Rapport Copie Finale PDF

MINISTERE DE L’ENSEIGNEMENT SUPERIEUR
UNIVERSITE DE SOUSSE
‫المعهد العالي لإلعالمية وتقنيات االتصال بحمام سوسة‬
Institut Supérieur D’informatique

Et Des Techniques De Communication – Hammam Sousse
MEMOIRE DE PROJET DE FIN D’ETUDES
Présenté en vue de l’obtention du Diplôme National d’Ingénieur

Spécialité : Téléinformatique
Performance Analysis and Fault detection tool for the
3G-UTRAN and 4G-EUTRAN based on Business Intelligence
Réalisé par :
Mlle Sourour MEJBRI
Encadré par :
Mme Hédia JEGHAM
Mr Jawhar BEN ABED
Année Universitaire 2016 - 2017

Acknowledgements
MINISTERE DE L’ENSEIGNEMENT SUPERIEUR

DE LA RECHERCHE SCIENTIFIQUE ET DE LA TECHNOLOGIE
UNIVERSITE DE SOUSSE
‫المعهد العالي لإلعالمية وتقنيات االتصال بحمام سوسة‬
INSTITUT SUPERIEUR D’INFORMATIQUE

ET DES TECHNOLOGIES DE COMMUNICATION – HAMMAM SOUSSE
MEMOIRE DE PROJET DE FIN D’ETUDES
Présenté en vue de l’obtention du Diplôme National d’ingénieur

Spécialité : Téléinformatique
Performance Analysis and Fault detection tool for the
3G-UTRAN and 4G-EUTRAN based on Business Intelligence
Réalisé par :
Encadrant : Date : …………………Signature :…………….
Encadrant : Date :……………………Signature :…………….
Année Universitaire 2016 - 2017

Acknowledgements
Acknowledgements
As the long journey towards my Thesis comes to an end, it is the time for me to
express my deep and sincere gratitude to everyone who gave me the possibility to complete
this dissertation. As always, I would like to devote my utmost gratefulness to the
incomparable professor at ISITCom Mme Hedia JEGHAM, who has been such a strong
and supportive advisor and turned me from a novice into a mature engineer in the last four
months. Her strong demand for excellence and relentless attention to detail will be guiding
me for the rest of my job as software engineer. I am also truly grateful for the infinite
patience and hours she spent on helping me improve my work, writing, and presentation
skills.
A special ‘thank you’ goes to Mr Jawhar BEN ABED the Quality Assurance Engineer
at HUAWEI TECHNOLOGIES. I want to express my deepest gratitude for considering me an
eligible candidate for this internship. Nothing I can say will ever convey the amount of
gratitude, I really appreciate you taking the time to discuss the advancement of my work.
Last, but not least, I gratefully acknowledge my family – from my father who is the
person I always look up to, to my mother who provides me infinite passion and energy, to my
sisters and brothers specially my little sister who brightened a couple of hard working days
with her smile.
Contents
Contents
Acknowledgements ........................................................................................................... i
Contents .......................................................................................................................... ii
Figures List ..................................................................................................................... iii
Tables List ...................................................................................................................... iv
Acronyms List .................................................................................................................. v
Introduction ..................................................................................................................... 1
Chapter 1 State Of Art ....................................................................................................... 3
1 Introduction .......................................................................................................... 4
2 Huawei presentation ............................................................................................. 4
3 Project presentation .............................................................................................. 4
4 Study of existing solutions ..................................................................................... 5
5 The adopted Methodology ...................................................................................... 7
6 Conclusion............................................................................................................10
Chapter 2 Theoretical concepts about Mobile Networks................................................... 11
1 Introduction ......................................................................................................... 12
2 Mobile network overview ...................................................................................... 12
2.1.1 GSM (Global System for Mobile communications) .......................................................... 13

2.1.2 GPRS (General Packet Radio Service) .............................................................................. 13
2.1.3 EDGE (Enhanced Data rates for GSM Evolution)............................................................ 13
2.2.1 UMTS (Universal Mobile Telecommunications System) ................................................. 13

2.2.2 HSDPA (High Speed Down-link Packet data Access) ..................................................... 15
2.2.3 HSUPA (High Speed Up-link Packet data Access ............................................................ 15
2.3.1 HSPA+ (Evolved High-Speed Packet Access) .................................................................. 16

2.3.2 LTE (Long Term Evolution) ............................................................................................. 17
2.4.1 LTE Advanced .................................................................................................................... 18

3 UMTS Network architecture ................................................................................. 19
4 LTE Network Architecture .................................................................................... 21
5 Performance Measurement ................................................................................. 25

Contents
5.3.1 Definition ...........................................................................................................................26

5.3.2 KPI’s categories ................................................................................................................. 27
5.4.1 Accessibility ....................................................................................................................... 28

5.4.2 Mobility ..............................................................................................................................29
5.4.3 Retainbility ....................................................................................................................... 30
5.5.1 Accessibility KPI................................................................................................................. 31

5.5.2 Retainability KPIs.............................................................................................................. 32
5.5.3 Mobility KPIs ..................................................................................................................... 33
5.5.4 Service Integrity KPIs........................................................................................................ 35
6 Fault Monitoring ................................................................................................. 35
7 Conclusion........................................................................................................... 38
Chapter 3 Business Intelligence & Data Mining............................................................... 39
1 Introduction ........................................................................................................ 40
2 Business Intelligence ........................................................................................... 40
2.2.1 Data Source Layer .............................................................................................................. 41

2.2.2 ETL (Extract-Transform-Load) Layer ..............................................................................42
2.2.3 Data Warehouse Layer ......................................................................................................43
2.2.4 End User Layer ..................................................................................................................43
3 Data Mining ......................................................................................................... 44
3.2.1 The classification ............................................................................................................... 45

3.2.2 The estimation ................................................................................................................... 45
3.2.3 The Segmentation ............................................................................................................. 45
3.2.4 Forecasting ........................................................................................................................46
3.2.5 Association.........................................................................................................................46
3.4.1 Choosing an Algorithm by Type ........................................................................................ 47

3.4.2 Choosing an Algorithm by Task ....................................................................................... 48
4 Comparative study about Business Intelligence tools ........................................... 48
Contents
4.3.1 Comparing Data Mining .................................................................................................... 51

4.3.2 Comparing Packaging and Pricing ................................................................................... 51
4.3.3 Summarizing the analysis ................................................................................................. 52
5 Conclusion........................................................................................................... 53
Chapter 4 Requirements Specification ........................................................................... 54
1 Introduction .........................................................................................................55
2 Actors Identification .............................................................................................55
3 Functional requirements ......................................................................................55
4 Global Use Case Diagram ..................................................................................... 56
5 Refinement of use cases ....................................................................................... 56
5.1.1 Use case “Performance Analysis” Diagram ....................................................................... 56

5.1.2 Textual Description of the use case “Analyze 3G Key Performance Indicator” .............. 57
5.2.1 Use case “Monitor alarms” Diagram .................................................................................58

5.2.2 Textual Description the use case “Monitor 4G EUTRAN alarms” .................................. 59
5.3.1 Use case “Perform Prediction” Diagram ........................................................................... 59

5.3.2 Textual Description of the use case “Predict UTRAN KPIs Behavior”........................... 60
5.3.3 Textual Description of the use case “Predict EUTRAN Alarms Behavior” ..................... 61
6 Package Diagram of Use cases ............................................................................... 61
7 Conclusion........................................................................................................... 62
Chapter 5 Conceptual Study........................................................................................... 63
1 Introduction ........................................................................................................ 64
2 Static modelization .............................................................................................. 64
3 Dynamic Modelization ..........................................................................................67
3.2.1 Sequence Diagram of “Predict 3G KPI Trend” use case .................................................. 68

3.2.2 Sequence Diagram of “Predict 4G alarm causes Trend” .................................................69
4 Conclusion........................................................................................................... 69
Chapter 6 Realization .................................................................................................... 70
1 Introduction ......................................................................................................... 71
2 Deployed Technologies ......................................................................................... 71
2.1.1 Microsoft SQL Server components .................................................................................... 71

2.1.2 Business Intelligence SQL server implementation process.............................................. 72
2.1.3 SQL Server data mining features ...................................................................................... 73
2.2.1 Data Definition Statements ............................................................................................... 74

2.2.2 Data Manipulation Statements ......................................................................................... 75
Contents
3 Application Demonstration ................................................................................... 77
4 Conclusion........................................................................................................... 84
General Conclusion ........................................................................................................ 84
Bibliography .................................................................................................................. 85
Webography .................................................................................................................. 86
Figures List
Figures List
Figure 1 Proposed solution diagram ................................................................................................ 6

Figure 2 RUP phases in the software process ................................................................................. 7
Figure 3 RUP Work flow Process ..................................................................................................... 8
Figure 4 Scrambling mechanism ................................................................................................... 14
Figure 5 UMTS Network Architecture .......................................................................................... 19
Figure 6: The EPS network elements ............................................................................................ 22
Figure 7: RRCConnectionRequest message .................................................................................. 31
Figure 8: ERAB message ................................................................................................................ 32
Figure 9 BI architecture ................................................................................................................. 41
Figure 10 example of the process of ETL on some unstructred data ........................................... 42
Figure 11 an example of The ETL process ..................................................................................... 42
Figure 12 Neural Network layers .................................................................................................. 47
Figure 13 largest BI vendors of the world IT market .................................................................... 49
Figure 14 Global use case diagram ................................................................................................ 56
Figure 15 Use case “Performance Analysis “ ................................................................................. 57
Figure 16 Use case “Alarm monitoring” ........................................................................................ 58
Figure 17 Use case “Perform Prediction” .......................................................................................60
Figure 18 Use Cases Package Diagram .......................................................................................... 62
Figure 19 Class Diagram ................................................................................................................. 65
Figure 20 Package Diagram ........................................................................................................... 66
Figure 21 Deployment Diagram ..................................................................................................... 67
Figure 22 Activity diagram of "Predict 3G KPI trend" use case ................................................... 67
Figure 23 Sequence Diagram of “Predict 3G KPI Trend” use case ..............................................68
Figure 24 Sequence Diagram of “Predict 4G alarm causes Trend” ............................................. 69
Figure 25 The core components of the SQL Server product ....................................................... 71
Figure 26 Login Interface ............................................................................................................... 77
Figure 27 3G UTRAN Performance Analysis Interface ................................................................ 78
Figure 28 Average Values of each category's KPIs ...................................................................... 79
Figure 29 3G KPIs of the Retainability category ........................................................................... 79
Figure 30 Retainability KPIs ......................................................................................................... 80
Figure 31 the Retainability KPI : "3G Call Drop Rate" Trend ..................................................... 80
Figure 32 UTRAN Alarm Monitoring interface ............................................................................ 81
Figure 33 filter information Interface ........................................................................................... 81
Figure 34 Key Performance Indicator Prediction Interface .........................................................82
Figure 35 Parametrizing the neural network results ....................................................................83
Figure 36 alarms causes distribution according to neural network model .................................83
Figure 37 another example of alarms causes distribution according to neural network model 83
i
Tables List
Tables List
Table 1 RUP Workflows description ................................................................................................ 8
Table 2 Workflows Chronogram ...................................................................................................... 9
Table 3 Differences between HSDPA and HSUPA ....................................................................... 16
Table 4 comparing data mining capabilities ................................................................................. 51
Table 5 Comparing Packaging and Pricing ................................................................................... 52
Table 6 Textual Description of the use case “Analyze 3G Key Performance Indicator” ............. 57
Table 7 Textual Description the use case “Monitor 4G EUTRAN alarms” .................................. 59
Table 8 Textual Description of the use case “Predict UTRAN KPIs Behavior” ...........................60
Table 9 Textual Description of the use case “Predict EUTRAN Alarms Behavior” .................... 61
Table 10 Components of Microsoft BI solution ............................................................................ 72
Table 11 DMX statements ............................................................................................................... 74
Table 12 Data Manipulation Statements ....................................................................................... 75
Table 13 Data Mining prediction queries ...................................................................................... 76
ii
Acronyms List
Acronyms List
FDT: Fault Detection Tool
FMA: Fault Manage Assistant
RUP: Rational Unified Process
BI: Business Intelligence
KPI: (Key Performance Indicator
3GPP: 3rd Generation Partnership Project
GSM: Global System for Mobile communications
GPRS: General Packet Radio Service
EDGE: Enhanced Data rates for GSM Evolution
UMTS: Universal Mobile Telecommunications System
HSDPA: High Speed Down-link Packet data Access
HSUPA: High Speed Up-link Packet data Access
HSPA+: Evolved High-Speed Packet Access
LTE: Long Term Evolution
RAN: Radio Access Network
UTRAN: Universal Terrestrial Radio Access Network
EUTRAN: Evolved Universal Terrestrial Radio Access Network
CN: Core Network
MSC: Mobile Switching Center
GGSN: Gateway GPRS Support Node
GMSC: Gateway MSC
SGSN: Serving GPRS Support Node
HLR: Home location register
EIR: Equipment identity register
AuC: Authentication center
RNC: Radio Network Controller
RRM: Radio resource management
P-GW: PDN Gateway
S-GW: Serving Gateway
MME: Mobility Management Entity
HSS: Home Subscriber Server
PCRF: Policy Control and Charging Rules Function
EPC: Evolved Packet Core
PDN: Packet data network
RNC: Radio Network Controller
iii
Acronyms List
RNS: Radio Network Subsystem

MIMO: Multiple Input Multiple Output
MSC: Mobile Switching Center
GMSC: Gateway MSC
OFDMA: Orthogonal Frequency Division Multiple Access
SC-FDMA: Single Carrier - Frequency Division Multiple Access
CDR: Call Detail Records
RLC: Radio Link Control
RAB: Radio Access Bearer
E-RAB: Evolved Radio Access Bearer
VoIP: Voice Over IP
CDR: Call Drop Rate
RRC_SR: Setup Success Rate
CS: Circuit System
PS: packet System
CSSR: Call Setup Success Rate
RRC: Radio Resource Control
HOIn_SR: Handover In Success Rate
IRATHO_E2W_SR: Inter_RAT Handover Out Success Rate (LTE to WCDMA)
IRATHO_E2G_SR: Inter_RAT Handover Out Success Rate (LTE to GSM)
ETL: Extract-Transform-Load
ODS: Operational Data Store
OLAP: Online Analytical Processing
SSAS: SQL Server Analysis Services
DMX: Data Mining Extensions
iv
General Introduction
From the early analogy mobile generation (1G) to the last implemented fourth
generation (4G) the paradigm has changed. The new mobile generations do not pretend to
improve the voice communication experience but try to give the user access to a new global
communication reality. The aim is to be able to reach communication every time and
everywhere and to provide users with a new set of services.
The growth of the number of mobile subscribers over the last years led to a saturation
of voice-oriented wireless telephony. It is now time to explore new demands and to find new
ways to extend the mobile concept. The first steps have already been taken by the 2.5G, which
gave users access to a data network (e.g. Internet access, MMS - Multimedia Message Service).
However, users and applications demanded more communication power. As a response to this
demand a new generation with new standards has been developed - 3G, moreover for more
sophisticated applications use and new generation of services the -4G technologies shall
respond to all the new requirements for best mobile radio technologies.
Today’s 3G specifications call for 144 Kb/s while the user is on the move, 384 Kb/s for
pedestrians, and ups to 2 Mb/s for stationary users. That is a big step up from 2G bandwidth
using 8 to 13 Kb/s per channel to transport speech signals. Besides UMTS introduce a new
access technique which is WCDMA. However, the 4G wireless uses Orthogonal Frequency
Division Multiplexing (OFDM), Ultra Wide Radio Band (UWB), and smart antenna. So it gave
more improvements where an average of data rate of 20mb/s is employed. Mobile speed will
be up to 200km/hr, the frequency band is between 2 and 8 GHz and it gives the ability for
worldwide roaming to access cell anywhere.
Hence, there are different types of challenges in 3G and 4G Radio Access Network. With
this evolution in term of user’s number and application diversity, there is the problem of
troubleshooting and fault management that should be taken in consideration to improve
customer satisfaction.
Business Intelligence and Data Mining analytical tools can be very effective for
analyzing Performance across network operations and describing fault diagnosis. Data Mining
technologies can be used for identifying faults by extracting knowledge from network data.
Under this context, this internship took place in designing faults detective tool based
on business intelligence and performance analysis.
In The first chapter we begin by putting the project in context, the host company. Then
we go through the presentation of the proposed solution. The first part of the report offers an
overview of the organization, followed by a presentation of our project. Following, it proceeds
to study some existing solution similar to our project. Then the proposed solution is described
1
in some detail. Finally, we end up defining the methodology that we adopted to manage our
project.
The second chapter, presents theoretical concepts about mobile networks and describes
how the different interfaces of the 3G and the 4G Radio Access Network can be monitored. It
also outlines which important performance parameters are required to check the performance
of the most essential Radio network functions.
The third chapter presents a literature review related to Business Intelligence and Data
Mining in Telecommunications, first we define the main areas of BI and Data Mining
applications. Last, we identify the most common Data Mining techniques and methods used.
The fourth chapter contains a complete description of the behavior of the system to be
developed. We begin by defining the actors then we present the functional requirement of our
project. In the next step we analyze our system requirements through the development of use
case diagrams and their detailed text descriptions.
The fifth chapter presents a set of conceptual models used to construct a simulation of
our application. We begin by a static modelization where will describe the static structure of
the system. We warp up by a dynamic modelization where we express and model the behavior
of our application system over time.
The last chapter is the realization. We begin by the presentation of the deployed
technologies, then we present the configuration of the data mining technique used in our
application. Finally we demonstrate the realization of our project by presenting the interfaces
of the application.
2
Chapter 1 State Of Art
1 Introduction............................................................................................................................ 4
2 Huawei presentation .............................................................................................................. 4
3 Project presentation ............................................................................................................... 4
4 Study of existing solutions ...................................................................................................... 5
5 The adopted Methodology ..................................................................................................... 7
6 Conclusion .............................................................................................................................10
3
1 Introduction
We begin in this chapter by putting the project in context, the host company. Then we
go through the presentation of the proposed solution. The first part of the report offers an
overview of the organization, followed by a presentation of our project. Following, it proceeds
to study some existing solution similar to our project. Then the proposed solution is described
in some detail. Finally, we end up defining the methodology that we adopted to manage our
project.
2 Huawei presentation
This project will be carried out at Huawei Technologies, a leading telecom solutions
provider. It has established end-to-end advantages in Telecom Networks, Global Services and
Devices. With comprehensive strengths in wire line, wireless and IP technologies, Huawei has
gained a leading position in the All-IP convergence age. Their products and solutions have been
deployed in over 140 countries and have served 45 of the world's top 50 telecom operators, as
well as one third of the world's population. The Huawei Company consists of three accounts
Tunisie Telecom, Tunisiana Oreedoo and Orange Tunisie.
3 Project presentation
Network data Telecommunication networks contain thousands of components, which
are interconnected. These components are capable of generating error and status messages
which leads to a large volume of network data. These network data are used for network
management functions like fault detection.
Therefore, manufacturers like HUAWEI are always looking for solutions to optimize
this task because it is essential to be efficient in the procedure of management of the
breakdowns.
The subject of this project, proposed and held by HAUWEI, is in this context. It is titled
"Performance Analysis and Fault Detection tool for the 3G UTRAN and 4G
EUTRAN based on Business Intelligence".
The aim of this project is to develop an application capable of analyzing raw logs of
multiple RNCs and eNodeBs. These raw logs could contain either values of Radio Access
Network (RAN) KPI (Key Performance indicators) or a number of generated alarms during a
specific date.
4
Users are able then to visualize the behavior of the RAN, to evaluate it performance in
terms of meeting mobile network’s strategic and operational goals. And to predict network faults
trend by using a data mining technique which will be explained in details in further chapter.
"3G, 4G Faults Detection Tool" is a desktop application that will help HUAWEI
engineers detect 3G UTRAN and 4G EUTRAN troubleshooting.
Operators can use these insights to increase network capacity, for example, finding the
right locations to deploy low-powered small cells to boost data speeds for subscribers inside
high-rise buildings. Alternatively, they can provide wider coverage areas and more reliable
voice services by automatically adjusting the angle of base station antennas.
4 Study of existing solutions

4.1 Study of the M2000 Server
The main work of HUAWEI is based on the M2000 server. In fact, The M2000 provides
centralized O&M functions for the HUAWEI mobile element management solution. The
M2000 provides mediations for connecting various types of NEs.
The M2000 provides basic functions such as configuration management, performance

management, fault management, security management, log management, topology
management, software management, and system management. It also provides various
optional functions.
The M2000 is a future-proof solution for managing mobile networks. It provides a

centralized network management platform for supporting users in their long-term network
evolution and shielding the differences between various network technologies. (1) The M2000
focuses on the continuous efforts of users toward better network OM and retains the O&M
experience of previous network management systems
4.2 Study of the FMA (Fault Manage Assistant)

To improve the efficiency of UMTS fault location, the GU Delivery Department
developed the Fault Manager Assistant (FMA) for fast fault location and troubleshooting.
Currently, the FMA analyzes faults based on the following data sources:
 Configuration script of the RNC.

 Original alarm files of the RNC.
 KPI file of the RNC.
 Original performance measurement of the RNC.
 Operation log of the RNC.
 TMF File Log
5
4.3 Critiques
The M2000 server is an online tool that can help HUAWEI engineers to detect troubleshooting
in the radio network such as: KPI degradation and Alarms. To use these functionalities users
should have access to it. But not all engineers in HUAWEI are allowed to connect to M2000.
However, our Fault Detection Tool will be an offline tool that can be used by everyone:
engineers, client and trainee also.
While FMA is only destined to manage 3G network. That’s why our application aims to
manage not only 3G network but also 4G network. FMA can’t be installed in the user laptop
unless he has Window 7 professional provided by HUAWEI to their employees for security
reasons. So only HUAWEI engineers can use this application. Our application FDT is open for
everyone that have a login.
4.4 Proposed solution

Our project consist of implementing a desktop application destined to HUAWEI
engineer, trainees or HUAWEI client.
This application gives users to ability to:
 Analyze Performance Measurement of the UMTS and LTE radio interface through a set of
Key Performance indicators (KPI).
 Detect, identify and interpret anomalies and suspicious behavior of the UMTS and LTE
radio interface through the analysis of current and historical alarms and cause diagnosis.
Our application utilizes clever data mining algorithm for much deeper and intelligent analysis
of raw data. Users are then capable of handling “what will happen?” matter to improve network
utilization and achieve best user experiences.
Finally, for more flexibility, our application provide an interface that helps the user generate
reports. So our user can create his own archive and he can also create a reference that can be
used by his colleagues in case they encounter the same problems.
Alarms
Performance Data
KPIs
Performance Analysis
Fault detection
Data Mining
Performance Degradation Prediction
Algorithm
Faults Cause Prediction
Figure 1 Proposed solution diagram

6
5 The adopted Methodology

The adoption of a development methodology is a necessary to ensure an acceptable level
of quality and to avoid any spillover in terms of deadlines. Because the project specification
given by Huawei does not require a specific methodology to use, we needs to take extra caution
before deciding on the ideal methodology which we can plan, structure and control the overall
process of developing our project.
We used The Rational Unified methodology. Throughout the next section we’ll explore
this method in detail, examine the life cycles that software will iteratively progress through,
then we outline a few advantages and potential disadvantages to using the Rational Unified
Process in our project.
5.1 RUP (Rational Unified Process) Presentation

RUP is a software development process from Rational, a division of IBM. It can be
described in two dimensions, or along two axis:
Dynamic perspective: RUP identifies four discrete phases in the software process:
Figure 2 RUP phases in the software process
1. Inception: The aim of this phase is to establish fundamental business requirement for the
system through a set of preliminary use cases.
2. Elaboration: The goals of the elaboration phase are to develop an understanding of the
problem domain, establish an architecture for the system and propose the project plan.
3. Construction: The construction phase involves system design, programming and testing.
Parts of the system are developed in parallel and integrated during this phase.
4. Transition: In this final phase the software is released to the public. Final adjustments or
updates are made based on feedback from end-users.
Static perspective: RUP focuses on the activities that take place during the development
process. These are called workflows. The RUP has been designed in conjunction with the UML,
7
so the workflow description is oriented around associated UML models such as sequence
models, object models, etc… The support workflows are shown in figure 3 and explained in
table 1.
Figure 3 RUP Work flow Process
Table 1 RUP Workflows description
Workflow Description
Business The business processes are modelled using business use cases.
modeling
Requirements Actors who interact with the system are identified and use cases are developed
to model the system requirements.
Analysis and A design model is created and documented using architectural models;
design component models; object models and sequence models.
Implementation The components in the system are implemented and structured into
implementation sub-systems.
Testing This an iterative process that performs an objective evaluation to ensure quality.
This includes finding defects, validating that system works as designed, and
verifying that the requirements are met.
Deployment A product release is created; distributed to users and installed in their work
place.
Configuration and This workflow manages changes to the system including tracking versions over
change time, controlling and managing changes.
management
8
Project This workflow manages the system development including managing risks,
management directing people (assigning tasks, tracking progress, etc.), and coordinating with
people and systems outside the scope of the project to be sure that it is delivered
on time and within budget.
Environment This workflow is concerned with making appropriate tools (hardware, software,
etc.) available to the team as needed.
5.2 Reasons for choosing RUP methodology

Our project is based on many unfamiliar aspects such as Performance analysis of the 3G and
4G Radio Access Network and the decision-making discipline that is based on predictions and
data mining. We had not been given specific tools to use with developing or with data mining
integration. So we decided to use RUP because with this methodology the development process
itself can be improved and refined along the way so we can learn along the way.
With RUP, the assessment at the end of iteration not only looks at the status of the project
from a product or schedule perspective, but also analyzes what should be changed in the
organization and in the process to make it perform better in the next iteration.
5.3 Workflows Chronogram

Table 2 Workflows Chronogram
workflows Week
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Study of the existing solution
Define requirements and specifications
Study the3G and 4G network features

Inception
Study the Performance Analysis of Radio Access Network
Discover the features of business intelligence
refine requirements and add the data mining functionalities
Searching for solution to implement data mining to our

Elaboration
application.
Familiarize with the SQL Server analysis Services (SSAS)
and it Data Mining features.
Application design
Construction
Programming basic functionalities
Implementing the Neural network Algorithm using the

DMX (Data Mining Extension) language
Generation of the .exe software
Transit
The application is given to end user for beta testing

ion
9
6 Conclusion
This chapter presented the general framework of this project carried out within Huawei
Technologies. Therefore, we highlighted the context of our application then we presented the
methodology chosen for the planning and the smooth progress of our project. The next chapter
will define the basic concepts of mobile networks, Performance analyzing and fault Monitoring
in the radio Access Network.
10
Chapter 2 Theoretical concepts about Mobile Networks
Chapter 2 Theoretical concepts about

Mobile Networks
1 Introduction...........................................................................................................................12
2 Mobile network overview ......................................................................................................12
2.1.1 GSM (Global System for Mobile communications) ...................................................... 13

2.1.2 GPRS (General Packet Radio Service) .......................................................................... 13
2.1.3 EDGE (Enhanced Data rates for GSM Evolution) ......................................................... 13
2.2.1 UMTS (Universal Mobile Telecommunications System) .............................................. 13

2.2.2 HSDPA (High Speed Down-link Packet data Access) ................................................... 15
2.2.3 HSUPA (High Speed Up-link Packet data Access ......................................................... 15
2.3.1 HSPA+ (Evolved High-Speed Packet Access) ................................................................ 16

2.3.2 LTE (Long Term Evolution) .......................................................................................... 17
2.4.1 LTE Advanced ............................................................................................................... 18

3 UMTS Network architecture ..................................................................................................19
4 LTE Network Architecture ......................................................................................................21
5 Performance Measurement ...................................................................................................25
5.3.1 Definition ..................................................................................................................... 26

5.3.2 KPI’s categories ........................................................................................................... 27
5.4.1 Accessibility ................................................................................................................. 28

5.4.2 Mobility ....................................................................................................................... 29
11
5.4.3 Retainbility .................................................................................................................. 30
5.5.1 Accessibility KPI............................................................................................................ 31

5.5.2 Retainability KPIs ......................................................................................................... 32
5.5.3 Mobility KPIs ................................................................................................................ 33
5.5.4 Service Integrity KPIs ................................................................................................... 35
6 fault Monitoring ....................................................................................................................35
7 Conclusion .............................................................................................................................38
1 Introduction
Operators need to understand, measure, and manage both quality and performance of the 3G
and 4G services they offer. This has to be fulfilled on the technical quality of service (QoS) as
well as on the quality of experience (QoE) level. This chapter describes how the different
interfaces of the Universal Terrestrial Radio Access Network (UTRAN) and the Evolved
Universal Terrestrial Radio Access Network (E-UTRAN) can be monitored. It also outlines
which important performance parameters are required to troubleshoot and optimize the
network and a set of counter-based Key Performance Indicators (KPIs) is proposed to check
the performance of the most essential Radio network functions.
2 Mobile network overview

The cellular technology evolution has been going on since the late 1950s, though the
first commercial systems came into being in the late 1970s and early 1980s. Here is a brief
overview of the cellular technologies and the networks that made an impact on the
development and the fast evolution of the mobile communications.
2.1 Second Generation Cellular Networks

2G comprised of the following Mobile technologies: General Packet Radio Service
(GPRS), Code Division Multiple Access (CDMA), Global System for Mobile Communication
(GSM) and Enhanced Data Rates for GSM Evolution (EDGE).
12
2.1.1 GSM (Global System for Mobile communications)

The well-known GSM (Global System for Mobile communications) entered into
commercial service in 1991. This technology uses a TDMA/FDMA (Time Division Multiple
Access/ Frequency Division Multiple Access) radio multiplexing scheme and GMSK (Gaussian
Minimum Shift Keying) radio modulation. In its early days, GSM was only providing a voice
service, SMS (Short Message Service) and low-rate circuit-switched data at 9.6 Kb/s.
2.1.2 GPRS (General Packet Radio Service)

GPRS (General Packet Radio Service) is an evolution of GSM, designed for packet data
communication. Based on the same GMSK modulation, GPRS provides, however, better user
rate by providing the least robust but fastest coding scheme (theoretical maximum bit rate is
around 20 Kb/s per radio time slot) and the possibility to aggregate several radio slots for data
transmission.
This system was intended to be more suitable and efficient for packet data applications,
as it relies on sharing a set of radio channels among several terminals, for both uplink and
downlink transmission.
2.1.3 EDGE (Enhanced Data rates for GSM Evolution)

EDGE (Enhanced Data rates for GSM Evolution) introduces the 8-PSK (8 Phase Shift
Keying) modulation, providing increased data rate. EDGE provides a maximum theoretical bit
rate of around 50 Kb/s per radio time slot. When combined with GPRS, it is possible to
aggregate several EDGE-coded radio slots to increase the peak data rate for packet applications
(2) . EDGE enhances the data throughput by a combination of different techniques:
 EDGE's 8-level phase-shift keying (8PSK) modulation scheme allows a raw radio
throughput increase of 3 times the raw radio throughput of GPRS with the same number
of timeslots.
 Nine new coding schemes using both Gaussian minimum shift keying (GMSK) and 8PSK
modulations and a link adaptation algorithm are introduced to automatically switch
between coding schemes based on the radio environment.
2.2 3G technologies
2.2.1 UMTS (Universal Mobile Telecommunications System)
UMTS is based on a standard developed by the 3GPP (3rd Generation Partnership
Project). UMTS relies on the CDMA (Code Division Multiple Access) multiplexing scheme
using a high chip rate direct spread sequence. In its first form, UMTS/FDD advantages were
limited to increased data rates (up to 384 Kb/s per user on a single channel), the possibility of
simultaneous packet and circuit applications.
13
A theoretical maximum bit rate of 2 Mb/s over dedicated channels has been defined
within the specification, but has never been deployed as such in commercial networks. UMTS
is often presented as the 3G evolution of GSM networks. Although the UMTS radio interface is
completely different from the GSM/EDGE one, a lot of architectural concepts and procedures
have been inherited from GSM.
UMTS uses Wideband CDMA (WCDMA) as the radio transmission standard. It

employs a 5 MHz channel bandwidth. Using this bandwidth, it has the capacity to carry over
100 simultaneous voice calls, or it is able to carry data at speeds up to 2 Mbps in its original
format.
A technique called scrambling is carried out by the transmitter makes it possible to

separate the various signals of the same base station or of the same terminal without modifying
either the bit rate or the bandwidth. This makes it possible to spread a signal by several
transmitters with the same spreading code without compromising the detection of the signals
by the receiver.
Downlink: Different cells have different downlink scrambling codes. Each cell is configured
with a unique downlink scramble. The UE identifies a cell based on the scramble. The OVSF
code is used to distinguish different users in a cell.
Uplink: Scrambles are used to distinguish different users. In one cell, each user is configured
with a unique uplink scrambling code. The OVSF code is used to distinguish the services of a
user. Channel Source RF Interleaving coding and Spreading Scrambling Modulation coding
transmission interleaving
3,84Mbps Chips after

Symbols
× × spreading and
scrambling
OVSF Scrambling
code code
Figure 4 Scrambling mechanism
The UMTS modulation format for the downlink is more straightforward than that used in
the uplink: The downlink uses quadrature phase shift keying, QPSK. The uplink uses
Differential phase shift keying, DQPSK.
Within UMTS it is possible to define a number of different types of UMTS handover.

With the advent of CDMA technology, new possibilities for effecting more reliable forms of
handover became possible depending upon the different circumstances.
14
 Hard handover: This form of handover is essentially the same as that used for 2G
networks where one link is broken and another established.
 Soft handover: This form of handover is a more gradual and the UE communicates
simultaneously with more than one Node B or base station during the handover process.
 Softer handover: Not a full form of UMTS handover, but the UE communicates with
more than one sector managed by the same NodeB.
 UMTS GSM inter RAT handover: This form of handover occurs when mobiles have to
change between Radio Access Technologies (W1)
2.2.2 HSDPA (High Speed Down-link Packet data Access)

HSDPA was introduced in year 2002 in 3GPP release 5. The key feature of HSDPA is
the concept of AM (Amplitude Modulation), where the modulation format (QPSK : Quadrature
Phase Shift Keying or 16-QAM : Quadrature Amplitude Modulation) and effective code rate are
changed by the network according to system load and channel conditions.
HSDPA was developed to support up to 14.4 Mbps in a single cell per user. Introduction
of new transport channel known as HS-DSCH (High Speed-Downlink Shared Channel), uplink
control channel and downlink control channel are the major enhancements to UTRAN as per
the HSDPA standard.
HSDPA selects coding rate and modulation method based on the channel conditions
reported by user equipment and Node-B, which is also known as AMC (Adaptive Modulation
and Coding) scheme.
From a radio interface perspective, HSDPA is based on a shared radio scheme and real
time (every 2 ms) evaluation and allocation of radio resources, allowing the system to quickly
react to data bursts. In addition, HSDPA implements a HARQ (Hybrid Automatic Repeat
Request) which is a fast packet retransmission scheme located in the Base Station as close as
possible to the radio interface. This allows fast adaptation to a change in radio transmission
characteristics.
Other than the QPSK (Quadrature Phase Shift Keying) used by WCDMA networks,
HSDPA supports 16QAM (Quadrature Amplitude Modulation) for data transmission under
good channel conditions.
2.2.3 HSUPA (High Speed Up-link Packet data Access

HSUPA was introduced with the 3GPP release 6 in year 2004, where Enhanced
Dedicated Channel (E-DCH) is used to improve the uplink of the radio interface. Maximum
theoretical uplink data rate that can be supported by a single cell as per the HSUPA
specification is 5.76 Mbps.
15
HSUPA relies on QPSK modulation scheme, which is already specified for WCDMA. It
also uses HARQ with incremental redundancy to make retransmissions more effective.
HSUPA uses uplink scheduler to control the transmit power to the individual E-DCH
users to mitigate the power overload at Node-B.
HSUPA also allows self-initiated transmission mode that is called as non-scheduled

transmission from UE to support services such as VoIP that need reduced Transmission Time
Interval (TTI) and constant bandwidth. E-DCH support both 2ms and 10ms TTI. Introduction
of E-DCH in HSUPA standard introduced new five physical layer channels. (W2)
Table 3 Differences between HSDPA and HSUPA
Characteristics HSDPA HSUPA

Maximum bit Rate 14,4 Mbps 5,7 Mbps
Modulation QPSK, 16 QAM (64 QAM) 8 PSK (QPSK)
TTI 2 ms 2ms /10ms
AMC Yes No
SF (Spreading 16(fixed) 2 to 256
Factor)
HARQ HARQ incremental redundancy HARQ incremental redundancy
Soft Handover No yes
High Speed Packet data Access (HSPA) has been an upgrade to WCDMA networks
(both FDD, and TDD) used to increase packet data performance. The introduction was done in
steps; High Speed Down Link Packet data Access (HSDPA) and Enhanced Up Link, also
referred to as High Speed UL Packet data Access (HSUPA).
The combination of HSDPA and Enhanced UL is referred to as HSPA. HSPA evolution

(also known as HSPA+ and evolved HSPA) came in Release 7 with further improvements in
later releases.
2.3 Pre-4G technologies

2.3.1 HSPA+ (Evolved High-Speed Packet Access)
HSPA+ stands for “Evolved High-Speed Packet Access.” It has been touted by some
carriers as a 4G mobile broadband network type, and the ITU recently changed its 4G
definition to put HPSA+ under the 4G umbrella, but really, it is merely an upgrade to existing
3G GSM technologies. Think of it as a stopgap or bridge between 3G and 4G.
HSPA+ uses the same 5MHz band of WCDMA spectrum and hence great ease for
operators. In the same 5MHz HSPA+ tries to increase data rate by using MIMO/higher order
modulation techniques. We present some features of HSPA+
16
 16QAM Uplink: With the possibility to use 16QAM on the E-DCH (Enhanced Dedicated
Channel) in the uplink, HSPA+ can achieve uplink peak data rates of 11.5 Mbits/s.
 64QAM In Downlink : With the possibility to use 64QAM in the downlink, HSPA+ can
achieve downlink data rates of 21 Mbits/s. 64QAM is a user equipment (UE) capability, i.e.,
not all UE will be able to support it.
 MIMO used with HSPA+ is able to provide significant increases in data throughput. It
exploits the multiple paths in ways that enable the system to become more resilient to
interference or to enable high data throughput rates by using spatial multiplexing.
 Continuous Packet Connectivity (CPC): CPC is a collection of enhancements that allow
more users to be continuously connected to the network and, at the same time, increase
UE battery life and improve the link quality, especially for low-data-rate services like Voice
over Internet Protocol (VoIP).
 Dual Carrier-HSDPA (DC-HSDPA): Other functions have been added to HSPA+ in later
releases. In Release 8 Dual Cell-HSDPA (also referred to as Dual Carrier-HSDPA, DC-
HSDPA), where carrier aggregation of two adjacent 5 MHz bands, covering the same area,
is used to increase the performance. (3)
2.3.2 LTE (Long Term Evolution)

LTE was introduced in 3GPP R8, is the access part of the Evolved Packet System (EPS).
The main requirements for the new access network are high spectral efficiency, high peak data
rates, short round trip time as well as flexibility in frequency and bandwidth.
The followed highlighted specifications give an overall view of the performance that
LTE offer. It meets the requirements of industry for high data download speeds as well as
reduced latency.
Tableau 1 LTE Basic Specifications
LTE BASIC SPECIFICATIONS
PARAMETER DETAILS
Peak downlink speed 64QAM (Mbps) 100 (SISO), 172 (2x2 MIMO), 326 (4x4 MIMO)
Peak uplink speeds (Mbps) 50 (QPSK), 57 (16QAM), 86 (64QAM)
All packet switched data (voice and data).

Data type
No circuit switched.
Channel bandwidths (MHz) 1.4, 3, 5, 10, 15, 20
Duplex schemes FDD and TDD
0 - 15 km/h (optimised),
Mobility
15 - 120 km/h (high performance)
17
Idle to active less than 100ms

Latency
Small packets ~10 ms
Downlink: 3 - 4 times Rel 6 HSDPA

Spectral efficiency
Uplink: 2 -3 x Rel 6 HSUPA
OFDMA (Downlink)
Access schemes
SC-FDMA (Uplink)
Modulation types supported QPSK, 16QAM, 64QAM (Uplink and downlink)
LTE has introduced a number of new technologies when compared to the previous
cellular systems. As an overview, we outline here three fundamental technologies that have
shaped the LTE radio interface design: multicarrier technology, multiple-antenna technology,
and the application of packet-switching.
Multicarrier Technology: OFDMA (Orthogonal Frequency Division Multiple Access) is

used in the downlink; while SC-FDMA(Single Carrier - Frequency Division Multiple Access) is
used in the uplink.
 OFDM has been incorporated into LTE because it enables high data bandwidths to be
transmitted efficiently while still providing a high degree of resilience to reflections
and interference.
 SC-FDMA is used in view of the fact that its peak to average power ratio is small and
the more constant power enables high RF power amplifier efficiency in the mobile
handsets - an important factor for battery power equipment. (4)
Multiple Antenna Technology: This also allows multiple users located in different
directions to be served simultaneously. The MIMO transmission modes (Multiple Input
Multiple Output) employs multiple antennas on the receiver and transmitter to utilise the
multi-path effects that always exist to transmit additional data, rather than causing
interference. Accordingly schemes using 2 x 2, 4 x 2, or 4 x 4 antenna matrices can be used.
User Equipment Categories: In practice it is important to recognize that the market for
UEs is large and diverse, and there is therefore a need for LTE to support a range of categories
of UE with different capabilities to satisfy different market segments.
2.4 4G technologies
2.4.1 LTE Advanced
With the standards definitions now available for LTE, eyes are now turning towards
the next development, that of the truly 4G technology. The new technology being developed
under the auspices of 3GPP to meet these requirements is often termed LTE Advanced.
Some of the main headline aims for LTE Advanced can be seen below:
18
 Peak data rates: downlink - 1 Gbps; uplink - 500 Mbps.
 Spectrum efficiency: 3 times greater than LTE.
 Peak spectrum efficiency: downlink - 30 bps/Hz; uplink - 15 bps/Hz.
 Spectrum use: the ability to support scalable bandwidth use and spectrum aggregation
where non-contiguous spectrum needs to be used.
 Latency: from Idle to Connected in less than 50 ms and then shorter than 5 ms one way
for individual packet transmission.
 Cell edge user throughput to be twice that of LTE.
 Average user throughput to be 3 times that of LTE.
 Mobility: Same as that in LTE
 Compatibility: LTE Advanced shall be capable of interworking with LTE and 3GPP legacy
systems.
There are a number of key technologies that will enable LTE Advanced to achieve the
high data throughput rates that are required. MIMO and OFDM are two of the base
technologies that will be enablers. Along with these there are a number of other techniques and
technologies that will be employed. (4)
3 UMTS Network architecture

The UMTS network architecture can be divided into three main elements:
User Equipment (UE): The User Equipment or UE is the name given to what was
previous termed the mobile, or cellphone. It could also be anything between a mobile phone
used for talking to a data terminal attached to a computer with no voice capability.
Radio Network Subsystem (RNS): The RNS also known as the UMTS Radio Access
Network, UTRAN, is the equivalent of the previous Base Station Subsystem or BSS in GSM. It
provides and manages the air interface for the overall network.
Figure 5 UMTS Network Architecture
19
Core Network: The core network provides all the central processing and management
for the system. It is the equivalent of the GSM Network Switching Subsystem or NSS. (5)
3.1 UMTS radio access network UTRAN

The UMTS Radio Access Network, UTRAN comprises two main components:
 Radio Network Controller (RNC): This element controls the NodeBs that are
connected to it. The RNC undertakes the radio resource management and some of the
mobility management functions. It is also the point at which the data encryption/
decryption is performed to protect the user data from eavesdropping.
 Node B: NodeB is the term used within UMTS to denote the base station transceiver.
This part of the UTRAN contains the transmitter and receiver to communicate with the
UEs within the cell. It participates with the RNC in the resource management.
3.2 UMTS Core Network

The 3G UMTS core network architecture is a migration of that used for GSM with
further elements overlaid to enable the additional functionality demanded by UMTS. The
UMTS core network may be split into two different areas:
 Circuit switched elements: These elements are primarily based on the GSM network
entities and carry data in a circuit switched manner, i.e. a permanent channel for the
duration of the call.
 Packet switched elements: These network entities are designed to carry packet data.
This enables much higher network usage as the capacity can be shared and data is carried
as packets which are routed according to their destination.
Some network elements, particularly those that are associated with registration are
shared by both domains and operate in the same way that they did with GSM.
Circuit switched elements: The circuit switched elements of the UMTS core network
architecture include the following network entities:
 Mobile switching Center (MSC): This is essentially the same as that within GSM, and it
manages the circuit switched calls under way.
 Gateway MSC (GMSC): This is effectively the interface to the external networks.
Packet switched elements: The packet switched elements of the 3G UMTS core network
architecture include the following network entities:
 Serving GPRS Support Node (SGSN): As the name implies, this entity was first developed
when GPRS was introduced, and its use has been carried over into the UMTS network
architecture. The SGSN provides a number of functions:
20
 Mobility management: Generates MM information based on the mobile's current

location.
 Session management: Management of the data sessions providing the required

quality of service.
 Interaction with other areas of the network: Management of its elements within
the network only by communicating with other areas of the network, e.g. MSC and
other circuit switched areas.
 Billing: Monitoring the flow of user data across the GPRS network. CDRs (Call
Detail Records) are generated by the SGSN before being transferred to the
charging entities (Charging Gateway Function, CGF). (5)
 Gateway GPRS Support Node (GGSN): Like the SGSN, this entity was also first
introduced into the GPRS network. It is the central element within the UMTS packet
switched network. It handles inter-working between the UMTS packet switched network
and external packet switched networks. In operation, when the GGSN receives data
addressed to a specific user, it checks if the user is active and then forwards the data to
the SGSN serving the particular UE.
Shared elements: The shared elements of the 3G UMTS core network architecture
include the following network entities:
 Home location register (HLR): This database contains all the administrative
information about each subscriber along with their last known location. In this way, the
UMTS network is able to route calls to the relevant RNC / Node B. When a user switches
on their UE, it registers with the network and from this it is possible to determine which
Node B it communicates with so that incoming calls can be routed appropriately
 Equipment identity register (EIR): The EIR is the entity that decides whether a given
UE equipment may be allowed onto the network. Each UE equipment has a number
known as the International Mobile Equipment Identity.
 Authentication center (AuC): The AuC is a protected database that contains the secret
key also contained in the user's USIM card.
4 LTE Network Architecture

In contrast to the circuit-switched model of previous cellular systems, Long Term
Evolution (LTE) has been designed to support only packet-switched services. It aims to provide
seamless Internet Protocol (IP) connectivity between user equipment (UE) and the packet data
network (PDN), without any disruption to the end users’ applications during mobility.
21
EPS uses the concept of EPS bearers to route IP traffic from a gateway in the PDN to
the UE. A bearer is an IP packet flow with a defined quality of service (QoS) between the
gateway and the UE. The E-UTRAN and EPC together set up and release bearers as required
by applications.
In this section we are going to give an overview of the functions provided by the core
network (CN) and E-UTRAN. The remainder presents the network interfaces in detail, with
particular focus on the E-UTRAN interfaces and the procedures used across these interfaces,
including those for the support of user mobility. (6)
Figure 6: The EPS network elements
4.1 Radio access network E-UTRAN

The access network of LTE, E-UTRAN, simply consists of a network of eNodeBs, as
illustrated in Figure 6.
Figure 6 The Radio Access Network (E-UTRAN)
The eNodeBs are normally interconnected with each other by means of an interface
known as X2 and to the EPC by means of the S1 interface. More specifically, to the MME by
means of the S1-MME interface and to the S-GW by means of the S1-U interface. (6)
22
The X2 interface: The main purpose of this interface is to minimize packet loss due
to user mobility. As the terminal moves across the access network, unsent or unacknowledged
packets stored in the old eNodeB queues can be forwarded or tunneled to the new eNodeB
thanks to the X2 interface. From a high-level perspective, the new E-UTRAN architecture is
actually moving towards WLAN network structures and Wifi or WiMAX Base Stations. (7)
S1-U interface: is the point for user plane traffic between eNodeB and serve GW
reference. The main activity via this benchmark is to transfer IP packets encapsulated users
arising from traffic or tunnel shape. Encapsulation is needed to realize the virtual IP link
between eNodeB and GW service, even during the movement of UE, and thus enable mobility.
S1-MME interface: is the point for the control plane between eNodeB and MME
reference. All control activities are carried out on it, for example, signaling for attachment,
detachment, and the establishment of the support of the change, safety procedures, etc. The E-
UTRAN is responsible for all radio-related functions, which can be summarized briefly as:
 Radio resource management (RRM): This covers all functions related to the radio bearers,
such as radio bearer control, radio admission control, radio mobility control, scheduling
and dynamic allocation of resources to UEs in both uplink and downlink.
 Header Compression: This helps to ensure efficient use of the radio interface by
compressing the IP packet headers that could otherwise represent a significant overhead,
especially for small packets such as VoIP.
 Security: All data sent over the radio interface is encrypted.
 Connectivity to the EPC: This consists of the signaling toward MME and the bearer path
toward the S-GW. (6)
4.2 EPC (The Evolved Packet Core)

The Evolved Packet Core (EPC) is the LTE core network. It is comprised of components
that have the following functions: mobility management, authentication, quality of service,
routing upload and download IP packets, IP address allocation, and more.
The EPC is responsible for the overall control of the UE and establishment of the
bearers. The main logical nodes of the EPC are:
 PDN Gateway (P-GW)

 Serving Gateway (S-GW)
 Mobility Management Entity (MME)
In addition to these nodes, EPC also includes other logical nodes and functions such as
the Home Subscriber Server (HSS) and the Policy Control and Charging Rules Function
(PCRF). The logical CN nodes are discussed in more detail below:
23
 PCRF: The Policy Control and Charging Rules Function provides the QoS authorization
(QoS class identifier [QCI] and bit rates) that decides how a certain data flow will be treated
in the PCEF and ensures that this is in accordance with the user’s subscription profile.
 P-GW: The PDN Gateway is responsible for IP address allocation for the UE, as well as
QoS enforcement and flow-based charging according to rules from the PCRF. It is
responsible for the filtering of downlink user IP packets into the different QoS-based
bearers. It also serves as the mobility anchor for interworking with non-3GPP technologies
such as CDMA2000 and WiMAX networks.
 S-GW: All user IP packets are transferred through the Serving Gateway, which serves as
the local mobility anchor for the data bearers when the UE moves between eNodeBs. It also
retains the information about the bearers when the UE is in the idle state and temporarily
buffers downlink data while the MME initiates paging of the UE to reestablish the bearers.
In addition, the S-GW performs some administrative functions in the visited network such
as collecting information for charging (for example, the volume of data sent to or received
from the user) and lawful interception. It also serves as the mobility anchor for
interworking with other 3GPP technologies such as general packet radio service (GPRS)
and UMTS.
 MME: The Mobility Management Entity (MME) is the control node that processes the
signaling between the UE and the CN (6). the MME supports the following:
 Security procedures: this relates to end-user authentication as well as initiation and
negotiation of ciphering and integrity protection algorithms.
 Terminal-to-network session handling: this relates to all the signalling procedures
used to set up Packet Data context and negotiate associated parameters like the Quality
of Service.
 Idle terminal location management: this relates to the tracking area update process
used in order for the network to be able to join terminals in case of incoming sessions.
The MME is linked through the S6 interface to the HSS which supports the database
containing all the user subscription information.
 HSS: The Home Subscriber Server contains user’s subscription data such as the EPS-
subscribed QoS profile and any access restrictions for roaming. It also holds information
about the PDNs to which the user can connect. In addition the HSS holds dynamic
information such as the identity of the MME to which the user is currently attached or
registered. The HSS may also integrate the authentication center (AUC), which generates
the vectors for authentication and security keys. (6)
24
5 Performance Measurement
5.1 Features of Performance Measurement
The term End-User Performance is used to refer to the Quality of Service (QoS) that the
user perceives when using an application. No matter which wireless network lies beneath the
transmission (whether it is GPRS, EDGE, CDMA2000, WCDMA or WiFi), the metrics which
define the performance of one application are directly linked to the characteristics of the
service itself. From the user point of view, the whole network can be considered as a data bearer
which provides certain transmission capabilities. These capabilities can be defined by the
combination of two basic parameters: throughput and latency. However, in order to fully
define the access network it is also necessary to consider other additional factors, like the
accessibility and the retainability of the connections. These terms can be described as follows.
 Throughput: is the amount of data transferred in one direction over a link divided by the
time taken to transfer it, expressed in bits or bytes per second. Generally the throughput in
the radio interface is measured as the effective Radio Link Control (RLC) payload
throughput. This indicator depends on many factors, such as the terminal capability,
cellular planning, load conditions or radio technology.
 Latency: is the time it takes for a packet to cross a network connection, from sender to
receiver. Latency is particularly important for a synchronous protocol where each packet
must be acknowledged before the next can be transmitted. For example, TCP
establishments and congestion control algorithms are very much affected by the latency of
the network.
 Accessibility: is the probability that the service is available when user requires it. In the
case of a wireless network it will depend on several factors, from the signal coverage to the
network connectivity with external nodes and application servers. The access to a service
may fail for any cause, some of them inside the operator’s network, but some others by the
external network itself (Internet). A wireless network operator shall take care of the
optimization of the part of the transmission that goes through its network: radio
accessibility and session accessibility.
 Retainability of the network will be given by the probability that data connections are
able to finish successfully, without transmission errors that may affect the performance.
5.2 Radio Access Network Performance Measurement

Performance measurement is an effective means of scanning the whole network at any
time and systematically searching for errors, bottlenecks and suspicious behavior. In fact, the
most critical part of mobile networks is the radio interface. The UTRAN controlled by RNCs
and the EUTRAN controlled by eNodeBs are an excellent place to collect data giving an
overview of radio interface quality that’s why monitoring the RAN performance is a very
25
important task for operation and maintenance personnel, network engineers, and
management.
Within UTRAN four interfaces exist where performance-related data can be captured:
 the Iub interface between Node Bs and RNC

 the Iur interface between different RNCs
 the IuCS interface between RNCs and the CS core network domain
 and the IuPS interface between RNCs and the PS core network domain
5.3 Key Performance Indicators (KPI)

5.3.1 Definition
One of the main functions of the OMC-R is performance management. Performance
measures are based on the collection of counters calculated by the entities of the network.
These measures are basically used for four types of need:
 Optimization and effective network planning

 Statistics
 Detailed investigation of a past problem
 Real-time analysis
Counter measures at the OMC level (reported by Nodes B and E-Node B to the
OMC-R) are collected over a precise time interval and are linked to an event occurring in the
network. They are used to calculate Key Performance Indicators (KPI) of the network by
combining these counters according to well-defined formulas. The analysis of these indicators
is very essential for the supervision of quality of service.
KPIs are a set of quantifiable measures used in GSM, UMTS, HSPA, and LTE
networks to compare performance in terms of meeting mobile network’s strategic and
operational goals. KPIs vary between management, marketing, operations and
network engineering people depending on their priorities, perspectives or performance
criteria.
Network KPIs provide understanding about how a mobile communication network is

performing from the system point of view. Various groups of KPIs can be defined which allow
to do the following classification according to different areas of interest.
 Radio Access Network KPIs: Such KPIs provide understanding about how the radio
access network is performing. Radio accessibility and retainability are two typical measures
which indicate how well a user can access and keep the radio resource. However, radio
performance indicators shall also be considered.
26
 Core Network KPIs: Such KPIs provide understanding about how the core network is
performing. They are normally focused on measuring the load of transport interfaces and
network elements, ensuring that the core network is able to support the demand of
resources from the radio access network. They also provide information that helps to detect
bottlenecks in the network and take decisions about dimensioning upgrades.
 Signaling KPIs: Such KPIs provide understanding about mobility management
procedures performance. Signaling load from different areas can be used to analyze load
distribution in the network and optimize network topology in a way that the impact of
mobility procedures is minimized. For instance, high number of handovers between cells
belonging to different routing areas may impact on the user performance and signaling
load, and might be improved by moving some cells to a different area.
5.3.2 KPI’s categories

KPI can be categorized into five fundamental subcategories:
 Mobility: these KPIs are used to monitor the successful ratio for many kinds of handover
features or service mode changing in different scenarios.
 Accessibility: is the ability of a user to obtain the requested service from the system. RRC
connection and RAB setup are the main procedures of accessibility.
 Integrity: Service Integrity KPIs mainly indicate the service capabilities for PS/HSPA
throughput during busy hours in each cell and the service UL Average BLER for evaluating
the UL BLER value of services in each cell.
 Retainability: is defined as the ability of a user to retain its requested service for the
required duration once connected. The RNC level KPIs can be calculated by aggregating all
the cell counters and Iur counters.
 Availability: Availability KPIs mainly indicate the utilization for several kinds of network
resources such as Radio, bandwidth or CPU Load.
Each constructor deals with a different range of KPI categories and KPI types
according to its needs for example KPI used by HUAWEI are different of KPI used by
Nokia or Ericsson. So in the next part of this chapter we illustrate the different
categories of KPI used by HUAWEI.
5.4 Huawei UTRAN KPIs

In this section we describe the UTRAN KPI performance monitoring and analysis of
problems that bad KPI values indicate. The following analysis contains a list of the most
common UTRAN KPIs used in Huawei networks. These KPIs are monitored constantly. When
the value of a KPI goes below the defined threshold, then detailed analysis should be performed
27
in order to identify the reasons of this deterioration. Once the reasons are found, proper
solutions will be proposed and implemented.
5.4.1 Accessibility
5.4.1.1 CSSR/CS
This KPI is used to check the Radio Access Success Ratio. The details of the Access
Failures caused by the SCCP congestion are not provided in this call setup procedure.
Call Setup Success Rate /CS:
𝑅𝑅𝐶 𝑆𝑒𝑡𝑢𝑝 𝑆𝑢𝑐𝑐𝑒𝑠𝑠 𝑅𝑎𝑡𝑖𝑜 ∗ 𝐶𝑆 𝑅𝐴𝐵 𝑆𝑒𝑡𝑢𝑝 𝑆𝑢𝑐𝑐𝑒𝑠𝑠 𝑅𝑎𝑡𝑖𝑜 ∗ 100%
5.4.1.2 CSSR/PS
This KPI is used to check the PS Radio Access Success Ratio. The details of the Access
Failures caused by the SCCP congestion are not provided in this call setup procedure.
Call Setup Success Rate /PS:
𝑅𝑅𝐶 𝑆𝑒𝑡𝑢𝑝 𝑆𝑢𝑐𝑐𝑒𝑠𝑠 𝑅𝑎𝑡𝑖𝑜 ∗ 𝑃𝑆 𝑅𝐴𝐵 𝑆𝑒𝑡𝑢𝑝 𝑆𝑢𝑐𝑐𝑒𝑠𝑠 𝑅𝑎𝑡𝑖𝑜 ∗ 100%
5.4.1.3 RRC Setup Success Ratio

The RRC connection attempt for service procedure is complete when the RNC receives
an RRC CONNECTION REQUEST message from the UE. The message contains information
about one of the following service types requested by the UE: Conversational Call, Streaming
Call, Background Call, Interactive Call, Originating Subscribed Traffic Call, Emergency Call,
High Priority Signaling, Low Priority Signaling, Cause Unknown, and Call Re-Establishment.
The RRC Setup Success for Service Procedure is complete when the RNC receives an
RRC CONNECTION SETUP COMPLETE message from the UE.
RRC Setup Success Ratio:
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑅𝑅𝐶 𝑆𝑒𝑡𝑢𝑝 𝑆𝑢𝑐𝑐𝑒𝑠𝑠

∗ 100%
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑅𝑅𝐶 𝐶𝑜𝑛𝑛𝑒𝑐𝑡𝑖𝑜𝑛 𝐴𝑡𝑡𝑒𝑚𝑝𝑡𝑠
5.4.1.4 RAB Setup Success Ratio

This KPI is used to check the RAB Setup Success Ratio of all CS and PS services in an
RNC or a cluster. The RAB Setup Attempt Procedure is complete when the RNC receives an
RAB ASSIGNMENT REQUEST message from the CN in the CS/PS domain. The message
contains information about one of the following service types: Conversational Services,
streaming Services.
28
The RAB Setup Success Procedure starts when the RNC receives a RADIO BEARER
SETUP COMPLETE message from UE. This procedure is complete when the RNC sends an
RAB ASSIGNMENT RESPONSE message to the CN in the CS domain.
RAB Setup Success Ratio:
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑅𝐴𝐵 𝑆𝑒𝑡𝑢𝑝 𝑆𝑢𝑐𝑐𝑒𝑠𝑠

∗ 100%
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑅𝐴𝐵 𝐶𝑜𝑛𝑛𝑒𝑐𝑡𝑖𝑜𝑛 𝐴𝑡𝑡𝑒𝑚𝑝𝑡𝑠
5.4.2 Mobility
5.4.2.1 IRAT/CS
This KPI is used to indicate the success ratio of CS Outgoing inter-RAT handovers in an
RNC or a Cluster.
The CS inter-RAT Handover Success procedure is complete when the RNC receives an
IU RELEASE COMMAND message with the cause value of "Successful Relocation", "Normal
Release", or "Network Optimization" after sending a HANDOVER FROM UTRAN COMMAND
message during the CS outgoing inter-RAT handover. In this case, the outgoing handover
succeeds and this counter is measured.
IRAT/CS (Inter-Rat Handover Success Ratio /CS):
𝐶𝑆 𝑖𝑛𝑡𝑒𝑟 − 𝑅𝐴𝑇 𝐻𝑎𝑛𝑑𝑜𝑣𝑒𝑟 𝑆𝑢𝑐𝑐𝑒𝑠𝑠

∗ 100%
𝐶𝑆 𝑖𝑛𝑡𝑒𝑟 − 𝑅𝐴𝑇 𝐻𝑎𝑛𝑑𝑜𝑣𝑒𝑟 𝐴𝑡𝑡𝑒𝑚𝑝𝑡
5.4.2.2 IRAT/PS
This KPI is used to indicate the success ratio of PS outgoing inter-RAT handover
initiated by the RNC or in the best cell.
The PS W2G Inter-RAT Outgoing Handover successes procedure is complete when the
RNC sends the RANAP IU RELEASE COMPLETE message after receiving the IU RELEASE
COMMAND message with the cause of "Successful Relocation", "Normal Release" or "Network
Optimization." In this case, the PS outgoing inter-RAT handover succeeds and the counter is
measured.
IRAT/PS (Inter-Rat Handover Success Ratio /PS):
𝑁𝑢𝑚𝑏𝑒𝑟 𝑂𝑓 𝑆𝑢𝑐𝑐𝑒𝑠𝑠𝑓𝑢𝑙 𝑃𝑆 𝐼𝑛𝑡𝑒𝑟 − 𝑅𝑎𝑡 𝑂𝑢𝑡𝑔𝑜𝑖𝑛𝑔 𝐻𝑎𝑛𝑑𝑜𝑣𝑒𝑟

∗ 100%
𝑁𝑢𝑚𝑏𝑒𝑟 𝑂𝑓 𝑃𝑆 𝐼𝑛𝑡𝑒𝑟 − 𝑅𝑎𝑡 𝑂𝑢𝑡𝑔𝑜𝑖𝑛𝑔 𝐻𝑎𝑛𝑑𝑜𝑣𝑒𝑟 𝐴𝑡𝑡𝑒𝑚𝑝𝑡
29
5.4.2.3 Soft/Softer Handover Success Rate

This KPI is used to check the soft/softer handover success ratio in an RNC or in a
Cluster, including softer handover.
 The Successful Soft Handover (including softer handovers) procedure is complete

when the RNC receives an ACTIVE SET UPDATE COMPLETE message from the UE
during the soft handover (including the softer handover) procedure.
 The attempt procedure is complete when the RNC sends an ACTIVE SET UPDATE
message to the UE.
Soft Handover Success Rate:
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑆𝑢𝑐𝑐𝑒𝑠𝑠𝑓𝑢𝑙 𝑆𝑜𝑓𝑡 𝐻𝑂

∗ 100%
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝐴𝑡𝑡𝑒𝑚𝑝𝑡𝑠 𝑜𝑓 𝑆𝑜𝑓𝑡 𝐻𝑂
5.4.2.4 Hard Handover Success Rate

This KPI provides the success rate of CS/PS inter-frequency hard handovers of a Cluster.
The number of CS/PS inter-frequency hard handover attempts is measured when the
RNC sends the UE a PHYSICAL CHANNEL RECONFIGURATION message during a CS/PS
inter-frequency hard handover.
The number of successful CS/PS inter-frequency hard handovers is measured when the
RNC receives from the UE a PHYSICAL CHANNEL RECONFIGURATION COMPLETE
message during a CS/PS inter-frequency hard handover.
5.4.3 Retainbility
5.4.3.1 CDR/CS
This KPI provides the ratio of the CS RAB abnormal Releases to the total CS RAB
Releases (Normal Release + Abnormal Release). This KPI is used to check the retainabililty of
CS Service within the UTRAN (RNC or Cluster).
Call Drop Rate /CS:

i.
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝐶𝑆 𝑅𝐴𝐵 𝐴𝑏𝑛𝑜𝑟𝑚𝑎𝑙 𝑅𝑒𝑙𝑒𝑎𝑠𝑒
∗ 100%
𝑇𝑜𝑡𝑎𝑙 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝐶𝑆 𝑅𝐴𝐵 𝑅𝑒𝑙𝑒𝑎𝑠𝑒𝑠
30
5.4.3.2 CDR/PS
Call Drop Rate /PS:
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑃𝑆 𝑅𝐴𝐵 𝐴𝑏𝑛𝑜𝑟𝑚𝑎𝑙 𝑅𝑒𝑙𝑒𝑎𝑠𝑒

∗ 100%
𝑇𝑜𝑡𝑎𝑙 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑃𝑆 𝑅𝐴𝐵 𝑅𝑒𝑙𝑒𝑎𝑠𝑒𝑠
This KPI provides the ratio of the PS RAB abnormal Releases to the total PSRAB
Releases (Normal Release + Abnormal Release) and is used to check the retainabililty of PS
Service within the UTRAN (RNC or Cluster).
5.5 Huawei EUTRAN KPIs

5.5.1 Accessibility KPI
Accessibility is a measurement that allows operators to know information related to the
mobile services accessibility for the subscriber. The measurement is performed through E-
UTRAN's E-RAB service.
5.5.1.1 RRC Setup Success Rate
RRC Setup_SR:
RRCConnectionSuccess
∗ 100%
RRCConnectionAttempt
This KPI evaluates the RRC connection setup success rate in a cell or cluster involved.
Figure 7: RRCConnectionRequest message
Possible failure reasons:
 Resource allocation failure (admission failure)

 No response from UE ( Poor coverage or terminal problem)
31
Figure 8: ERAB message
5.5.1.2 E-RAB Setup Success Rate (VoIP)

There is two type of bearers:
 Default bear: With default QCI which is specified in HSS

 Dedicate bear: With dedicated QCI which is negotiated by PDN-GW and PCRF
An E-RAB is the access layer bearer for carrying service data of users. The E-RAB setup
success rate in a cell directly represents the capability of the cell to provide E-RAB connection
setups for users.
VoIP_ERABS_SR:
VoIPERABSetupSuccess
∗ 100%
VoIPERABSetupAttempt
This KPI is used to evaluate the E-RAB setup success rate of the VoIP service in a cell
or a cluster.
5.5.1.3 E-RAB Setup Success Rate (All Service)
ERABS_SR:
ERABSetupSuccess
∗ 100%
ERABSetupAttempt
5.5.1.4 Call Setup Success Rate
CSSR:
RRCConnectionSuccess𝑆𝑒𝑟𝑣𝑖𝑐𝑒𝑠 S1SigConnectionEstablishSuccess 𝐸𝑅𝐴𝐵𝑆𝑒𝑡𝑢𝑝𝑆𝑢𝑐𝑐𝑒𝑠𝑠

× × 𝐸𝑅𝐴𝐵𝑆𝑒𝑡𝑢𝑝𝐴𝑡𝑡𝑒𝑚𝑝𝑡 × 100%
RRCConnectionAttempt𝑆𝑒𝑟𝑣𝑖𝑐𝑒𝑠 1SigConnectionEstablish
5.5.2 Retainability KPIs

Retainability measures how many times a service was interrupted or dropped during
use, thus preventing the subscriber to benefit from it or making it difficult for the operator to
32
charge for it. Therefore, a high retainability is very important from a business stand point. The
measurement is performed through E-UTRAN's E-RAB service.
5.5.2.1 VoIP Call Drop Rate
VoIP_CDR:
VoIPERABABnormalRelease
∗ 100%
VoIPERABRelease
This KPI can be used to evaluate the call drop rate of the VoIP service in a cell or a
cluster. The call drop rate is calculated by monitoring the VoIP ERAB abnormal release rate.
Each E-RAB is associated with the QoS information.
5.5.2.2 Service Drop Rate (All)
Service_CDR:
ERABABnormalRelease
∗ 100%
ERABRelease
This KPI can be used to evaluate the call drop rate of all services in a cell or a cluster,
including VoIP service. E-RAB Release can be either Normal release or Abnormal release.
Normal release are released with the normal procedure, the statistic is based on each QCI
index. Abnormal release are considered as service (call) drop, the statistic can be based on
different release cause and each QCI index. (8) The possible causes could be:
 EPC failure
 Radio link failure
 Handover failure
 Transport failure
 Network congestion
5.5.3 Mobility KPIs

Mobility measures how many times a service was interrupted or dropped during a
subscriber's handover or mobility from on cell to another.
The measurement is performed in the E-UTRAN and will include Intra E-UTRAN and
Inter RAT handovers.
Following table summarizes KPIs for LTE RAN (Radio Access Network). This is used
to measure contribution to subscriber perceived quality and system performance.
33
RAN performance monitoring and control is a very important task for O&M team as well as
network engineers.
5.5.3.1 Intra/Inter-frequency Handover Out Success Rate
Intra-frequency Handover Out Success Rate:
𝐼𝑛𝑡𝑟𝑎𝐹𝑟𝑒𝑞𝐻𝑂𝑂𝑢𝑡𝑆𝑢𝑐𝑐𝑒𝑠𝑠
∗ 100%
𝐼𝑛𝑡𝑟𝑎𝐹𝑟𝑒𝑞𝐻𝑂𝑂𝑢𝑡𝐴𝑡𝑡𝑒𝑚𝑝𝑡𝑠
Inter-frequency Handover Out Success Rate:
𝐼𝑛𝑡𝑒𝑟𝐹𝑟𝑒𝑞𝐻𝑂𝑂𝑢𝑡𝑆𝑢𝑐𝑐𝑒𝑠𝑠
∗ 100%
𝐼𝑛𝑡𝑒𝑟𝐹𝑟𝑒𝑞𝐻𝑂𝑂𝑢𝑡𝐴𝑡𝑡𝑒𝑚𝑝𝑡𝑠
This KPI can be used to evaluate the intra/inter-frequency Handover out success rate
in a cell or a cluster. The intra-frequency handover (HO) includes both inter-eNodeB and intra-
eNodeB.
5.5.3.2 Handover In Success Rate
HOIn_SR:
HOInSuccess
∗ 100
HOInAttempt
This KPI can be used to evaluate the handover in success rate in a cell or a cluster.
5.5.3.3 Inter_RAT Handover Out Success Rate (LTE to WCDMA)
IRATHO_E2W_SR:
IRATHO_E2W_Success
∗ 100
IRATHO_E2W_Attempt
This KPI can be used to evaluate the inter-RAT handover out success rate in a cell or a
cluster. As suggested by the name, this KPI measures the performance of the inter-RAT
handover from LTE to WCDMA.
5.5.3.4 Inter_RAT Handover Out Success Rate (LTE to GSM)
IRATHO_E2G_SR:
IRATHO_E2G_Success
∗ 100
IRATHO_E2G_Attempt
34
This KPI can be used to evaluate the inter-RAT handover out success rate in a cell or a
cluster. As suggested by the name, this KPI measures the performance of the inter-RAT
handover from LTE to GSM. (8)
5.5.4 Service Integrity KPIs

Integrity measures the high or low quality of a service while the subscriber is using it.
The measurement is performed through E-UTRAN's delivery of IP packets.
5.5.4.1 Cell Downlink Average Throughput

This KPI evaluates the cell downlink average throughput when there are data
transferring at downlink. It reflects the cell capacity. (8)
Cell Downlink Average Throughput:
𝐶𝑒𝑙𝑙𝐷𝐿𝑇𝑟𝑎𝑓𝑓𝑖𝑐𝑉𝑜𝑙𝑢𝑚𝑒
(𝑏𝑝𝑠)
𝐶𝑒𝑙𝑙𝐷𝐿𝑇𝑟𝑎𝑛𝑠𝑓𝑒𝑟𝑇𝑖𝑚𝑒
5.5.4.2 Cell Uplink Average Troughput

This KPI evaluates the cell uplink average throughput when there are data transferring
at uplink. (8)
Cell Uplink Average Throughput:
𝐶𝑒𝑙𝑙𝐷𝐿𝑇𝑟𝑎𝑓𝑓𝑖𝑐𝑉𝑜𝑙𝑢𝑚𝑒
(𝑏𝑝𝑠)
𝐶𝑒𝑙𝑙𝐷𝐿𝑇𝑟𝑎𝑛𝑠𝑓𝑒𝑟𝑇𝑖𝑚𝑒
6 Fault Monitoring
6.1 Features of fault monitoring
Network fault monitoring deals with measuring the problems in the network. This
procedure looks at various layers of the network, so when a problem occurs, it is also necessary
to differentiate which layers of the network are affected.
Additionally, in order to effectively detect problems, it is required to understand what

the normal behavior of the network is during an extended period of time. It is also important
to understand that there are always punctual errors in the network, i.e. due to the variability
of the radio conditions and network load, but it does not mean that the network is having
persistent problems. It is then necessary to look not only at specific values, but also at the
duration of abnormal states, and having a clear reference of what the normal expected behavior
would be.
35
6.2 Alarm definition

Faults in a telecommunication network are reported to management centers in the form
of alarms. An alarm is a message emitted by a network element, typically when a problem is
encountered. Unfortunately, a network element has a very narrow view to the network, and
can therefore only report the symptoms of the fault from its limited viewpoint. On the other
hand, one fault can result in a number of different alarms from several network elements. (9)
Through alarms, network administrators can observe undesirable states such as link
failures, power problems and high dropped-call rates, and thereby make a hypothesis of the
underlying problem.
6.3 Alarm management

The massive amount of network elements in a mobile system and the variety of network
elements and infrastructure equipment creates huge amount of alarms saturating operators
alarm management systems. In parallel the numbers of types of alarms have increased to
overwhelming proportions.
Major mobility network incident management center can count alarms in n*100 000
per day. Handling of n* 1000 different types of alarms. The most important problem is that
the network administrators are flooded with alarms with often poor quality.
Poor alarm quality means:
 Nuisance alarms (repeating and fleeting alarms, redundant and cascading alarms).
 Alarm floods.
 Alarms without response.
 Alarms with the wrong priority.
 Out-of-Service alarms.
 Redundant alarms.
So the alarm management is a huge problem for constructors like HUAWEI that why
we decided to create an Alarm interface in our Fault Management tool (FDT) to help HUAWEI
engineers to handle with alarms easily.
To develop this interface we should know how HUAWEI classify the different alarms in
the 3G radio access network. So in the next section I will illustrate the classification of alarms
according to HUAWEI.
6.4 Alarm Parameters

An alarm is defined by:
 Alarm ID: Unique identifier of an alarm for one product.
36
 Alarm Name: Unique name of an alarm for one product. Alarm names clearly and
accurately indicate alarm meanings. There is a one-to-one mapping between alarm
names and IDs.
 Network Element: Network element where an alarm is generated.
 Alarm Severity
 Alarm Category
 Alarm Description: Conditions in which an alarm is generated.
 Impact on System: Impact of an alarm on the system or services.
 System Action: Actions taken by the system when an alarm is generated.
 Alarm start date / Alarm end date
 Occurrence number
6.5 Alarm Severity

The Alarm Severity is the Impact of an alarm on service quality. There are four alarm
severity levels: critical, major, minor, and warning.
Critical: A critical alarm affects system services. If a critical alarm is generated, immediate
actions are necessary even when the fault occurs during non-working hours.
Major: A major alarm affects service quality and requires immediate actions during working
hours.
Minor: A minor alarm generally does not affect service quality. It requires handling at an
appropriate time or further observation to avoid more serious faults.
Warning: A warning alarm indicates a potential error that may affect service quality. It
requires different actions depending on errors.
6.6 Alarm Categories

There are 16 alarm categories based on monitored objects.
 Power alarm: relates to the power system.

 Environment alarm: relates to equipment room environment variables, such as
temperature, humidity, and door status control.
 Signaling alarm: relates to channel associated signaling and common channel signaling
(for example, SS7).
 Trunk alarm: relates to the trunk system, including trunk circuits and boards.
 Hardware alarm: relates to hardware, such as the clock unit and CPU.
 Software alarm: relates to software.
 System alarm: relates to system operations.
 Communication alarm: relates to communication.
 Service quality alarm: relates to service quality.
37
 Unexpected operation alarm: generated when an exception occurs.

 OMC alarm: generated when the OMC is not working properly.
 Integrity alarm: generated when information is illegally added, modified, or deleted.
 Operation alarm: generated when service is unavailable or inaccessible because of
inappropriate operations.
 Physical resource alarm: generated when physical resources are damaged by a suspected
security attack.
 Security alarm: generated when security services or a security mechanism has detected
that the system encounters a security attack.
 Time domain alarm: generated when an event occurs during an unexpected or forbidden
time.
7 Conclusion
In this chapter we gave a general overview of the 3G and 4G mobile Networks. We
highlighted the concept of the performance Analysis in radio access network. Aside from that,
we presented the feature of fault monitoring. The next chapter we introduce the concepts of
business intelligence and data mining.
38
Chapter 3 Business Intelligence & Data Mining
Chapter 3 Business Intelligence &

Data Mining
1 Introduction...........................................................................................................................40
2 Business Intelligence ..............................................................................................................40
2.2.1 Data Source Layer ........................................................................................................ 41

2.2.2 ETL (Extract-Transform-Load) Layer ............................................................................ 42
2.2.3 Data Warehouse Layer ................................................................................................ 43
2.2.4 End User Layer ............................................................................................................ 43
3 Data Mining ...........................................................................................................................44
3.2.1 The classification ......................................................................................................... 45

3.2.2 The estimation ............................................................................................................ 45
3.2.3 The Segmentation ....................................................................................................... 45
3.2.4 Forecasting .................................................................................................................. 46
3.2.5 Association .................................................................................................................. 46
3.4.1 Choosing an Algorithm by Type ................................................................................... 47

3.4.2 Choosing an Algorithm by Task ................................................................................... 48
4 Comparative study about Business Intelligence tools ............................................................48
4.3.1 Comparing Data Mining ............................................................................................... 51

4.3.2 Comparing Packaging and Pricing ............................................................................... 51
4.3.3 Summarizing the analysis ............................................................................................ 52
5 Conclusion .............................................................................................................................53
39
1 Introduction
Telecommunication companies are operating today in an extremely challenging
business environment. The Telecommunication industry is in possession of large quantities of
data, generated from numerous operational systems, and is confronted with many business
problems that need urgent handling. Naturally, it has been among the first to adopt Business
Intelligence (BI) and Data Mining technologies. The main purpose of this chapter is to present
a literature review related to BI and Data Mining in Telecommunications, first we define the
main areas of BI and Data Mining applications. Last, we identify the most common Data
Mining techniques and methods used.
2 Business Intelligence
2.1 What is Business Intelligence
In today’s world data is spread across the inside as well as outside of the organization
and doesn’t provide any insight about the organization. Due to this, business executives,
managers and senior officials used to make strategic decisions based on their gut feeling rather
than fact; which is not good for any organization in the long term.
Business Intelligence (BI) helps business owners/officials to make important decisions

based on data or more precisely facts. It is a method to make the right information available at
the right time to the right people.
So, BI is a set of processes, tools, technologies and methodologies that combine data
from various data sources and make the single source of data available at right time so that
timely, better informed and strategic decisions can be made. BI also provides different tools to
quickly analyze the consolidated information in a variety of data presenting tools. (10)
2.2 Business Intelligence architecture

A typical BI architecture consists of four layers: data source, ETL, data warehouse and
end user layers. These four layers are essential to ensure high data quality and smooth
information flow in a BI system.
40
Figure 9 BI architecture
The rest of this section describes each of the layers.
2.2.1 Data Source Layer

Nowadays, many application domains require the use of structured data as well as
unstructured and semi-structured data to make effective and timely decision. All these data
can be acquired from two types of sources: internal and external.
Internal data source refers to data that is captured and maintained by operational
systems inside an organization. It includes the data related to business operations (i.e.,
customers, products, and sales data).
41
External data source refers to those that originate outside an organization. This type of
data can be collected from external sources such as business partners, syndicate data suppliers,
the Internet, governments, and market research organizations (11)
2.2.2 ETL (Extract-Transform-Load) Layer

This layer focuses on three main processes: extraction, transformation and loading
Extraction is the process of identifying and collecting relevant data from different sources.
Data source systems can take many shapes.it may be contained in Excel files, Access databases,
simple text files, and some other form of database. Usually, these data are not integrated,
incomplete, and may be duplicated. Therefore, the extraction process is needed to select data
that are significant in supporting organizational decision making.
Figure 10 example of the process of ETL on some unstructred data
After that, the data will go through the transformation and the cleansing process.
Transformation is the process of converting data using a set of business rules (such as
aggregation functions) into consistent formats for reporting and analysis. (12)
As for data cleansing, it refers to the process of identifying and correcting data errors
based on pre-specified rules. Loading is the last phase of the ETL process. The data in staging
area are loaded into target repository.
Figure 11 an example of The ETL process
42
2.2.3 Data Warehouse Layer

There are three components in the data warehouse layer, they are; operational data
store, data warehouse, and data marts.
Operational Data Store (ODS) integrates all data from the ETL layer and load them
into data warehouses. It is a database that stores subject-oriented, detailed, and current data
from multiple sources to support decision making. It provides an integrated view of near real-
time data such as transactions.
Data Warehouse is one of the most important components in BI architecture. It a

central storage that collects and stores data from different sources for the purpose of strategic
decision making, queries, and analysis. It also stores large amount of historical data for long
term analysis. (11)
Data Mart is a subset of the data warehouse that is used to support analytical needs of
a particular business function or department.
2.2.4 End User Layer

This layer consists of tools that display information in different formats to different
users. These tools can be grouped hierarchically in a pyramid shape (as shown in Figure 16).
Basically, it support have the following functionalities:
Query and Reporting Tools allow end users to access and query data, and to produce
reports for decision making and management purposes.
OLAP (Online Analytical Processing) is a “data manipulation engine that is designed

to support multidimensional data structures” (13) . In OLAP server data are organized in the
form of data cubes instead of tables (rows and columns) as in relational data model. Data cubes
are dimensional models that contains fact and dimensional tables to store and manage multi-
dimensional data so that users can analyze data easily and in a faster manner.
Data Mining is a process that automatically identifies useful information such as

unusual patterns, trends, and relationships that are hidden within large amount of data.
Data Visualization Tools such as dashboard and scorecards are provided to give closely
and more effectively an overall view of the business performance.
Analytical Applications provide functionalities such as modeling, forecasting, analysis,

and what-if scenarios. Applications that are equipped with analytical capabilities allow users
to gain insights into improving the performance of business operations (14)
For a BI system to work smoothly, all four layers described above have to be linked
together in a systematic manner. Data originating from internal and external sources have to
be extracted, transformed, and loaded into the data warehouse layer. When the data passes
43
through the ETL layer, it can flow to both directions, either to ODS (and then to data
warehouse) or to data warehouse directly. Since data warehouse is developed for the usage of
the entire organization, data from the warehouse is sent to data marts to fulfill specific
operational needs. At end user layer, data in the ODS, data warehouse, and data marts can be
accessed by using a variety of tools such as query and reporting tools, data visualization tools,
and analytical applications.
The next decade of competitive advantage revolves around the ability to make
predictions and discover patterns in data. Data mining is at the center of this revolution. It
combines, machine learning, and statistical methodologies to extract knowledge and leverage
predictions from data. The next section begins with introducing the term Data mining and
prominent aspects of machine learning, predictive analytics, and business intelligence.
3 Data Mining
3.1 Data Mining Concepts
Quite fashionable lately, we hear more and more the term Data Mining in several fields!
It is a significant discovery process of new correlations, patterns and trends by sifting through
large amounts of data stored in repositories, using pattern recognition technology, statistical
and mathematical techniques.
According to the online news magazine ZDNet technologies, data mining is "one of the
most revolutionary science of the next decade."
The prestigious journal of the American University MIT "Technology Review"

announced Date Mining as one of the 10 technologies that will change the world.
It can also be defined as an "interdisciplinary field that uses machine learning

techniques, pattern recognition, statistics, databases and visualization for extracting
information from large databases"
The Data Mining is very recent. It is the process of discovering actionable information
from large sets of data. It uses mathematical analysis to derive patterns and trends that exist
in data.
Data mining is very different from query, reporting and statistical where the user has
to create and execute queries based on hypotheses. Unlike this, data mining searches for
answers to questions that may have not been previously asked. This discovery could take the
form of finding significance in relationships between certain data elements, a clustering
together of specific data elements, or other patterns in the usage of specific sets of data
elements. After finding these patterns, the algorithms can infer rules. These rules can then be
used to generate a model that can predict a desired behavior, identify relationships among the
data, discover patterns, and group clusters of records with similar attributes.
44
Data mining is most typically used for statistical data analysis and knowledge discovery.
Statistical data analysis detects unusual patterns in data and applies statistical and
mathematical modeling techniques to explain the patterns. The models are then used to
forecast and predict. Types of statistical data analysis techniques include linear and nonlinear
analysis, regression analysis, multivariant analysis, and time series analysis. Knowledge
discovery extracts implicit, previously unknown information from the data. This often results
in uncovering unknown business facts.
3.2 Data mining Tasks

Currently, Data Mining methodology is more or less established, it is expressed in terms
of 7 spots that are:
 The classification
 The prediction
 The estimation
 The segmentation
 The association
3.2.1 The classification

Classification is supposedly the most popular Data Mining tasks. Its main purpose is to
classify one or more data samples that may consist of few or many features (dimensions). The
actual number of classes is not always given or obvious in a classification task. Therefore, it is
possible to distinguish between supervised and unsupervised classification. For supervised
classification the number of classes is known along with the properties of each class. Neither
of these is given in unsupervised classification which makes this task the more challenging one
of the two.
3.2.2 The estimation

The estimation is a task similar to the classification except that the output variable is
digital rather than categorical. Depending on other fields of the record estimation is to
complete a missing value in a particular field. In other words, estimation does not deal with
determining a class for a particular data sample. Instead, it tries to predict a certain measure
for a given data sample.
3.2.3 The Segmentation

Segmentation basically deals with the task of grouping a given data set into a few main
groups (clusters). The task of describing a large multidimensional data set (say customers) will
therefore benefit from the use of segmentation. Moreover, many algorithm types can be used
in segmentation systems.
45
3.2.4 Forecasting
Forecasting is another important Data Mining task that is used for predicting future
data values given a time series of prior data. Forecasting is a popular task often performed
using simple statistical methods. However, forecasting done in the Data Mining domain uses
advanced (learning) methods (e.g. Neural Networks, Hidden Markov Models) that in many
cases are more accurate and informative than the standard statistical methods (e.g. moving
averages).
3.2.5 Association
Association deals with task of locating events that are frequently occurring together and
benefiting from this knowledge. One of the most popular examples of association is probably
Amazon.com’s web shop that is able to recommend related products to customers. The list
below further exemplifies the use of the association task.
3.3 The Neural network Algorithm

A neural network is a computation model whose operation is to simulate the
functioning of biological neurons, it consists of a large number of neurons called units each
having a small local memory and interconnected by communication channels that carry digital
data. These units can only act on their local data and on the inputs, they receive from their
connections. Neural networks are able to predict new observations on specific variables from
other observations after executing a learning process on existing data. (15)
The learning phase of a neural network is an iterative process for adjusting the weights
of the network for optimizing the prediction of data samples on which learning was done. After
the training phase the neural network becomes able to generalize (16)
We present now how this Algorithm Works. The Microsoft Neural Network algorithm
creates a network that is composed of up to three layers of nodes. These layers are the input
layer, the hidden layer, and the output layer.
Input layer: Input nodes define all the input attribute values for the data mining
model, and their probabilities.
Hidden layer: Hidden nodes receive inputs from input nodes and provide outputs to
output nodes. The hidden layer is where the various probabilities of the inputs are assigned
weights. A weight present the importance of a particular input to the hidden node. The greater
the weight that is assigned to an input, the more important the value of that input is. Weights
can be negative, which means that the input can inhibit, rather than favor, a specific result. (17)
Output layer: Output nodes represent predictable attribute values for the data mining
model.
46
Figure 12 Neural Network layers
3.4 Choosing the Right Algorithm

Choosing the best algorithm to use for a specific analytical task can be a challenge.
While you can use different algorithms to perform the same business task, each algorithm
produces a different result, and some algorithms can produce more than one type of result.
3.4.1 Choosing an Algorithm by Type

The following are the major classifications of the data mining algorithms:
Classification algorithms: These algorithms preliminarily help in determining the

values of the discrete values in a dataset based on other attributes in the dataset. The included
algorithm most often used to implement this technique is Decision Trees. Naïve Bayes and the
Neural Network algorithms can also be used.
The Naïve Bayes algorithm is so named because it assumes all input columns are completely
independent (or equally weighted).
The Neural Network algorithm is often used with very large volumes of data that have very
complex relationships. With this type of source data, Neural Network will often produce the
most meaningful results of all of the possible algorithms.
Regression algorithms. These algorithms preliminarily help in determining the

values of continuous variables in a dataset based on other attributes in the dataset. The
included algorithm most often used to implement this technique is the Time Series. The Linear
and Logistical Regression algorithms can also be used.
Segmentation algorithms: These help in understanding the relationship between

different attributes by classifying them based on similarity of properties. The segmentation
algorithm that is available in Analysis Services is the clustering algorithm.
47
Association algorithms: The most common usage of these algorithms is to

determine the relationship between different attributes in a dataset. The included algorithm
most often used to implement this technique is that is available in Analysis Services is
Association and Decision Trees.
Sequence analysis algorithms: These algorithms are used to analyze the sequence
path or flow of user interaction with a particular system. In terms of data, these determine the
hierarchical relationship between different datasets. The sequence algorithm that is available
in Analysis Services is Sequence Clustering.
3.4.2 Choosing an Algorithm by Task

The following table provides suggestions for the types of tasks for which each algorithm
is traditionally used.
Examples of tasks Algorithms to use
Decision Trees Algorithm

Naive Bayes Algorithm
Predicting a discrete attribute
Clustering Algorithm
Neural Network Algorithm
Decision Trees Algorithm

Predicting a continuous attribute Time Series Algorithm
Linear Regression Algorithm
Predicting a sequence Sequence Clustering Algorithm
Finding groups of common items in Association Algorithm

transactions Decision Trees Algorithm
Clustering Algorithm
Finding groups of similar items
Sequence Clustering Algorithm
4 Comparative study about Business Intelligence tools

The aim of this section is to perform an effective comparison among the most popular
platforms on the market, so we are going to evaluates and compares the business intelligence
platform strategies and business intelligence platform components of some BI suppliers.
4.1 The Leading Suppliers of Business Intelligence Platforms

The BI market has matured to the point where many vendors now offer both BI and
performance management tools. The figure represents the largest BI vendors of the world IT
market.
48
Figure 13 largest BI vendors of the world IT market
There are a plenty of business intelligence tools on the market that address the
increasing analytics needs of businesses of all sizes and industries such as: IBM, Information
Builders, Logi Analytics, Microsoft, Micro Strategy, Oracle, SAP, SAS, Qlik Technologies Inc.
In the next section, we present BI analytics tools from three leading vendors: IBM, Microsoft
and Oracle. Let's examine which product can best meet our project's needs.
4.2 Selection Criteria of BI Tools

While choosing a BI tool, it is necessary to take the following criteria into consideration:
functionality, complexity of solutions, and compatibility.
This comparative evaluation is made through the analysis of the components of

business intelligence platforms: data warehousing databases, OLAP, data mining, interfaces,
and build and manage capabilities. All three vendors have characteristic strengths and
limitations. Relative to its competitors, Microsoft has a few disadvantages and several
significant advantages.
49
Data Warehouse Databases: BI tools should support both relational and

multidimensional data warehousing databases.
OLAP: is a Critical business intelligence platform component. Business intelligence

platforms must provide: OLAP support and functionality within their databases.
Data Mining: BI tools should provide data mining functionality that offers a range of mining
algorithms that can operate on data warehouse data.
Interfaces: BI platforms should provide open interfaces to data warehouse databases. In

addition, OLAP, and data mining interfaces should comply with standards.
Build and Manage Capabilities: Business intelligence platforms should provide the
capabilities to build and manage data warehouses in their data warehouse databases. In
addition, a wide range of data sources should be supported including databases, files, and the
data of popular packaged software. Another important condition should be provided; a
coherent platform, not a set of diverse and heterogeneous technologies. For example, a single
toolset should provide build and manage capabilities across both relational and
multidimensional data warehouses.
4.3 Comparing Business Intelligence Platforms Strategies

Microsoft, Oracle, and IBM all offer Data Warehouse databases as the foundation for
their business intelligence platforms:
Table 2 Comparing Business Intelligence Platforms
capabilities Microsoft Oracle IBM
LTE (Extract, SQL Server Oracle

DB2 Warehouse
transformation, and Integration Services Warehouse
Manager
load) (SSIS) Builder (OWB)
DB2 Universal
Data Warehouse SQL Server Oracle9i
Database
DB2 OLAP
SSAS (SQL Server
OLAP Oracle9i OLAP Administrative
Analysis Services)
Services
SSAS (SQL Server Oracle9i Data DB2 Intelligent

Data Mining
Analysis Services) Mining Miner
All these BI platforms provide excellent relational data warehousing capabilities as well
as integrated OLAP data warehousing capabilities, data mining capabilities, and build and
50
manage capabilities. However, their approach to integration of other business intelligence

platform functions and their implementations of those functions are the areas where
significant differentiation becomes apparent. Now we’ll get into the differentiated areas
between these technologies. (18)
4.3.1 Comparing Data Mining

The business intelligence platforms of Microsoft, Oracle, and IBM all integrate data
mining capabilities. We can differentiate these offerings by packaging, tools, implementation
of data mining functions, algorithms, and input.
Table 4 comparing data mining capabilities
product Description
 Microsoft includes data mining capabilities to the OLAP functionality

packaged in SQL Server Analysis Services.
 The best features integration within Microsoft’s BI platform, wizard-driven
Microsoft
model building, and the capability to mine relational, OLAP, or external OLE
DB data.
 Analysis Services implements a wide range of data mining algorithms
 Oracle Data Mining, a separately packaged and priced product, provides data
mining capabilities.
 Based on technology called Darwin that Oracle recently purchased from
Oracle Thinking Machines Corporation, this product implements a broad range of
data mining algorithms and can apply those algorithms to data within
Oracle8i.
 Oracle Data Mining does not integrate with Express Edition.
 Intelligent Miner provides data mining capabilities. It’s a separately sold and
separately packaged product that is tightly integrated with DB2.
 It can useDB2 data as input and also can create its output within the
IBM
database. Originally developed by IBM Research, Intelligent Miner is a
widely-used and well-proven data mining product that implements a broad
range of algorithms.
4.3.2 Comparing Packaging and Pricing

When packaging and pricing are factored into the analysis, Microsoft leaves IBM and
Oracle behind. It delivers at least five to eight times the value of IBM or Oracle. But, that’s not
a good assumption because Microsoft provides at least equivalent capabilities for most
51
platform components, lesser functionality only in data mining and only compared to IBM, and
greater functionality in build and manage, OLAP, and application interfaces. (18)
Table 5 Comparing Packaging and Pricing
Product Price What We Get

SQL Server Enterprise Edition
Microsoft $19,999 OLAP / data mining / build & manage
capabilities
Oracle $95,000 All BI Platform
IBM $138,600 Without ETL & Data cleansing
4.3.3 Summarizing the analysis

Microsoft, IBM, and Oracle address all of our business intelligence platform
requirements. They provide relational data warehousing, build and manage facilities, OLAP,
data mining, and application interfaces to relational data warehouses. Summarizing our
analysis of each business intelligence platform:
 Microsoft provides a comprehensive business intelligence platform. Build and manage

capabilities, OLAP capabilities, and application interfaces are its key strengths. Data
mining is very new, although data mining integration and data mining tools are quite
good. (18)
 Oracle provides a comprehensive business intelligence platform. While this platform has
a complete set of components, OLAP and data mining functionality are unproven, data
mining tools are low level, and build and manage capabilities are not consistently
implemented for relational and OLAP data.
 IBM provides a rich business intelligence platform. Relational data warehousing and data
mining are key strengths. OLAP capabilities are very good, but the platform is not well
integrated. Build and Manage capabilities require too many toolsets and there’s a
disconnection between managing relational data and managing OLAP data.
4.4 Why choosing Microsoft SQL server Business intelligence

solution
Business intelligence (BI) delivers on a simple promise to improve business
performance by delivering better decisions. Unfortunately, companies are struggling to extract
data from the islands of separately acquired business applications. There has not been an easy
way to obtain a holistic view across different data sources. In order to solve this problem, both
Microsoft and Oracle offer end-to-end solutions for Business Intelligence
52
On the surface, these feature offerings may appear to be quite similar. However, the
integration of these products within their respective BI suites can differ dramatically. For
example, the Oracle BI product line was created and expanded through its acquisition of Siebel
Systems Inc., Sunopsis Inc., and Hyperion Solutions Corp. As a result, Oracle is still working
on integrating the metadata model, user interfaces, and even the packaging of these initially
separately developed solutions. In contrast, Microsoft BI solutions have been built to work
together as the result of a unified design. The key benefits of Microsoft BI solutions are their
tight integration with and between SQL Server and Office products, their strong development
capabilities, and their acquisition cost and packaging.
The Microsoft BI platform provides a professional grade integrated development

environment. The SQL Server BI Development Studio is based on the highly popular
development tool, Microsoft Visual Studio. In contrast, Oracle BI development tools can vary
by product and may not integrate with other Oracle development tools.
5 Conclusion
In this chapter, we presented a literature review related to BI and Data Mining in
Telecommunications, first we defined the main areas of BI and Data Mining applications. Last,
we identified the most common Data Mining techniques and methods used. The next chapter
will be the first step toward the identification of the requirements specification of our
application
53
Chapter 4 Requirements Specification
1 Introduction............................................................................................................................55
2 Actors Identification ................................................................................................................55
3 Functional requirements..........................................................................................................55
4 Global Use Case Diagram .........................................................................................................56
5 Refinement of use cases ..........................................................................................................56
5.1.1 Use case “Performance Analysis” Diagram .......................................................................56

5.1.2 Textual Description of the use case “Analyze 3G Key Performance Indicator”..............57
5.2.1 Use case “Monitor alarms” Diagram ................................................................................58

5.2.2 Textual Description the use case “Monitor 4G EUTRAN alarms” ...................................59
5.3.1 Use case “Perform Prediction” Diagram...........................................................................59

5.3.2 Textual Description of the use case “Predict UTRAN KPIs Behavior” .............................60
5.3.3 Textual Description of the use case “Predict EUTRAN Alarms Behavior” ......................61
6 Package Diagram of Use cases .................................................................................................61
7 Conclusion ..............................................................................................................................62
54
1 Introduction
This chapter contains a complete description of the behavior of the system to be
developed. We begin by defining the actors then we present the functional requirement of our
project. In the next step we analyze our system requirements through the development of use
case diagrams and their detailed text descriptions
2 Actors Identification
Our application could be used by either HUAWEI engineer or trainees. Because there
is no limitation in access permission to the application except the login; one type of user is the
business owner of the projected application is named FDT-user.
3 Functional requirements
1- Performance Analysis of the Radio Access Network: With the UTRAN/ EUTRAN
Performance Analysis interface, the user can:
 Visualize the statue of the radio access network based either on the counter files
provided by him or on historical measures.
 Display the distribution of the statistics results in different objects within the specified
time.
 Detect Key Performance Indicator (KPI) (cf. Page 26 Paragraph 4) degradation,
discover the causes and the possible solutions to fix this troubleshooting, hence the
user can get many details about this degradation: date of start, date of end, severity
level, location and the value of this degradation.
 Export result data to files in Excel format or pdf format.
2- Fault Alarm monitoring of the Radio Access Network: With the fault alarm monitoring
interface, the user can:
 Display the fault alarms) (cf. Page 49 Paragraph 2) stored in the application database.
 Select an alarm from the list to query the alarm details, such as the location
information, the alarm processing and the alarm cause.
 Query alarm According to several criteria (date of occurrence, Severity, Statuses,
location).
 Identify the most occurred fault cause, which location has the greatest number of faults
(Top worst Cells), the distribution of failure cause between cells.
3- Prediction of the Radio Access Network behavior: With the behavior Prediction
interface, the user can:
 Perform Prediction about the Trend of alarm causes and its distributions between
cells.
55
 Perform Prediction about the behavior of the Radio Access Network (RAN) in term of
Accessibility, Mobility, Retainability.
4 Global Use Case Diagram

The aim of this section is to describe the expected behavior of the application. For this
purpose, we base on the use case diagram to model the functionalities of the application in user
point of view.
Figure 14 Global use case diagram
5 Refinement of use cases

We detail in this section some use cases for better visibility of the application.
5.1 Use case “Performance Analysis”

5.1.1 Use case “Performance Analysis” Diagram
The user can
 Visualize the statue of the radio access network based either on the counter files
provided by him or on historical measures.
 Display the distribution of the statistics results in different objects within the specified
time.
 Detect Key Performance Indicator (KPI) degradation, discover the causes and the
possible solutions to fix this troubleshooting, hence the user can get many details about
56
this degradation: date of start, date of end, severity level, location and the value of this
degradation.
Figure 15 Use case “Performance Analysis “
5.1.2 Textual Description of the use case “Analyze 3G Key

Performance Indicator”
Table 6 Textual Description of the use case “Analyze 3G Key Performance Indicator”
Name Analyze 3G Key Performance Indicator
Summary After entering the counter file, the period, the type of RNC and after had
chosen the category of KPIs, the user can get details about the trend of the
chosen KPI during the selected period. He can visualize average values,
charts of the selected KPI.
Actor 3g,4g FDT User
Precondition The user must login to the application.

The user can enter a counter file that contains 3G KPIs measurement during
a period of time.
Nominal 1- The user chooses a RNC to analyze it behavior.

scenario 2- The user select the period of time he wants to analyze it.
3- An analysis of the selected KPI is shown.
3-1- The user choose the category of KPIs that he wants to analyze it.
57
3-2- The user can have more details about the category of KPIs chosen.
3-3- The user visualize details about the trend, average values, and
charts of the selected KPI.
4- The user can switch between categories and between KPIs.
5- The user can apply a filter.
6- The user can export result data to files in Excel format or pdf format.
Postcondition A report is generated with the select format.

Exception  The user enters a wrong file (not Excel file).
 The user doesn’t choose the RNC
 The user doesn’t choose the period of analysis.
 The user enters a wrong period (start date or end date doesn’t exist in
the counter file or in database).
5.2 Use case “Alarm monitoring”

5.2.1 Use case “Monitor alarms” Diagram
The user can
 Display the fault alarms stored in the application database.

 Select an alarm from the list to query the alarm details, such as the location
information, the alarm processing and the alarm cause.
 Query alarm According to several criteria (date of occurrence, Severity, Statuses,
location).
 Identify the most occurred fault cause, which location has the greatest number of faults
(Top worst Cells), the distribution of failure cause between cells.
Figure 16 Use case “Alarm monitoring”
58
5.2.2 Textual Description the use case “Monitor 4G EUTRAN alarms”

Table 7 Textual Description the use case “Monitor 4G EUTRAN alarms”
Name Monitor 4G EUTRAN alarms
Summary The user is able to monitor the behavior the EUTRAN through the generated
alarms in the network. He can get details about the selected alarm during the
defined period. He can visualize and display the fault alarms.

The user can enter a log file that contains detected alarms during a period of
time.
Nominal 1- The user chooses a RNC to analyze it behavior.

scenario 2- The user selects the period of time he wants to analyze it.
3- A diagnosis of current alarms is shown.
4- The user can select an alarm from the list to query the alarm details,
such as: location information, alarm processing and alarm cause.
5- The user can query results according to several criteria (date of
occurrence, Severity, Statuses, location).
6- The user can identify the most occurred fault cause, which location has
the greatest number of faults (Top worst Cells), the distribution of
failure cause between cells.
7- The user can export result data to files in Excel format or pdf format.

Exception  The user enters a wrong file (not Excel file).
 The user doesn’t choose the period of analysis.
 The user enters a wrong period (start date or end date doesn’t exist in
the counter file or in database).
5.3 Use case “Perform Prediction”

5.3.1 Use case “Perform Prediction” Diagram
The user can:
 Perform Prediction about the Trend of alarm causes and its distributions between cells.
 Perform Prediction about the behavior of the Radio Access Network (RAN) in term of
Accessibility, Mobility, Retainability.
59
Figure 17 Use case “Perform Prediction”
5.3.2 Textual Description of the use case “Predict UTRAN KPIs

Behavior”
Table 8 Textual Description of the use case “Predict UTRAN KPIs Behavior”
Name Predict UTRAN KPIs Behavior
Summary The user is able to use a forecasting algorithm embedded in the application
to perform prediction of the degradation of UTRAN Performance through
charts that present the evolution of a selected KPI during future date.
Nominal 1- The user chooses a RNC to predict it behavior.

scenario 2- The user selects a KPI name he want to forecast it degradation.
3- The user selects the number of step he want to predict.
4- A graph is shown result of the forecasting algorithm
5- Results show both a table of the historical and predicted values and a
chart that expresses this data.

Exception  The user doesn’t choose a RNC
 The user doesn’t choose a NodeB
60
 The user doesn’t choose a cell

 The user doesn’t choose the number of steps he want to predict
5.3.3 Textual Description of the use case “Predict EUTRAN Alarms

Behavior”
Table 9 Textual Description of the use case “Predict EUTRAN Alarms Behavior”
Name Predict EUTRAN Alarms Behavior
Summary The user is able to use a neural network algorithm embedded in the
application to perform prediction of the distribution of alarms ‘causes in the
EUTRAN.
Nominal 1- The user chooses an eNodeB to predict it behavior.

scenario 2- Results show both a table of the predicted values and a chart that
expresses this data.

Exception The user doesn’t choose a eNodeB
6 Package Diagram of Use cases

After we presenting our project requirements using use cases we create package
diagrams to organize these usage requirements. As it is shown in figure below, we split the
global use cases diagram into three package; one for analyzing Radio interface performance,
the second present the fault monitoring and anomalies detection and the third is associated
with prediction the radio interface behavior in UMTS and LTE networks.
61
Figure 18 Use Cases Package Diagram
7 Conclusion
In this chapter, we defined the scope of our application by identifying actors, presenting
functional requirements then we detailed some important use cases. In the following chapter
we extend this analysis to a conceptual study where we develop the application design from the
developer's point of view.
62
Chapter 5 Conceptual Study
Chapter 5 Conceptual
Study
1 Introduction...........................................................................................................................64
2 Static modelization ................................................................................................................64
3 Dynamic Modelization ...........................................................................................................67
4 Conclusion .............................................................................................................................69
63
1 Introduction
This chapter presents a set of conceptual models that can be used to construct a
simulation of our application. We begin by a static modelization where will describe the static
structure of the system. We warp up by a dynamic Modelization where we express and model
the behavior of our application system over time.
2 Static modelization
In this section, we present a static view classes that make up the design/analysis space.
The definition of use cases set out in the previous chapter will allow us to find the main
candidate classes of our application. Using the class diagrams we will describe the static
structure of the system.
64
2.1 Class Diagram
Figure 19 Class Diagram
65
2.2 Package Diagram

The division of the problem into specialized subparts makes it possible to refine the
technical specification. Thus, the adequate distribution of the load between the different layers
makes it possible to better organize the treatments and the interactions between the classes,
which improves the quality and the handling of the components of the system. Figure 20
models the various packages that make up our application and their dependencies.
Figure 20 Package Diagram
2.3 Deployment Diagram

The deployment model is used to represent the physical architecture of a system. It
illustrates the distribution of software components based on threads (nodes). Therefore, it
proposes a static view of the topology of the hardware on which the system is running.
The deployment diagram in Figure 23 models the hardware components used to

implement our solution and the association between these components.
66
Figure 21 Deployment Diagram
3 Dynamic Modelization
We use the dynamic modeling to represent the behavior of the static constituents of our
application. Activity diagram and Sequence diagrams represents the interaction workflow, and
different states of the static constituents of the application.
3.1 Activity Diagram

We present the activity diagrams of one of the main use cases of our application to
visualize the dynamic nature of our system. The figure below present the activity diagram”
Predict 3G KPI Trend” use case:
Figure 22 Activity diagram of "Predict 3G KPI trend" use case

67
3.2 Sequence Diagram

We present the sequence diagram of two main uses cases of our application to captures the
time sequence of operation flow from one object to another.
3.2.1 Sequence Diagram of “Predict 3G KPI Trend” use case
Figure 23 Sequence Diagram of “Predict 3G KPI Trend” use case
68
3.2.2 Sequence Diagram of “Predict 4G alarm causes Trend”
Figure 24 Sequence Diagram of “Predict 4G alarm causes Trend”
4 Conclusion
In this chapter, we have highlighted the conceptual modelization of our application. In the
following chapter, we will discuss the choice of development technology used to implement
each functionality and the realization steps of our application.
69
Chapter 6 Realization
1 Introduction ......................................................................................................... 71
2 Deployed Technologies ......................................................................................... 71
2.1.1 Microsoft SQL Server components .............................................................................. 71

2.1.2 Business Intelligence SQL server implementation process .......................................... 72
2.1.3 SQL Server data mining features ................................................................................. 73
2.2.1 Data Definition Statements ......................................................................................... 74

2.2.2 Data Manipulation Statements ................................................................................... 75
3 Application Demonstration ................................................................................... 77
4 Conclusion............................................................................................................ 84
70
1 Introduction
We have proceeded in the last chapters to analyze of requirements specification and the
conception of our application. The next step is the realization. We begin by the presentation of
the deployed technologies, then we present the configuration of the data mining technique
used in our application. Finally we demonstrate the realization of our project by presenting the
interfaces of the application.
2 Deployed Technologies
2.1 Microsoft SQL server for Business Intelligence
Microsoft BI combines BI concepts with the built-in features of SQL Server product. As
Microsoft technology advances, the company has taken a head-on approach improving
business intelligence.
2.1.1 Microsoft SQL Server components

The core of Microsoft BI consists of the components that make up Microsoft SQL
Server, as shown in Figure
SQL Server Reporting SQL Server Integration SQL Server Analysis

Services (SSIS)
Services (SSRS) Services (SSAS)
SQL Server Database Engine
Figure 25 The core components of the SQL Server product
When SQL Server come to mind, many people think only of the database engine
because it does the most basic database task: It creates regular relational databases. In fact,
SQL Server contains functionality apart from the database engine, including some highly
useful BI tools. Table 10 outlines these core components and what they do.
71
Table 10 Components of Microsoft BI solution
Product Description
SQL Server
The core program used to create standard relational databases, including
Database
data warehouses and data marts
Engine
SQL Server
This is the primary reporting component in Microsoft Business
Reporting
Intelligence. SSRS that can incorporate data from a number of different
Services
data sources into reports.
(SSRS)
Used for connecting to a multitude of data sources. Before data can be

SQL Server turned into useful information, it has to be collected and transformed into
Integration a standard format. In business intelligence, the tool that speaks those
Services (SSIS) multiple languages is called an Extract Transform Load (ETL) tool. The
ETL tool is called SQL Server Integration Services (SSIS).
SQL Server This is the component of SQL Server that you use to analyze data, usually
Analysis after it’s been transformed. The SSAS component is made up of an OnLine
Services Analytical Processing (OLAP) engine and a Data Mining engine.
(SSAS)
In our project we used the SSAS Tool to implement the intelligent engine into our
application. So we are going to focus on SAAS Component and how we used in our project.
OLAP Engine refers to optimizing a database for analytical activities such as aggregation,
grouping, and slicing and dicing data in various ways. OLAP databases are known as cubes, the
numerical data that is analyzed is called a fact or measure, and the way of grouping the data is
called dimension.
Data Mining Engine contains a number of mathematical algorithms that can be used on data.
Data mining is also called predictive analytics or machine learning.
2.1.2 Business Intelligence SQL server implementation process

We used a Microsoft BI component to import raw Performance data, to transform it
into a usable format, to organize it, and to load it into a data warehouse witch is the SQL Server
Integration Services (SSIS). When the data are stored and organized, they’re ready to be loaded
into the intelligent analysis engine to run the data mining algorithm via the SQL Server
Analysis Services (SSAS). Finally, SQL Server Reporting Services (SSRS) is used to display
results to the decision-maker who is in our case the 3g-4g-user.
72
We present the steps that we followed to implement the intelligent engine into our
application:
1- KPI log files and alarms log files are loaded to the SQL Server Integration Services (SSIS)
2- SSIS cleans up the data and transforms it into a format that’s usable across the application.
For example, the date format may differ between different log file. SSIS transforms these
dates into a standard format before loading it into the data warehouse.
3- SQL Server Analysis Services (SSAS) is used to build analytics model based on data mining
algorithms. An analytic model is created and parameterized according to user inputs.
4- SQL Server Reporting Services (SSRS) is used to build reports, using the data in the data
warehouse or results of data mining model as a source.
5- Our application connects to the SSAS (data mining models) to take advantage of the data-
mining capabilities. Reports and charts can then be created and published.
2.1.3 SQL Server data mining features

SQL Server has been a leader in predictive analytics since the 2000 release, by
providing data mining in Analysis Services. The combination of Integration Services,
Reporting Services, and SQL Server Data Mining provides an integrated platform for predictive
analytics that encompasses data cleansing and preparation, machine learning, and reporting.
SQL Server supports the entire development lifecycle of data in a project, which is referred to
as Integrate, Analyze, and Report.
SQL Server Data Mining is a platform for developing intelligent applications. It includes
multiple standard algorithms, including neural networks and the time series algorithm that we
implemented them into our application.
Table 1: SQL Server Data Mining algorithms used in the FDT application
Model Description
The Time Series algorithm is used to analyze and forecast time-based data. Key
Performance Indicator (KPI) are analyzed and predicted using the Time Series
Time Series
algorithm. This algorithm looks for patterns across multiple data series so that
businesses can determine how different elements affect the analyzed series.
Neural networks are the core of artificial intelligence. They seek to uncover
Neural
relationships in data that other algorithms miss. While it tends to be slower than
Network
the other algorithms, it finds relationships that may be non-intuitive.
73
2.2 Data Mining Extensions (DMX)

Data Mining Extensions (DMX) is a language that we can use to create and work with
data mining models in Microsoft SQL Server Analysis Services. DMX is used to create the
structure of new data mining models, to train these models, and to browse, manage, and
predict against them.
2.2.1 Data Definition Statements

We use DMX to create and define new mining structures and models, to import them
into our application, and to drop existing models. DMX allows us to:
Table 11 DMX statements
statement Description
CREATE MINING STRUCTURE Creates a new mining structure in the SQL Server database
ALTER MINING STRUCTURE Adds a mining model to an existing mining structure.
Creates a new mining structure and mining model in the

CREATE MINING MODEL
database.
DROP MINING MODEL Deletes a mining model from the SQL Server database.
DROP MINING STRUCTURE Deletes a mining structure from the SQL Server database.
1- Creates a new mining structure: the statement below creates a mining structure.
CREATE [SESSION] MINING STRUCTURE <structure>

(
[(<column definition list>)]
)
Arguments:
- Structure: A unique name for the structure.

- Column definition list: A comma-separated list of column definitions. A single column
can be defined as: column name><data type> [<Distribution>] [<Modeling Flags>]
<Content Type> [<column relationship>]. (19)
2- Creates a new mining model that is based on an existing mining structure
ALTER MINING STRUCTURE <structure>

ADD MINING MODEL <model>
(
<column definition list>
[(<nested column definition list>) [WITH FILTER (<nested filter criteria>)]]
)
74
USING <algorithm> [(<parameter list>)]
[WITH DRILLTHROUGH]
[,FILTER(<filter criteria>)]
Arguments:
- Structure: The name of the mining structure to which the mining model will be added.
- Model: A unique name for the mining model.
- column definition list: A comma-separated list of column definitions.
- nested column definition list: A comma-separated list of columns from a nested table,
if applicable.
- nested filter criteria: A filter expression that is applied to the columns in a nested table.
- Algorithm: The name of a data mining algorithm, as defined by the provider.
- parameter list: Optional. A comma-separated list of provider-defined parameters for
the algorithm.
- filter criteria: A filter expression that is applied to the columns in the case table. (19)
2.2.2 Data Manipulation Statements

We used data manipulation statements in DMX manipulate already crated mining
models, with these statement we can browse the models and to create predictions against them.
We can perform the following tasks with the data manipulation statements in DMX:
 Train a mining model by using the INSERT INTO statement.
 Browse information that is calculated during model training and stored in the data
mining model, such as statistics of the source data by using the SELECT statement.
The following table lists the statements that are part of the data mining data manipulation:
Table 12 Data Manipulation Statements
command Description
DELETE Clears the trained content from a mining model.
INSERT INTO Trains a mining model.
SELECT Browses a mining model.
<source data query> Queries data sources for both INSERT
INTO and SELECT statements.
UPDATE Changes the content in the mining model.
1- Training data mining model
Before this step data mining model is just a container that specifies the columns used
for input, the attribute that we are predicting, and parameters that tell the algorithm how to
process the data. Training refers to the process of applying a specific mathematical algorithm
to the data in the structure in order to extract patterns. The patterns that we find in the training
process depend on the selection of training data and the chosen algorithm.
75
Syntax:
INSERT INTO [MINING MODEL]|[MINING STRUCTURE] <model>|<structure>.COLUMN_VALUES
(<mapped model columns>) <source data query> USING <algorithm> [(<parameter
list>)]
Arguments:
- Model: A model identifier.

- Structure: A structure identifier.
- mapped model columns: A comma-separated list of column identifiers and nested
identifiers.
- source data query: The source query in the provider-defined format.
2- Browses a mining model.
SELECT [FLATTENED] DISTINCT [TOP <n>] <expression list> FROM <model>

[WHERE <condition list>][ORDER BY <expression>]
Arguments
- n: Optional. An integer specifying how many rows to return.

- expression list: A comma-separated list of related column identifiers (derived from the
model) or expressions.
- Model: A model identifier.
- condition list: A condition to restrict the values that are returned from the column list.
- Expression: Optional. An expression that returns a scalar value. (19)
3- Predict against mining model: we can perform predictions based on a mining model
by using the following queries:
Table 13 Data Mining prediction queries
Query Description
Returns a prediction that is created by joining the columns in the

SELECT FROM
mining model to the columns of an internal data source.
[NATURAL]
The domain for this query type is the predictable columns from the
PREDICTION JOIN
model and the columns from the input data source.
Returns the most likely state of the predictable column, based only
on the mining model. This query type is a shortcut for creating a
SELECT
prediction with an empty prediction join.
FROM <model>
The domain for this query type is the predictable columns from the
model. (19)
76
2.3 Neural Network configuration

Using The DMX (Data Mining Extension) language, we embedded the Microsoft Neural
Network algorithm into our application to find a pattern in existing raw data. These data
contains triggered fault alarms of the Radio Access Network (RAN). Neural Network algorithm
accept inputs with different data types. These inputs are the main factors that will affect the
predictions that we want to make to enrich our analysis.
We want to find out the distribution of alarm causes based on the previous triggered alarms.
However, it is not just about the causes. We also want to have a prediction on alarm location
and severity.
For predicting the causes of alarms, many factors will affect such as historical alarms and the
related KPIs that provoke trigging of alarms when they exceed a certain threshold.
Using the Microsoft Business Intelligence solution, SSAS (SQL Server Analysis Services) we
succeeded to create and to train the neural network mining model and finally to create the
prediction queries and export results.
Creating a data source and a data source view is straightforward. We begin by defining a data
source to our application database, and then we create an associated data source view for the
related tables.
3 Application Demonstration
3.1 Login Interface
Figure 26 Login Interface
77
3.2 3G UTRAN Performance Analysis Interface
Figure 27 3G UTRAN Performance Analysis Interface
78
The user can view the average values of each KPI by clicking on the name of the associated
category.
Figure 29 3G KPIs of the Retainability category
Figure 28 Average Values of each category's KPIs
The user can display details about each KPI and analyze it degradation by clicking on a KPI
category. A new form is shown to display a dashboard with the selected KPI. User can navigate
between categories and automatically between KPIs. User can display a graph or a table that
shows all KPIs values during the selected period.
79
Figure 30 Retainability KPIs
Figure 31 the Retainability KPI : "3G Call Drop Rate" Trend
3.3 UTRAN Alarm Monitoring interface

Within the UTRAN Alarm Monitoring Interface, users can display the fault alarms
stored in the application database. Then he can select an alarm from the list to query the alarm
details. In addition, he can identify the most occurred fault cause, which location has the
greatest number of faults (Top worst Cells) and the distribution of failure cause between cells.
As shown in figure below alarms are shown with different colors to identify them by severity
80
Figure 32 UTRAN Alarm Monitoring interface
Users are able to create a filter criteria to search for specific alarms by clicking on the Query
button:
Figure 33 filter information Interface
81
3.4 Key Performance Indicator (KPI) Prediction interface

To forecast future values and to predict the trend of Key Performance indicators, the user
can use Key Performance Indicator interface to choose a RNC to predict it behavior. After
selecting a KPI associated to its category, the user selects the number of step he want to predict.
A graph shows result of the forecasting algorithm applied on historical data.
As demonstrated in figure 34 Results show both a table and chart of the historical and
predicted values of the 3G Radio Drop Call Rate-Speech.
Figure 34 Key Performance Indicator Prediction Interface
this KPI can be defined as rate of calls which end due to technical reasons, were cut off
before the speaking parties had finished their conversation and before one of them had hung
up. This type of call termination is regarded as dropped calls. The dropped-call rate is
measured as a percentage of all calls. It is in between 0.1% and a few percent.
3.5 EUTRAN Alarm Cause Prediction interface

With the Alarm cause Prediction Interface the user chooses an eNodeB to predict it flow
behavior. Results shows the distribution of alarms. Because Alarms can be triggered when
there is a material failure or when certain Key Performance indicators exceed some thresholds.
The data required for the neural Network algorithm are the historical alarms (source, severity,
location, date of occurrence, cause) and the key performance indicators related to the
triggering of a specific alarm.
82
The picture below shows the results of the execution of the neural network. This algorithm will
identify patterns that will successfully predict the target alarm cause. It discovers recurrent
patterns of alarms in database and determines relationships between these alarms and KPI.
Figure 35 Parameterizing the neural network results
Results show then the correlation, in terms of alarm sources between the different
detected causes of alarms using the neural network algorithm.
Figure 36 alarms causes distribution according to neural network model
Figure 37 another example of alarms causes distribution according to neural network model
83
General Conclusion
4 Conclusion
In this last chapter we presents the different technologies that we used to implement
our application. Then we demonstrated the realization through presenting the application’s
interfaces associated to each functionality.
General Conclusion
Telecommunications industry has undergone intensive growth and development

during the last decade. Telecom operators and carriers are operating today in an extremely
challenging business environment, due to the increasing customer dissatisfaction with existing
services.
Network data Telecommunication networks contain thousands of components, which

are interconnected. Hence, these components are capable of generating error and, which leads
to a large volume of network data. These network data are used for network management
functions like Performance analysis and fault diagnosis.
We proposed a fault detection tool the analysis of the Radio Access Network (RAN) Key
Performance indicators and the monitoring of current alarms. In addition, our application
attempts to find relationships among historical data by sifting through databases
automatically, seeking for regularities or patterns. Strong patterns will likely generalize to
make accurate predictions on future data;
We implemented an application that enable users to profits from raw data (such as
collected Key Performance indicator log files and triggered alarms log files) and effectively used
to solve some of the important business problems and for getting competitive advantage, by
turning it into useful information and knowledge using some data mining techniques.
Time Series algorithm is used to predict future trends and performance using he
gathered raw data. Allowing decision makers to prevent undesirable states, to take
consideration of these diagnosis in the process of network planning and dimensioning.
Neural network algorithm is used for the identification of network faults by

automatically extracting knowledge from alarms log files. In order to detect alarm causes more
efficiently than traditional techniques.
84
Bibliography
Bibliography
1. Rule Discovery in Telecommunication Alarm Data. Mika Klemettinen, Heikki Mannila and Hannu
Toivonen. Journal of Network and Systems Management.
2. Inmon, W.H. Building Data Warehouse 4th Edition. 2005.
3. ETL and its impact on Business Intelligence. Anand, Nitin. February 2014.
4. Reinschmidt, J & Francoise, L. Business Intelligence Certification Guide. 2000.
5. Hobbs, L. Warehousing, Elsevier Digital Press, Burlington, Massachusetts. 2007.
6. G. CALAS. Études des principaux algorithmes de data mining, Spécialisation Sciences Cognitives
et Informatique Avancée. France. : s.n.
7. M. J. BERRY, G. S. LINOFF. neural networks and data mining concepts.
8. Microsoft Neural Network Algorithm.
9. Kramer, Mitch. A Comparison of Business Intelligence Strategies and Platforms. s.l. : Green Hill
Analysis.
10. SELECTING A DEVELOPMENT APPROACH. United States Department of Health and Human
Services (HHS). March 27, 2008.
85
Webography
Webography
W1. Huawei enterprise.
http://enterprise.huawei.com/ilink/usenterprise/products/network/wireless/lte/HW_198009.
W2. Lucidarme, Pierre Lescuyer Thierry. Evolved Packet System (EPS). worldcat. 2008.
http://www.worldcat.org/wcpa/servlet/DCARead.
W3. Wannstrom, Jeanette. HSPA. 3GPP. http://www.3gpp.org/technologies/keywords-acronyms/99-

hspa.
W4. UMTS network. docshare. http://docshare04.docshare.tips/files/20598/205986417.pdf.
W5. Poole, Ian. 3G UMTS / WCDMA Network Architecture. http://www.radio-

electronics.com/info/cellulartelecomms/umts/umts-wcdma-network-architecture.php.
W6. The LTE Network Architecture. http://computerscience.engineering.unt.edu.
W7. LTE basics. http://lteabi.blogspot.com/2012/12/enodeb-functions.html.
W8. KPI in LTE Radio Network (Huawei Based). slide share.

https://www.slideshare.net/RayKhastur/kpi-in-lte-radio-network-huawei-based.
W9. Anoop Kumar. Business Intelligence – An Overview. developer. August 4, 2014.

http://www.developer.com/db/business-intelligence-an-overview.html.
W10. Data Mining Extensions (DMX) Reference.

https://opbuildstorageprod.blob.core.windows.net/output-pdf-files/en-us/SQL.sql-
content/live/dmx.pdf.
W11. aminuddin, muhamad. Quality of Service (QoS) in LTE. telcoaaminto. [Online] March 3, 2015.
http://telcoaaminto.blogspot.com.
W12. Difference Between . HSDPA vs HSUPA.

http://www.differencebetween.net/technology/difference-between-hsdpa-and-hsupa/.
86

Rapport Copie Finale PDF

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Rapport Copie Finale PDF

Hochgeladen von

Copyright:

Verfügbare Formate

MINISTERE DE L’ENSEIGNEMENT SUPERIEUR

‫المعهد العالي لإلعالمية وتقنيات االتصال بحمام سوسة‬

Institut Supérieur D’informatique

MEMOIRE DE PROJET DE FIN D’ETUDES

Présenté en vue de l’obtention du Diplôme National d’Ingénieur

Performance Analysis and Fault detection tool for the

3G-UTRAN and 4G-EUTRAN based on Business Intelligence

Année Universitaire 2016 - 2017

MINISTERE DE L’ENSEIGNEMENT SUPERIEUR

‫المعهد العالي لإلعالمية وتقنيات االتصال بحمام سوسة‬

INSTITUT SUPERIEUR D’INFORMATIQUE

MEMOIRE DE PROJET DE FIN D’ETUDES

Présenté en vue de l’obtention du Diplôme National d’ingénieur

Performance Analysis and Fault detection tool for the

3G-UTRAN and 4G-EUTRAN based on Business Intelligence

Encadrant : Date : …………………Signature :…………….

Encadrant : Date :……………………Signature :…………….

Année Universitaire 2016 - 2017

5 The adopted Methodology ...................................................................................... 7

2.1.1 GSM (Global System for Mobile communications) .......................................................... 13

2.2.1 UMTS (Universal Mobile Telecommunications System) ................................................. 13

2.3.1 HSPA+ (Evolved High-Speed Packet Access) .................................................................. 16

2.4.1 LTE Advanced .................................................................................................................... 18

4 LTE Network Architecture .................................................................................... 21

5 Performance Measurement ................................................................................. 25

5.3.1 Definition ...........................................................................................................................26

5.4.1 Accessibility ....................................................................................................................... 28

5.5.1 Accessibility KPI................................................................................................................. 31

2.2.1 Data Source Layer .............................................................................................................. 41

3.2.1 The classification ............................................................................................................... 45

3.4.1 Choosing an Algorithm by Type ........................................................................................ 47

4.3.1 Comparing Data Mining .................................................................................................... 51

5.1.1 Use case “Performance Analysis” Diagram ....................................................................... 56

5.2.1 Use case “Monitor alarms” Diagram .................................................................................58

5.3.1 Use case “Perform Prediction” Diagram ........................................................................... 59

3 Dynamic Modelization ..........................................................................................67

3.2.1 Sequence Diagram of “Predict 3G KPI Trend” use case .................................................. 68

2.1.1 Microsoft SQL Server components .................................................................................... 71

2.2.1 Data Definition Statements ............................................................................................... 74

3 Application Demonstration ................................................................................... 77

Figure 1 Proposed solution diagram ................................................................................................ 6

Table 1 RUP Workflows description ................................................................................................ 8

Table 2 Workflows Chronogram ...................................................................................................... 9

Table 3 Differences between HSDPA and HSUPA ....................................................................... 16

Table 4 comparing data mining capabilities ................................................................................. 51

Table 5 Comparing Packaging and Pricing ................................................................................... 52

Table 10 Components of Microsoft BI solution ............................................................................ 72

Table 11 DMX statements ............................................................................................................... 74

Table 12 Data Manipulation Statements ....................................................................................... 75

Table 13 Data Mining prediction queries ...................................................................................... 76

RNS: Radio Network Subsystem

Chapter 1 State Of Art

5 The adopted Methodology ..................................................................................................... 7

4 Study of existing solutions

The M2000 provides basic functions such as configuration management, performance

The M2000 is a future-proof solution for managing mobile networks. It provides a

4.2 Study of the FMA (Fault Manage Assistant)

 Configuration script of the RNC.

4.4 Proposed solution

This application gives users to ability to:

Figure 1 Proposed solution diagram

5 The adopted Methodology

5.1 RUP (Rational Unified Process) Presentation