Wudneh Alebign

ADDIS ABABA UNIVERSITY
ADDIS ABABA INSTITUTE OF TECHNOLOGY

SCHOOL OF ELECTRICAL AND COMPUTER ENGINEERING
Mobile Data Offloading using Wi-Fi for the case of

Ethio Telecom in Addis Ketema Area
By
Wudneh Alebign
Advisor
Dr. Yihenew Wondie
A Thesis Submitted to the School of Graduate Studies of Addis Ababa

University in Partial Fulfillment of the Requirements for the Degree of
Masters of Science in Communication Engineering
June, 2018
Addis Ababa, Ethiopia
ADDIS ABABA UNIVERSITY
ADDIS ABABA INSTITUTE OF TECHNOLOGY
SCHOOL OF ELECTRICAL AND COMPUTER ENGINEERING
Mobile Data Offloading using Wi-Fi for the case of
Ethio Telecom in Addis Ketema Area
By: - Wudneh Alebign
Approved by Board of Examiners
________________________________
Chairman, School of Graduate Studies Signature
Committee
Dr. Yihenew_Wondie ________________
Advisor Signature
________________ ________________
Internal Examiner Signature
________________ ________________
Internal Examiner Signature
II | P a g e
Declaration
I, the undersigned, declare that this thesis is my original work, has not been presented for a degree
in this or any other university, and all sources of materials used for the thesis have been fully
acknowledged.
Wudneh Alebign __________________

Name Signature
Place: Addis Ababa

Date of Submission: July 11, 2018
This thesis has been submitted for examination with my approval as a university advisor.
Dr. Yihenew Wendie __________________

Advisor Signature
II | P a g e
Abstract
One of the most difficult task for any mobile network operator is to efficiently manage an
increasing number of mobile data users especially in a congested urban areas. These high number
of data users require higher bandwidth and high speed throughput. To overcome these problems,
high capital investments are required which are not economically viable. Mobile network operators
are forced to see another alternative to solve the problem by using mobile data offloading
technologies. Mobile data offloading is one of the technique to efficiently handle the growing
mobile data traffic by offloading cellular data traffic to other complementary networks. This will
decrease the burden on mobile network operators and increase the capacity of data network.
Collected data from ethio telecom shows that there is high number of mobile data traffic usage in
3G network that creates network congestion. Currently in Ethiopia ethio telecom solves the
problem by upgrading cellular network to 4G/LTE, using smaller cell/sector method and upgrade
cellular network by adding base stations/transmitters and enhance software resources. Since the
number of peoples in the urban areas like Addis Ababa are increasing dynamically, it is difficult
task for ethio telecom to manage rapid growth of mobile data usage using the above techniques
only. Hence offloading cellular network to other complementary network provides ethio telecom
an advantage to enhance the data network capacity easily.
This thesis also proposes Wi-Fi offloading technique as one of the alternative way to solve mobile
network congestion problems in a congested urban areas in Addis Ababa city Addis Ketema area.
The thesis also shows how Wi-Fi offloading will improve the capacity of the whole mobile
network using Atoll simulation tools and presents options on the technical integration of Wi-Fi
and cellular network. And finally the paper shows the challenges and future works of Wi-Fi
offloading in Ethiopia. The simulation result shows that after offloading a percentage of 3G data
traffic to Wi-Fi network the required grade of service for both voice and data improved.
Keywords: Mobile Data Offloading, Wi-Fi, Wi-Fi Offloading
III | P a g e
Acknowledgement
First and foremost, I would like to thank Almighty God who provides me strength this far, and
supports me all the way to this phase of my life.
I wish to express my gratitude to my university supervisor, Dr. Yihenew Wondie for his invaluable
assistance, support and guidance. Additionally I would like to thank AAIT communication
department staffs for their assistance and guidance.
I would like to thank ethio telecom staffs especially Fitsum Mergia and Yibeltal who supports and
provides me necessary information. Without their assistance this study would not have been
successful.
I would like to thank my family and friends who supports me, especially Tadesse Alebign, who
encourages and motivates me beyond my expectation.
IV | P a g e
Table of Contents
1 Introduction.................................................................................................................. 1
1.2 Statement of the problem ................................................................................................................................. 2
1.3 Objective ................................................................................................................................................................ 3
1.3.1 General objective ................................................................................................................................. 3
1.3.2 Specific objective ................................................................................................................................ 3
1.4 Literature Review ............................................................................................................................................... 3
1.5 Scope and Limitation......................................................................................................................................... 4
1.6 Methodology ......................................................................................................................................................... 4
1.7 Contribution.......................................................................................................................................................... 5
1.8 Thesis Layout........................................................................................................................................................ 5
2.0 Background ............................................................................................................... 7

2.1 Cellular network ................................................................................................................................................. 7
2.2 Wi-Fi network ...................................................................................................................................................... 9
2.2.1 Describing WLAN Topologies........................................................................................................ 9
2.2.2 Wi-Fi Standards ................................................................................................................................ 10
2.2.3 Service Set Identifier ....................................................................................................................... 12
2.2.4 Roaming technology in WLAN ................................................................................................... 12
2.2.5 WLAN Security ................................................................................................................................ 13
2.3 Mobile data offloading .................................................................................................................................. 14
2.3.1 3GPP Release 6 ................................................................................................................................. 14
2.3.2 3GPP Release 7 ................................................................................................................................. 21
2.3.3 3GPP Release 8 ................................................................................................................................. 22
2.3.4 3GPP Release 9 ................................................................................................................................. 25
2.3.5 3GPP Release 10 .............................................................................................................................. 25
2.3.6 3GPP Release 11 .............................................................................................................................. 29
2.3.7 3GPP Release 12 .............................................................................................................................. 30
2.3.8 3GPP Release 13 .............................................................................................................................. 31
2.3.9 3GPP Release 14 .............................................................................................................................. 33
3.0 WLAN design.......................................................................................................... 34

3.1 Wi-Fi Coverage and capacity ..................................................................................................................... 34
3.2 Identify application throughput.................................................................................................................. 36
3.3 Key performance indicators ........................................................................................................................ 36
P a g e V | 81
3.4 Estimating number of Client and AP Counts ........................................................................................ 37
4.3.1 Estimate access point throughput ................................................................................................ 37
4.0 Implementation consideration ............................................................................ 41

4.1 Cellular network monitoring ....................................................................................................................... 41
4.2 Resource of 3G network ................................................................................................................................ 41
4.3 3G Resource usage metrics.......................................................................................................................... 42
4.3.1 Uplink Load ....................................................................................................................................... 42
4.3.2 Downlink Load ................................................................................................................................. 43
4.3.3 Channel Element resource usage ................................................................................................. 43
4.3.4 OVSF code usage ............................................................................................................................. 44
5.0 Simulation................................................................................................................ 45
5.1 Simulation Tool ................................................................................................................................................ 45
5.2 Planning and design ....................................................................................................................................... 45
5.2.1 Addis Ketema site Information .................................................................................................... 45
5.2.2 Deployment Steps ............................................................................................................................ 47
5.2.3 Traffic parameters ............................................................................................................................ 52
5.2.4 Propagation model ........................................................................................................................... 54
5.2.5 Deployment target ............................................................................................................................ 55
5.2.6 Simulation Steps ............................................................................................................................... 55
6.0 Result ........................................................................................................................ 57

6.1 3G Network before Wi-Fi offloading ....................................................................................................... 57
6.1.1 Study coverage signal level ........................................................................................................... 57
6.1.2 Study transmitter coverage ............................................................................................................ 58
6.1.3 3G Capacity analysis ....................................................................................................................... 58
6.1.3.1 3G network capacity Before Wi-Fi offload .............................................................................................................58
6.1.4 Result Analysis before Wi-Fi offloading................................................................................... 60
6.2 UMTS coverage After Wi-Fi offloading .................................................................................................. 61
6.3 UMTS and Wi-Fi network integration ..................................................................................................... 62
6.3.1 Users Demand ................................................................................................................................... 63
6.3.2 Result ................................................................................................................................................... 64
7.0 Conclusion and future works .............................................................................. 65
Reference ........................................................................................................................ 67
VI | P a g e
List of Figures
Figure 1 Ad hoc mode WLAN connectivity [5] ............................................................................................ 9
Figure 2 Basic Service Set WLAN connectivity [5] ................................................................................. 10
Figure 3 Extended Services Set WLAN connectivity [5] ........................................................................ 10
Figure 4 IEEE 802.11b frequency band allocation [5]. ............................................................................ 11
Figure 5 Scenario 2 3GPP-WLAN interworking architecture [11] ....................................................... 16
Figure 6 End to end EAP authentication [11] .............................................................................................. 17
Figure 7 Authorization and authentication using diameter. [11] ........................................................... 18
Figure 8 authentication based on EAP AKA scheme [12] ...................................................................... 19
Figure 9 3GPP-WLAN interconnection architecture [12]. ...................................................................... 20
Figure 10 ANDSF architecture [16] ............................................................................................................... 22
Figure 11 Non-Roaming Architecture within EPS using S5, S2a, S2b [16] ...................................... 24
Figure 12 multi access PDN connectivity [15] ........................................................................................... 26
Figure 13 IP Flow mobility [15] ...................................................................................................................... 27
Figure 14 WLAN offloading for different data flow using IFOM [18] ............................................... 28
Figure 15 non-seamless IP mobility [18] ...................................................................................................... 28
Figure 16 111053_Addis Ketema site map .................................................................................................. 46
Figure 17 Study level of signal coverage ...................................................................................................... 57
Figure 18 Study transmitter coverage ............................................................................................................ 58
Figure 19 3G Capacity Result before Offloading ....................................................................................... 59
Figure 20 3G network simulation after Wi-Fi offloading ........................................................................ 62
Figure 21 Cumulative Cellular and Wi-Fi network status ....................................................................... 63
VII | P a g e
List of Tables
Table 1 coverage vs capacity [33]................................................................................................................... 36
Table 2 application requirement [33] ............................................................................................................. 36
Table 3 Wi-Fi capability [33] ........................................................................................................................... 37
Table 4 Addis Ketema Site information ........................................................................................................ 46
Table 5 1 Site parameter for 3G network...................................................................................................... 48
Table 6 Transmitter parameter for 3G network .......................................................................................... 49
Table 7 Cell parameters for 3G network ....................................................................................................... 50
Table 8 Wi-Fi site parameter ............................................................................................................................ 51
Table 9 Wi-Fi Transmitter parameter ............................................................................................................ 51
Table 10 3G user server profile ....................................................................................................................... 52
Table 11 Wi-Fi user server profile .................................................................................................................. 52
Table 12 Terminal profile .................................................................................................................................. 53
Table 13 Environment profile .......................................................................................................................... 53
Table 14 User profile for 3G ............................................................................................................................ 54
Table 15 User profile for 3G ............................................................................................................................ 54
Table 16 deployment target .............................................................................................................................. 55
Table 17 user distribution per service ............................................................................................................ 56
Table 18 3G network capacity analysis ......................................................................................................... 59
VIII | P a g e
List of abbreviations
1G First Generation Mobile Networks
2G Second Generation Mobile Networks
3G Third Generation Mobile Networks
3GPP Third Generation Partnership Project
4G Fourth Generation Mobile Networks
AAA Authentication, Authorization and Accounting
ANDSF Access Network Discovery and Selection Function
AP Access Point
BSS Basic Service Set
CAGR Compound Annual Growth Rate
CAPEX Capital Expense
CDMA Code Division Multiple Access
DL Downlink
DSMIPv6 Dual-stack Mobile IPv6
DSSS Direct-Sequence Spread Spectrum
EAP-AKA Extensible Authentication Protocol-Authentication and Key Agreement
EDGE Enhance Data rates for GSM Evolution
eNB evolved NodeB
EPC Evolved Packet Core
EPS Evolved Packet System
ESS Extended Service Set
FACoA Foreign Agent Care of Address
FAST Flexible Authentication via Secure Tunneling
FDD Frequency Division Duplex
FQDN Fully Qualified Domain Name
GPRS General Packet Radio Service
GSM Global System for Mobile communications
IX | P a g e
GTP GPRS Tunneling Protocol
HSDPA High Speed Downlink Packet Access
HSPA High Speed Packet Access
HSS Home Subscriber Server
HSUPA High-speed Uplink Packet Access
IBSS Independent Basic Service Set
IEEE Institute of Electrical and Electronics Engineers
IFOM IP Flow Mobility
IKEv2 Internet Key Exchange version 2
IMS Internet multimedia server
IP Internet Protocol
ISMP Inter-System Mobility Policy
ISRP Inter-System Routing Policy
I-WLAN Interworking-WLAN
LAN Local Area Network
LEAP Lightweight Extensible Authentication Protocol
LTE Long Term Evolution
MAC Media Access Control
MAPCON Multiple Access PDN Connectivity
MD5 Message Digest 5
MIMO Multiple-Input and Multiple-Output
MIPv4 Mobile IPv4
OVSF Orthogonal Variable Spreading Factor
PCC Policy and Charging Control
PDA Personal Digital Assistant
PDN Packet Data Network
PDN GW GW Packet Data Network Gateway
PEAP Protected Extensible Authentication Protocol
X|Page
PLMN Public Land Mobile Network
PMIP Proxy Mobile IP
PMIPv6 Proxy Mobile IPv6
QoE Quality of Experience
QoS Quality of Service
RAN Radio Access Network
RFC Request for Comments
RNC Radio Network Controller
RSSI Received Signal Strength Indicator
RTWP Received Total Wideband Power
SIM Subscriber Identity Module
SINR Signal to Interference-plus-Noise Ratio
SSID Service Set Identifier
TCP Transmission Control Protocol
TLS Transport Layer Security
TTLS Tunneled Transport Layer Security
UDP User Datagram Protocol
UE User Equipment
UL Uplink
UMTS Universal Mobile Telecommunications System
VoIP Voice over IP
VPLMN Visited Public Land Mobile Network
VPN Virtual Private Network
WCDMA Wideband Code Division Multiple Access
Wi-Fi Wireless Fidelity
WiMAX Worldwide Interoperability for Microwave Access
WLAN Wireless Local Area Network
XI | P a g e
1 Introduction
Our world is connected through globalization and becoming one village through Internet. Internet
plays a crucial role in the current dynamically growing world by connecting peoples and providing
instant access to global information. The number of Internet users using mobile data network is
also increasing from time to time in the world. The introduction of smart phones and tablets in the
market and the people’s habit in the use of Internet for social networking, web surfing, instant
messaging, video communications are increasing from time to time and it will have impact on the
current mobile networks. Mobile network operators are expected to ensure that their networks are
able to handle an increase number of mobile data network usage.
From Cisco visual networking index [1], global mobile data traffic will increase nearly seven fold
between 2016 and 2021. Mobile data traffic will grow at a compound annual growth rate (CAGR)
of 47 % from 2016 to 2021, reaching 49 Exabyte per month by 2019. Further the study shows that
by 2016, 63 percent of all traffic from mobile-connected devices (almost 84 Exabyte) will be
offloaded to the fixed network by means of Wi-Fi devices and femtocells each month. Of all IP
traffic (fixed and mobile) in 2021, 50% will be Wi-Fi, 30% will be wired, and 20% will be mobile.
In Ethiopia also the number of users who have access to mobile Internet especially in the urban
areas are increasing from time to time. In May 2017 the GlobaStats Stat Couter studies shows that,
from total Ethiopia Internet connection 80.08% of Internet connection come from mobile network,
whereas 17.74% come from desktop usage and the rest 1.18% Internet usage come from tablet
users [2]. The number shows that there are high number of mobile data usage in Ethiopia compared
with desktop.
Additionally in Ethiopia ethio telecom provides EVDO, 3G and 4G data VPN service for banking,
public and private sectors to connect their branches using mobile network. All the increase in
mobile data usage consumes the mobile data network and obviously will create network
congestions [43].
Hence mobile network operators have to plan forward and expected to provide good performance
and higher capacity data service to their customers in parallel before network congestion is created
There are a number of ways to address the network congestions problems, these are by creating
smaller cells, using heterogeneous network and by using mobile data offloading. Mobile data
P a g e 1 | 81
offloading is a technique that offloads data connection from cellular network to another
complementary network. Currently mobile data offloading becomes interesting and cost effective
technology compared with the other methods to solve mobile network congestions by integrating
both 3GPP standard (cellular) network with non-3GPP network [27].
The thesis also focuses and describes how mobile data offloading using Wi-Fi technology will
reduce network congestion problems and increases the overall capacity of the wireless network for
the case of ethio telecom in Ethiopia, Addis Ababa, Addis Ketema area. As we know Wi-Fi uses
unlicensed band spectrum that is working in the band of 2.5 GHz and 5GHz Frequency. Wi-Fi has
a capacity that can provide a data rate service up to 300Mbps based on IEEE 802.11n standard and
even up to 1Gbps speed with a new IEEE 802.11ac standard with low price scheme compared with
other wireless technologies [44]. This will enable the mobile network operators to deliver
broadband Internet service to their customers with satisfied data rate, with a good indoor coverage
and increase its income.
1.2 Statement of the problem

Mobile network congestion is one of the most critical issue for mobile network operators. Data’s
from ethio telecom shows that there is 3G network congestion problems mainly comes from data
connection. The problem also imposes a burden on ethio telecom to provide reliable service to its
customers. There are a number of factors that enables network congestion to occur on mobile
networks. But mainly network congestion occurs due to a rising number of users that uses mobile
data services. The increase usage of high speed Internet, network intensive mobile applications,
gaming and videos are the main source of mobile data network congestions.
This problem will have negative effect on the satisfaction of subscribers by decreasing the traffic
intensity and performance of mobile service. It will also create a burden for ethio telecom to handle
all mobile data services using cellular network only.
Currently in Ethiopia ethio telecom tries to solve the problem by upgrading cellular network from
3G to 4G network, using additional transmitters, using smaller cells/sectoring mechanism and
enhancing software resources. But this option is not a cost-effective way to solve the mobile data
congestion problems in most congested urban areas. Therefore, this paper proposes Wi-Fi
2|Page
technology as an alternative way to solve this problem by offloading mobile data traffic in to Wi-
Fi network.
The main purpose of the paper is to solve the current network congestion problem in most
congested urban areas in Addis Ababa like stadium, market places and in the densely populated
areas in a cost-effective way using Wi-Fi network compared with other methods.
Moreover, this paper shows how mobile data offloading using Wi-Fi increases the data coverage
of indoor areas and traffic intensity of the cellular network. In another way, it will increase the
capacity of the cellular network. Hence it will increase the satisfaction of its customers, increase
the income of ethio telecom. Wi-Fi network integration with cellular networks. Finally the paper
discusses on the challenges and futures works on the deployment of Wi-Fi offloading in Ethiopia.
1.3 Objective
1.3.1 General objective
The main objective of the project is to solve the mobile network congestion problems for the case
of ethio telecom in Addis Ababa, Ethiopia particularly in congested areas like stadium, market
places, and densely populated areas by offloading mobile data traffic in to Wi-Fi network.
1.3.2 Specific objective

Specifically this paper addresses the following three points.
1. Reduce mobile data traffic to complementary Wi-Fi network using atoll simulation tool.
2. Increase the capacity of mobile network in the integration with Wi-Fi network.
3. Increase the satisfaction of the mobile network customer by increasing data rate service.
1.4 Literature Review

There are different literatures studies about mobile data offloading advantage, best technical
approaches and Wi-Fi offloading feasibility status and maturity status.
Rebecchi, et al. presents advantage of mobile data offloading and different approach on technical
implementation [45]. They shows that the rapid increase in video traffic creates burden on cellular
network and discussed that shifting the network load to another wireless technology improves the
3|Page
overall throughput, network coverage and increase network availability. Additionally they classify
offloading techniques using fixed Wi-Fi access points (AP) and using Terminal-to-terminal
approaches. Where in fixed access point mode the users’ data traffic will be offloaded to fixed Wi-
Fi AP. while in Terminal-to-Terminal mode users use another subscriber mobile phones as
intermediate or bridge device. I.e. users are not expected to directly connect to Wi-Fi AP.
Beneyam, et al. presents mobile data offloading business model for the case of Africa market [46].
They covers different approaches of mobile data offloading techniques including current vendor
solutions that are available in the market. Additionally they showed standardization on the
integration of cellular and Wi-Fi network become mature and analyze in depth from political,
economic, social and technical approach and proposes best business model approach for the case
of Africa.
Rajavelsamy, et al. reviews the evolution of 3GPP integration with WLAN network [47]. Based
on their study starting from 3GPP release 6 up to release 13, the integration between cellular and
WLAN shifted from loose integration to more tight and seamless integration. Additionally they
forecasted that leveraging WLAN cost-effectiveness and its unlicensed radio access technology it
will maximize the revenue of mobile network operators.
Aijaz, et al. analyzes practical aspects of mobile data offloading to Wi-Fi network [48]. They
discussed that although the integration of cellular and WLAN network become mature through
time, still there are some challenges that needs to be addressed. Hence they proposes different
recommendation for different levels of challenges.
1.5 Scope and Limitation

The scope of the thesis is limited to show how Wi-Fi offloading reduce mobile network congestion
problems by integrating cellular network and Wi-Fi network together from technical aspect only.
The thesis work simulation focuses in Ethiopia, Addis Ababa city, Addis Ketema area only. The
paper doesn’t cover from legal, social and business angles.
1.6 Methodology
The thesis work started by reading about mobile network congestions, the reason behind mobile
network congestions, and isolate different options to solve the problem. After communicating ethio
4|Page
telecom, it is clear that the problem is visible in Addis Ababa, Ethiopia area especially in congested
places. And then study about basic mobile network technologies and do literature reviews related
to cellular network, Wi-Fi network and about mobile data offloading techniques. Additionally in
order to make the study visualize Atoll simulator is selected which is one of popular planning tool
which can simulate both cellular and Wi-Fi network together. The next step was gathering
necessary data like identifying places where areas that are most congested, gather cellular design
parameters and information of site, transmitter and cell parameters, avail Atoll simulator and Addis
Ababa digital maps. It takes some time to gather all necessary data’s and to familiarize with Atoll
simulation tools. Since in Addis Ababa area most of users are using 3G network, we have covered
in simulation part 3G and Wi-Fi network technology only. Then using gathered 3G network design
parameters from ethio telecom, the first simulation is started on selected congested sites. And see
the results of simulation, identify the reasons behind network congestions. The next simulation is
done by offloading some percentage of cellular network in to Wi-Fi network and see the result and
identify the changes. Then compare the simulation result before and after offloading from
coverage, capacity and performance level. Finally based on the outcomes of simulation result the
paper will do recommendation, conclusion and future works to do.
1.7 Contribution
Although there are different literature reviews on the advantage of mobile data offloading and also
known that some telecom companies already implemented as alternative way to manage mobile
data traffic, ethio telecom still not leverage the advantage. Hence this thesis work is mainly to
recommend ethio telecom to see mobile data offloading using Wi-Fi as an alternative way to solve
network congestion problems in urban areas.
1.8 Thesis Layout

The thesis work is organized in order to give a clear understanding regarding the subject matter.
Chapter one contains the introduction, statement of the problems, objective of the thesis, literature
review, scope of the thesis, methodologies, contribution and thesis layout. Chapter two presents
basic study behind the thesis works that includes cellular network, Wi-Fi network and about mobile
data offloading. Chapter three discusses about Wi-Fi design considerations in very congested
places and calculates the required number of access points that are needed to cover a given area.
5|Page
Chapter four presents considerations for 3G network before implementation. Chapter five contains
the simulation part and shows gathered data from ethio telecom and also contains planning and
design of 3G and Wi-Fi network. Chapter six present the result based on simulation output. Mainly
the result contains 3G network before offloading, after offloading and in the integration of 3G and
Wi-Fi network. Finally conclusion, recommendation and future works on the thesis works.
6|Page
2.0 Background
2.1 Cellular network
Mobile network is evolved drastically for the last four decades. Starting from first generation 1G
network to fourth generation 4G LTE networks, the mobile network is increased in capacity and
service types. The growth in information technology and internet has direct impact on the
advancement of mobile network. Next we will revise the mobile generations from different
perspectives.
First Generation (1G) wireless mobile network
The 1G is known as the first generation mobile telecommunication was introduced in 1980s and
mainly uses to give analog based voice service to customers [38]. Technically the 1G mobile
network uses frequency modulation technique with 150 MHz frequency bandwidth and had
2.4kbps speeds [39]. There were different standards of 1G mobile network like Advance Mobile
Phone Service (AMPS), Nordic Mobile Telephone (NMT), and Total Access Communication
System (TACS). 1G mobile network had some drawbacks regarding security, voice quality and
use of large cell phones.
Second Generation (2G) wireless mobile network
As peoples need grow to use better voice quality over wireless network and additional application
requirements, the mobile technology evolved to 2G mobile network. Basically the 2G network is
target to address the drawback of 1G network and introduce additionally technologies like text and
image messaging. 2G has higher data rate speed and have better voice quality compared with 1G
mobile network. GSM (Global system for mobile communication) is one of the most popular
technology which is implemented widely in the world. GSM operates on 900 MHz and 1800 MHz
frequency band and uses TDMA (Time division multiple access) technique for transmitting and
receiving. The GSM network mainly provides voice service and have drawbacks on providing data
service. And because of that 2G network is evolved to 2.5G (second and half generation wireless
mobile network) to support data service. The 2.5G mobile network uses packet-switching
technology for data communication and works on top of 2G network and provides data rate serves
up to 144kbps [40]. The three popular 2.5G technologies are GPRS (General Packet Radio
7|Page
Service), (HSCSD) High-Speed Circuit-Switched Data and Enhanced Data Rates for GSM
Evolution (EDGE). 2.5G mobile network is transition period for 3G network which provides more
data rate service to customers.
Third Generation (3G) wireless mobile network
3G network is the third generation wireless mobile telecommunication technology introduced in

2000s which provides data rate service more than 144kbps speed. CDMA2000 and UMTS
(Universal Mobile Telecommunications Systems) are most popular standards of 3G network. 3G
network introduces voice call and mobile TVs and high speed Internet services. 3G network
architecture constitutes User Equipment (UE), Radio access network subsystem and Core network.
HSPA and HSPA+ provides additional capacity over previous 3G technology. HSPA provides
14.4 Mbps downlink and 5.76 Mbps uplink while Release 11 HSPA+ provides 336Mbps and
69Mbps respectively [41]. Both HSPA and HSPA+ technologies were the transition to 4G LTE
network which provides additional data rate speed.
Fourth Generation (4G) LTE wireless mobile network
As bandwidth intensive Internet applications, videos and gaming usage increase 3G network
evolved to 4G LTE network. 4G LTE is fourth generation cellular network based on long term
evolution technology. All connections between 4G LTE interfaces are all IP based technology
which provides seamless mobility to users. 4G LTE network is capable of providing more than
100Mbps bandwidth to users which utilizes MIMO and OFDM technology to provide high data
rate service. 4G LTE comprises The User Equipment (UE), The Evolved UMTS Terrestrial Radio
Access Network (E-UTRAN) and The Evolved Packet Core (EPC). Compared from previous 2G
and 3G technology 4G LTE technology provides higher capacity, performance, more security and
low latency communication.
Fifth Generation (5G) wireless mobile network
Still it is expected that network bandwidth intensive Internet applications will continue and 5G
network is in the way parallely to satisfy peoples need. 5G network is successor of 4G LTE
expected to provide more than 1Gbps speed with efficient bandwidth utilization. 5G standards are
still in the development stage and it is expected to be implemented in 2020.
8|Page
2.2 Wi-Fi network
Wi-Fi is an acronym for Wireless Fidelity and is most commonly used to describe a wireless local
area network based on the IEEE 802.11 series of standards, which is a set of wireless technical
specifications issued by the Institute of Electrical and Electronic Engineers (IEEE) [4].
Wi-Fi uses radio waves technology to transmit and receive data between client devices and access
point or between client devices. And works mainly in the 2.4 GHz and 5 GHz frequency band. Wi-
Fi becomes well known technology as an alternative to wired connections, in connecting Laptops,
Tablets, printers, scanners and different Wi-Fi supported devices. Wi-Fi technologies provide
wireless connectivity with in a building, campus and city wide environments.
2.2.1 Describing WLAN Topologies

Mainly there are two types of WLAN topology these are ad hoc mode and infrastructure mode.
Infrastructure mode consists of basic service set and extended service set [5].
 Ad hoc mode: in this mode direct peer to peer connection established between two WLAN
enabled devices without intermediate
Figure 1 Ad hoc mode WLAN connectivity [5]
 Infrastructure mode: In this mode WLAN enabled devices connected through additional
3rd party wireless access point. There are two types of infrastructure mode
o Basic Service Set: in this mode a single wireless access point devices interconnect
WLAN enable devices.
9|Page
Figure 2 Basic Service Set WLAN connectivity [5]
o Extended Services Set: in this mode different wireless access point devices are
interconnected through additional network devices.
Figure 3 Extended Services Set WLAN connectivity [5]
2.2.2 Wi-Fi Standards

Wi-Fi operates on the unlicensed band of frequency and this will give flexible and easy to deploy
without control of telecommunication companies. Some of main standards are described below.
10 | P a g e
IEEE 802.11a:- Operates in the 5 GHz frequency range with maximum data rates service up to 54
Mbps and uses encoding scheme of orthogonal frequency-division multiplexing (OFDM). This
technology supports 23 non overlapping frequency channels. Because it works in the 5 GHz
frequency range it is less susceptible to radio frequency interference.
IEEE 802.11b:- Operates in the 2.4 GHz frequency range with maximum data rate service up to
11Mbps. Uses direct sequence spread spectrum (DSSS) modulation technique and provides three
overlapping frequency channels. This means three access points can operates in the same cell area
simultaneously using three different frequency channels. The following figure shows frequency
allocation of IEEE 802.11b protocol [5]
Figure 4 IEEE 802.11b frequency band allocation [5].
IEEE 802.11g:- Operates in the 2.4 GHz frequency bands and provides up to 54Mbps maximum
data rate. This standard uses OFDM technology for transmission. And provides maximum three
non-overlapping frequency channels.
IEEE 802.11i:- Security is one of main concern for Wi-Fi networks. This standard addresses
security issues using different security mechanisms. In order to secure Wi-Fi network there should
be valid authentication to ensure that legitimate users are connected to wireless network. And
encryption in order to give privacy and confidentiality.
IEEE 802.11n:- Based on MIMO (multiple input/multiple output) OFDM technology provides
high data rate service up to 600Mbps [6].
11 | P a g e
IEEE 802.11ac:- A new gigabit Wi-Fi standard and provides up to 1.73Gbps data rate service. It
supports MIMO and higher modulation techniques in order to get high data rate. This standard will
be ideal solution for video, data sync and backup [3].
IEEE 802.11u:- The main reason to introduce this standard is to make an easy integration between
Wi-Fi network and cellular network [7]. Mainly this standard specifies,
 The discovery of suitable networks through the advertisement of access network

type, roaming consortium and venue information
 Generic Advertisement Service (GAS) provides for layer 2 transport of an
advertisement protocol’s frames between a mobile device and a server in the
network prior to authentication. The AP is responsible for the relay of a mobile
device’s query to a server in the carrier’s network and for delivering the server’s
response back to the mobile
 Access Network Query Protocol (ANQP), which is a query and response based
protocol used by a mobile device to discover a range of information, like roaming
partners accessible via the hotspot along with their credential type including the
hotspot operator’s domain name; and extensible authentication protocol (EAP)
method supported for authentication; IP address type availability and other
metadata useful in a mobile device’s network selection process
2.2.3 Service Set Identifier

The SSID is the name of the wireless cell. It is used to logically separate WLANs. It must match
exactly between the client and the access point. The access point broadcasts the SSID in the
beacons. Beacons are broadcasts that the access points send to announce the available services [5].
2.2.4 Roaming technology in WLAN

Roaming is one of most important parts of WLAN technology for users’ mobility. Users can walk
through from one access point to another without network disconnection. The WLAN consists of
microcells, and the user has the ability to move freely anywhere that the RF coverage permits.
Roaming is enabled by complete coverage with wireless cells. Seamless roaming across access
12 | P a g e
points allows users to maintain a connection while moving around the facility. The wireless client
initiates the roaming if one of these conditions is detected [5]:
 The maximum data retry count is exceeded.

 The client has missed too many beacons from the access point.
 The client has reduced the data rate.
 The client intends to search for a new access point at periodic intervals.
2.2.5 WLAN Security

Since wireless network open with in coverage area any person can have access in to wireless
network and unencrypted traffic will be sniffed and captured in to 3rd party person. Wireless LAN
security is the protection from unauthorized access to computers using wireless networks. There
are different standards and protocols to protect WLAN network [8]:
Wired Equivalent Privacy (WEP): a method to make a wireless link more like or equivalent to
a wired connection. WEP aims to provide security by encrypting data over radio waves so that it
is protected as it is transmitted from one end point to another. WEP is highly susceptible to a
variety of “man-in-the-middle” attacks and session hijacks and has many vulnerabilities. Currently
WEP is very weak security standards not used widely.
IEEE 802.11i and IEEE 802.1X: IEEE 802.11i is a secure authentication mechanisms based on
IEEE 802.1X which is an encryption key management protocol and provides stronger data
encryption mechanisms. IEEE 802.1X uses EAP in order to provide strong authentication between
the client and the WLAN network. Thus the user must authenticate to radius server before allowed
to access the WLAN network.
EAP: as we have seen above IEEE 802.11X requires EAP to provide both authentication and key
management. Some of most popular types of EAP methods are Lightweight Extensible
Authentication Protocol (LEAP), Protected Extensible Authentication Protocol (PEAP), Transport
Layer Security (TLS), and Tunneled Transport Layer Security (TTLS). Each EAP types have their
own security strength and weakness so in the design and implementation process we have to
consider it properly [8].
 Message Digest 5 (MD5): one of the weakest EAP methods, and it doesn’t provide more
security to the WLAN infrastructure.
13 | P a g e
 LEAP: is the easiest two way authentication methods. But since it doesn’t use certificate.
it is susceptible to dictionary attacks.
 TLS: in the case of TLS there is exchange of certificate between client and access server.
A TLS tunnel is then built and the client can be authenticated and encryption key material
can be securely exchanged. TLS is one of the most secure WLAN authentication methods.
 PEAP does not require a certificate on the client, but will use a server-side certificate.
 TTLS: Very secure solution. It is very similar to PEAP using TLS to create a tunnel to
avoid using certificates on the client.
 Flexible Authentication via Secure Tunneling (FAST): Very secure, creates a secure
tunnel, and then uses a RADIUS server to authenticate the server and client.
WPA and WPA2: are two security protocols developed by the Wi-Fi Alliance to secure wireless
LAN networks. WPA is intend to provide more security compared with WEP. WPA and WPA2
uses IEEE 802.1X for authentication and key management; and TKIP or AES cipher suites for
encryption.
2.3 Mobile data offloading

Mobile data offloading is a mechanism to offload a cellular network to other complementary
network technology in order to reduce data usage of cellular network. There are different
technologies to complement mobile data traffic. Mainly Wi-Fi and femtocell are popular ones.
Since Wi-Fi network is cost effective network infrastructure and uses unlicensed frequency
spectrum it becomes popular offloading mechanism compared with other methods. Additionally,
since Wi-Fi technology is integrated to every smart phone devices, it is easy to integrate cellular
network and Wi-Fi network.
Third Generation Partnership Project (3GPP) has started to develop a 3GPP WLAN interworking
architecture to allow 3GPP service providers to offload data traffic from cellular network to
WLANs in indoor areas, hotspots, and other areas where high user density exist.
2.3.1 3GPP Release 6

Interconnection between WLAN and UMTS
Focuses on the interworking between WLAN and UMTS network and prepares a new specification
on integration, security, charging and management. The I-WLAN work item defines the
14 | P a g e
interworking between 3GPP systems and Wireless Local Area Networks (WLANs). For this
purpose, “3GPP - WLAN interworking” refers to the utilization of resources and access to services
within the 3GPP system by respectively the WLAN UE and user. The intent of 3GPP - WLAN
Interworking is to extend 3GPP services and functionality to the WLAN access environment. Thus
the WLAN effectively becomes a complementary radio access technology to the 3GPP system [9].
This release provides a number of different scenarios of 3GPP-WLAN interworking ranging from
common billing to the provision of services seamlessly between the WLAN and the 3GPP system.
Additionally describes an approach for a flexible, general, scalable and future proof 3GPP - WLAN
interworking [10]. 3GPP-WLAN integration can be deployed from simple architecture to scalable
and more advanced seamless interconnection. Next we will see different network integration
scenarios which are presented on 3GPP Release 6.
Interworking scenarios
There are different proposed scenarios and in each scenario there are detail steps in integrating
3GPP and WLAN network. Figure 2 shows a general simplified integration model.
Scenario 1 - Common Billing and Customer Care: - in this scenario common billing system
will be installed, where both 3GPP and WLAN network treats as a single customer. But other
services like security may be independent. This scenario only provides simplified customer care
service by subscribe a single billing system. I.e the customer does not required to subscribe
independent billing system for WLAN and UMTS network.
Scenario 2 - 3GPP system based Access Control and Charging: - in this scenario WLAN
network access control (authentication, authorization and accounting) mechanism will be the same
as 3GPP network. Since this scenario uses the 3GPP access control and charging system both the
mobile operator and customer will be easy for operation.
15 | P a g e
IP Network
3GPP Visited Network
WLAN Access Wa
WLAN 3GPP AAA
UE Network Proxy
Wd
3GPP AAA
Server HSS
3GPP Home Network
Figure 5 Scenario 2 3GPP-WLAN interworking architecture [11]
UICC smart cards with SIM/USIM applications (WLAN UE) used to authenticate the WLAN
interworking terminals. In this scenario authentication, authorization will be delivered by a
centralized AAA server which is connected to HSS home network in order to access user
information. IEEE 802.11i provides confidentiality, integrity and access control between WLAN
UE and WLAN access network. Therefore after the WLAN UE authenticated and authorized by
the 3GPP network the Internet data traffic will be redirected to WLAN network in order to access
Internet.
Figure 6 show how end to end EAP authentication works.
16 | P a g e
EAP-SIM /EAP AKA authentication
WLAN Wa Wd
AAA Proxy AAA Server HSS/HLR
802.11X
EAPOL
Diameter/MAP Authorization
WLAN information and
authentication vectors
UE needed in the
EAP over Diameter/Radius authentication
protocols
are stored by the HSS.
Figure 6 End to end EAP authentication [11]
Protocols used for Authentication
The 802.1X is a specification that defines EAP (Extensible Authentication Protocol) over LAN.
This is also known as EAPOL. EAP is a method of conducting an authentication conversation
between a user and an authentication server. 802.11x employs the EAP as authentication
framework [13]. There are different methods of implementing EAP. For the case of 3GPP-WLAN
interconnection it is specified that EAP-AKA and EAP-SIM shall be used.
EAP-SIM is an Extensible Authentication Protocol (EAP) [14] mechanism for authentication and
session key distribution using the Global System for Mobile communications (GSM) Subscriber
Identity Module (SIM). EAP-SIM uses a SIM authentication algorithm between the client and an
Authentication, Authorization and Accounting (AAA) server. EAP SIM also includes mechanisms
for identity hiding using temporary identifiers, or pseudonyms, and a fast re-authentication
procedure
EAP-AKA specifies an EAP method that is based on the Authentication and Key Agreement
(AKA) mechanism used in 3rd generation mobile networks Universal Mobile
Telecommunications System (UMTS) and CDMA2000. EAP AKA includes the same identity
hiding and fast re-authentication functions as EAP SIM.
The two main protocols used for Authentication, Authorization and Accounting service in 3GPP-
WLAN are RADIUS and Diameter. Both protocols provides the same functionality and can
encapsulate EAP messages. But Diameter is the successor of RADIUS and was developed to overcome
several limitations of RADIUS. Some improvements of Diameter are Application-layer
17 | P a g e
acknowledgments and failover algorithms, Mandatory IPsec and optional TLS supports, Reliable
transport mechanisms (TCP, SCTP), Support for server-initiated messages, Data object security is
supported but not mandatory, Capability negotiation between clients and servers, Peer discovery and
configuration.
Figure 7 Authorization and authentication using diameter. [11]
18 | P a g e
Figure 8 authentication based on EAP AKA scheme [12]
Scenario 3: Access to 3GPP system PS based services:-This scenario is to allow the mobile
operator to spread out access to 3GPP services to WLAN access network. E.g. IMS based services,
19 | P a g e
location based services, instant messaging, presence based services. And in order to have access
to the 3GPP packet-switched service from WLAN network, the user data traffic have to be tunneled
between WLAN and 3GPP network. In this scenario service continuity between WLAN and 3GPP
is not required [12]. The figure below shows that a complete end to end IPsec tunnel using IKEv2
protocol between WLAN UE and Packet Data Gateway will be implemented to protect user data
traffic. IKEv2 have important features that EAP-SIM or EAP-AKA can be used for user
authentication.
Internet
3GPP AAA
Proxy
Wa
Wg
WAG
Wd 3GPP Visited Network
Ww WLAN Access Network Wn
WLAN UE
Wx
HSS
Wp
3GPP AAA
Wm
Server Dw
Packet Data SLF

Wu
Gateway
D’/Dr’
Wi HLR
3GPP Home Network
IP Network
Figure 9 3GPP-WLAN interconnection architecture [12].
Scenario 4: Service Continuity:-here the main purpose is to provide service continuity between
3GPP and WLAN network. The user notice the translation but it is not required to reconnect the
connection again.
20 | P a g e
Scenario 5: Seamless services:-in this scenario seamless service continuity between 3GPP and
WLAN network will be provided. That is not data loss or break time service when a user changes
from one network type to the other.
Scenario 6: Access to 3GPP CS Services:- in this scenario services provided by 3GPP circuit
switch core network will be provided by WLAN network. Where any circuit-switched
characteristics will not be shown to WLAN network.
Ownership:- the owner ship of WLAN network can be the following [10]
1. 3GPP system operator.

2. A public network operator who is not a 3GPP system operator.
3. The WLAN owner is an entity providing WLAN access in a local area (i.e. building
manager/owner or airport authority) but who is otherwise not a public network operator.
4. The WLAN owner is a business entity that may be providing a WLAN for its internal use
that also wishes to allow interconnection, and possibly visitor use, for some or all of their
WLANs.
Note: The scenarios 4, 5 and 6 are not part of 3GPP Release 6 standardization, and they are covered
on the next release of 3GPP standardization.
Interworking Architecture [10]
1. Non-roaming reference model

2. Roaming reference model - 3GPP PS based services provided via the 3GPP Home Network
3. Roaming reference model - 3GPP PS based services provided via the 3GPP Visited
Network

Release 7 is mainly focused on the usage of quality of services (QOS) for 3GPP-WLAN network.
3GPP Release 6 discussed in detail when using 3GPP IP Access, a tunnel from UE to PDG is
established for carrying PS based services traffic. While accessing PS based services, it is possible
that data for more than one IP flow and for different services is carried in one tunnel. And since
the data is encrypted inside the tunnel it is not possible to differentiate individual IP flow at
21 | P a g e
intermediate devices. Therefore the only way to apply QOS is by using DiffServ to mark different
color the DS field of outer IP header at the WLAN UE or PDG [15].

Introduction of ANDSF
In Release 8 specification Access Network Discovery and Selection Function framework
(ANDSF) provides access network information and to enhance the way UE discovers a new non-
3GPP Access Networks. It also provides mobility policies in order for the operator to guide the
UE to select the proper radio technology in any given location at any given time. The ANDSF
shall respond to UE requests for access network discovery information and may be able to initiate
data transfer to the UE, based on network triggers [16]. In this release simultaneous connection to
both WLAN and 3GPP network is not allowed. I.e. a user connected either WLAN network to
offload to 3GPP network not both network. The WLAN and PLMN network access selection in
Release 8 replaces pre-release 8 specification. ANDSF provides the following information [16]
1. Inter-system mobility policy: The inter-system mobility policy is a set of operator-defined
rules and preferences that affect the inter-system mobility decisions taken by the UE. The
UE uses the inter-system mobility policy to:
(I) Decide when inter-system mobility is allowed or restricted; and
(II) To select the most preferable access technology type or access network that should be used
to access EPC.
2. Access network discovery information: when a user request network access ANDSF
provides a list of available access network technology and the radio access network
identifier or SSID for the case of WLAN.
As shown figure 10 ANDSF server is one of the components of EPC network and the
communication between UE and ANDSF is over the S14 interface.
3GPP IP Access or S14

Trusted/Untrusted ANDSF
UE
Non-3GPP IP
Access
Figure 10 ANDSF architecture [16]
22 | P a g e
EPC-WLAN architecture.
Release 8 introduce the integration of EPC and WLAN network. The Evolved Packet System is a
higher-data-rate, lower-latency, packet-optimized system that supports LTE, UMTS and other
different RATs. And supports service continuity of subscriber IP sessions during in UE handovers
from one IP access network to another IP access network, regardless of whether the new IP access
network supports the same version of IP as the old IP access network. [17].
There are two different IP access procedures a WLAN or non-3GPP network connect to EPC
network. These are trusted and non-trusted access. It is up to the operator to decide whether it is
trusted or not [18].
Trusted WLAN access
Trusted WLAN access assumes that there is a secure encrypted communication with encryption
between WLAN radio access network gateway and EPC including secure authentication
mechanism. Hence it is not necessary to to set up an additional IPSec tunnel between the UE and
the EPC network. In this architecture communication to PDN-GW is connected through S2a
interface and connection to AAA server is connected through STa interface.
Non-trusted WLAN access
Non-trusted EPC access is the evolution of the previous release 6 3GPP-WLAN interworking,
where the functionality of PDG is evolved to ePDG. In the untrusted access case the user set up
secure IPsec tunnel using IKEv2 between UE and ePDG. And the ePDG is connected to the PDN
GW through S2b interface, where each user session is transported through a secure tunnel (GTP
or PMIP). The PDN-GW provides a common data service gateway service WLAN and 3GPP
network and performs policy enforcement, packet filtering for each user, charging support, lawful
interception and packet screening. Another key role of the PGW is interface with the outside packet
data networks, for example the IMS (Internet multimedia server) and the Internet [19]. Therefore
As per request of the services, PDN-GW sends the traffic to the operator IP services (IMS, PSS)
or redirects to the Internet. Other Wi-Fi traffic may be locally broken out at the WLAN. Before a
data traffic communication started access authentication signalling should be implemented
between UE and 3GPP AAA server or Proxy AAA server in case of roaming scenario. This 3GPP
based access authentication is implemented across a SWa reference point and the Tunnel
authentication used to create IPsec tunnel between UE and ePDG is connected across SWm
reference point as shown in the figure 11 [16].
23 | P a g e
HSS
SWx
S6a PCRF
Gxc Rx
Gx
Operator's IP
SGi Services
3GPP Serving PDN (e.g. IMS, PSS
Access Gateway Gateway etc.)
S5
S6b
S2b
Gxb
SWm
S2a ePDG 3GPP AAA
Server
HPLMN SWn
Non-3GPP Gxa
Networks SWu
Trusted Untrusted
Non-3GPP IP Non-3GPP IP
Access Access SWa
STa
UE
Figure 11 Non-Roaming Architecture within EPS using S5, S2a, S2b [16]
IP Mobility between 3GPP-WLAN

IP mobility between 3GPP and non-3GPP network is first implemented in release 8 specification.
IP Mobility between 3GPP and WLAN network provides uninterrupted services for users moving
from one RAN to another. The main target of IP mobility is to preserve the IP address of UE in
order to continue the service without downtime. There are two types of IP mobility. These are
Network Based Mobility (NBM) and Host based mobility (HBM).
Network Based Mobility:

NBM mechanism is used for establishing connectivity in the target access upon inter-access
mobility, IP address preservation for session continuity based on NBM may take place as per
PMIPv6 specification, [20] and additionally based on the knowledge in the network of UE's
capability (if available) to support NBM. Such knowledge may be based on an explicit indication
from the UE upon handover that IP address preservation based on NBM management can be
provided [16].
24 | P a g e
Host based mobility
In the HBM IP address preservation take place when the network is aware of the UE capability to
support DSMIPv6 or MIPv4. Such knowledge may be based on an indication to the target trusted
non-3GPP access or ePDG from the HSS/AAA (e.g. in case of DSMIPv6, the UE performed S2c
bootstrap before moving to the target trusted non-3GPP access or ePDG) [16]. In such a case, the
trusted non-3GPP access network or ePDG provides the UE with a new IP address, local to the
access network if IP mobility management protocol selected is DSMIPv6. In that case, in order to
get IP address preservation for session continuity, the UE shall use DSMIPv6 over S2c reference
point. This IP address shall be used as a care-of address for DSMIPv6. If the IP mobility
management protocol selected is MIPv4, the address provided to the UE by the non-3GPP access
network is a FACoA and IP address preservation is performed over S2a using MIPv4 FACoA
procedures [16].
The final decision on the mobility management mechanism is made by the HSS/AAA upon UE
authentication in the trusted non-3GPP access system or ePDG (both at initial attachment and
handover), based on the information it has regarding the UE, local/home network capabilities and
local/home network policies. If the UE provided an explicit indication of the supported mobility
mechanisms, the network shall provide an indication to the UE identifying the selected mobility
management mechanism [16].

ANDSF enhancement
Release 9 mainly provides enhancement for Release 8 ANDSF discovery and communication.
This enhancement provides Release 8 ANDSF procedures to cover also roaming scenarios, i.e. the
discovery and communication with ANDSF server while UE is attached in VPLMN. Additionally
ANDSF security architecture is enhanced by using of GBA-Push mechanism for push-based
ANDSF security establishment. This requires changes to the push-based security establishment
ANDSF procedures for Rel-9 [21].

When the UE have both WLAN and 3GPP network coverage, it is important for the operator to
offload some traffic (e.g. best effort) to the WLAN access. At the same time it may be beneficial
25 | P a g e
to still keep some traffic (e.g. VoIP flow) in the cellular access. With this IP flow mobility solution
the operator can lower it data access costs while the subscriber just experiences maximised
bandwidth without any service disruption or interruption [22]. Therefore the main objectives of
release 10 is to provide simultaneous network connections to multiple radio access technologies
using Multi Access PDN Connectivity (MAPCON), IP Flow Mobility (IFOM) and non-seamless
Wi-Fi offload and to extend ANDSF framework to handle these simultaneous connections.
MAPCON
Multi Access PDN Connectivity (MAPCON) provides simultaneous network connections to

multiple radio access technologies managed by multiple PDN connections and when the UE has
multiple IP addresses for each PDN. Using MAPCON mobile offloading will take place easily
instead of requiring the UE to support multiple client-based mobility management like DSMIPv6.
One of the example of MAPCON is when a UE can transfers FTP and other best effort Internet
connections using WLAN network and other web and VOIP connections will be transferred to
3GPP networks. The following figure shows MAPCON connectivity [15]
Packet Data
Network Internet
WLAN AP
UE-1
;APN
1; Int
ernet
Untrusted/
Trusted
Network
l Serving PDN
IP cal
2; Vo Gateway Gateway
E -1 ;APN
U
Core Network
eNB
Simultaneous Multiple PDN connectivity to different Packet Data
Network(operator IP
APNs
Service)
Figure 12 multi access PDN connectivity [15]
26 | P a g e
IP flow mobility and seamless WLAN offload (IFOM)
IP flow mobility provides Simultaneous connection to 3GPP access and WLAN and exchange
different IP flows belonging to the same PDN and provides movement of IP flow from one access
network to another. The solution is based on DSMIPv6 and IP address preservation and session
continuity is provided when moving IP flows from one access network to another. By using Policy
provisioning the operator informs the UE which network access technology the IP flow is routed.
And By using DSMIPv6 IP address processing, the session can be maintained without knowledge
of the different network paths. In the same way as MAPCON, downloading large files using FTP
is performed via the WLAN network while voice calls using VoLTE and video calls are handled
via the 3GPP network.
Packet Data
Network Internet
WLAN AP
UE-1
;A
Inter PN1;
net(F
TP)
Untrusted/
Trusted
Network
PDN
1; Serving
;APN Gateway
UE-1 Stream) Gateway
net(
Inter
Core Network
eNB
Simultaneous multi-RAT access PDN connection to
the same APN
Figure 13 IP Flow mobility [15]
Using IFOM extensions, the offloading of different data flows described in Figure 14 can be
realized: Depending on the availability and quality of the access technologies, different flows can
be offloaded to WLAN while keeping the LTE connection running. In this example, the (real time)
27 | P a g e
video stream is kept on LTE, while the VoIP, Web and FTP connections are offloaded to WLAN
[18].
Figure 14 WLAN offloading for different data flow using IFOM [18]
Non-seamless offloading.
In Non-seamless offloading UE that is connected to WLAN access, route specific IP flows via the
WLAN access without traversing the EPC. For such IP flows the UE uses the local IP address
allocated by the WLAN access network and no IP address preservation is provided between
WLAN and 3GPP accesses. The following figure shows show that a subscriber who is using 3GPP
network is offloaded to WLAN network.
Figure 15 non-seamless IP mobility [18]
28 | P a g e
Extensions to the ANDSF framework
Generally the information addressed by ANDSF to UE are inter-system mobility policy (ISMP),
the access network discovery information and the inter-system routing policy (ISRP). The ANDSF
may provide all types of information or only one of them. Both ISMP and access network
discovery information are covered in release 8 specification. In Rel-10 ANDSF provides a list of
ISRP to UE in order to transmit IP traffic simultaneously over multiple radio access interfaces.
The ISRP contains the following three information
1. IFOM rule: provides filter rule identifying prioritized access network and also identifies
which radio accesses are restricted for traffic that matches specific IP filters on a specific
APN (e.g. WLAN is not allowed for RTP/RTCP traffic flows on APN-x) or on any APN
[23].
2. MAPCON: provides filter rule identifying prioritized access network and a filter rule also
identifies which radio accesses are restricted for PDN connections to specific APNs (e.g.
WLAN is not allowed for PDN connection to APN-x) [23].
3. Non-seamless WLAN offload: Filter Rules, each one identifying which traffic shall or
shall not be non-seamlessly offloaded to a WLAN when available.
After operators indicate different preferred or forbidden radio access technologies as a function of
the type of traffic the UE sends. Specifically an ISRP can be based on:
 The PDN identifier (i.e. APN) the UE uses for a given connection;
 The destination IP address the UE sends traffic to;
 The destination port number the UE connects to;
 A combination of the three elements above.

Release 11 introduce the support of Broadband Forum Accesses network Interworking. The
interworking between a 3GPP system and a Fixed Broadband Access network defined by
Broadband Forum to provide the IP connectivity to a 3GPP UE using a WLAN and a H(e)NB
connected to a Fixed Broadband Access network. The specification also provides mobility, Policy,
29 | P a g e
QoS aspects between 3GPP and a Fixed Broadband Access network as well as the respective
interactions with the PCC frameworks [24].
There are some enhancements done in Rel-11 specifications. The first enhancement is on Data
Identification in ANDSF. Since Rel-10 specification have limitation on how the traffic is identified
due to the clear of aggregation of the Internet traffic into few transport ports. For example the
operator with the Rel-10 framework is not able to discriminate between video streaming (e.g.
www.youtube.com) and web browsing (e.g. www.google.com) [25].
Therefore this release specify ANDSF extensions to provide to operators a better control of the
network resources used for each application or IP flow. Specifically this WID will specify:
 Additional ways to identify classes of traffic an ISRP applies to.

 Extensions to the ANDSF MO to convey these additional ways over S14.
Additionally FQDN representation of traffic and Application-ID representation of traffic is

included in ANDSF.
Based on the requirements from the operators to use GPRS Tunnelling Protocol (GTP) based S2a
for WLAN to access EPC, 3GPP supported GTPv2 over S2a, under S2a Mobility based on GTP
(SaMOG) work item. The use of GTP based S2a or PMIPv6 based S2a for the WLAN to access
the EPC depends on the operator’s network deployment policy [15].

In Prerelease 12 specifications UEs using WLAN networks controlled by 3GPP operators and their
partners often make suboptimal offload to/from WLAN decisions resulting in poor user experience
and inefficient resource utilization of operator's networks [26]. Therefore in order to improve
WLAN utilization when it is available and not congested, in this release proposed a solution to
improve WLAN/3GPP access network selection and traffic steering that addresses requirements
from all operators. The solution supports deployments with and without ANDSF and supports co-
existence of ANDSF with RAN rules when both are deployed.
Additionally The Hotspot 2.0 solution developed by WFA builds on the architecture and set of
protocols defined by IEEE 802.11u and develops key capabilities for network discovery and
selection of WLAN terminals based on the ANQP (Access Network Query Protocol) defined in
30 | P a g e
IEEE 802.11u. Hotspot 2.0 is focused on enabling a mobile device to automatically “discover”
APs that have a roaming arrangement with the user‟s home network and then securely connect.
Hotspot 2.0 facilitates the seamless mobility of users from one WLAN to another or from Cellular
RAN to WLAN and vice-versa, with minimal or no user intervention [27]. Release 12 proposed
3GPP network support for network selection for WLAN networks taking into account WFA
Hotspot 2.0 solutions.
For both Access Network Selection and Traffic Routing Selected cases RAN assistance parameters
transferred to UE via system broadcast and/or dedicated signaling and the UE decides accessing
WLAN or 3GPP network by RAN assistance information, UE measurements, Information
provided by WLAN, Policies obtained via the ANDSF or via existing OMA-DM mechanisms or
pre-configured at the UE.
RAN Assistance Information contains WLAN and RAN Thresholds, Load Information, Resource
Allocation and Traffic Information ( LTE-WLAN Interworking: 3GPP Release-12 Perspective).

Release 13 specification introduces new technologies and enhancement from release 12. Specific
to WLAN offloading Rel-13 shows Improvements in Radio Access Network (RAN)
implementation makes LTE and WLAN Aggregation (LWA) and where Wi-Fi can now be
supported by a radio bearer and aggregated with an LTE radio bearer. Additionally Rel-13
introduces Wi-Fi integration enhancements to support Network-Based IP Flow Mobility
(NBIFOM) enhancements to harmonize the support of voice and video services over Wi-Fi.
Pre Rel-13 specification support that, a UE can connect to both WLAN and 3GPP network
simultaneously using different PDN network. And 3GPP defined the capability for DSMIPv6
capable UEs to allow seamless offload of individual IP flows corresponding to a PDN connection
to WLAN by introducing IP flow mobility (IFOM) support to the EPC. Since many mobile
operators started deploying network-based mobility protocols (e.g. GTP and PMIP). Hence in Rel-
13 seamless offload and flow mobility using network-based protocol, PMIP and GTP based S2a
and S2b over WLAN is introduced.
31 | P a g e
LTE-WLAN RADIO LEVEL INTEGRATION
3GPP/WLAN radio interworking Release-12 solution enhances core network based WLAN
offload by improving user QoE and network utilization and providing more control to operators.
And these improvements can be further enhanced by LTE/WLAN integration at radio level [29].
The following LTE-WLAN Radio Level Integration features are standardized in Rel-13
specification.
LTE-WLAN Aggregation (LWA): LTE WLAN Ag (LWA) allows the integration of WLAN and
LTE network at RAN level. I.e. WLAN access point interacts to LTE eNB and no direct interaction
with core 3GPP network [30]. LWA enables LTE and WLAN interworking with data aggregation
at the radio access network, using an LTE dual-connectivity like framework. Here an eNB
schedules packets to be served an LTE and Wi-Fi radio links. In essence, to achieve enhanced
performance, the LTE data payload is split, with some traffic tunneled over Wi-Fi and some
transmitted over LTE [31]. LWA enables LTE and WLAN interworking with data aggregation at
the radio access network, using an LTE dual-connectivity like framework.
LTE-WLAN Radio Level Integration with IPsec Tunnel (LWIP): this will provide direct IPsec
tunnel between UE and LTE eNB. This IPsec tunnel is important to support any legacy WLAN
network infrastructure by hiding WLAN access point [32].
RAN Controlled LTE WLAN Interworking (RCLWI): RCLWI still uses CN-based offloading but
the eNB can make the decision to steer traffic between LTE and WLAN which can provide better
performance compared to previous CN based solutions. RCLWI is also based on WT and Xw
interface upgrade of the WiFi network for control signaling, however, the User Plane (UP) bearers
instead of going through the LTE eNB are routed through a CN with WiFi legacy link. This is
rather a bearer handover (or an offload) than an aggregation compared to LWA, however still the
UE is controlled by the network to receive the data from WiFi link, instead of taking this decision
by itself. Compared to LWA, this solution doesn’t require the UE to be upgraded with LWAAP
[32].
32 | P a g e
As we have seen above, 3GPP Release-13 (Rel-13) introduces interworking or aggregation
between WLAN and LTE using LWA, LWIP and RCLWI. Rel-14 builds based on Rel-13 LWA
framework and architecture, introduces enhanced LAA (eLAA), enhanced LWA (eLWA) which
optimizes the WLAN-LTE integration. Mainly eLWA focuses on the following points [32]
1. Uplink data transmission on WLAN, including uplink bearer switch and bearer split
2. Mobility optimizations, e.g., intra and inter eNB handover without WT change and
improvements for Change of WT
3. Potential enhancements to support 60 GHz new band and channels (e.g., in measurements) and
increased data rates for 802.11ax, 802.11ad, and 802.11ay (e.g., by PDCP optimizations)
4. Additional information collection and feedback e.g., for better estimation of available WLAN
capacity (by additional signaling on both Uu and Xw) to improve LWA performance
5. Automatic Neighbor Relation (ANR) for LWA e.g., for discovery of WLANs under eNB
coverage.
33 | P a g e
3.0 WLAN design
Since the introduction of WLAN 1997 [4], its capability and availability is growing for indoor and
outdoor areas. WLAN introduced to provide any service through the air from anywhere in the
range of wireless access point equivalently that a wired LAN can provides. From time to time more
number of Wi-Fi enabled devices enter the market and also the average number of users in a given
access point increases. In designing and planning WLAN network it is important to consider
different factors. Some of the things we need to consider are the types of traffic, amount of traffic
throughput, the number of clients in the network, number of access points that provides required
area coverage and bandwidth. Especially in a very highly densely environment it is very important
to consider the WLAN coverage, performance, management and security issues properly.
Additionally satisfactory performance and throughput for all users and applications need to take
account. This chapter also focus designing WLAN network in a very high density environment.
3.1 Wi-Fi Coverage and capacity

Most Wi-Fi networks are designed and implemented for indoor environment like office building,
schools, cafes, hospital etc. The wireless coverage that a user access should not compromise the
total capacity of WLAN network. In designing WLAN in densely populated public area is a little
different compared with indoor places. Highly densely environments are wide in coverage and
more peoples are entertained in a given access point.
In highly densely environments access points are deployed close to each other and WLAN
performance can be degraded by RF interferences. In order to mitigate RF interference and meet
the design targets, it is important to consider the following points [33]
 Increase the use of 5GHz RF spectrum: Wi-Fi network connections use either 2.4 GHz or
5GHz frequency band. 5 GHz has a shorter range than 2.4 GHz but its frequency reuse
factor reaches up to 12 and this provides more capacity can support much higher speeds
and also less susceptible to interference due to the small amount of devices that use this
range. The Wi-Fi trend shows that 5 GHz RF spectrum usage is increasing, currently
manufactured WLAN devices also support 5GHz RF spectrum. Especially in a very
34 | P a g e
densely environment frequency re-use factor is important in order to mitigate co-channel
interference.
 Decrease cell size as much as possible: decreasing cell size reduces the co-channel
interference between WLAN access points that use the same frequency band and increase
the capacity of the whole WLAN network. In highly densely environment it is important
and recommended to decrease the cell size as much as possible, and since there will be
more number of users in a small area, capacity is more important compared with coverage.
 Increase Access point number: proper estimation of access points is important to increase
the WLAN coverage and capacity, but it is important to plan the optimum number of access
points in a given area.
 Frequency re-use: it is a good practice to increase the frequency re-use in order to increase
the WLAN capacity. As we have seen above 5 GHz frequency spectrum has more
frequency re-use factor than 2.4GHz frequency spectrum. This will reduce the co-channel
interference between access points.
Antenna selection is also very important for both coverage and capacity. We need to select antenna
type and pattern by not compromising both coverage and capacity.
RF simulation tool and site survey: a good RF simulation tool and site survey is important to assess
all the above factors before implementing the WLAN. This will provide necessary information
about Wi-Fi coverage area, RF interferences in the area and to identify the placement of access
points. The following table shows coverage vs capacity comparison.
Factors Coverage Capacity

Number of APs Prefer Low Prefer high
Limiting factor Path loss Interference
Obstacle Bad Good
RF frequency Lower is better Higher is better
Antenna pattern Omni is better Sector is better
AP placement Higher is better Lower is better
Design metric SNA area SINR area
35 | P a g e
Table 1 coverage vs capacity [33]
3.2 Identify application throughput

Identifying types of devices and applications that will be running over the wireless network is
helpful in planning for the WLAN. Different types of applications have varying response time.
Some applications are very delay sensitive and others are not. Interactive applications like voice
and video are very delay sensitive with round trip delay of less than 150ms. Other applications like
email, web browsing and file downloading can tolerate delays of a few seconds. So in the design
of WLAN network we have to consider different types of applications and their bandwidth usage.
Application type Bandwidth

Web – Casual 500 kilobits per second (Kbps)
Audio – Casual 100 Kbps
Audio - Instructional 1 Mbps
On-demand or Streaming Video - Casual 1 Mbps
File Sharing - Casual 1 Mbps
Table 2 application requirement [33]
3.3 Key performance indicators

In designing WLAN network the following KPI (key performance indicator) metrics are important
in order to measure a successful deployment. If one of KPIs is not properly considered poor
performance will be resulted and users will not be satisfied. Recommended WLAN KPIs are the
following [33],
 Type of application supported
 Minimum bandwidth required by a given application
 Min, average and maximum number of Wi-Fi enabled devices
 Expected number of active Wi-Fi users
 Quality of service
 Service areas
36 | P a g e
3.4 Estimating number of Client and AP Counts
Estimating the number of users that connect to a single access points and the access point numbers
for a given coverage area depends on the factors that we have on the above sections. Different Wi-
Fi protocols have different performance and capability. An 802.11ac and 802.11n capable devices
can provide higher throughput compared with older 802.11a/b/g standards. The following table
shows a summarized maximum throughput statics that Wi-Fi enabled device can support.
Wi-Fi standard Channel Max Physical
bandwidth Rate (Mbps)
(MHz)
802.11b 20 11
802.11g 20 54
802.11n 20 54
802.11n (1x1:1) 20 72.2
802.11n (1x1:1) 40 150
802.11n (2x2:2) 40 300
Table 3 Wi-Fi capability [33]
Physical rate is the maximum throughput of raw data. Physical rate works in the data link layer of
TCP/IP model and different form application throughput. Application throughput decreases by
additional datalink headers. Other WLAN management traffic also decreases the total performance
of WLAN. In the design of WLAN, it is important to consider the application throughput not the
physical data rate. As an example 802.11g standards have 54Mbs physical data rate speed but all
the TCP/IP overhead decrease the overall throughput decreases to 20Mbps. Some protocols like
UDP have lower TCP/IP header so it provides higher throughput. Other encryption methods also
decreases the throughput. There is no precise calculation to know the exact throughput of WLAN
standards. It can vary based on the protocol the client uses. Further it can also vary because of
congestion and interference [33].
4.3.1 Estimate access point throughput

Identifying the maximum data rate of a given AP is important for determining how fast a client
devices can transmit and receive data. This will have a big impact on the amount of airtime utilized
37 | P a g e
to achieve the target application throughput level and overall network capacity planning. Since
WLAN protocol incorporates different network overhead, the theoretical maximum Wi-Fi data
rate does not represent the actual application throughput that a given user achieve. To estimate the
maximum amount of TCP/IP throughput that a client is capable of achieving, the amount of
network overhead must be determined either through live network testing under load or through
an educated assumption. It is common for Wi-Fi networks to have between 40-60% overhead [42].
The estimated access point throughput can be calculated as shown below [33]
𝑦 = 𝑥 - (𝑥 ∗ 0.4) = 0.6 𝑥 (3.1)
𝑛 = 𝑦 * 0.35 = (0.6 ∗0.35) 𝑥 = 0.21 𝑥 (3.2)
z = y – n = 0.6 𝑥 – 0.21 𝑥 = 0.39 𝑥 (3.3)
Where z is estimated access point throughput, y is TCP/IP throughput, 𝑥 is maximum PHY Rate,
n is loss from interference
As we have seen above TCP/IP header can have 40% overhead, which is subtracted from physical
data rate. Additionally the throughput in densely populated public environments can be reduced
due to network collision. It is very difficult to know the exact number but in densely environment
it is important to use a lower number with 35% of the throughput.
Currently there are different 802.11 standards but implementing 802.11n standards have benefit
from both capacity and widely integrated in to WLAN equipment’s. Although new standards like
802.11ac provide a very high bandwidth compared with all the previous WLAN standards, it is
difficult to consider it in the design at this time because of most smart phones doesn’t support it
except newest one.
Since in highly densely environment there will be more access points; co-channel interference
between access points will be high. The same kind of interference can be created because of many
number of WLAN users. Higher modulation techniques offer higher data rate and spectral
efficiency. But higher order modulations are more affected by noise and interference. And because
of that not all WLAN users can’t use higher modulations rates always. Hence in the design we will
assume a WLAN users will use average modulation rates. Therefore a single access point using
weighted average modulation rate will have the following throughput.
38 | P a g e
Using equation 3.1, the TCP/IP throughput become
y = 72.2 Mbps - (72.2Mbps* 0.40%) = 43.3Mbps
Using equation 3.2, the loss from interference is
n = y*0.35= 15.16 Mbps
Using equation 3.3, therefore the total access point throughput is
z = 43.3 – 15.16 = 28.12 Mbps
Dual-band access points with two radios can simultaneously support both 2.4 GHz and 5 GHz RF
bands. And this will increase the total throughput a given access point can deliver.
Estimating clients per AP capacity:
The maximum number of clients in a single AP can be calculated below [33]
k=l/n (3.4)
Where k is number of client per AP, l is access point aggregate throughput and n is minimum
bandwidth per client
Therefore the total number of APs required to cover a given area will be
s=t/k (3.5)
Where s is the total number of access point and t is total number of active users.
The total number of APs required can be affected by the distance that a client is placed, access
point coverage and But for the case of very highly densely areas it is very recommended to use
small coverage and use directional antenna when needed in order to boast the overall performance.
Access point location

As a good practice access points should be mounted close to clients as much as possible. The RF
signal strength is proportional to Inverse Square of distance. Hence when a client become far away
from AP the RF signal strength degrades very fast. A client which is 20m far away from AP receive
a signal ¼ (-6dB) of a client with 10m faraway [33].
39 | P a g e
In highly densely environment the APs coverage area should be small in order to increase the
capacity. Additionally this increases the performance and allows for the implementation of
narrower beam antennas that can increase signal gain. More APs also increases the receive signal
for clients since there are more APs closer to any client location. This has the benefit of better
SNR, which is required for high performance and capacity. In order to have higher signal gain
directional antennas are recommended in the public areas. This will help us to reduce the RF
interference also.
40 | P a g e
4.0 Implementation consideration
As we have seen on mobile data offloading section, Wi-Fi network can be integrated and work
with all cellular network technologies including LTE advanced. In Ethiopia currently 4G LTE
coverage and usage is not widened enough compared with 3G UMTS network. The 3G UMTS
network have good coverage on all over the country. Additionally almost all current smart phones
have built in 3G network integrated. Although ethio telecom started 4G LTE network service in
main city areas in Addis Ababa, Ethiopia; subscribers are mainly using 3G network for Internet
and data connection. Therefore the implementation part particularly focuses and works on the
integration of 3G and Wi-Fi.
4.1 Cellular network monitoring

There are two ways to monitor cellular network system resources and congestions. The first one
is proactively analyze network quality status, follow up performance thresholds and provide a
solution to secure subscriber services. In the proactive approach different types of critical resources
are monitored and when the resource usage become above a threshold level system congestion will
occur. And immediate capacity expansions required such as adding resource license and cell
redesign will take place [34]. Proactive approach method is very efficient way to handle issues
before the problem reaches to subscribers and it is good practice to do in a daily basis.
The second method is reactive approach where appropriate actions will be taken after the problem
or the issue happen. As an example the call block rate become higher than expected threshold
value and because of that subscribers service quality affected. Using this approach it is very
difficult to address the issue immediately and subscribers may be affected if actions are not taken
on time.
4.2 Resource of 3G network

The following resources are mainly monitored in ethio telecom 3G network [34]
 Received Total Wideband Power (RTWP): total uplink power received by the base station.
41 | P a g e
 Transmitting Carrier Power (TCP): total downlink transmitting power of a cell. This
measures the downlink load of a cell. The downlink capacity is limited by the base station
amplifier capability.
 Orthogonal Variable Spreading Factor (OVSF): codes used by WCDMA as channelization
codes. For one cell, only one OVSF code tree is available.
 Channel Element (CE): base station baseband processing resources. The CE is managed
and shared on a NodeB level. Normally, a newly launched network will start with a very
low CE configuration to save CAPEX. The CE would be the most likely resource that to
run out first.
 Iub transmission resource: In IP RAN, the RNC and NodeB can dynamically adapt to the
transmission network bandwidth. The Iub interface between the RNC and the NodeB will
not become a bottleneck if the transmission capability of the interface boards is sufficient.
Only the transmission network itself should be monitored from transmission network
performance results.
 SPU: The RNC has several types of hardware boards. The Signaling Processing Unit (SPU)
is the most likely bottleneck according to the configuration principles of the currently
designed RNC. The SPU performs all air-interface signaling and all transmission resource
management.
4.3 3G Resource usage metrics

Resource usages thresholds are defined in order to proactively monitor the resource status. In
resource metric it is important to focus on busy hour time ranges. Busy hours can be an hour the
resource usage is the highest of the day.
The main source of network congestion in 3G networks are TCP (transmission control power), CE
(channel element), UL (uplink power) and code ratio. So we focus on these parameters. For the
simulation purpose we choose most congested areas based on the above factors.
4.3.1 Uplink Load

In a CDMA network, the cell radio capacity is limited by the interference rise over the background
noise. In this manner, the total interference (or total power) can be used to measure the cell
42 | P a g e
capacity. In a WCDMA system, the total power is defined as RTWP by base station. The system
target load (maximum load considered normal) is 75%. The corresponding normal RTWP value
should be below -100 dBm [34].
4.3.2 Downlink Load

The total available transmit power from base station affects the downlink capacity of the 3G
network. When the downlink load become congested the following parameters will be affected
 The cell coverage will be limited.

 The data throughput will be limited.
 The call quality will be degraded.
 New calls will be blocked
The total downlink power also affected by user location and cell size. The larger the cell size or
the user far away it consumes more power. TCP is used to measure the downlink total transmitting
power in WCDMA. The mean power measurement is used as the load indication. If the mean TCP
is constantly higher than a threshold (such as 85%), it indicates that the cell downlink is overloaded
[34].
4.3.3 Channel Element resource usage

Channel Elements are one of the main resource in NodeB that to be properly planned and managed.
A CE (channel Element) is a baseband resource in a NodeB. Different services have different
requirement of CEs. One CE corresponds to the resource consumed by a 12.2 kbit/s voice call. If
a new call is arrives but there is no enough CE to entertain the required resource the call will be
dropped or blocked.
The total available CE resources are limited by both the installed hardware and the configured
software licenses. If the hardware resources in the current installation are sufficient and the CEs
are only limited by licenses, then the corrective action is to modify the license file to expand the
cell capacity [34].
Monitoring CE resource carefully is very important using threshold level. It is recommended that
if CE resource usage is above 70%, it means that the CE resource is congested and It is important
to take actions accordingly immediately.
43 | P a g e
4.3.4 OVSF code usage
Orthogonal Variable Spreading Factor (OVSF) are channelization codes used to separate data and
control channels from same UE. In WCDMA network each channels are uniquely identified by
codes. There are two types of codes used by WCDMA networks, these are scramble code and
OVSF code. In the uplink, each user is allocated and identified by a unique scrambling code. In
the downlink, the same scrambling code is used for users in the same cell, and users are identified
by OVSF codes allocated to them. Each user is assigned a unique OVSF code. A threshold (such
as 70%) can be defined for the DCH_OVSF_Utilization, to judge whether the system is
overloaded.
When the OVSF resource is overloaded, capacity expansion procedures such as cell splitting or
adding a new carrier are methods of corrective actions [34].
44 | P a g e
5.0 Simulation
5.1 Simulation Tool
Atoll simulation tools will be used to simulate both cellular and Wi-Fi network. It is known that
Atoll is a 64-bit multi-technology wireless network design and optimization platform that supports
wireless operators throughout the network lifecycle, from initial design to densification and
Optimization. Atoll 3.2 includes integrated single RAN – multiple RAT network design
capabilities for both 3GPP (GSM/UMTS/LTE) and 3GPP2 (CDMA/LTE) technology streams. It
provides operators and vendors with a powerful native 64-bit framework for designing and
optimizing current and future integrated multi-technology networks. Atoll 3.2 supports the latest
technology advances such as HetNets and Wi-Fi offloading [35].
5.2 Planning and design
5.2.1 Addis Ketema site Information

Currently most ethio telecom subscriber’s uses 3G network as a primary data usage compared with
4G network, hence the simulation part focuses on the integration of 3G and Wi-Fi network. As we
have seen above in the implementation consideration section, there are many reasons we say
cellular network become congested. Data’s from ethio telecom shows that there are many sites that
are congested because of CE and TCP shortages. The following section shows collected data from
ethio telecom about congested sites.
Ethio telecom gathered data
1. What are the main call/data rejection/congestion Reason
There are many reasons, but mainly uplink Received Total Wideband Power (RTWP),
downlink TCP, CE and scramble code are source of network congestion.
2. Congested Site
There are many sites that are congested because of CE and TCP shortages, from these site
we selected Addis-Ketema Merkato area where there a lot of business and trade transaction
45 | P a g e
takes places. Table 4 shows that 111053_Addis Ketema site has resource usage with CE %
level of 86.95% and TCP level of 98% which is above the threshold level.
Site name Congestion Status

111053_Addis Ketema CE % level: 86.95%
TCP % level: 98 %
Table 4 Addis Ketema Site information
3. Digital MAP information
Figure 16 shows 111053_Addis Ketema site digital map information.
Figure 16 111053_Addis Ketema site map
4. Area coverage of Cell radius

It depends based on coverage of morphology. But when designed on average in Addis
Ababa area 0.19KM is considered for cell radius.
5. Max number of users per Cell
For the case of Addis Ababa city, a 3G network designed to serve a maximum of 224
people per cell or sector.
6. Max number of people in the congested circumstance per cell/transmitter
46 | P a g e
For the case of 3G network it depends on the type of traffic profile of each user. It is
difficult to say this is the maximum number of users that makes the network congested. As
we have seen above there are different parameters that makes 3G network congestion.
Normally the admission threshold for HSUPA and HSDPA is 10/64 user respectively. Even
though we don’t have admission control threshold for DCH user, 35-40 user can be
considered as load margin for expansion.
7. Recommended rejected call% and data%

Out target is to maintain the rejection below 2-3 % for voice traffic and below 10% for data
traffic.
5.2.2 Deployment Steps

1. The first step is to import digital map of Addis Ababa where the resolution of the maps that we
use is 25 m. The map contains heights and altimetry and topographic relief of the Addis Ketema
area. Some of Digital map information’s are [37]
Clutter refers to a Land Use/Land Cover classification of surface features which impact on radio
wave propagation. These features are classed according to their physical and electrical properties.
Clutter altitude: Digital Terrain Model (DTM) is a continuous model of ground-level land
surface, represented by a digital raster grid with each grid cell holding an elevation value. A DTM
is a fundamental data input for radio propagation studies in that terrain blocks and reflects radio
waves.
Clutter height: the average building height on clutter.
2. Define Atoll Radio Parameters. Atoll radio parameters contains
Site information
Characterized by their X (longitude) and Y (Latitude) coordinates
Transmitter information
Transmitter
 Activity
47 | P a g e
 Antenna configuration (model, height, azimuth, mechanical/electrical tilts )
 UL/DL losses /UL noise figure
 Propagation (model , radius and resolution)
Cell Parameter
 Frequency band and channel
 Layer
 Physical cell ID
 Power definition of DL channels
 Min. RSRP
 DL and UL traffic loads
 Diversity support (MIMO)
 Neighbors
The following table shows radio parameters for the selected sites
1. Site parameter for 3G network
Name Longitude Latitude Altitude (m) Max No. Max UP Max No. Max DL
of UL throughpu of DL through
CEs t Kbps CEs put
Kbps
111053_Addi 38.74158 9.02763 32 256 12,288 256 12,288

333 8889
s Ketema
Table 5 1 Site parameter for 3G network
2. Transmitter parameter for 3G network

There are three transmitter with each 3 cell/ sectors. Costa-Hata propagation model will be used
both for 3G and WLAN network.
Site Transmit Freq Antenn Height(m Azimut Mechanical Receiver Main

ter uency a ) h(°) Downtilt (°) antenna Propagation
diversity Model
gain(dB)
111053_Ad 11105 Band 65deg 32 325 2 degree 18 Costa-Hata

dis Ketema 3_1 I 18dBi model
48 | P a g e
Site Transmit Freq Antenn Height(m Azimut Mechanical Receiver Main
ter uency a ) h(°) Downtilt (°) antenna Propagation
diversity Model
gain(dB)
2Tilt210
0MHz

dis Ketema 3_2 II 18dBi model
2Tilt210
0MHz

dis Ketema 3_3 III 18dBi model
2Tilt210
0MHz

2Tilt210
0MHz

2Tilt210
0MHz

2Tilt210
0MHz

2Tilt210
0MHz

2Tilt210
0MHz

2Tilt210
0MHz
Table 6 Transmitter parameter for 3G network
49 | P a g e
3. Cell parameters for 3G network
Cell ID Transmitter ID
36601 111053_1
36602 111053_2
36603 111053_3
36604 111053_4
36605 111053_5
36606 111053_6
36607 111053_7
36608 111053_8
36609 111053_9
Table 7 Cell parameters for 3G network
4. Site parameter for Wi-Fi network
In the design phase it is assumed that a cell or sector is expected to handle maximum of 220 users.
Hence for the case of 111053_Addis Ketema site, since we have a nine cells it is expected the site
serves a total of 1980 users. The thesis assumes that 20% of active users from a total 3G network
subscribers. Out of 20% active 3G network users for this simulation purpose, we have offloaded
40% of users to Wi-Fi network, so the total number of users offloaded to Wi-Fi network is 159.
Therefore we need to calculate the number of Wi-Fi access points needed to cover the required
users. As we have seen in the Wi-Fi design section, the maximum number of clients in a single AP
is
Number of client = AP aggregate throughput / Minimum bandwidth per client

= 16.9/0.5
= 33.8 users per AP
Total Number of Aps = Total number of active user/ Number of client per AP
= 159/33.8
= 5 APs
50 | P a g e
Therefore from Capacity perspective minimum 5 APs is required to cover the 159 users with
minimum 500kbps throughput traffic usage. The following table summarizes site parameter for
Wi-Fi network.
Name Longitude Latitude
Site0 "38°44'22.97""E" "9°1'42.82""N"
Site1 "38°44'21.66""E" "9°1'35.66""N"
Site2 "38°44'29.56""E" "9°1'33.25""N"
Site3 "38°44'34.6""E" "9°1'38.43""N"
Site4 "38°44'27.92""E" "9°1'40.22""N"
Table 8 Wi-Fi site parameter
5. Transmitter parameter for Wi-Fi network

For our simulation purpose we use Omni directionally antenna with Costa-Hata
propagation model.
Height Main Propagation

Site Transmitter Antenna (m) Model
2450MHz Costa-Hata model
Site0 Site0_1 Omni 5.8dBi 22
Table 9 Wi-Fi Transmitter parameter
51 | P a g e
5.2.3 Traffic parameters
A user equipment (UE) profile describes the terminal equipment type, traffic type, and mobility
characteristics which is defined in the UE Profile. UE profile has to be defined before starting the
simulation. The following section describes UE profiles used for our simulation.
Service
It is expected that in each site, users generate different 3G traffics. This research considers only
voice, Mobile Internet, Internet service and high-speed Internet service for traffic modeling and
uses Atoll default values for simulation purpose.
Service for 3G
Name Average throughput UL Average throughput DL
High Speed Internet 32 kbps 256 kbps

Mobile Internet 32 kbps 64kbps
voice 12.2 kbps 12,2 kbps
Table 10 3G user server profile
Service for Wi-Fi
Name Average throughput UL Average throughput DL
High Speed Internet 64 kbps 500 kbps

Internet service 64 kbps 128 kbps
Table 11 Wi-Fi user server profile
Mobility
The thesis mainly focuses on dense urban areas, hence it is not expected that users will travel at
high speed, therefore we considers only pedestrian and max 50 km/h speed.
52 | P a g e
Terminal
From atoll default terminal device types the paper considers HSPA and Mobile Terminal device,
and also it is expected that most of the traffic will be created by mobile terminal devices and high
throughput traffic will be created by HSPA terminal.
Name Min Power Max Power

HSPA terminal -50dBm 24dBm
Mobile terminal -50dBm 25dBm
Wi-Fi terminal -50dBm 25dBm
Table 12 Terminal profile
Environment
There are different environment models, but for this thesis we will considers only dense urban area
only. At the stage of the design it is expected that hexagonal cells, with 190 m cell radius will be
used for 3G network
User profile Mobile percent

user pedestrian 80
user 50Km/hour 20
Table 13 Environment profile
User profile by service
User profile for 3G
Service Terminal Call/hour Call UL DL

duration(sec) Volume(KB) Volume(KB)
Mobile HSPA 0.1 500 2,500

Internet
Access
53 | P a g e
Service Terminal Call/hour Call UL DL
duration(sec) Volume(KB) Volume(KB)
High HSPA 0.05 2,000

Speed 10,000
Internet
Voice Mobile 0.2 240
Table 14 User profile for 3G
User profile for Wi-Fi
UL DL
User Duration Volume Volume
Profile Service Terminal Calls/hour (sec.) (KBytes) (KBytes)
Business High Speed 802.11n

User Internet Terminal 0.01 2,000 10,000
Business Web 802.11n

User Browsing Terminal 0.1 700 4,500
Table 15 User profile for 3G
5.2.4 Propagation model

There are different propagation models to model wireless network. Costa-Hata model is best for
frequency range above 1500 MHz, Therefore Costa-Hata will be used for both 3G and Wi-Fi
network.
The default prediction resolution of atoll is 50m, so to make the result more accurate 20m
resolution will be used.
Monte-Carlo Simulation
54 | P a g e
We use atoll Monte Carlo simulator for both 3G and Wi-Fi simulations. Since the user distributions
of traffic map Atoll generates a population of users on the map and for each of these users the
simulator executes a power control algorithm for the uplink and downlink. The objective of the
algorithm is to minimize interference and maximize network capacity. This will restrict the
connection to the network users who use low-priority services and generate a lot of interference.
This process creates a snapshot of the UMTS network, the result is a distribution of users with
different network parameters: level of interference, the terminal state (connected, connection
refused ...), load factor for each cell, etc [37].
5.2.5 Deployment target

After we modeled the traffic of Addis Ababa city, it is possible to start the simulation to meet the
target quality objective. The simulation is expected to meet the following target level.
Service Service Probability of service rejection
Voice 2-3 %
Internet service 10%
High speed Internet service 10%
Table 16 deployment target
5.2.6 Simulation Steps

As we have seen above Atoll simulations are based on a Monte Carlo simulator and we use both
for UMTS and Wi-Fi network.
In UMTS each mobile station receives interference from base stations other than their own cells,
but not other phones, and all base station receives interference from their cell phones and other
cells, but not the other base stations [37].
The UMTS capacity depends on the total received interference. Atoll simulates the power control
mechanism using an iterative algorithm in each iteration, all the population of mobile users
generated try to be connected, one by one, to the network.
55 | P a g e
The simulation assumes that there will be a total of 220 people density with in a cell per a single
carrier. Since we have a total of 9 carrier, the total number of people is 1980. Additionally the
simulation considers that 20% of a total number population which is 396 people is actively
participate both in the data and voice connection.
The following table shows user distribution per service
Number of People Type of Service Percentage

40 High speed Internet 10%
119 Mobile Internet 30%
237 Voice 60%
Table 17 user distribution per service
56 | P a g e
6.0 Result
The following section shows simulation results for both 3G and Wi-Fi network for 111053_Addis
Ketema site. The result first shows 3G network coverage and capacity analysis before offloading
and then after Wi-Fi offloading.
6.1 3G Network before Wi-Fi offloading

In this section we simulate all users connecting to 3G UMTS network only. We see coverage signal
level, coverage area and its quality of service from users’ perspective for different services.
6.1.1 Study coverage signal level

Converge by signal level shows graphical representation of the signal level received by the
terminal or downlink coverage. Figure 17 shows coverage signal level for 3G network. The figure
shows that centrally we have good signal level coverage of >-80 dBm. Additionally over the entire
target area it is shown that a good signal coverage of >=90dBm. Since our mobile terminal typical
sensitivity is -105 dBm, the overall signal level coverage shown in the figure is good.
Figure 17 Study level of signal coverage
57 | P a g e
6.1.2 Study transmitter coverage
Transmitter coverage shows each transmitter coverage with in target area. Figure 18 shows that all
three transmitters’ coverage using different color and shows that signal coverage have good
coverage with in computational zone.
Figure 18 Study transmitter coverage
6.1.3 3G Capacity analysis

The following section shows 3G users’ connection for different services accepted by UE terminal.
The simulation is based on Monte Carlo simulator, where Using Monte Carlo simulator a user
connects randomly to one of 3G services. The UE is capable of accessing all 3G network for high
speed Internet, mobile Internet and voice traffic.
6.1.3.1 3G network capacity Before Wi-Fi offload

This section shows the actual 3G network capacity analysis simulation result for Addis Ketema
area before Wi-Fi network offload. Table 18 shows that out of 381 total user 175 or 45.9% become
rejected. The table shows that there are many reasons a user become rejected. But mainly channel
element (CE) and Downlink load saturation are the main reasons.
58 | P a g e
Total number of rejected users: 175 (45.9%)
Pmob > PmobMax: 16

Ptch > PtchMax: 0
Ec/Io < (Ec/Io) min: 5
UL Load Saturation: 0
Ch. Elts saturation: 111
DL Load Saturation: 41
OVSF Code Saturation: 2
Table 18 3G network capacity analysis
Further Figure 19 shows in detail users 3G service connection at the time of simulation where
trying to access 3G network for high speed Internet, mobile Internet and voice services.
Figure 19 3G Capacity Result before Offloading
6.1.3.2 3G user connection by Service
In this section we see 3G network simulation results by different services. Out of a total 381 users
there are 77 inactive users
Users: 381
59 | P a g e
Active: Downlink: 79 Uplink: 100 Downlink + Uplink: 125
Inactive: 77
Throughput: Downlink: 37.23 Mbps Uplink: 9.53 Mbps
Breakdown per service:
High Speed Internet:

Users: 119
Inactive: 6
Mobile Internet Access:
Users: 53
Inactive: 4
Voice:
Users: 209
Inactive: 67
Throughput: Downlink: 939.4 kbps Uplink: 1.15 Mbps
6.1.4 Result Analysis before Wi-Fi offloading

The simulation results are far from the target grade of service objectives which is 2-3% for voice
and 10% for Internet services. More users are rejected because of channel element (CE) shortage
and downlink (DL) power saturation which is the actual ethio telecom current problem for the case
of 3G network. As users request web browsing and high speed Internet service, which is higher
date rate services, more channels elements required and additionally it requires more downlink
power in order to cover the request services.
Therefore without Wi-Fi offloading, ethio telecom has to do the following measurements in order
to solve the above issue
60 | P a g e
1. For the shortage of Channel Element (CE):
a. Channel element is Software licensed capacity resource which is required for traffic
channel. So if ethio telecom has remaining CE license, it can be upgraded to meet
the requested traffic.
b. If the upgraded software license is not enough to cover the requested traffic
channel, it is required to add new carrier
2. For the DL power shortage: It is required to add more transmitters in order to distribute
more power for the requested traffic. This provide us additional benefit to improve the
signal quality and improve the carrier interference pilot channel (DL) threshold level (E c
/ I o <E c / I or min).
6.2 UMTS coverage After Wi-Fi offloading

The UMTS coverage after Wi-Fi offloading considers 42% of 3G network users offloaded to Wi-
Fi network, which is the remaining 222 user’s statistics. As we can see from figure 20, the total
number of rejected user is 2.7%, which meets the target value.
Total number of rejected users: 6 (2.7%)

Figure 20 shows the result in detail that the total number of rejected users is reduced to 2.7%. This
provides higher satisfaction to the mobile users and reduces ethio telecom cellular network burden.
61 | P a g e
Figure 20 3G network simulation after Wi-Fi offloading
6.3 UMTS and Wi-Fi network integration

In this section we see cumulative result of 3G and Wi-Fi network. In the simulation part we see
the whole coverage and capacity status together. Figure 21 shows traffic statistics that are
offloaded to Wi-Fi network only. Out of 159 users, 91.8 % of users become successfully connected
to Wi-Fi network and get the required services from Wi-Fi, which meets the design objective for
data traffic. The simulation uses IEEE 802.11n protocol which provides higher data rate service
compared with older IEEE 802.11a and IEEE 802.11b version. Other Wi-Fi protocols like
802.11ac provides more data rate compared with all other protocols. But since currently most smart
phones doesn’t support IEEE 802.11ac protocol we used IEEE 802.11n protocol for the simulation
purpose.
62 | P a g e
Figure 21 Cumulative Cellular and Wi-Fi network status
The following section shows Wi-Fi user connections by different services. All users are active and
can access web browsing and high speed Internet. And from a total of 159 users only 8.9% are
rejected, which meets deployment target.
6.3.1 Users Demand

Total number of users trying to connect
The following result shows summarized traffic demand for 159 users. All users have both
downlink and uplink demand with 49.74Mbps and 15.23Mbps throughput respectively.
Users: 159
Inactive: 0
DL: Max Throughput Demand (DL): 49.74 Mbps
Min Throughput Demand (DL): 0 bps
UL: Max Throughput Demand (UL): 15.23 Mbps
Min Throughput Demand (UL): 0 bps
63 | P a g e
Breakdown per service:
The following section shows users traffic demand per services.
High Speed Internet:

There are 79 high speed Internet users that demand both downlink and uplink data with 39.5Mbps
and 10.11Mbps throughput.
Users: 79
Inactive: 0
Web Browsing:
There are 80 high speed Internet users that demand both downlink and uplink data with 10.24Mbps
and 5.12Mbps throughput.
Users: 80
Inactive: 0
6.3.2 Result
The result shows that out of 159 users only 13 or 8.2% of users are not connected to Wi-Fi network
because of no signal coverage and no service on that area. The rest 146 users are effectively get
the requested uplink and downlink traffic request. The result meets designed objective for data
traffic.
64 | P a g e
Total number of users not connected (rejected): 13 (8.2%)
No Coverage: 1
No Service: 12
7.0 Conclusion and future works

The thesis work is mainly focuses on offloading cellular network in to complementary Wi-Fi
network to reduce cellular network congestion. As the main intention of the paper is to recommend
65 | P a g e
ethio telecom to look alternative way to solve the current cellular congestions, we see in the
simulation part that after offloading 42% of 3G network to Wi-Fi network, the number of rejected
users reduced from 45.9% to 2.3% for voice and 8.2% for data users. Additionally the simulation
result shows that, data traffic is the main source of cellular network congestion. And offloading
data traffic to Wi-Fi network benefits ethio telecom by reducing the burden of cellular network
and insures that subscribers are satisfied by both voice and data service.
Starting from 3GPP release 6 the integration of cellular and Wi-Fi network shows more
improvement and user equipment’s (UE) can connect both 3G and Wi-Fi network simultaneously
and can seamlessly transition data traffics with or without user’s interactions.
Wi-Fi is an inexpensive technology compared with other wireless technology which provides
higher data rate services. Since Wi-Fi technology is integrated to every smart phone and laptop
devices, it easy to integrate with cellular network and can use SIM based authentication and secure
the traffic using IPsec tunnel. Without Wi-Fi offloading ethio telecom has to plan additional
separate carrier to meet the subscriber’s request.
The integration of cellular and Wi-Fi network is new concept for ethio telecom and its subscribers.
Additional works have to done on the market interest, from legal perspectives and price
perspective.
66 | P a g e
Reference
1. Cisco visual networking index: Global Mobile Data Traffic Forecast Update, 2016-2021,
February 2017.
https://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-
vni/mobile-white-paper-c11-520862.pdf
2. Desktop vs Mobile vs Tablet Market Share Ethiopia, January, 2017.
http://gs.statcounter.com/platform-market-share/desktop-mobile-tablet/Ethiopia
3. Md. Alimul Haque, Pritam Kumar, Amrendra Kumar Singh, “IEEE 802.11ac: 5th
Generation WiFi Networking” World Applied Programming, Vol (2), No (4), April 2012.
4. Janice Reynolds “Going Wi-Fi-A Practical Guide to Planning and Building an 802.11
Network”, 2003
5. Cisco Systems, Inc. Building Cisco Multilayer Switched Networks, Volume 2 Version 3.0,
2006
6. Albert K. Ansah q, IAENG, T.Kwantwi, and Agangiba W.Akotam “Comparing Wireless N
(IEEE 802.11n) and Wireless G (IEEE 802.11g) Standards in terms of Performance and
Reliability”, July 2011
7. Sourangsu Banerji, Rahul Singha Chowdhyry “On IEEE 802.11: Wireless LAN Technology”
2013
8. Research in Motion Limited, “Considerations for Planning and Deploying a Wireless LAN,
Technical White Paper”, 295 Phillip Street Waterloo, ON N2L 3W8 Canada, 2010
9. ETSI Mobile Competence Centre “Overview of 3GPP Release 6 Summary of all Release 6
Features Version TSG #33”, 2006
10. 3GPP TR 22.934 V6.2.0 “Feasibility study on 3GPP system to Wireless Local Area
Network (WLAN) interworking (Release 6)”, Sep. 2009.
11. Tatiana Issaeva, 3G – WLAN Systems Interworking Architecture, February 2017.
http://www.tcs.hut.fi/Studies/T 79.5401/2005AUT/ 3G_WLAN_ Interworking .pdf
12. 3GPP TS 33.234 V6.9.0, “3G Security; Wireless Local Area Network (WLAN) interworking
security (Release 6)”Mar. 2007
67 | P a g e
13. A.S.Rumale, Dr.D.N.Chaudhari, IEEE 802.11x, and WEP, EAP,WPA/WPA2,
International Journal of Computer Technology and Appli-cations,pp. 2229-6093 Vol.2(6)
Nov-Dec 2011.
14. IETF RFC 3748 – Extensible Authentication Protocol (EAP), 2004
15. Rajavelsamy, Manoj Choudhary and Debabrata Das “A Review on Evolution of 3GPP Systems
Interworking with WLAN” 19 March 2016.
16. 3GPP TS 23.402 V8.10.0 “Architecture enhancements for non-3GPP accesses (Release 8)”
Mar 2012
17. 3GPP TS 22.278, “Service requirements for the Evolved Service requirements for the Evolved
Packet System (EPS) (Release 8)” 2009.
18. A. Schumacher / J. Schlienz “WLAN Traffic Offload in LTE”, White Paper Nov. 2012
19. Warda Ahmed, Sidra Anwar and M. Junaid Arshad “Security Architecture of 3GPP LTE and LTE-
A Network: A Review”
20. IETF RFC 5213: Network Working Group, “Proxy Mobile IPv6”, August 2008.
21. ETSI MCC, “Overview of 3GPP Release 9 V0.3.” 2014
22. ETSI MCC, “Overview of 3GPP Release 10 V0.2” Jun 2014
23. 3GPP TS 23.402 “3rd Generation Partnership Project; Technical Specification Group Services and
System Aspects; Architecture enhancements for non-3GPP accesses (Release 10)” 2011
24. 3GPP TS 23.139 “3GPP System-Fixed Broadband Access Network Interworking stage 2 (Release
11)
25. QUALCOMM Incorporated, “A 3G/LTE Wi-Fi Offload Framework: Connectivity Engine (CnE)
to Manage Inter-System Radio Connections and Applications” Jun. 2011.
26. ETSI MCC, “Overview of 3GPP Release 12 V0.2” 2015.
27. Mobile Data Offload- Wi-Fi Offload – TEC, February 2017.
http://tec.gov.in/pdf/Studypaper/Mobile%20data%20offload.pdf
28. Nadeem Akhtar, “LTE-WLAN Interworking: 3GPP Release-12 Perspective” IEEE standards
workshop on last mile Connectivity”, Bangalore, India, July. 2013.
29. ETSI MCC, “Release 13 analytical view version “, Sept. 2015.
30. Richard Burbidge, 3GPP RAN2 WG chair “LTE-WLAN Aggregation (LWA) and LTE WLAN
Radio Level Integration with IPsec Tunnel (LWIP)”, IEEE meeting in Macao, March 13-18 2016.
31. Intel Corporation, “Alternative LTE Solutions in Unlicensed Spectrum: Overview of LWA, LTE-
LAA and beyond”, whitepaper, 2016
68 | P a g e
32. 5G America, “Mobile Broadband Evolution towards 5G: Rel-12 & Rel-13 and Beyond” Feb. 2017.
33. Ruckus Wireless, Inc “Deploying Very High Density Wi--‐Fi DESIGN AND CONFIGURATION
GUIDE FOR STADIUMS”, Best Practices v1.0, 2012.
34. HUAWEI TECHNOLOGIES CO., LTD, “UMTS RAN13.0 Capacity Monitoring Guideline”,
Draft A, Feb. 2011.
35. Forsk, “Atoll wireless network engineering group version 3.2 brochure”, February 2013
36. Forsk, “LTE Radio Network Design Tool – Atoll”
37. https://www.slideshare.net/mobile/kaboudei/rf-planning-umts-with-atoll1, February, 2017)
38. Sapna Shukla, Varsha Khare, Shubhanshi Garg, Paramanand Sharma, “Comparative Study of
1G, 2G, 3G and 4G” April 2013
39. Jyotsna Agrawal, Rakesh Patel, Dr. P.Mor, Dr. P.Dubey and Dr. J.M.keller, “Evolution of
Mobile Communication Network: from 1G to 4G” Nov 2015
40. Martyn Mallick, “Mobile and Wireless Design Essentials” 2003
41. Rysavy Research/4G Americas “Mobile Broadband Explosion”, August 2013
42. Aerohive Networks “High-Density Wi-Fi Design Principles” 2012
43. Mobile Broadband VPN service, March, 2017
http://www.ethionet.et/node/580396?q=mobilebroadbandvpn
44. Ramia Babiker Mohammed Abdelrahman, Amin Babiker A. Mustafa, Ashraf A. Osman“A
Comparison between IEEE 802.11a, b, g and ac Standards” 2015
45. Rebecchi, F., De Amorim, M. D., Conan, V., Passarella, A., Bruno, R., & Conti, M. (2015).
Data offloading techniques in cellular networks: A survey. IEEE Communications Surveys
and Tutorials, 17(2), 580–603.
46. Beneyam H., Haile B., Edward Mutafungwa and Warma, “Development of the Wi-Fi
Offloading Business Concept within the African Market Context”, 2012.
47. R. Rajavelsamy, M. Choudhary, and D. Das, “A Review on Evolution of 3GPP Systems
Interworking with WLAN,” Journal of ICT Standardization, vol. 3, no. 2, pp. 133–156, 2015.
48. A. Aijaz, N. Uddin, O. Holland, and A. H. Aghvami, “On Practical Aspects of Mobile Data
Offloading to Wi-Fi Networks,” Ithaca, NY, USA, 2015.
69 | P a g e

Wudneh Alebign

Uploaded by

Copyright:

Available Formats

Wudneh Alebign

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wudneh Alebign

Uploaded by

Copyright:

Available Formats

ADDIS ABABA UNIVERSITY

ADDIS ABABA INSTITUTE OF TECHNOLOGY

Mobile Data Offloading using Wi-Fi for the case of

A Thesis Submitted to the School of Graduate Studies of Addis Ababa

By: - Wudneh Alebign

Approved by Board of Examiners

Chairman, School of Graduate Studies Signature

Dr. Yihenew_Wondie ________________

Internal Examiner Signature

Internal Examiner Signature

Wudneh Alebign __________________

Place: Addis Ababa

Dr. Yihenew Wendie __________________

Keywords: Mobile Data Offloading, Wi-Fi, Wi-Fi Offloading

2.0 Background ............................................................................................................... 7

3.0 WLAN design.......................................................................................................... 34

4.0 Implementation consideration ............................................................................ 41

6.0 Result ........................................................................................................................ 57

7.0 Conclusion and future works .............................................................................. 65

1.2 Statement of the problem

1.3.2 Specific objective

1.4 Literature Review

1.5 Scope and Limitation

1.8 Thesis Layout

First Generation (1G) wireless mobile network

Second Generation (2G) wireless mobile network

Third Generation (3G) wireless mobile network

3G network is the third generation wireless mobile telecommunication technology introduced in

Fourth Generation (4G) LTE wireless mobile network

Fifth Generation (5G) wireless mobile network

2.2.1 Describing WLAN Topologies

Figure 1 Ad hoc mode WLAN connectivity [5]

Figure 3 Extended Services Set WLAN connectivity [5]

2.2.2 Wi-Fi Standards

Figure 4 IEEE 802.11b frequency band allocation [5].

 The discovery of suitable networks through the advertisement of access network

2.2.3 Service Set Identifier

2.2.4 Roaming technology in WLAN

 The maximum data retry count is exceeded.

2.2.5 WLAN Security

2.3 Mobile data offloading

2.3.1 3GPP Release 6

3GPP Visited Network

3GPP Home Network

Figure 5 Scenario 2 3GPP-WLAN interworking architecture [11]

Figure 6 show how end to end EAP authentication works.

Figure 6 End to end EAP authentication [11]

Protocols used for Authentication

Figure 7 Authorization and authentication using diameter. [11]

Packet Data SLF

Figure 9 3GPP-WLAN interconnection architecture [12].

1. 3GPP system operator.

Interworking Architecture [10]

1. Non-roaming reference model

2.3.2 3GPP Release 7

2.3.3 3GPP Release 8

3GPP IP Access or S14

Figure 10 ANDSF architecture [16]

IP Mobility between 3GPP-WLAN

Network Based Mobility: