.g., by repairing the error involved). A true detection is a detection which is reported by the cell outage detection mechanism during the failure duration

• interval. In contrast a false detection is reported outside the failure duration interval.

The following abbreviations are used: Tfail is the time instant when the failure occurred Tdetect is the time instant when the failure is detected Ndetect is the number of detections within the failure duration interval (true detection) Nfail is number of failures observed during the simulation Nfalse is the number of detections outside the failure duration interval (false detection) Signalling Overhead can be divided into two parts:

• Transport network signalling overhead captures all data that is transmitted over the transport network,

e.g., between eNodeBs (X2), between eNodeB and MME (S1), and eNodeB and OSS (Itf-S) measured in either number of messages sent / time unit / eNodeB or bytes sent / time unit / eNodeB.

• Radio interface signalling overhead captures the communication needed between a UE and an eNodeB

in order to facilitate outage detection. This is measured using the number of resource blocks / time unit / eNodeB scheduled for cell outage detection specific communication. Processing overhead refers to the amount of processing needed to detect the outage. The execution time

• f a particular algorithm typically depends on the size n of the input data, e.g., number of measurements,

and can be asymptotically logarithmic (log n), polynomial (na), or even exponential (an). An analysis of asymptotic execution time gives an insight in the processing demand of an algorithm. Assume that there are nL levels of severity, where each severity is characterized by multiple factors, e.g., reduction in QoS, capacity, and coverage. Let Ldetector denote the severity indication level estimated by the detection algorithm and Ltrue be the true severity indication level, i.e., the level classified if true network conditions are known. The level estimation accuracy AL is a function of Ldetector - Ltrue, e.g., AL = |Ldetector - Ltrue|. Non-linear variations of AL may also be considered, e.g., AL = (Ldetector - Ltrue)2.

13

WWW.FP7-SOCRATES.EU

14/20 14/20

Assessment Criteria – Cell Outage Compensation

Call blocking ratio; Call Dropping ratio; Throughput, etc. Coverage [%] = (Nbin – Nbin_outage)/Nbin Other metrics such as convergence time per cell

Margin 1 Ti TSS2 TSS1 Time Margin 2 Coverage Throughput Convergence Time Network Performance (Coverage, Throughput)

The following definitions are used:

• Call blocking ratio is the ratio of the number of blocked calls to the number of calls that attempt to access

the network

• Call dropping ratio is the probability that an existing call is dropped before it is finished i.e. the ratio of

the number of dropped calls to the number of calls that are accepted by the network.

• Throughput is the rate of successful data delivery measured in bits per second or data packets per

second, and calculated as the number of bits (or packets) that are successfully delivered in a certain time period, divided by the length of that time period.

• Nbin is the number of pixels within the cell
• Nbin_outage is the number of pixels in outage, i.e. pixels having average SINR or data rate lower than a pre-

defined threshold

• Ti is the outage detection time
• TSS1 is the time instant of steady state for the cell throughput
• TSS2 is the time instant of steady state for the cell coverage

The convergence time is defined as the difference between the time the outage compensation algorithm reaches a steady state on the network performance and the time outage is detected. This is measure of how fast the compensation algorithm reconfigures the network. Different network performance metrics can be considered here. In the figure above, coverage and throughput performance metrics are chosen for illustration. The convergence time of coverage is obtained by TSS2 - Ti for one cell (similar definition is used for, e.g. throughput). The margins (Margin 1 and 2) presented in the figure are used to set some allowed deviation to the final state of performance. Margin 1 can be considered e.g. ±1%, meaning that coverage can

• scillate between the steady state plus or minus 1%. TSS2 is set to the earliest time after which coverage is

within margin 1. The convergence time in the area under consideration is obtained as the maximum of convergence times of all the cells involved in the cell outage compensation.

14

WWW.FP7-SOCRATES.EU

15/20 15/20

Simulator

Monte-Carlo based static simulator

will be used for cell outage compensation due to

– Simpler modeling and – Faster execution time

At each iteration an eNodeB

(sector):

– Samples/gathers performance – Updates radio parameters

The time between two iterations is

assumed to be in the order of minutes or tens of minutes

– Small correlation between the

samples

– Correlation is ignored

UE generation Simulate PHY, RRM etc Sample Performance Compensation Algorithm Final snapshot? No Round of algorithm iteration Pathloss (G-matrix) Cell Selection It is assumed that a compensation algorithm runs over several iterations. For each snapshot the following steps are carried out:

• UEs are distributed randomly across the network
• The pathloss between each UE and base station is calculated, adding the lognormal fading, and stored to

a so-called G-matrix. Distance attenuation and lognormal fading are kept constant during the execution of a snapshot

• UEs perform cell selection based on best server (RSRP)
• Physical layer and radio resource management (RRM), e.g., scheduler and power control, models are

executed.

• The performance is sampled and statistics is collected.
• The cell outage compensation algorithm is executed. If needed the algorithm is executed every X

snapshots depending on whether sufficient statistics is gathered

15

WWW.FP7-SOCRATES.EU

16/20 16/20

Socrates will continue the work on cell outage detection and compensation Next steps involve:

– Simulator implementation (finished by April 2009) – Controllability studies (finished by May 2009) – Observability studies (finished by May 2009) – Intermediate report D4.2A (finished by Oct 2009) – Algorithm development and evaluation (beginning of 2010)

Publications and reports

Future Work

Next steps involve modelling various aspects of LTE and implementing simulators for cell outage compensation and coverage map estimation. The simulator implementation is planned to be finished by April 2009. A controllability study will be carried out in order to see to what degree an outage can be compensated (by manual means) as well as to understand the relation between control parameters and overall goal and assessment criteria of cell outage compensation. For example, a study will be carried out to understand the impact of antenna tilt on coverage and quality. The controllability study is planned to be finished by May 2009. The questions that will be addressed in the observability study deal mostly with to what degree measurements, counters etc from the network can aid in

• detecting an outage,
• estimating the impact of compensation measures and the performance of the compensation function,

and

• estimating the coverage in the network

The observability study is planned to be finished by May 2009. There will be an intermediate report finalized by Oct 2009. Algorithms for cell outage detection and compensation will be developed using the knowledge obtained in the controllability and observability

• study. The algorithm development and evaluation is planned to be finished by beginning 2010. Socrates

will publish results in project deliverables, conferences, and journals.

16

WWW.FP7-SOCRATES.EU

17/20 17/20

• M. Amirijoo (Ericsson), mehdi.amirijoo@ericsson.com

(presenting author)

• L. Jorguseski (TNO ICT), ljupco.jorguseski@tno.nl
• T. Kürner (TU Braunschweig), t.kuerner@tu-bs.de
• R. Litjens (TNO ICT), remco.litjens@tno.nl
• M. Neuland (TU Braunschweig), m.neuland@tu-bs.de
• L. C. Schmelz (Nokia Siemens Networks), lars.schmelz@nsn.com
• U. Türke (Atesio), tuerke@atesio.de

Contact

Thank you very much for your attention

FP7 ICT-SOCRATES