Last Submission

Federal Technical and Vocational Education and Training Institute
Faculty of Electrical Electronics Department of Information and
Communication Technology Management
Enhancing end-to- end Quality of Service of Border Gateway Protocol Multiprotocol
Label Switching Virtual Private Network of EthioTelecom Service Level Agreements
By
Anwar Kade
Advisor
Dr. Sreenivas Nune
Thesis Submitted to the Faculty of Electrical Electronics in Partial Fulfillment of the
Requirements for the Degree of Master of Science in Information and Communication
Technology Management
October 2020
Addis Ababa, Ethiopia
Enhancing end- to- end Quality of Service of Border Gateway Protocol
Multiprotocol Label Switching Virtual Private Network of EthioTelecom
Service Level Agreements
By
Anwar Kade
This thesis has been approved in partial fulfillment of the requirements for the Degree
of Master of Science in Information and Communication Technology Management
Faculty of Electrical Electronics Department of Information and Communication
Technology Management
Thesis Examination Committee:
Full Name Signature Date
October 2020
Addis Ababa, Ethiopia
DECLARATION
Word Template by Friedman & Morgan 2014
2
I, the undersigned, declare that this thesis work is my original work, has not been presented
for a degree in this or any other universities, and all sources of materials used for the thesis
work have been duly acknowledged.
Anwar Kade Hassen
Full Name of Student
____________________
Signature
October 2020
Addis Ababa
Ethiopia
This thesis has been submitted for examination with my approval as an advisor.
Dr. Sreenivas Nune
Full Name of Advisor
__________________
Signature
October 2020
Addis Ababa Ethiopia

Word Template by Friedman & Morgan 2014
3
Enhancing end-to- end Quality of Service of Border Gateway Protocol Multiprotocol Label Switching Virtual
Private Network of EthioTelecom Service Level Agreements
Acknowledgment
First and foremost, my sincere gratitude goes to Almighty ALLAH for giving me the
enablement and capability physically, mentally and spiritually for the completion of this
thesis work.
I would like also to express my sincere thanks to my advisor Dr. Sreenivas Nune for his
unreserved encouragement, excellent advice, and continuous support during the work of this
thesis. Without his comment, information, guidance, and suggestion of several breakthroughs
in this thesis would be impossible to be achieved.
I also want to thank EthioTelecom IP Quality of Service Management Section staffs for their
provision of necessary data to prepare this thesis. At last, I would also like to take this
opportunity to acknowledge Mr. Nuredin Mohammed for this support and encouragement. I
am especially indebted to his support.
1
Table of Contents
Acknowledgment................................................................................................................3
Table of Contents...............................................................................................................4
List of Acronyms................................................................................................................7
List of Tables....................................................................................................................10
List of Figures..................................................................................................................11
Abstract............................................................................................................................14
Chapter One.....................................................................................................................15
1 Introduction.................................................................................................................15
1.1 Background..........................................................................................................15
1.2 Statement of the Problem....................................................................................20
1.3 Objectives.............................................................................................................22
1.3.1 General Objective....................................................................................22
1.3.2 Specific Objectives...................................................................................22
1.4 Methodology.........................................................................................................22
1.4.1 General Approach and specific research method.................................22
1.4.2 Data type and source...............................................................................23
1.4.3 Sample size and sampling Techniques...................................................23
1.4.4 Instruments and procedures, Process, and Analysis............................24
1.4.5 Design and Evolution Procedures..........................................................25
1.5 The significance of the study...............................................................................25
1.6 Contributions.......................................................................................................25
1.7 Scope and Limitations.........................................................................................26
1.7.1 The scope of the Study.............................................................................26
1.7.2 Limitations of the Study..........................................................................26
1.8 Organization of the Thesis..................................................................................27
2
Chapter Two.....................................................................................................................28
2 Review of Literature and Related Works.................................................................28
2.1 Review of Literature............................................................................................28
2.1.1 Border Gateway Protocol (BGP)............................................................29
2.1.2 Multi-Protocol Label Switching.............................................................30
2.1.3 Multiprotocol Label Switching and Virtual Private Network.............33
2.1.4 MP BGP MPLS VPN...............................................................................35
2.1.5 Quality of Service (QoS)..........................................................................36
2.1.6 QoS Models...............................................................................................42
2.1.7 DiffServ QoS Implementation over MPLS VPN..................................44
2.1.8 Congestion Management Mechanisms..................................................48
2.1.9 Congestion Avoidance Mechanisms.......................................................50
2.2 Related Works......................................................................................................51
Chapter Three..................................................................................................................54
3 Proposed Network Architectures..............................................................................54
3.1 Introduction..........................................................................................................54
3.2 Designed BGP MPLS VPN.................................................................................55
3.2.1 Network IP Address................................................................................55
3.2.2 Interfaces..................................................................................................57
3.2.3 Interior Gateway Protocol (IGP) Interconnection...............................57
3.2.4 MPLS and MP BGP Interconnection....................................................58
3.2.5 Configuring MPLS RSVP-TE................................................................58
3.2.6 Configure VPN instances on PEs...........................................................60
3.2.7 Creating EBGP peer relationship between the PE and CE routers. . .60
3.3 Designed QoS of Proposed network architectures...........................................61
3.3.1 Define Access Control List rules............................................................62
3.3.2 Define traffic classifiers...........................................................................63
3
3.3.3 Define traffic behavior............................................................................64
3.3.4 Define and apply traffic policies.............................................................65
3.4 Applying the predefined policies to the inbound interfaces routers...............65
3.5 Experimental Results of Proposed Architecture..............................................66
3.5.1 IGP protocol.............................................................................................67
3.5.2 Signaling protocol RSVP-TE..................................................................70
3.5.3 MPLS TE Tunnel.....................................................................................70
3.5.4 MPLS Operation......................................................................................72
3.5.5 BGP Protocol............................................................................................74
3.5.6 Performance of established L3VPN Service..........................................75
3.5.7 Quality of Service of proposed network architectures.........................76
3.6 Discussions............................................................................................................77
Chapter Four....................................................................................................................86
4 Conclusions and Future Works.................................................................................86
4.1 Conclusions...........................................................................................................86
4.2 Future Works.......................................................................................................88
Reference..........................................................................................................................90
Appendices........................................................................................................................94
4
1 List of Acronyms
BECN Backward Explicit Congestion Notification

BGP Border Gateway Protocol
BoS Bottom of the Stack
BRAS Broadband Remote Access Server
CAR Committed Access Rate
CBQ Class-based Queueing
CDMA Code Division Multiple Access
CE Customer Edge
CEF Cisco Express Forwarding
CoS Class of Service
CPE Custom Premises Equipment
CPU Central Processing Unit
CQ Custom Queueing
DE Discard Eligible
DiffServ Differentiated Services
DSCP Differentiated service code point

ADSL Asymmetric Digital Subscriber Line
DSLAM Digital Subscriber Line Multiplexer
eBGP Exterior Border Gateway Protocol
EF Expedited Forwarding
EPON Ethernet Passive Optical Network
ER Edge Router
FECN Forward Explicit Congestion Notification
FIFO First in Frist Out
GPON Gigabit Passive Optical Network
GSM Global System for Mobile System
GTSM Generalized TTL Security Mechanism
iBGP Interior Border Gateway Protocol
IEEE Institute of Electrical and Electronics Engineers
IETF Internet Engineering Task Force
IGP Interior Gateway Protocol
IntServ Integrated Services
IP Internet Protocol
IS-IS Intermediate System to Intermediate System
ITU International Telecommunication Union
L3VPN Layer 3 Virtual Private Network
LAN Local Area Network
LFB Label Information Base
LFIB Label Forwarding Information Base
LSP Label Switched Path
LSR Label Switch Router
MP Multi-Protocol
MPLS Multiprotocol Label Switching
MSAG Multiple Service Access Gateway
MSAN Multiple Service Access Node
NGN New Generation Network
5
OSPF Open Shortest Path First

P Provider
PE Provider Edge
PHB Per-Hop-Behavior
PQ Priority Queuing
OPNET Optimized Network Engineering Tools
QoS Quality of Service
RD Route Distinguisher
RED Random Early Detection
RFC Request for Comment
RSVP Resource Reservation Protocol
RT Route Target
SLA Service Level Agreement
SP Service Provider
TE Traffic Engineering
TP Traffic policing
TS Traffic shaping
TTL Time to Live
UDP User Datagram Protocol
VDSL Very high bit rate Digital Subscriber Line
VPN Virtual Private Network
VRF Virtual Routing and Forwarding
ToS Type of Services
WFQ Weight Fair Queueing
WRED Weighted Random Early Detection
6
List of Tables
Table 1-1 Total BGP MPS VPN EthioTelecom SLA customers and Sample size [24]..........24
Table 2-1 Y.1541 (2011) recommended QoS targets [34].......................................................40
Table 2-2 EthioTelecom recommended QoS targets [20]........................................................41
Table 3-1 the connection between routers and IP address assignment....................................55
Table 3-2 Network Routers Loopback IP Addresses...............................................................56
Table 3-3 Values and parameters of the traffic model.............................................................77
Table 3-4 Simulation configuration.........................................................................................77
Table 3-5 5 the similarities and differences between existing and proposed network
architecture...............................................................................................................................78
Table 3-6 Exist and proposed network architecture numerical QoS results............................85
7
List of Figures
Figure 1:1 BGP MPLS VPN components and working principles [3]....................................18
Figure 1:2 QoS differentiated service model [3]......................................................................19
Figure 1:3 Existing BGP MPLS VPN network architecture....................................................21
Figure 2:1 MPLS label position in different technologies [30]...............................................30
Figure 2:2 MPLS network [31]................................................................................................31
Figure 2:3 User perception of end-to-end QoS delivery framework [25]................................39
Figure 2:4 QoS viewpoints framework [9]..............................................................................39
Figure 2:5 Traffic classification [36].......................................................................................45
Figure 2:6 Token Bucket implementation [26]........................................................................48
Figure 2:7 Implementation of a FIFO scheduling [28]............................................................49
Figure 2:8 Implementation of a WFQ scheduling [29]............................................................49
Figure 3:1 Simplified Proposed BGP MPLS VPN network architecture with end-to-end QoS.
..................................................................................................................................................54
Figure 3:2 OSPF route information.........................................................................................67
Figure 3:3 Ospf neighbor relation............................................................................................68
Figure 3:4 Ospf link state data base.........................................................................................68
Figure 3:5 interface ospf enabled.............................................................................................69
Figure 3:6 OSPF brief information..........................................................................................69
Figure 3:7 RSVP-TE detail information..................................................................................70
Figure 3:8 MPLS TE tunnel information.................................................................................71
8
Figure 3:9 MPLS TE tunnel status...........................................................................................71
Figure 3:10 MPLS TE tunnels constraint shortest path database............................................72
Figure 3:11 MPLS routing information...................................................................................72
Figure 3:12 MPLS link state protocol information..................................................................73
Figure 3:13 MPLS adjacency information...............................................................................73
Figure 3:14 BGP neighbor relationship...................................................................................74
Figure 3:15 VPN instances routing information......................................................................75
Figure 3:16 VPN instances reachability to access router.........................................................75
Figure 3:17 VPN instances reachability to aggregation router................................................76
Figure 3:18 user-defined QoS..................................................................................................76
Figure 3:19 Defined QoS.........................................................................................................77
Figure 3:20 FTP response time with light load (Blue - MPLS, Red - MPLS /DiffServ).........80
Figure 3:21 FTP traffic response time with heavy Load (Blue - DiffServ; Red- MPLS).......80
Figure 3:22 End-to-end traffic delay Voice with light load (Blue MPLS; Red / DiffServ).....81
Figure 3:23 Delayed end-to-end traffic Voice with heavy load (Blue - MPLS / DiffServ; Red-
MPLS)......................................................................................................................................81
Figure 3:24 Variation of delay (jitter) of traffic Voice with light load (Blue - MPLS; Red -
MPLS / DiffServ).....................................................................................................................82
Figure 3:25 Variation of the Delay Delay (Jitter) of Voice heavy load traffic (Blue - MPLS,
Red-MPLS / DiffServ).............................................................................................................82
Figure 3:26 Video End-to-End Delay with Light Load (Blue - MPLS; Red - MPLS /
DiffServ)..................................................................................................................................83
9
Figure 3:27 Video End-to-End Delays with heavy Load (Blue - MPLS / DiffServ; Red-
MPLS)......................................................................................................................................83
Figure 3:28 Variation of the delay time (jitter) of video traffic with light load (Blue - MPLS;
Red - MPLS / DiffServ)...........................................................................................................84
Figure 3:29 Variation of delay of video traffic with heavy Load (Blue - MPLS / DiffServ;
Red - MPLS)............................................................................................................................84
10
Abstract
The primary goals of Quality of Service are bandwidth management, controlled jitter,
latency, and improved packet loss characteristics to provide satisfactory services for users.
Shaping network optimization is crucial for the service provider. To implement the network
QoS, optimizing the current network physical and logical architectures is among the best
practice.
The approach used in this work was to design and build an operator network type backbone
to simulate a real scenario that conveys different types of traffic. The results of the work are
presented according to the simulation time and the network load. The results of the
comparison demonstrate the advantage over the performance of MPLS networks with
DiffServ compared to MPLS networks with the best- effort model.
In this research work, an attempt has been made to investigate the end-to-end QoS parameters
of the Ethio-Telecom service level agreement (SLA) customer's network by using the
differentiated service (DiffServ) model, to manage end-to-end traffic delay, jitter, and packet
loss. The traffic is classified and marked depending on their priorities assigned. The proposed
network architecture has used weighted fair queuing (WFQ) for congestion management and
weighted random early detection (WRED) for the congestion avoidance method.
The eNSP and Wireshark have used been to design, demonstrate, and evaluate the existing
and proposed network architectures. Once the results of the existing network are compared
with the proposed network architecture that is designed using the DiffServ model; delay,
jitter, and packet loss have decreased whereas the traffic bandwidth utilization increased.
Keywords: Traffic engineering, CISCO, IP SLAs, MPLS, IP networking, GNS3,
network simulator.
11
Chapter One
2 Introduction
2.1 Background
The last decade has witnessed a major change in the types of traffic scaling with the
development of real-time applications which result in several challenges within traditional IP
networks. Some of these challenges are better end-to-end bandwidth, delay, jitter, and packet
loss, etc. To combat these challenges, researchers have steered towards finding alternate
solutions. One of the alternate solutions found were to use Multiprotocol Label Switching
(MPLS) in the network[1]. MPLS architectures can be deployed on existing service provider
backbones reducing infrastructure costs and increase network performance. MPLS can also
be used to interconnect geographically diverse sites while at the same time, reduce the delay
found in traditional IP networks.
Now a day because of a different kind of traffic such as voice, video, and data are sent over
the same network infrastructure quality of service is a big issue for enterprise. Multiprotocol
label switching (MPLS) virtual private networks (VPN) are the new alternatives to private
wide area networks (WAN) services[2] [3]. Due to the effectiveness of MPLS VPN
enterprise customers are moving to service providers that offer MPLS VPNs. The reason for
this shifting is the capability of MPLS VPN to provide built-in security features and end-to-
end connectivity. QoS is the most important factor for enterprise networks. Using the
Multiprotocol border gateway protocol (MP BGP) MPLS VPNs assures the quality of
services for these enterprises. To guarantee the quality of service for different types of traffic,
the differential service (DiffServ) QoS model can be used with MP BGP MPLS VPN[4] [5].
It provides a good performance service in terms of better end-to-end bandwidth, delay, jitter,
and packet loss.
12
In the case of EthioTelecom SLA customers, MP BGP MPLS VPNs services subscription the
customers have a service level agreement (SLA) with the company on end-to-end QoS to
guarantee the quality and performance of service provided. To achieve the intended Qos the
Company is also working on it by setting SLA targets. But still, there is a gap between the
company's SLA targets and what SLA enterprise customers have been getting[6] [7].
Quality of Service (QoS) is a set of specific requirements provided by a network provider to
users, which are necessary to achieve the required functionality of service[8] [9]. The users
specify their performance requirements in form of Quality of Service parameters such as
delay or packet loss and the network commits its bandwidth making use of different QoS
schemes to satisfy the request. Each service model has its own QoS parameters. Its
parameters and measures are necessary to indicate how well a service is and therefore, is an
important point when selecting services offered by different service providers. If service
features price are similar, quality becomes the differentiator for users, as well as, service
providers can make use of quality to have an image of a “respected” provider[8] [10].
VPN is a network connection between devices that do not share a physical cable. Virtual
Private Networks (VPNs) are a method of interconnecting multiple sites belonging to a
customer using a Service Provider backbone network in place of dedicated leased lines. Each
customer site is directly connected to the SP backbone[3]. The SP can offer a VPN service
more economically than if dedicated private WANs are built by each customer because the
SP can share the same backbone network resources (bandwidth, redundant links) between
many customers. The customer also gains by outsourcing the complex task of planning,
provisioning, and managing a geographically distributed network to the SP.
There are two different methods to construct VPNs across the IP backbone that is custom
premises equipment (CPE) based and network-based[11]. Most current VPN implementations
are based on CPE equipment. VPN capabilities are being integrated into a wide variety of
13
CPE devices, ranging from firewalls to WAN edge routers. On the other hand, there is a
significant interest in network-based VPNs where the operation of the VPN is outsourced to
service providers[12] [2].
Multiprotocol label switching (MPLS) is a versatile solution to address the problems faced by
present-day networks—speed, scalability, quality-of-service (QoS) management, and traffic
engineering. MPLS has emerged as an elegant solution to meet the bandwidth-management
and service requirements for next-generation Internet protocol (IP)–based backbone
networks. MPLS addresses issues related to scalability and routing (based on QoS and service
quality metrics[1] [12].
BGP is a standardized exterior gateway protocol (EGP), as opposed to RIP, OSPF, and
EIGRP which are interior gateway protocols (IGP’s) and considered as “Path Vector” routing
protocol. BGP is not built to route within an Autonomous System (AS), but rather to route
between AS's. BGP maintains a separate routing table based on the shortest AS Path and
various other attributes, as opposed to IGP metrics like distance or cost[13] [14].
BGP MPLS VPN is a layer 3 Provider-provisioned VPN where Enterprises have sites spread
across distant locations that need to be interconnected[15] [5]. Instead of having fully
dedicated links between their sites, many enterprises prefer to contract a Virtual Private
Network (VPN) service from a VPN service provider, thereby reducing the connection costs.
This service model is known as the provider-provisioned VPN service. In this model, the
VPN provider shares its physical network infrastructure among multiple enterprises,
guaranteeing isolation of virtual networks. This technology uses BGP as a control plane to
provide VPN routing and MPLS as a transport technique to achieve isolation between
customer traffic. Its popularity as a result of the high number of customers supported. The
key core network elements of a provider-provisioned BGP/MPLS VPN network are provider
edge (PE) and provider core (P) routers as shown in Fig 1.1[3], whereas the customer edge
14
(CE) router is not considered part of the provider’s core network. It acts as a peer of the PE
router, but not a peer to other CE routers.
Figure 1:1 BGP MPLS VPN components and working principles [3].
The routers that link the customer sites to the provider network are called customer edge (CE)
routers, whereas the service provider routers to which the CE routers are connected are called
provider edge (PE) routers. In most cases, the provider network is made up of more than just
the PE routers; those other routers are called P devices [10].PE routers take the charge of both
accessing VPN service and forwarding packets from private intranet to the public internet,
whereas P routers only have basic forwarding and typically are not have directly connecting
customer access circuits[16] [17]. All PE and P routers run label switching so that they can
build MPLS label switched paths (LSPs) from each PE to each other PE. This is achieved
through the use of the label distribution protocol (LDP) in conjunction with the interior
gateway protocol, such as open shortest path first (OSPF)[5] [2].
When a PE forwards a VPN-addressed packet across the core it adds two MPLS labels, one
external which identifies the PE in the provider backbone, and the other internal which
identifies the interface inside the PE. Any intermediate P or PE routers switch the packet to
the egress PE using the outer label only[5] [3]. The inner label is used by the egress PE to
determine the VPN port to which the packet should be forwarded. Each PE router supports
multiple routing and forwarding tables called virtual route forwarding (VRF) tables. Every
site to which the PE is attached must be mapped to one of those forwarding tables. When a
15
packet is received from a particular site, the forwarding table associated with that site is
consulted to determine how to route the packet outstands of VPN sites.
BGP MPLS VPN service models are provided for user services to ensure QoS according to
the user’s requirements and the quality of the network[15] [18]. The common service models
are as follows:
 Best Effort service model
 Integrated service model and
 Differentiated service model.
The Best Effort is a traditional Internet model, without any QoS guarantees. The IP networks
just route packets until they reach the destination. It is realized through the FIFO mechanism.
The integrated service model is the first standardized mechanism by IETF, which is based on
resource reservation in routers on the path by using signaling that means, does not send
packets until the arrival of confirmation that the network has reserved resources for it. It is an
end-to-end QoS mechanism[19]. The Differentiated service model is the most commonly
used method for traffic differentiation on the Internet, in which all packets are classified into
a limited number of classes, So routers have to store only information per class (not per
connection, i.e., per-flow). It is a hop-by-hop QoS mechanism.
Figure 1:2 QoS differentiated service model [3].
16
The main goal of improving QoS is to guarantee end-to-end service delivery. QoS includes
end-user perception, service provider perception, and network performance issues.
Optimization of the network by using the different queuing algorithm is the best suit to
increase the network QoS. Increasing network performance increases end-user perception and
service provider perception[2] [18].
2.2 Statement of the Problem
EthioTelecom is one of the enterprises that offer BGP MPLS VPN in Ethiopia for its SLA
customers. These services are widely used in IP MPLS networks for connecting customers’
remote VPN sites. According to the literature review done on the company's QoS level in[8]
[20] [21], service provider faced many challenges such as low bandwidth, high jitter, high
packet drops, and high packet delay which degrade the quality of service and overall network
performances to offer these services. In addition to this, as in the questionnaires conducted
from the company's BGP MPLS VPN SLA customers, 41 % of their connection has QoS
problems
EthioTelecom has done continuous optimization on MPLS VPN SLA customer’s network, to
provide QoS for its customers. Moreover, the company did continuous expansion projects in
its existing networks[21] [3]. For example, it has recently expanded the existing network of
IP backhaul, multiple service access gateways (MSAG), and multiple service access nodes
(MSAN) plantation projects for broadband VPN and Internet customers, and still an end-to-
end QoS is impossible for the company to guarantee.
There is a detailed analysis done by its IP QoS management team[2] [22] [20], on QoS of
BGP MPLS VPN of SLA customers. The analysis result shows that there are some gaps
between the company’s SLA targets and analysis results which what customers are getting.
This might be caused by improper network optimization and customer LAN side problems.
17
Moreover, there are several VPN complaints from end-users across the country upon poor
bandwidth, high packet drops, high packet delay, and high jitter as suggested by the team in a
fixed access network, operation, and maintenance department of the company[22] [20]. It has
happened because there were QoS problems. The cause of these problems is EthioTelecom
hasn’t considered the optimization of their network architecture logically as shown in
Fig.1.3.In the existing network architecture of the company, there are three types of routers,
two P, four PE, and two CE routers. These routers have their functionalities in the
optimization of BGP MPLS VPN traffic flow and network performance.
Figure 1:3 Existing BGP MPLS VPN network architecture
EthioTelecom Currently treaties all customers which are SLA, major and residential network
equally which lead to QoS and overall network performance problems. These problems arise
because of the company using the Best Effort QoS model, FIFO for congestion management,
and tail drop for congestion avoidance. Because of the above-mentioned gap, the research is
motivated to address the above problems by optimizing the network logically using a
differentiated Service QoS model.
18
2.3 Objectives
2.3.1 General Objective
The general objective of this research study is to identify EthioTelecom BGP MPLS VPN of
SLA customer’s QoS complaints, analyze the result to the company's SLA targets and ITU
threshold values, design solution to improve the service, and evaluate the proposed solution.
2.3.2 Specific Objectives
The specific objectives of the research are summarized as follows:
 Propose the solution to improve QoS of BGP MPLS VPN of the company's SLA
customer's problems.
 Develop the artifact (prototype) for the proposed solution.
 Design, demonstrate, and evaluate the solution to improve the QoS of BGP MPLS
VPN of SLA customers.
2.4 Methodology
For the sec of effectiveness, this section of the research thesis work includes the general
approach and specific method, type of data and source, sample size and sampling techniques,
instruments and procedures for data collection, process and analysis, methods of data analysis
and evaluation approaches are stated hereunder.
2.4.1 General Approach and specific research method
Design science was chosen, as a research approach, because it was particularly suitable to
conduct “practical” research by combining Information System research with the process of
design and development[23]. Design science is a research method where the researcher
interacts with the participant through questionnaires to gather data regarding the current
status and conditions of the services. Moreover, literature review and related work have been
19
made to decide the state of art in the area. Then the design, development, demonstration, and
evaluation of improved QoS of MPLS VPN of EthioTelecom SLA customers have been
done.
The questionnaire approach helps to collect precise information concerning the status of QoS
of MPLS VPN of EthioTelecom SLA customers. In addition to that, it helps to demonstrate
the link between QoS of MPLS VPN and quality of service parameter (bandwidth, delay,
jitter, and packet loss ratio). A literature review is the best suit to compare and contrast the
QoS of EthioTelecom that has been promised in its SLA and what the customers are getting.
It also helps to cross-check with ITU QoS threshold values.
In general, a literature review is the best method to identify and analyze the existing
conditions of QoS of BGP MPLS VPN of SLA customers, compare the existing conditions
with SLA and ITU target values shortly and briefly. Because of this, the researcher was
attracted to use these research methods.
2.4.2 Data type and source
The primary and secondary data sources were used for the study. Concerning the primary
source, data was collected randomly from seventeen EthioTelecom BGP MPLS VPN of SLA
customers out of one hundred seven as a sample from different parts of the country.
Additionally, secondary data was gathered from applicable documents such as SLA charter,
QoS guide, network element configuration guide, achieved configuration, and empirical
studies through literature review.
2.4.3 Sample size and sampling Techniques
The total sample size involved in this study was seventeen EthioTelecom BGP MPLS VPN
of SLA customers from different parts of the country as shown in Table 1.1. The researcher
takes a sample of 16 % SLA customers from those that use different types of networks using
a random sampling technique.

20
Table 1-1 Total BGP MPS VPN EthioTelecom SLA customers and Sample size [24].
Types of Network Total Customers Sample Size Selected
ADSL/VDSL 3 2
EPON 31 4
GPON 23 4
DIRECT FIBER 29 4
AIRONET 17 3
TOTAL 107 17
2.4.4 Instruments and procedures, Process, and Analysis
The data for this study were collected from both primary and secondary sources. As a
primary source, a questionnaire was used. The questionnaire is a close-ended type. Before
distributing the questionnaire, to determine the quality and reliability of the questionnaire, the
researcher distributed the questionnaire to four participants who were not included in the
actual study to check if there is any unclear idea or statement. As a result, based on the
feedback obtained some questions were rephrased or rewritten which lacked clarity. The
secondary data was conducted from archived database management of EthioTelecom SLA
charter, QoS guide, network element configuration guide, and CPE configuration guide
through literature review.
The collected data through close-ended questionnaires were analyzed and compared against
QoS of EthioTelecom SLA targets and ITU threshold values then by taking the analyzed data
as input, the researcher has modeled the traffic flows which is traffic flow architecture by
applying different network optimization modeling and network simulation tool.
2.4.5 Design and Evolution Procedures
Taking the gap of analyzed data as input, the researcher has modeled the QoS of
EthioTelecom BGP MPLS VPN of SLA customers. The researcher also used the DiffServ
model by giving priority to the class of services at network and data link layers. The
21
researcher used weighted fair queueing and weighted random early detection algorithms for
congestion management and congestion avoidance respectively. The traffic flows were
studied using computer network modeling and simulations tools such as Wire shark and
eNSP. The simulation process covered end-to-end QoS delivery. Bandwidth and other QoS
parameters latency, jitter, and packet delay have been reviewed to the QoS of EthioTelecom’s
SLA targets and ITU QoS threshold values.
2.5 The significance of the study
This study has shown the analysis of QoS of BGP MPLS VPN of EthioTelecom SLA
customer’s status and the numerical result has compared with the SLA targets and ITU
standard threshold values. Taking the numerical results as input the study has proposed a way
of improving QoS of BGP MPLS VPN of the company's SLA customer’s network. For the
proposed solution practical modeling and demonstration of the EthioTelecom high speed and
optimized network usage have been conducted. Moreover, the study has designed,
demonstrated, and evaluated the solution to improve the QoS of BGP MPLS VPN of SLA
customer’s network.
2.6 Contributions
QoS of BGP MPLS VPN is one of the areas that need very strict follow up in the
telecommunication sector. This is because every SLA customer needs uninterrupted services
to support their day to day activities. This, in turn, demands end-to-end network traffic
optimization and attention must be given to improving QoS of BGP MPLS VPN. To exploit
the maximum possible capacity of the network and understand its usage after installing the
network, there should be continuous and organized traffic optimization on end-to-end
networks.
22
This research work contributed to improving the QoS of BGP MPLS VPN of EthioTelecom’s
SLA customer’s network connection. This is done by traffic classification, marking, shaping,
and policing using different KPIs by the Differentiated Service DiffServ QoS model. The
proposed solution has been designed, developed, demonstrated, and evaluated using
computer-aided tools.
2.7 Scope and Limitations

2.7.1 The scope of the Study
This study has evaluated the existing quality of services of BGP MPLS VPN of seventeen
EthioTelecom SLA customer’s connection moreover compare the level of existing quality of
services to the company's SLA targets and ITU standard threshold values. After evaluation
and comparison, the study has taken the gap in existing QoS as input and developed a logical
architecture design to improve QoS. The proposed solution has been designed, demonstrated,
and evaluated using computer-aided tools in a detailed manner. However, the overall process
of improving the QoS of the existing infrastructure is done by using traffic management and
queuing algorithms.
2.7.2 Limitations of the Study
Implementing end-to-end QoS of BGP MPLS VPN could be done by optimizing both the
current network physical and logical architectures. But this study has not focused on the
physical design architectures to improve the existing QoS of BGP MPLS VPN of SLA
customers. Moreover, the simulation has used the developed sample models and randomly
selected variables for the packet arrival rate. But on the real traffic, there is an additional
configuration such as for bidirectional forwarding, time synchronization, and fast route
recovery. These variables are generated after close inspection of the real network scenario.
23
2.8 Organization of the Thesis
This thesis paper contains four chapters. Chapter one deals with the introduction of the whole
thesis. It mentions to the statement of the problem, objectives, methodology, thesis
contribution, scopes, and limitations. Chapter two presents the MPLS, BGP, VPN, and QoS
model used. It also showed some light on what other authors and researchers have forward
their ideas on the area of improving QoS of MP BGP MPLS VPNs were presented.
The proposed network architecture was presented in chapter three. This includes how BGP,
MPLS, VPN, and QoS were designed, demonstrated, and evaluated with the DiffServ model.
The experimental results and discussions were also presented. Finally, the chapter four
concluded the paper by presenting the conclusions and future recommendations.
24
Chapter Two
3 Review of Literature and Related Works
3.1 Review of Literature

Several abstracts of research papers, articles, journals, and student thesis were studied to get a
rough review of the study. This review leads to a better understanding of the thesis contents.
This section of the study reviews former contributions on improving QoS of BGP MPLS
VPN using different approaches and different models such as best effort, integrated service,
differential service models. The goal of this section is to understand the problem and
emphasize the research gap related to the study. Moreover, a brief review has been made on
BGP, MPLS, VPN, QoS threshold and model, and traffic shaping (policing), and congestion
management.
QoS technologies refer to the set of tools and techniques to manage network resources and
are considered as the key enabling technology for network convergence. The objective of
QoS technologies is to make voice, video, and data convergence appear transparent to end-
users [25] [26]. QoS technologies allow different types of traffic to contend inequitably for
network resources. Voice, video, and critical data applications may be granted priority or
preferential services from network devices so that the quality of these strategic applications
does not degrade to the point of being unusable. Therefore, QoS is a critical, intrinsic element
for successful network quality and performance [8].
QoS of any service is acceptable when they fulfill SLA and leads to proper customer
satisfaction [7]. To guarantee the full throughput, a specific level of assurance is required
over the traffic load to reduce losses, jitters, and delays. QoS has several tools which include
[8] [27]
 Classification and Marking tools
 Policing and Markdown tools

25
 Scheduling tools.
3.1.1 Border Gateway Protocol (BGP)
Border Gateway Protocol (BGP) is the prevalent wide-area routing protocol, the internet
composed of Autonomous System (AS's) that use BGP to implement inter-AS and intra-AS
IP routing based on a set of attributes (Weight, Local preference, Multi-exit discriminator,
Origin, AS path, Next-hop, Community). Routing involves two basic tasks, the first task is
the determination of optimal routing paths, which is the complex, and the second is the
transport of information groups (Packets) through an internetwork, here the researcher uses
BGP to address the task of path determination [13] [15].
BGP design motivated by three important goals, first one is scalability through dividing
internet to AS under independent administration, the second is Policy in which AS can
implement and enforce various forms of routing policies, and the last one is cooperation
under competitive circumstances in which the structure allow AS to determine among any set
of choices [13].
AS which is owned and administered by a single commercial entity, implements a set of
policies in deciding how to route its packets to the rest of the Internet, and how to export its
routes to other AS, and identified by a unique 16-bit number (the new is a 32-bit number).
Inside AS operates different routing protocols (Interior Gateway Protocols – IGPs) which
includes (RIP, OSPF, IS-IS, E-IGRP), and in contrast inter-domain protocols like BGP are
called Exterior Gateway Protocols (EGP) [28].
The function of the BGP System is to exchange reachability information including
information about a list of AS paths with other BGP systems. From this information it
constructs AS connectivity graph so loops pruned and AS policy decisions enforced, each
BGP router maintains a routing table that lists all feasible paths to a particular network,
routing in [15] formation received from peer routers is retained until an incremental update is
26
received, BGP routers exchange routing information upon initial data exchange and after
incremental updates, when an update occurs routers send the portion of their routing table
that has changed [13] [3].
3.1.2 Multi-Protocol Label Switching
MPLS is an Internet Engineering Task Force specified framework that provides efficient,
routing, and switching of traffic flow through the network. It is a technology for the delivery
of IP services. It gives the ability to offer highly scalable, advanced IP services end-to-end
with simpler configuration and management for both service providers and customers.
Enterprises and corporations use these technologies to create their service simple and price
economical. Instead of the IP address or MAC address, MPLS works on small labels. These
labels are inserted between layer 2 and layer 3 of OSI. Forwarding decisions are based on
these labels instead of having to look at complex IP tables [1] [17] [12].
MPLS is either layer 2 or layer 3 technology. It is referred to as layer 2.5. It gives the speed
of layer 2 and the dynamics of layer 3 technologies. The interesting factor is its flexibility of
use as completely layer 2 or layer 3 technologies. It can also be used as a combination of both
layers. In MPLS VPN it is solely used as layer 3 technology making labels based on IP
addresses. In the case of VPLS, it makes labels based on MAC addresses, making it a layer 2
technology. Thus, the adaptability of MPLS makes it charismatic for professionals. The
MPLS label position in different types of technologies can be seen in Figure 2.1 [29].
Figure 2:4 MPLS label position in different technologies [30].
MPLS labels consist of four parts:
27
 Label value: 20 bits
 Experimental: 3 bits
 Bottom of the stack: 1 bit
 Time to live: 8 bits
3.1.2.1 How an MPLS network works
In an MPLS network, each packet gets labeled on entry into the service provider's network by
the ingress router, also known as the label edge router (LER). This is also the router that
decides the LSP the packet will take until it reaches its destination address. All the
subsequent label-switching routers (LSRs) perform packet forwarding based only on those
MPLS labels. Finally, the egress router removes the labels and forwards the original IP
packet toward its final destination. When an LSR receives a packet, it performs one or more
of the following actions [4] [9]:
Push: Adds a label. This is typically performed by the ingress router.
Swap: Replaces a label. This is usually performed by LSRs between the ingress and egress
routers.
Pop: Removes a label. This is most often done by the egress router.
This diagram illustrates how a simple MPLS network works:
Figure 2:5 MPLS network [31].
28
3.1.2.2 MPLS Architecture
The MPLS architecture describes the mechanisms to perform label switching, which
combines the benefits of packet forwarding based on Layer 2 switching with the benefits of
Layer 3 routing [17]. MPLS assigns labels to packets for transport across packet- or cell-
based networks. The forwarding mechanism throughout the network is label swapping, in
which units of data (for example, a packet or a cell) carry a short, fixed-length label that tells
switching nodes along the path of the packet how to process and forward the data [32].
3.1.2.3 Control Plane
Control Plane collects the information that is used to forward the incoming packets; While
Forwarding Plane decides how to switch the incoming packets after being received at the
inbound interface. Control planes perform information exchange between neighboring
devices by the use of different protocols namely routing protocols (e.g., RIP, EIGRP, OSPF,
and BGP) and label exchange information protocols (e.g., LDP, TDP, RSVP, etc.) [16] [17].
3.1.2.4 Data Plane
Data Plane has a forwarding plane that is based on the information attached to labels. There
are two types of tables, namely LIB and LFIB. Label Forwarding Information Base (LFIB) is
used by the data plane to forward the labeled packets. The Local Information Base (LIB)
table contains all the local labels and the mapping of the labels which is received from the
adjacent routers. The information in LFIB and label value is used by the MPLS-enabled
routers to make forwarding decisions [3] [17].
29
3.1.2.5 MPLS Label Distribution
So that LSPs can be used, the forwarding tables at each LSR must be populated with the
mappings from {incoming interface, label value} to {outgoing interface, label value}. This
process is called LSP setup, or Label Distribution [32].
The packet in the MPLS network that must pass through the network is forwarded over the
label switch path (LSP) tunnel. When the packet reaches the MPLS network then the Ingress
router receives the packet and puts the MPLS label in the packet and sends it to the next hop
according to the destination address in the packet. There can be many LSRs between Ingress
and Egress routers, so when the packet reaches an LSR it swaps the labels and sends it to the
next LSR. When the packet reaches the egress router, it strips off all the labels and sends it to
the outgo. All the LSRs have interior gateway routing (IGP) [17] [12].
3.1.3 Multiprotocol Label Switching and Virtual Private Network
MPLS VPN is a network that connects private networks over the public network.
Multiprotocol Label Switching (MPLS) Virtual Private Network (VPN) technology has
enabled the service provider network and enterprise network to agree on common terms for
the provision of end-to-end service levels [5] [16]. Service providers use VPNs to
interconnect different sites that belong to the same corporation. At the IP layer, VPN models
might require that different VPNs are required to connect and also provide connectivity to the
internet. Moreover, the MPLS VPN provides this functionality [5].
 L2VPN and
 L3VPN
3.1.3.1 MPLS VPN Architecture
There are some basic building blocks for the MPLS VPN at Provider edge routers. These are
given below [1].

30
 Virtual Routing Forwarding (VRF)
 Route Targets (RT) and
 Route Distinguisher (RD).
3.1.3.2 Virtual Routing Forwarding
The combination of the VPN IP routing table and the associated VPN IP forwarding table is
called the VPN routing and forwarding instance (VRF) [5]. VRF is used to make the MPLS
VPN networks private. The VRF makes sure that the routing information is kept separate
from different customers and that the backbone of the MPLS network makes sure that the
packet forwarding is based on label information and not on the information in the IP header.
On PE routers each VPN has its separate routing table and this routing table is called the VRF
routing table. A PE router interface that is towards the CE router only has one VRF, so that
all IP packets coming to that interface will be considered as they are belonging to that VRF. It
is because there is a separate routing table per VPN. An interface can only assign to one
VRF, but several interfaces can be assigned to the same VRF [17] [4].
3.1.3.3 Route Distinguisher
Multiprotocol BGP (MP-BGP) is used by the VPN to propagate its prefixes over the MPLS
VPN networks. The IPv4 prefixes carried by BGP across the service providers' network
should be unique. If there is overlapping in the customers' IP addresses, there will be a
problem in routing. To overcome this problem, the route distinguisher concept was developed
to make IPv4 prefixes unique. The idea is that a unique identifier is received from each
customer with each prefix to differentiate the same prefix from other customers. The IPv4
prefix and RD combine to give the vpnv4 prefix. VPNv4 prefixes are carried by MP-BGP
between the PE routers [15] [2].
31
The route distinguisher (RD) is a 64-bit field that makes the VRF prefixes unique [31]. This
64-bit value can be in two formats: ASN: xx or IP-address: xx where xx is a number, and
ASN stands for autonomous system number. ASN: xx is the popular format used by most
service providers. Internet Assigned Numbers Authority (IANA) assigns ASN to the service
providers, and xx is the number uniquely assigned to VRF by the service provider. RD
combines with IPv4 prefixes to form a VPNv4 prefix that is 96 bit long [17].
3.1.3.4 Route Target
A Route Target is the feature of MPLS VPN which controls the communication between
different VPN sites. The Route Target (RT) was introduced to overcome the drawbacks of the
route distinguisher (RD) since the RD can only communicate with one VPN; whereas RT can
communicate with complex VPN topologies [30].
Route Target (RT) is the BGP extended member which indicates the route that should be
imported from MPLS BGP into VRF. The RT attached with the vpnv4 route is called the
exported route and configured under the IP vrf command separately for each virtual routing
table on the PE router. The vpnv4 route received from MPLS BGP is examined for a
matching extended community which is the route-target. This procedure is called importing
an RT. If the result matches, the prefix is inserted into the VRF routing table as the vpnv4
route. Otherwise, the prefix is rejected [13] [17].
3.1.4 MP BGP MPLS VPN
MP-BGP (Multi-Protocol) BGP is an extension to BGP which allows BGP to advertise
different types of addresses besides IPv4 unicast [15]. MP-BGP supports IPv4 and IPv6
unicast, IPv4, and IPv6 multicast and also VPN labels that are used in MPLS-VPN. MP-BGP
is required only within the service provider's backbone. Therefore, all MP-BGP sessions are
internal BGP sessions, internal because the session is between two routers that belong to the
same autonomous system [13].

32
3.1.5 Quality of Service (QoS)
Quality of service (QoS) is a generic term capability of providing improved services to
specific network traffic using different technologies like ATM, SONET, and MPLS, etc. The
main purpose of QoS is to prioritize specific traffic over another i.e. to take into consideration
Jitter, Latency, Packet Loss, and Burst of Jitter and Loss and minimize all these factors for
that flow specifically. It should also be considered that prioritizing one traffic flow must not
make another fail. QoS of any service is acceptable when they fulfill SLA and leads to proper
customer satisfaction [9] [25].
It provides a different level of treatment to the different types of traffic or applications that
flow over the network. QoS is required to provide the good management of network
resources that makes the sophisticated usage of resources and gives comfort to the network
user. There are mainly four measurement units for QoS [8]:
1. Available Bandwidth
2. Latency
3. Jitter or Variable Delay
4. Packet Loss
3.1.5.1 Bandwidth
Network bandwidth is the capacity of a wired or wireless network communications link to
transmit the maximum amount of data from one point to another over a computer network or
internet connection in a given amount of time. On the network, IP Packets travel through the
best route [1].
The bandwidth requirements are met for those applications which need it. If the network
contains low bandwidth and voice traffic is transported over it then users will experience
broken sentences. Similarly, when the video traffic is transported over the low bandwidth
33
network then the picture will start sticking and the enjoyment of the video is adversely
affected. Therefore, proper network planning and other factors need to be considered while
attempt to achieve end-to-end QoS. For example [17] [31]
Increase link bandwidth: - This is effective but costly.
Classify and mark traffic and apply to the queue: - Forward important packet first.
Use Compression technique: - Layer 2 payload compression, TCP header compression, and
compressed RTP (cRTP) are some examples. Usage of hardware compression is preferable
over software-based compression because compressions are CPU intensive and create a
delay.
3.1.5.2 Latency
Latency is a time that a packet takes in traveling from a sender node to a receiver node.
Sometimes, it is also considered a time a packet takes to make a round trip from the source to
the destination and back to the source. Various parameters, such as propagation time,
transmission media, and processing time at each network node, affect the delay. The voice
and video data are very sensitive to latency. Little more delay in transporting voice and video
data harms the Quality of Service to the application. End-to-end delay is the sum of all the
following delays [16].
 Processing delay
 Queuing delay
 Serialization delay and
 Propagation delay.
3.1.5.3 Jitter
Jitter is a delay between two packets at the receiving end. When there is a heavy load in the
network, the data must be buffered and queued in any given network node. As a result, the
34
amount of delay between two packets is inconsistent (variable). This variable delay is called
Jitter. The voice traffic is very sensitive to Jitter because the inconsistent delay will cause a
breaking voice [17].
3.1.5.4 Packet loss
In a data network, packets may get lost or dropped due to so several reasons. One common
reason is higher network utilization or congestion. In this case, the drop in voice or video
packets creates unrecognizable sentences to the listener. Therefore, it is very critical to
maintaining lower packet loss [4].
Packet loss occurs due to the low buffer space. When the buffers space of the interface full
then packets are dropped. In queue scheduling, packet loss will occur if the queue is full.
Packet loss creates extended delays and jitter. Packet loss can be controlled by applying some
techniques such as tail drop, random early detection, weighted random early detection, and
traffic shaping and policing [3].
3.1.5.5 QoS viewpoints
QoS is technically dependent on the four-pillar which are bandwidth, delay, packet loss, and
jitter. It also depends on the end-user perception of telecommunication services such as
trends, advertising, tariffs, and costs which are interrelated to the customer expectation of the
QoS [5].
35
Figure 2:6 User perception of end-to-end QoS delivery framework [25].
QoS can be divided into two viewpoints Customer viewpoints and Service provider
viewpoints. Customer viewpoints include QoS requirements and perception whereas service
provider viewpoints include QoS offered and QoS achieved as shown in detail in the below
framework.
Figure 2:7 QoS viewpoints framework [9].
Network QoS is not well defined by itself. It is best described by including network
performance and quality of experience. For example, quality of experience impacts QoS and
network performance even though end-user subjective [25].
Generally, the network performance and quality of experience are very much interrelated. If
network performance was well optimized, the service provider viewpoint reaches a high
level. If a Service provider affords quality services to its customer, the customer viewpoint
was reached at a high level which increases the quality of experience [33] [26].
3.1.5.6 ITU-T Y.1541 (2011) Recommended QoS Targets
This Recommendation defines classes of network Quality of Service (QoS) and specifies
provisional objectives for Internet Protocol network performance parameters [34]. These
classes are intended to be the basis for agreements among network providers and between
end-users and their network providers. By considering a range of applications involving the
media such as, voice, video, image, and data the parameters that govern end-user satisfaction
for these applications and a broad classification of end-user QoS categories are determined. It
36
is intended that these categories are used as the basis for deriving realistic QoS classes and
associated QoS control mechanisms for the underlying networks.
A typical user is not concerned with how a particular service is implemented. However, the
user is interested in comparing the same service offered by different providers in terms of
universal, user-oriented performance parameters. This implies that performance should be
expressed by parameters.
Table 2-2 Y.1541 (2011) recommended QoS targets [34]

Key performance parameters and target
values
Delay variation
Information loss
The typical amount of data
One-way delay (Note)

Application
Medium
Preferred < 2 s /page
Acceptable < 4 s
Data Web-browsing– HTML ~10 KB /page N.A. Zero
Bulk data 10 KB -10 Preferred < 15 s Acceptable <
Data transfer/retrieval MB 60 s N.A. Zero

Transaction Preferred < 2 s Acceptable < 4
Data services – high priority < 10 KB s N.A. Zero

Preferred < 2 s Acceptable < 4
Data E-mail (server access) < 10 KB s N.A. Zero

Low priority
Data transactions 10 KB < 30 s N.A. Zero
37
3.1.5.7 Recommended IP QoS in EthioTelecom network
QoS is a configuration that prioritizes data traffic based on a traffic type or destination. So
that in the event of congestion on a network, a site's critical traffic has higher priority over
other traffic [22] [20] [35]. Currently, in the EthioTelecom network, all packets from all
customers are treated equitably; thereby generalized IP network performance targets are set
(recommended) as shown in the table below.
Table 2-3 EthioTelecom recommended QoS targets [20]

QoS Across End to end (CPE to Internet connection as measured
parameters backbone (ER CPE across backbone) from the connected BRAS or ER
to ER) (or from speedtest.net)
Latency 50ms or less 200ms or less 150ms or less
Jitter 15ms or less 50ms or less N.A.
Packet loss 0.1% or less 2% or less 1% or less
Availability 99.9% or more 90% or more 90% or more
Throughput N.A. 75% or more of 75% or more of subscribed BW
subscribed BW
3.1.6 QoS Models
QoS Models Describes a set of end-to-end QoS capabilities. End-to-end QoS is the ability of
the network to deliver service required by specific network traffic from one end of the
network to another. To provide end-to-end QoS, an overall network deployment is required.
Service models are used to provide an end-to-end QoS guarantee based on specific
requirements. the Three QoS models are [26] [25]:
 Best-Effort Model
 Integrated Services (IntServ) Model and
 Differentiated Services (DiffServ) model.

38
3.1.6.1 Best-Effort Model
The best effort is a single service model in which an application sends data whenever it must,
in any quantity, and without requesting permission or first informing the network [7]. For
best-effort service, the network delivers data if it can, without any assurance of reliability,
delay bounds, or throughput. Best-effort service is suitable for a wide range of networked
applications such as general file transfers or e-mail. It can be concluded by the various
analysis of performance evaluation that best-effort service is not best suited for the
application which needs a specific level of quality of service.
3.1.6.2 Integrated Service Model (IntServ)
Integrated service is a multiple service model that can accommodate multiple QoS
requirements [10]. In this model, the application requests a specific kind of service from the
network before it sends data. The request is made by explicit signaling; the application
informs the network of its traffic profile and requests a particular kind of service that can
encompass its bandwidth and delay requirements. The application is expected to send data
only after it gets a confirmation from the network. It is also expected to send data that lies
within its described traffic profile [27].
3.1.6.3 Differentiated Service Model (DiffServ)
Differentiated service is a multiple service model that can satisfy differing QoS requirements.
However, unlike in the integrated service model, an application using differentiated service
does not explicitly signal the router before sending data [30].
For differentiated service, the network tries to deliver a particular kind of service based on the
QoS specified by each packet. This specification can occur in different ways. Using the IP
Precedence bit settings in IP packets or source and destination addresses. The network uses
39
the QoS specification to classify, mark, shape, and police traffic and to perform intelligent
queuing [5].
The differentiated service model is used for several mission-critical applications and for
providing end-to-end QoS. Typically, this service model is appropriate for aggregate flows
because it performs a relatively coarse level of traffic classification.
The differentiated service model QoS includes the following features that support [4] [30]:
Committed access rate (CAR), which performs metering and policing of traffic, providing
bandwidth management.
Intelligent queuing schemes such as WRED and WFQ and their equivalent features on the
Versatile Interface Processor (VIP), which are distributed WRED (DWRED) and distributed
WFQ. These features can be used with CAR to deliver differentiated services.
Differentiated Service Model comes after the IntServ QoS model. It overcomes the limitation
of the IntServ model. DiffServ is also called the “Soft QoS” model. IntServ model guarantees
for the end-to-end resource reservation before the application takes the start. It uses the RSVP
for signaling and end-to-end resource reservation.
3.1.7 DiffServ QoS Implementation over MPLS VPN
The DiffServ model is aimed at supporting service differentiation for aggregated traffic in a
scalable manner and to achieve service quality in the MPLS VPN environment. DiffServ QoS
model is widely used in industry due to its scalability [33]. DiffServ model consists of four
components Traffic classification, marking, congestion management, and congestion
avoidance. These were used to control network traffic, resource allocation in different ways,
and allow the system to provide differentiated services [8] [9].
40
3.1.7.1 Traffic Classification
Classification tools sort packets into different traffic types, to which different policies can be
applied. Classification can be done at every node in the network or be implemented at the
edge of the network when the packet enters the network [16]. Classification of packets can
happen without marking the packets. Classification inspects one or more fields in the packet
to identify the type of traffic that the packet is carrying. After the identification process, the
traffic is handed to the treatment application such as marking, remarking, queuing, policing,
shaping, or a combination of these. Classification can be done by [16]:
 Incoming interface
 IP precedence
 Differentiated service code point (DSCP)
 Source or destination IP address
QoS classification is implemented by marking the type of service (TOS) field in the IP packet
precedence as shown in fig.2.7. IP data stream can be classified based on the different RFC
standards. RFC 791 defines the IP precedence field to divide the IP application into 8
categories. The TOS field is divided into 16 categories. RFC 2472 redefines TOS to divide
services into 64 categories (DSCP) [10] [19].
Figure 2:8 Traffic classification [36].
RFC 1349 defines bits in the TOS; bits 0 to 2 refer to precedence. The value ranges from 0 to
7. The larger the value is the higher the precedence. The D bit refers to the delay, T bit refers
41
to the throughput, R bit refers to the reliability, and C bit refers to the monetary cost. Bits 6
and 7 are reserved [36].
3.1.7.2 Traffic Marking
Marking writes a field within the packet, frame, cell, or label to preserve the classification
decision that was reached during the classification process. Marking is also known as
coloring the packet, which involves marking each packet as a member of a network class so
all devices throughout the rest of the network can quickly recognize the packet class. The
marking process sets bits in the DSCP or IP Precedence field of each IP packet according to
the class that the packet is in. Packets that are marked as a high priority, such as a voice
packet, will generally never be dropped by congestion avoidance mechanisms. On the other
hand, if packets are marked as a low priority they will be dropped when congestion occurs
[25] [9].
3.1.7.3 Per-Hop Behavior (PHB)
In Diffserv terminology, the forwarding behavior that is assigned to a DSCP is called the per-
hop behavior (PHB) [3]. The PHB defines the forwarding precedence that a marked packet
receives with other traffic on the Diffserv-aware system. This precedence ultimately
determines whether the IPQoS-enabled system or Diffserv router forwards or drops the
marked packet.
The goal of a PHB is to provide a specified amount of network resources to a class of traffic
on the contiguous network. In the QoS policy, DSCPs indicate the precedence levels for
traffic classes when the traffic flows leave the IPQoS-enabled system. Precedence can range
from high-precedence/low-drop probability to low-precedence/high-drop probability. IETF
defines the following PHBs [5] [4]:
 Best Effort (BE) PHB: Default PHB, which is used for best-effort service.
 Expedited Forwarding (EF) PHB: Used for low-delay service.

42
 Assured Forwarding (AF) PHB: Used for guaranteed bandwidth service.
 Class-selector (CS) PHB: Used for backward compatibility with the non-DiffServ
compliant device.
3.1.7.4 Traffic Shaping and Policy
These methods are often necessary on the edge separating a customer's network from a
provider's network. Providers often force the customer to adhere to a specific policy of
service (or committed rate). This policy is referred to as the Service Level Agreement (SLA)
between the customer and the provider. Shaping and policing mechanisms differ in how each
handles violations of the SLA. There are two methods for managing traffic that exceeds a
specified rate [14] [2]:
 Traffic shaping
 Traffic policing
Shaping is usually implemented on the customer side and will buffer traffic that exceeds the
provider's committed rate. Thus, shaping can slow the traffic rate and siphon out traffic in
compliance with the provider's SLA. Buffering traffic will often create delay and jitter, which
can negatively impact sensitive traffic types [18]. Shaping also requires sufficient memory to
queue buffered traffic. Shaping provides no mechanism to re-mark traffic that exceeds the
committed rate. Policing is usually implemented on the provider side, and will either drop or
re-mark traffic that exceeds the provider’s committed rate [31].
3.1.7.5 The Token Bucket System
Transfer rate or a token bucket system, which is comprised of three parts [11] [14]:
1. Committed Information Rate (CIR) – specifies the traffic rate dictated by the SLA,
measured in bits per second (bps).
43
2. Burst Rate (Bc) – specifies the amount of traffic to be sent within a given time
interval, measured in bits.
3. Time Interval (Tc) – identifies the time interval for each burst, measured in seconds or
sometimes milliseconds.
The CIR is calculated using the formula: CIR (bps) = Bc (bits) / Tc (seconds)
Both, policing and shaping, use a token bucket as a traffic meter. The token bucket usually
works in the following way [15]:
Tokens are put into the bucket at a given data rate.
Each token contains the length of the packet (in bits) that it allows the node to send.
The traffic meter checks the packet size and then pulls out the number of tokens required to
send that packet.
If the bucket does not contain enough tokens to send a packet, according to traffic policing,
the packet is dropped. However, in the same situation, the traffic shaping mechanism stores
the excess packets into a buffer and wait for enough tokens to send out the packets.
In the case when the token bucket is full of tokens than more tokens are discarded.
Figure 2:9 Token Bucket implementation [26].
3.1.8 Congestion Management Mechanisms
Congestion can happen in computer networks when the offered load and the demands for
network resources exceed the network capacity. During congestion, router buffers are
44
overflowed causing large queuing delays, and high packet loss. Congestion reduces the
effective utilization of network resources and causes degradation in the performance
experienced by network users [7]. Therefore, it is worthwhile to reduce the occurrence of
congestion situations in a network to optimize the utilization of network resources to
Congestion management tools are implemented on all output interfaces in a QoS enabled
network the following congestion management queuing methods are used: provide the
network users with the suitable performance [18] [2].
 FIFO (First in First Out)
 Priority Queuing (PQ), Custom Queuing (CQ)
 Weighted Fair Queuing (WFQ)
 Class-Based Weighted Fair Queuing (CBWFQ)
First-In-First-Out (FIFO) is the simplest modality of queuing. The incoming packets are put
in a single queue and are processed in the order of receiving them. (Figure 2) Packets are
dropped when the FIFO queue is full (tail drop). This queuing type requires little computation
and its behavior is very predictable, i.e. the delay of the packet is a direct function of the
queue size FIFO the packet arriving first is always processed first [1] [28]. When the capacity
of the queue is achieved, the excess incoming packets are dropped.
Figure 2:10 Implementation of a FIFO scheduling [28].
In PQ In the priority queuing, each packet is assigned a priority based on the level of service
requested. All the packets marked with higher priority go in the different queues then the
packets marked as a lower priority. There could be many priority queues. All of the packets
45
in the higher priority queue are attended first. The packets in the lower priority queue are
processed only when there are no packets in the higher priority queue.
WFQ is a type of queuing methodology having the purpose to allow fair access for each
incoming flow and to prevent a burst flow from consuming all the output bandwidth. WFQ
includes a queue for each distinct flow and packets from each flow is put into its appropriate
queue [29]. ‘Weighted Fair Queuing (WFQ) supports flows with different bandwidth
requirements. Most obviously we can mention that WFQ does not allow classification options
to be configured. Based on flows, WFQ classifies packets automatically, with each flow
being placed into a separate queue.
Figure 2:11 Implementation of a WFQ scheduling [29].
For WFQ, a flow can be described as all packets with the same values for source IP address,
destination IP address, transport layer protocol, TCP or UDP source port, TCP or UDP
destination port, and IP Precedence. Because WFQ puts packets of different flows in different
queues, must have a greater number of queues than all of the non-flow-based queuing
instruments [27] [10] [19].
3.1.9 Congestion Avoidance Mechanisms
Congestion is said to occur in the network when the resource demands exceed the capacity
and packets are lost due to too much queuing in the network [10]. During congestion, the
network throughput may drop to zero and the path delay may become very high. A
congestion control scheme helps the network to recover from the congestion state [25].
46
Congestion avoidance is a traffic control mechanism that uses traffic scheduling to prevent
the network from being overloaded. With this mechanism, the device can monitor the usage
of network resources such as queues and buffer areas in the memory and discard packets
when network congestion is intensifying. To avoid congestion, the following algorithms are
introduced [11] [9]:
 Tail Drop
 Random Early Detection (RED) and
 Weighted Random Early Detection (WRED).
Traditionally, when the queue is reached to its limit and the buffer is full, the tail drop occurs
and all the additional packets are dropped [7] [2]. Due to packet drops, the TCP hosts in the
network will reduce the transmission rate and try to synchronize with each other. All TCP
hosts slow down the transmission rate until the congestion is completely cleared. Once the
congestion is cleared, all TCP hosts start increasing transmission rates. As a result, the
transmission rate is going up and down which leaves the transmission links underutilized
during some periods [18].
RED works with the TCP transport protocol [37]. The main goal of RED is to
 Reduce the number of packets dropped at the router
 Avoid global synchronizations of TCP hosts
 Provide congestion avoidance employing early dropping packets and controlling
average queue size
 RED introduces a minimum threshold (MinTh) and maximum threshold (MaxTh)
values for the queue size.
All traffic below MinTh is transmitted without observing any drop packets. All traffic above
MaxTh is dropped. The probability of traffic between MinTh and MaxTh being dropped is
based on the number of packet increases. RED randomly drops the packets without
47
considering the QoS parameters of the packet to protect the queue from being fully utilized.
As a result, it is not useful for traffic with hard QoS requirements [17] [28].
WRED discards packets based on its QoS requirements and importance. WRED considers the
MPLS EXP bits to prioritize the packets. The packet with bits 000 in EXP bits (or IP
precedence bits) is considered as lower priority traffic and is more likely to be discarded. The
packet with bits 111 is considered as higher priority traffic and is less likely to be discarded.
The bits in the voice and video packets can be set to 1 1 1 to minimize the probability of
being dropped [25] [3].
3.2 Related Works
Research literature addressing Improving Quality of Service of Border Gateway Protocol
Multiprotocol Label Switching Virtual Private Network is extensive, relatively little work has
been done on developing solution methodologies directly related to improving end- to -end
QoS of MP BGP MPLS VPN traffic using the DiffServ model in Ethio Telecom in general.
The researcher briefly summarize a sampling of research focusing on improving QoS of MP
BGP MPLS VPN traffic using DiffServ research addressing a variety of QoS related issues,
and provide references to point out interested readers to appropriate sources of additional
information. Most QoS of MP BGP MPLS VPN traffic-related literature focuses on traffic
classification and architectural or protocol related topics associated with MP BGP MPLS
VPN traffic, and Some of them have tried to describe the QoS of MPLS VPN from the
customer LAN side, provider edge (PE) to the customer side, network backbone, and others
end-to-end QoS perspectives. In this chapter, notable related works are reviewed to lay down
the foundation for this research study.
D. Adami et al., [38] proposed a new ns-2 module to speed-up the design, development, and
deployment of DiffServe-aware MPLS network. MPLS DiffServ-aware allows network
48
operators to provide services that require strict QoS performance guarantees. The new
software module is used to simulate the RSVP-TE protocol using the ns-2 simulator.
N. Aslam [39] compared performance of MPLS networks and IP networks. A network
topology is designed and a MATLAB based simulation tool is used to send bulk data within a
network. Network performance is measured with MPLS enabled or disabled. The author
illustrates that MPLS network may perform better than the traditional IP networks. Sending
data file from a source to destination does not require traffic engineering.
D. Adami et al., [40] presented an overview of MPLS networks and compared performance
of IP routing and the MPLS forwarding mechanism. The authors did not discuss TE, which is
an essential part of MPLS.
Sebastian N. and Desta D., in [41] worked on quality of service of access layer networks.
They have optimized the network through physical and logical architectures to improve the
end-to-end QoS. They have used weighted random early detection (WRED) algorithms for
the logical architecture and minimum spanning tree for the physical and data link layers’
architecture to increase effective bandwidth utilization and to improve the performances of
the networks. According to their result, VPN technology can be used to improving the quality
of service for customers. They advocated that the BGP MPLS protocol has its own benefits in
network speed, stability, quality of services. They indicated that it can be used to detect
denial of services (DoS) attacks.
Farsin S. et al., in [42], articulated that VPN in MPLS network with MP BGP to assure end-
to-end QoS. They conducted simulation using GNS3 by configuring two companies with
different VPN sites on the same backbone. They used Wireshark to monitor the traffic flow
and quality of service. From their investigation, they identified that MP BGP MPLS VPN is
the most popular standard [23]. This technology uses BGP as a control plane to provide VPN
routing and MPLS as a transport technique to achieve isolation between customers traffic. Its
49
popularity is from the fact that its capability to support quality of service, traffic engineering
and a high number of customer’s support (thousands of customers and hundreds of thousands
of VPNs sites).
A detailed review was done by Kanchan D. and Alam S. in [43] on quality of service
improvement with MPLS mechanism in the next generation networks (NGNs). It is stated
that the increased demand for new and broaden network in terms of quality of service plays
an important role in providing better services for consumers. Hence, MPLS VPN techniques
enabled to improve the quality of service. MPLS VPN is one of the simplest, scalable,
flexible and dynamic ways to provide a better quality of service to users in a degraded
condition, with fast failure node recovery and traffic engineering. These improved network
performances and an end-to-end quality of services.
According to the literature reviewed, the researcher tried to combine different methods and
procedures that are used in the above-mentioned studies as input and tries to work on
improving the end-to-end quality of services of EthioTelecom MP BGP MPLS VPN SLA
customers using consistent queue, scheduling, and congestion avoidance mechanism.
50
Chapter Three
4 Proposed Network Architectures
4.1 Introduction
This section of the study covers the main steps in designing QoS of BGP MPLS VPN
network logically in general and specifically include IP address allocation, interface
connections, and configuration, as well as the QOS have been applied on the proposed
network. The configuration of DiffServ QoS model over MPLS VPNs on the proposed
network been achieved by using the network topology shown in figure 3.1.The eNSP network
architecture was chosen according to the requirements for the design of networks with service
provisioning and implemented end-to-end QoS. The proposed eNSP network architecture can
be easily scaled with simply appending new devices in the network.
There are two types of routers in the service providers’ network, i.e., Provider Edge (PE) and
Provide (P) routers. The Provider Edge router is directly connected to the Customer Edge
(CE) router of the customer networks. Both Provider Edge (PE) and Provider (P) routers
should run MPLS so that they can distribute labels to each other and forward labeled packets
to subsequent routers.
Figure 3:12 Simplified Proposed BGP MPLS VPN network architecture end-to-end
QoS.
51
The proposed network model is slightly modified for the purposes of evaluating VPNs and
testing congestion analysis. The VPN A and VPN B routers are used as traffic generators.
Both VPN are MPLS based with RSVP-TE signaling and tunneling and uses OSPF for IGP
interconnection and use the same networking equipment. The interconnection links and
interfaces are similar in both VPN models including The QoS applied to the traffic running
through the network.
The core network is realized as a core router and route reflector. The core device is logically
divided into two logical systems and acting like separate routers; as well they have full
functional capabilities of two separate hardware devices. Peering interfaces is used in the
connections between the two logical systems. The links with the other devices in the network
are recognized by general Gigabit Ethernet interfaces.
4.2 Designed BGP MPLS VPN

4.2.1 Network IP Address
IP addressing scheme in the proposed network uses different private IP address. The class A
network is assigned for the connections between provider routers, between provider and
provider edge routers, between PE and PE routers, between PE and CE routers and for users in
VPN A and VPN B interconnection. This address space is split into IP address spaces
between different interfaces of the core, aggregation and access device and Loopback IP
Addresses.
IP address spaces between provider, provider edge and customer edge routers
Table 3-4 the connection between routers and IP address assignment

Broadcast Network
Router Host range address address

P1 and P2 10.0.13.1 – 10.0.13.6 10.0.13.7 10.0.13.0
P1 and P2 10.0.14.1 – 10.0.14.6 10.0.14.7 10.0.14.0
P1 and PE1 10.0.11.1 – 10.0.11.6 10.0.11.7 10.0.11.0
P2 and PE2 10.0.12.1 – 10.0.12.6 10.0.12.7 10.0.12.0
52
P1 and PE3 10.0.15.1 – 10.0.15.6 10.0.15.7 10.0.15.0

P2 and PE4 10.0.16.1– 10.0.16.6 10.0.16.7 10.0.16.0
PE1and PE2 10.0.10.1 – 10.0.10.6 10.0.10.7 10.0.10.0
PE3and PE4 10.0.17.1 – 10.0.17.6 10.0.17.7 10.0.17.0
PE1andCE1 10.0.20.1 – 10.0.20.6 10.0.20.7 10.0.20.0
PE1andCE2 10.0.30.1 – 10.0.30.6 10.0.30.7 10.0.30.0
PE3and E3 10.0.40.1 – 10.0.40.6 10.0.40.7 10.0.40.0
PE3andCE4 10.0.50.1 – 10.0.50.6 10.0.50.7 10.0.50.0
Table 3-5 Network Routers Loopback IP Addresses
LoopBack IP Address
P1 Loopback 10.0.1.1/32
P2 Loopback 10.0.2.2/32
PE1 Loopback 10.0.3.3/32
4.2.2 Interfaces
The configuration of the interfaces in the proposed network architecture is almost common
for all interfaces. Only loopback have distinct differences in the way of their configuration
and it can’t contain Mpls, Mpls LDP, Mpls te and Mpls rsvp-te configuration. The following
command is the most common format to configure given Interfaces:
[PE1] interface gigabitethernet1/0/0
[PE1- gigabitethernet1/0/0] undo shutdown
[PE1- gigabitethernet1/0/0] mpls
[PE1- gigabitethernet1/0/0] mpls te
[PE1- gigabitethernet1/0/0] mpls rsvp-te
[PE1 gigabitethernet1/0/0] quit

53
4.2.3 Interior Gateway Protocol (IGP) Interconnection
An interior gateway protocol (IGP) is a routing protocol that is used to exchange routing
information within an autonomous system (AS). The OSPF protocol has been used. This is
because of OSPF protocol is more convergent and secure. The following command is the
most common format to configure OSPF protocol.
[PE1] ospf
[PE1-ospf-1] area 0.0.0.0
[PE1-ospf-1-area-0.0.0.0] network 10.0.3.3 0.0.0.0
[PE1-ospf-1-area-0.0.0.0] network 10.0.1.1 0.0.0.255
[PE1-ospf-1-area-0.0.0.0] mpls-te enable
[PE1-ospf-1-area-0.0.0.0] quit
4.2.4 MPLS and MP BGP Interconnection
MPLS protocol is used for label switching and distribution. The following command is the
most common format to configure MPLS globally.
[PE1] Mpls user-id 10.0.3.3
[PE1] mpls
[PE1-mpls] quit
[PE1] Mpls ldp
[LSRA-mpls-ldp] quit
MP BGP protocol is used to create the peer relationship between different types of routers.
The following command is the most common format to configure MP BGP.
[PE1] bgp 100
[PE1-bgp] peer 10.0.5.5 as-number 100
[PE1-bgp] peer 10.0.5.5 Connect-interface loopback 0
[PE1-bgp] ipv4-family vpnv4

54
[PE1-bgp-af-vpnv4] peer 10.0.5.5 enable
[PE1-bgp-af-vpnv4] quit
[PE1-bgp] quit
4.2.5 Configuring MPLS RSVP-TE
MPLS RSVP-TE is used to establish a TE tunnel from one router to another. It limits the
maximum receivable bandwidth for links along the tunnel. TE tunnel has bandwidth
constraints model such bandwidth allocation method.
To configure TE tunnel firstly enable MPLS, MPLS TE, and RSVP-TE globally on each
router, enable MPLS, MPLS TE, and RSVP-TE on all tunnel interfaces, and enable CSPF in
the system on the ingress routers. The following command is the most common format to
enable MPLS, MPLS TE, and RSVP-TE globally and interfaces.
[PE1] mpls lsr-id 10.0.3.3
[PE1] mpls
[PE1-mpls] mpls te
[PE1-mpls] mpls rsvp-te
[PE1-mpls] mpls te cspf
[PE1-mpls] quit
Configure OSPF TE and maximum receivable bandwidth and the maximum usable (BC0)
bandwidth, creates the TE tunnel and limits the maximum receivable bandwidth on all tunnel
interfaces. The following command is the most common format to configure OSPF TE and
maximum receivable bandwidth.
[PE1] interface gigabitethernet 0/0/1
[PE1-GigabitEthernet0/0/1] mpls te bandwidth max-reservable-bandwidth 100000
[PE1-GigabitEthernet0/0/1] mpls te bandwidth bc0 100000
[PE1-GigabitEthernet0/0/1] quit
55
TO create tunnel interfaces on the ingress routers the following command is the most
common format to create a tunnel and configure IP addresses for the tunnel.
[PE1] interface tunnel 0/0/0
[PE1-Tunnel0/0/0] ip address unnumbered interface loopback 1
[PE1-Tunnel0/0/0] tunnel-protocol mpls te
[PE1-Tunnel0/0/0] destination 10.0.5.5
[PE1-Tunnel0/0/0] mpls te tunnel-id 100
[PE1-Tunnel0/0/0] mpls te signal-protocol rsvp-te
[PE1-Tunnel0/0/0] mpls te bandwidth ct0 20000
[PE1-Tunnel0/0/0] mpls te commit
[PE1-Tunnel0/0/0] quit
4.2.6 Configure VPN instances on PEs
In BGP/MPLS VPN application, each VPN has an instance to maintain forwarding
information of the local VPN. Such an instance is called a VPN instance or VPN routing and
forwarding table (VRF).
The following command is the most common format to configure VPN instance and bind the
instance to the interfaces.
[PE1] ip vpn-instance vpnaa
[PE1-vpn-instance-vpna] ipv4-family
[PE1-vpn-instance-vpna-af-ipv4] route-distinguisher 100:1
[PE1-vpn-instance-vpna-af-ipv4] vpn-target 111:1 both
[PE1-vpn-instance-vpna-af-ipv4] quit
[PE1-vpn-instance-vpna] quit [PE1] ip vpn-instance vpnbb
[PE1-vpn-instance-vpnb] ipv4-family
[PE1-vpn-instance-vpnb-af-ipv4] route-distinguisher 100:2
56
[PE1-vpn-instance-vpnb-af-ipv4] vpn-target 222:2 both
[PE1-vpn-instance-vpnb-af-ipv4] quit
[PE1-vpn-instance-vpnb] quit
4.2.7 Creating EBGP peer relationship between the PE and CE routers
In BGP/MPLS IP VPN, a routing protocol or static routes must be configured between a PE
and a CE to allow them to communicate and allow the CE to obtain routes to other CEs. The
routing protocol can be EBGP (External/Exterior BGP), IBGP (Internal/Interior BGP), RIP
(Routing Information Protocol), OSPF (Open Shortest Path First), or IS-IS (Intermediate
System to Intermediate System). For the proposed system OSPF is chosen to establish the
EBGP peer relationship between the PE and CE to import VPN routes on CEs routers and
PEs routers. The following commands are the most common format to enable EBGP peer
relationship.
[CE1] bgp 65410
[CE1-bgp] peer 20.1.1.2 as-number 100
[CE1-bgp] import-route direct
[CE1-bgp] quit
[PE1] bgp 65007
[PE1-bgp] ipv4-family vpn-instance vpnaa
[PE1-bgp-vpna] peer 10.0.20.2 as-number 65410
[PE1-bgp-vpna] import-route direct
[PE1-bgp-vpna] quit
[PE1-bgp] ipv4-family vpn-instance vpnbb
[PE1-bgp-vpnb] peer 10.0.30.2 as-number 65420
[PE1-bgp-vpna] import-route direct
[PE1-bgp-vpna] quit
57
4.3 Designed QoS of Proposed network architectures
The designed QoS is to provide different levels of service quality based on different
requirements to meet SLA targets and ITU threshold quality requirements of different VPNs.
Managing maximum receivable bandwidth, reducing transmission, queuing and processing
delay, managing jitter and packet loss are the main focuses of the design.
Quality of service assurance is designed based on the existing resources by using rational
scheduling and congestion avoidance methods. Differentiated service model (DiffServ) have
been used to classify, mark and shape the networks based on the existing SLA agreements.
This has been applied by the following step by step processes.
 Define access control list (ACL) rules
 Define traffic classifiers
 Define traffic behaviors
 Define traffic policies and
 Apply traffic policies to interfaces.
4.3.1 Define Access Control List rules
ACLs are used specify which VPNs are granted to guarantee the required services quality
within the time. Define ACL rules Configure complex traffic classification on CE routers to
control the traffic that accesses CEs from the local networks. The following command is the
most common format
[CE1] acl number 2001
[CE1-acl-basic-2001] rule permit source 1.1.1.1 0.0.0.255
[CE1-acl-basic-2001] quit

58
[CE1-acl-basic-3001] rule 0 permit UDP destination-port eq DNS
[CE1-acl-basic-3001] rule 1 permit UDP destination-port eq snmp
[CE1-acl-basic-3001] rule 2 permit UDP destination-port eq SNMP trap
[CE1-acl-basic-3001] rules 3 permit UDP destination-port eq Syslog
[CE1-acl-basic-3002] rules 4 permit up
[CE1-acl-basic-3002] qui
4.3.2 Define traffic classifiers
In the proposed network QoS classification is implemented by marking types of the services.
Classification classifies the packets while packets unchanged. It is based on the DiffServ
Code Point (DSCP) values of IP packets. The traffic of different service level can be
identified. Then the defined ACL can have applied to it. The following command is the most
common format to define traffic classifiers.
[CE1] traffic classifier a
[CE1-classifier-a] if-match all 2001
[CE1-classifier-a] quit
[CE1] traffic classifier b
[CE1-classifier-b] if-match all 2002
[CE1-classifier-b] quit
59
[CE1] traffic classifier up limit
[CE1-classifier-udplimit] if-match all 3001
[CE1-classifier-udplimit] quit
[CE1] traffic classifier udplimit1
[CE1-classifier-udplimit1] if-match all 3002
[CE1-classifier-udplimit1] quit
4.3.3 Define traffic behavior
In the proposed network traffic behavior is used to ensure the capability of the devices
support DSCP. It includes configuring traffic policing and re-marks DSCP values.
Committed information rate (CIR) indicates the rate at which the tokens are placed into the
bucket. Committed burst size (CBS) indicates the capacity of the bucket, whereas packet
burst size (PBS), shows the maximum volume of the traffic that can the interface allows
passing through the traffic burst. Then the Committed access rate (CAR) used to policing
specific excess traffics are dropped or remarked. All packets are marked according to
predefined traffic classifier match rule. The following command is the most common format
to define traffic behavior.
[CE1] traffic behavior e
[CE1-behavior-e] car car 10000 CBS 150000 PBS 0
[CE1-behavior-e] remark dscp 40
[CE1-behavior-e] quit
[CE1] traffic behavior f
[CE1-behavior-f] car car 5000 CBS 100000 PBS 0
[CE1-behavior-f] remark dscp 26
[CE1-behavior-f] quit
[CE1] traffic behavior g
[CE1-behavior-g] car car 2000 CBS 100000 PBS 0

60
[CE1-behavior-g] remark dscp 0
[CE1-behavior-g] quit
[CE1] traffic behavior up limit
[CE1-behavior-udplimit] permit
[CE1-behavior-udplimit] quit
[CE1] traffic behavior udplimit1
[CE1-behavior-udplimit1] car 5000 CBS 100000 PBS 150000 green pass yellow discard red discard
[CE1-behavior-udplimit1] quit
4.3.4 Define and apply traffic policies
Traffic policing control the rate of the incoming packet to ensure the network resources are
properly allocated. When the traffic rate of the connection exceeds the specification on an
interface, it allows remarking the excess packet depending on the priority to maximize the
network resource usage.
Traffic policy implements on the QoS requirement defined in the SLA. The SLA contains
parameters such as CIR, CAR, PBS and CBS which are predefined on the traffic behavior.
The following command is the most common format to define traffic policies.
[CE1] traffic policy 1
[CE1-trafficpolicy-1] classifier a behavior e
[CE1-trafficpolicy-1] quit
[CE1-trafficpolicy-2] classifier b behavior f
[CE1-trafficpolicy-3] classifier c behavior g
[CE1] traffic policy up limit
61
[CE1-trafficpolicy-udplimit] classifier UDP limit behavior up limit
[CE1-trafficpolicy-udplimit] classifier udplimit1 behavior udplimit1
4.4 Applying the predefined policies to the inbound interfaces routers
The predefined policies are used to guarantee the service requirements of SLA. The following
command is the most common format to apply traffic policies to the inbound interfaces.
[CE1] interface gigabitethernet 0/0/0
[CE1-GigabitEthernet0/0/0] undo shutdown
[CE1-GigabitEthernet0/0/0] traffic-policy 1 inbound
[CE1-GigabitEthernet0/0/0] quit
[CE1-GigabitEthernet0/0/1] quit
[CE1] interface gigabitEthernet 10/0/0
[CE1-GigabitEthernet10/0/0] traffic-policy UDP limit outbound
4.5 Experimental Results of Proposed Architecture
The proper functioning of the designed end to end QoS of BGP MPLS VPN network
architectures include:
 All protocols are fully operating
 Proper implementation of the designed QoS
 Provisioning of the necessary services ensuring L3VPN operation and

62
 Redundancy of network resources which includes rerouting in case of link or node

failure.
The necessities for fulfilling these requirements have discussed with the relevant tests for
each of them. To be entrusted the proper functioning of the network first the basic
components have checked.
4.5.1 IGP protocol
In proposed architectures, since it is one of the basic components of the designed models
OSPF operation is checked first. Checking OSPF routing protocol involves testing the routing
information, established neighbors, link state database, and interface enabled with OSPF
protocol.
To check the OSPF routing information “ip routing –table protocol ospf” command is used. It
checks whether routes are learned by other routers. Route information includes all direct
routes and the routes to loopback interfaces.
Figure 3:13 OSPF route information
63
To check the OSPF neighbor relationship information “display OSPF peer” command is
used. It checks the adjacency relationship.
Figure 3:14 Ospf neighbor relation
To check the OSPF link state database information “display OSPF lsdb” command is used.\
Figure 3:15 Ospf link state data base

To check the IS-IS interface information “display OSPF interface” command is used.
64
Figure 3:16 interface ospf enabled
65
To check the OSPF overall information “display OSPF brief” command is used.
Figure 3:17 OSPF brief information.
From the output of these commands, each router is connected to the other devices loopback
addresses which are an important prerequisite for the proper functioning of the other
components of the proposed network.
The outcome of the routers in Fig.3.2 means that the OSPF protocol successfully established
its link-state database of the network and built its routing table. The information about OSPF
interfaces (Fig.3.3) is important for updating routing information when there is a change of
the network topology.
The outcome of the routers shows that the routers made a neighbor relationship with each
other, and the links between them are functioning normally. From the brief information of the
protocol (Fig.3.4) can be seen that the establishment of the routing table has passed. It is
further understood that the protocol is configured to work with the signaling protocol RSVP-
TE and MPLS TE.
66
4.5.2 Signaling protocol RSVP-TE
Resource reservation setup protocol with traffic engineering is used for signaling in the
proposed architectures. To check the RSVP-TE overall information “display Mpls rsvp-te”
command is used.
4.5.3 MPLS TE Tunnel
Figure 3:18 RSVP-TE detail information.

MPLS TE is used to control the excess traffic and avoid congestion in the proposed network.
It solves the process of the certain link being overloaded where the other idle. It also ensures
full utilization of bandwidth resources. Checking MPLS TE functionality includes tunnel
information, tunnel status of MPLS TE constraint shortest path database and MPLS TE
session. To check the MPLS TE tunnel information “display interface tunnel” command is
used.
67
Figure 3:19 MPLS TE tunnel information

To check the MPLS TE tunnel status “display Mpls te tunnel-interface tunnel 0/0/0”
command is used.
Figure 3:20 MPLS TE tunnel status
To check the MPLS TE constraint shortest path database “display Mpls te cspf tedb all”
command is used.
4.5.4
MPLS Figure 3:21 MPLS TE tunnels constraint shortest path database
Operation
Checking the operation of MPLS involves testing its routing information, MPLS link state
protocol, and MPLS adjacency. To check the MPLS routing information “display Mpls route-
state” command is used.
68
Figure 3:22 MPLS routing information.

To check
the MPLS Figure 3:23 MPLS link state protocol information
link state protocol “display Mpls LDP LSP all” command is used.
To check the MPLS adjacency “display Mpls LDP adjacency” command is used.
69
Figure 3:24 MPLS adjacency information
To check the BGP neighbor relationship information “display BGP peer” command is used.
70
Figure 3:25 BGP neighbor relationship

Fig.3.14 shows that BGP is fully operational and has established a neighbor relationship.
BGP sessions are established. The L3VPN groups are properly signaled. The end routers
traffic is properly forwarded and there is communication between the routers in the L3VPN
services.
4.5.6 Performance of established L3VPN Service
The two L3VPN services are fully functional. To check detail routing information of the two
L3VPN “display ip VPN-instance verbose” command is used.
Figure 3:26 VPN instances routing information.
71
When the interfaces on routers are bound to the same VPN instance can exchange the
information. The same VPN instance can ping each other. To check the reachability of the
same VPN instance “ping VPN-instance CBE 20.1.1.2” command is used
Figure 3:27 VPN instances reachability to access router
Figure 3:28 VPN instances reachability to aggregation router.

4.5.7
Quality of Service of proposed network architectures
The applied QoS in proposed network architectures is fully functional. to view the traffic
policies, traffic classifiers defined in the traffic policies, and the traffic behaviors associated
with the traffic classifiers “display traffic policy user-defined” command is used.
Figure 3:29 user-defined QoS.
72
When the QoS of the proposed network architecture was verified from Wireshark, different
parameters such as a frame, Ethernet, MPLS, TCP, BGP, and IP are functioning fully. Under
IP the defined QoS, differentiated service field is operational with its parameter.
Figure 3:30 Defined QoS.
4.6 Discussions
In this study, the researcher evaluated the QOS performance metrics such as delay variation,
delay, response time ,throughput for different traffic types (voice, data and video) for both
platform the existing network and the proposed network.
The existing and proposed network architectures are the same in physical devices used and
interconnection. But they have major differences, especially in QoS designing.
Source of traffic
The traffic sources are generated by the client nodes
Table 3-6 Values and parameters of the traffic model

Traffic type DSCP
FTP AF11
VOICE AF31
VIDEO AF41
Table 3-7 Simulation configuration

Packet payload 180 bytes
No packets 1000
73
Packet interval 20ms

Frequency 60000ms
Graph result is plotted for ever 300 seconds
Table 3.5 shows the similarities and differences between existing and
proposed network architecture in detail manner.
Exit Network Architecture Proposed Network Architecture

Traffic Type BGP MPLS VPN BGP MPLS VPN
Service Type L3VPN L3VPN
IGP Routing Protocol IS-IS OSPF
NGN Backbone MPLS MPLS
QoS Model Best effort model Differentiated services model
Congestion Management FIFO LLWFQ
Congestion Avoidance Tail Drop Weighted random early detection
Table 3-8 5 the similarities and differences between existing and proposed network
architecture.
The researcher approach in this study is to design and build an operator network type back
bone to simulate a real scenario that convey different types of traffic (voice, data and
video).the result of the work are presented according to the simulation time and the network
load. The result of the comparison demonstrates the advantage over the performance of
MPLS networks with Diffserv compared to the existing network.
QoS in the proposed network architectures is tested using Wireshark modeler, using latest
simulation technique, where different QOS parameters can be measured to compare the
performance of networks. A couple of network simulation scenarios (Fig.1:3 and Fig.3.1) are
tested with different types of traffic (voice, data and video) with different parameters and
speeds. In the first scenario, the existing network (Fig.1:3) performance is tested. The
existing network architecture uses the best effort QoS model; which means all traffic has
equal priorities. The architecture uses a FIFO algorithm for congestion management and tail
74
drop algorithm for congestion avoidance. In this scenario, the generated traffic consists of
two VPN instance application traffic streams. The traffic sources are generated by the client
nodes of the objects (node) from which we associate the application profile that presents the
configuration of multiple applications.
In the second scenario (Fig.3:1) the proposed network architecture uses DiffServ QoS. The
traffic has different priorities. The higher the traffic processed first. The architecture uses a
LLWFQ queuing algorithm for congestion management and weighted random early detection
algorithm for congestion avoidance. In this case, the traffics were classified and priority is
given to it depending on their SLA levels. Then traffic policies were defined and applied on
an aggregation router outbound interface. In this case, the generated traffic consists of two
VPN instance application traffic streams. The traffic sources are generated by the client nodes
of the objects (node) from which we associate the application profile that presents the
configuration of multiple applications. We used Ftp, voice and video in our simulation. The
first test is made between CE1 and CE3 and the second test made between CE2 and CE4
routers. The results of both experiments are given as the following parameter comparison
 FTP traffic response time (light / Heavy Load)
 End-to-End Voice Traffic Delay (light / Heavy Load)
 Variation of Voice traffic delay (light / Heavy Load)
 End-to-End Video Traffic Delay (light / Heavy Load)
 Variation of delay with video traffic (light / Heavy Load)
Traffic Analysis Data (FTP)
75
Figure 3:31 FTP response time with light load (Blue - MPLS, Red - MPLS
/DiffServ)
The MPLS response time is higher than the MPLS / DiffServ network. We notice that the
MPLS network with DiffServ provides better performance in response times for FTP based
traffic.
Figure 3:32 FTP traffic response time with heavy Load (Blue - DiffServ; Red-
MPLS)
After increasing the load on the network (heavy load), the response time for MPLS changes
to a very high value, but DiffServ remains about the same as shown in Figure 3:21.
Traffic Analysis Voice:

76
We will compare between the two scenarios MPLS and MPLS_ DiffServ. Figure 3:22 and
Figure 3:23 illustrate the end-to-end delay parameters for light load and heavy load traffic.
Figure 3:33 End-to-end traffic delay Voice with light load (Blue MPLS; Red /
DiffServ)
Figures 3:24 and 3:25, shows the delay (jitter) and variations of voice traffic for light load
Figure 3:34 Delayed end-to-end traffic Voice with heavy load (Blue - MPLS /
DiffServ; Red-MPLS)
and heavy load traffic
77
Figure 3:35 Variation of delay (jitter) of traffic Voice with light load (Blue - MPLS;
Red - MPLS / DiffServ)
Figure 3:36 Variation of the Delay Delay (Jitter) of Voice heavy load traffic (Blue -
MPLS, Red-MPLS / DiffServ)
Traffic analysis Video:
Then we study the performance of the network when the traffic is the Video. We will
compare between the two scenarios MPLS and MPLS_ DiffServ. Below, illustrate the end-to-
end delay settings for light load and heavy load traffic.
78
Figure 3:37 Video End-to-End Delay with Light Load (Blue - MPLS; Red - MPLS /
DiffServ)
Figure 3:38 Video End-to-End Delays with heavy Load (Blue - MPLS / DiffServ;
Red- MPLS)
The end-to-end delay for video stream is higher for MPLS than MPLS / DiffServ. After the
load increases, the end-to-end delay for MPLS becomes very high and keeps increasing.
While the MPLS / DiffServ delay remains at a low level. This shows that MPLS / DiffServ
offers better quality even in congested network with higher loads, while in MPLS, the delay
increases to a very high value.
79
Figure 3:39 Variation of the delay time (jitter) of video traffic with light load (Blue -
MPLS; Red - MPLS / DiffServ)
The variation of the delay in FIG. 13 demonstrates the best quality of the MPLS /
The variation of the delay in FIG.3:29 demonstrate the best quality of the MPLS / DiffServ
service.
80
After the
increase in
traffic load,
the results of
FIG 3:29 are
evident that
Figure 3:40 Variation of delay of video traffic with heavy Load (Blue - MPLS /
DiffServ; Red - MPLS)
the delay variation has increased to a very high value for MPLS, while the variation of the
MPLS_ DiffServ delay has remained at a very low value, keeping the quality of service
performance to the required level.
When we look at the numerical results obtained from both the existing and proposed network
is shown in Table 3.1. Most of the results were as expected. The difference between packet
loss and bandwidth in existing and proposed network architecture was visible. But the
difference between end-to-end delay and jitter was not that much visible. This happened
because we have used ten routers only on both network architectures. This reduces the
transmission, serialization, queuing and processing delay. The difference was visible if the
number of routers (nodes) increased.
Table 3-9 Exist and proposed network architecture numerical QoS results
Exit Network (Best Effort) Proposed Network (DiffServ)
Parameters Result SLA Targets ITU threshold Result SLA Targets ITU threshold
Packet loss (%) 1.897% Out of Range Out of Range 0.026% Within Range Out of Range
Delay(sec) 0.169 Within Range Out of Range 0.14132 Within Range Within Range
Jitter (sec) 0.001 Within Range Out of Range 0.0007747 Within Range Within Range
Bandwidth (bit/sec) 15068 Out of Range Out of Range 15320 Within Range Out of Range
81
Chapter Four
5 Conclusions and Future Works
5.1 Conclusions
In this thesis quality of service (QOS) analysis of an end-to-end MPLS BGP VPN network
architecture with the DiffServ model is done in comparison with the existing MPLS network
with the Best-effort model against four QoS parameters (Bandwidth utilization, packet loss,
latency, and jitter). This work has investigated the limitations with the existing network
architecture of ETHIO Telecom and explores the possibility of optimizing the end to end
quality of service (QOS) using the DiffServ model and OSPF routing protocol. Simplified
network topology was created and two network architectures were designed, built, and
evaluated with generic telecommunication equipment. Firstly, the existing BGP MPLS VPN
network which used the best-effort QoS model was implemented and tested. Secondly, the
proposed BGP MPLS VPN architecture which used the DiffServ QoS model was designed
and tested. End-to-end QoS was designed and implemented in both network models. The
implemented services were Layer 3 VPN services to handle traffic from end nodes in the
proposed architecture. From the study and simulation results the following conclusions can
be drawn:
 Both network architectures were fully functional and Verification of the applied end-
to-end QoS parameters was made and results were obtained.
 The proposed BGP MPLS VPN network architecture has many more benefits than the
existing BGP MPLS VPN network architecture. This is due to the opportunity for the
class of services and traffic-engineering in the network, which helps for better traffic
management and provisioning of suitable end-to-end QoS.
82
 The proposed BGP MPLS VPN architecture which used DiffServ QoS model
architecture could be used in many mission-critical applications. The opportunity for
easy scalability of the network is in great help in today's rapidly growing VPN
networks. This approach is suitable for higher-priority services, because of the low
latency and low packet loss across the network.
 The designed BGP MPLS VPN architecture which used DiffServ QoS model network
architecture is easy to scale and troubleshoot. The problem with the fast ceasing
number of available ASs is evaded by using a single AS number in the core network
architecture. Because of the implemented failure mechanisms in case of a link failure,
the impact on network flow is ceased. The traffic which entered the network is
delivered through backup routes to minimize the traffic loss, while new paths are
built. With the careful design of the applied QoS, the traffic requirements of the
implemented applications are served.
 In the proposed BGP MPLS VPN architecture, which used DiffServ QoS model
architecture, all services got the required traffic handling. End-to-end MPLS solutions
for the NGN applications are smoothly served.
 The proposed approach provides more efficient use of network resources and a
reduction of the number of network nodes.
 It relies on a single MPLS forwarding scheme, which simplifies traffic management
in the network. This way of service provisioning offers simplicity to the end nodes
and depends more on the intelligent nodes in the core network. At the same time, its
implementation and maintenance are also simplified. The designed BGP MPLS VPN
network can easily have implemented in the operator core network. Then the
83
architecture is simply managed, configured, and scaled with the least efforts and
almost without any operational costs.
Generally, based on the analysis and results gained, we conclude that the DiffServ QoS
model was more reliable than the best effort QoS model for the EthioTelecom BGP MPLS
VPN network. As the whole research process, the main work goes through traffic
engineering, network optimization, and proper network utilization. The designed QoS used
the DiffServ model that has been guaranteeing all company's SLA QoS threshold. In a
conclusion, the designed network provides a way of increasing network performance based
on the DiffServ QoS model. High network performance indicates a high QoS service
provider. A high QoS service provider creates a satisfied and high quality of experience’s
customers.
5.2 Future Works
Although the thesis has achieved the objectives set in Chapter one and the QoS has been a
guarantee for the company's SLA QoS target, there are some issues related to the network
extendibility with more reliability functions to be addressed in the future. These issues are:
 Clustering for access and aggregation devices, implementation of high availability
features and implementation of LDP for MPLS label down streaming on demand.
 Extended DiffServ services with more application-specific QoS can be implemented.
Layer 2 VPNs and VPLS can be included as a service in the network architecture.
This can increase the scalability, availability, and security of layer 2 VPNs.
 The implementation of self-organizing network architecture, such as self-learning,
self-configuration and self-management, self-optimization, prediction of network
congestion, prediction of traffic loops. To implement advanced extensions there are
algorithms for prediction. Algorithms for adaptive training of the network such as the
84
Widrow-Hoff algorithm can be of great use for process predictions in operating
networks [40] [41]. This way the designed proposed BGP MPLS VPN architecture
which used DiffServ QoS model network architecture can become optimal, which
saves operational and maintenance costs.
85
Reference
[1] S. Gurung, “IMPLEMENTATION OF MPLS VPN,” p. 40.
[2] A. M. Beyene and S. A. Argaw, “Improving Quality of Service of Border Gateway
Protocol Multiprotocol Label Switching Virtual Private Network of EthioTelecom Service
Level Agreements,” in Information and Communication Technology for Development for
Africa, vol. 1026, F. Mekuria, E. Nigussie, and T. Tegegne, Eds. Cham: Springer
International Publishing, 2019, pp. 278–288.
[3] H. T. Co, “Configuration Guide - VPN,” no. 08, p. 1069.
[4] S. Maheshwarn, S. Lillypet, and C. Vennila, “QOS Capabilities for Building MPLS
VPN,” vol. 5, no. 5, p. 5, 2013.
[5] R. Kumar CV, D. C, B. A, and B. Kp, “Architecture for MPLS L3 VPN Deployment
in Service Provider Network,” J. Telecommun. Syst. Manag., vol. 06, no. 01, 2017,
[6] EthioTelecom, “QoS Document” MPLS VPN Services Quality and Customer
Experience Related Issues and Complaint Analysis, Version 02. 2017.
[7] A. Akbari-Moghanjoughi, “Service Level Agreements for Communication Networks:
A Survey,” vol. 18, no. 1, p. 25.
[8] A. El-Atawy and T. Samak, “End-to-end verification of QoS policies,” in 2012 IEEE
Network Operations and Management Symposium, Maui, HI, Apr. 2012, pp. 426–434,
[9] S. Thukral and B. Chadha, “A Survey on QoS Behavior in MPLS Networks,”
IJARCCE, pp. 289–293, Mar. 2015,
86
[10] T. Szigeti and C. Hattingh, End-to-end QoS network design. Indianapolis, IN: Cisco
Press, 2005.
[11] H. Dransfeld and S. Young, “VPNs and Application Optimization,” p. 9.
[12] L. Cittadini, G. D. Battista, and M. Patrignani, “MPLS Virtual Private Networks,” p.
30.
[13] N. Zidan and M. Hamarsheh, “Implementation of Border Gateway Protocol (BGP)
Attributes,” vol. 5, no. 2, p. 7, 2016.
[14] R. V. R. Kumar and C. S. Kumar, “Design and Implementation of a Differentiated
Service based QoS Model for Real-Time Interactive Traffic on Constrained Bandwidth IP
Networks,” p. 108.
[15] N. H. Almofary, H. S. Moustafa, and F. W. Zaki, “Scalability Aspects in BGP/MPLS
VPN,” p. 12, 2013.
[16] J. Guichard, I. Pepelnjak, and J. Apcar, MPLS, and VPN architectures. Indianapolis,
IN Cisco Press, 2001.
[17] “MPLS and VPN Architectures,” p. 336.
[18] M. Kolhar, M. M Abualhaj, and F. Rizwan, “QoS Design Consideration for Enterprise
and Provider’s Network at Ingress and Egress Router for VoIP protocols,” Int. J. Electr.
Comput. Eng. IJECE, vol. 6, no. 1, p. 235, Feb. 2016,
[19] L. Cittadini and G. D. Battista, “Design of Virtual Private Networks with MPLS,” p.
29.
[20] Solomon T. and Hilina T., "EthioTelecom IP QoS parameters”, Recommended IP-
QOS performance targets. 2015.

87
[21] ITU-T, “Network-based IP VPN over MPLS architecture”, Global information
infrastructure and internet protocol aspects. 2011.
[22] Ethio Telecom, ”High-Level Design for NGN”, designed network, vol. I. 2006.
[23] K. Peffers, T. Tuunanen, M. A. Rothenberger, and S. Chatterjee, “A Design Science
Research Methodology for Information Systems Research,” J. Manag. Inf. Syst., vol. 24, no.
3, pp. 45–77, Dec. 2007,
[24] EthioTelecom, “‘Service Provisioning Manual’, Provisioning guide,” Jul. 2014.
[25] R. Shinde, “QoS Deployment Experiences,” p. 35.
[26] “Quality of service regulation manual,” p. 176.
[27] A. Joshi, "Multi-Protocol Label Switching: Quality of Service, Traffic Engineering
application, and Virtual Private Network application," p. 137.
[28] A. Srikitja and D. Tipper, “QoS-based Virtual Private Network Design for an MPLS
network,” p. 7.
[29] G. Kaur and D. Kumar, “MPLS Technology on IP Backbone Network,” Int. J.
Comput. Appl., vol. 5, no. 1, pp. 13–16, Aug. 2010,
[30] A. S. Khan and B. Afzal, “MPLS VPNs with DiffServ – A QoS Performance study,”
p. 82.
[31] “Multiprotocol Label Switching (MPLS),” p. 24.
[32] P. Brittain, “MPLS Virtual Private Networks,” p. 42.
88
[33] N.-E. Rikli and S. Almogari, “Efficient priority schemes for the provision of end-to-
end quality of service for multimedia traffic over MPLS VPN networks,” J. King Saud Univ.
- Comput. Inf. Sci., vol. 25, no. 1, pp. 89–98, Jan. 2013,
[34] ITU, “QoS Parameters”, ITU-T Y.1541 Recommended QoS Target. 2012.
[35] EthioTelecom, Low-Level Design for IP Backhaul, Final version. 2013.
[36] S. K. Nair and D. C. Novak, “A traffic shaping model for optimizing network
operations,” Eur. J. Oper. Res., vol. 180, no. 3, pp. 1358–1380, Aug. 2007.
[37] D. A. S. Ahmad, D. T. Alatky, and M. Jafar, “Performance Analysis DiffServ based
Quality of Service in MPLS Network’s,” vol. 6, no. 9, p. 9, 2015.
[38] D. Adami, “"Signaling protocols in DiffServ-aware MPLS networks: Design and
Implementation of RSVP-TE network simulator,”,” St. Louis, MO, USA, pp. 792–796.
[39] N. Aslam, “Traffic engineering with MPLS,” Master Thesis, School of Engineering
Science, Blekine Institute of Technology, Ronneby, Sweden. .
[40] D. Adami, "A new ns2 module for the simulation of MPLS networks with point-to-
multipoint LSPs support,” IEEE International Conference on Communications (ICC 2009),
Dresden, Germany,. .
[41] Dr. Sebastian N. and Desta D. “‘Modeling Network Optimization by Optimize the
Current Network by physical and logical architectures to improve the QoS’, International
Journal of Engineering Science and Computing,” p. Volume 7.
[42] Farsin S.et al. “VPN in MPLS network with MP BGP”, International Journal of
Electronics and Communication Engineering,. .
[43] K. D. and A. S., “‘Review on QoS Improvement with MPLS Mechanism in NGN’,
International Journal of Innovative Research in Science,” vol. Vol. 3.
89
Appendices
<PE1>display current-configuration <PE3>display current-configuration

[V200R003C00] [V200R003C00]
# #
sysname PE1 sysname PE3
# #
clock timezone China-Standard-Time minus 08:00:00 clock timezone China-Standard-Time minus 08:00:00
# #
portal local-server load flash:/portalpage.zip portal local-server load flash:/portalpage.zip
# #
drop illegal-mac alarm drop illegal-mac alarm
# #
wlan ac-global carrier id other ac id 0 wlan ac-global carrier id other ac id 0
# #
set cpu-usage threshold 80 restore 75 set cpu-usage threshold 80 restore 75
# #
ip vpn-instance vpnaa ip vpn-instance vpnaa
ipv4-family ipv4-family
route-distinguisher 100:1 route-distinguisher 200:1
vpn-target 111:1 export-extcommunity vpn-target 111:1 export-extcommunity
vpn-target 111:1 import-extcommunity vpn-target 111:1 import-extcommunity
# #
ip vpn-instance vpnbb ip vpn-instance vpnbb
ipv4-family ipv4-family
route-distinguisher 100:2 route-distinguisher 200:2
vpn-target 222:2 export-extcommunity vpn-target 222:2 export-extcommunity
vpn-target 222:2 import-extcommunity vpn-target 222:2 import-extcommunity
# #
mpls lsr-id 10.0.3.3 mpls lsr-id 10.0.5.5
mpls mpls
mpls te mpls te
mpls rsvp-te mpls rsvp-te
mpls te cspf mpls te cspf
# #
mpls ldp mpls ldp
# #
# #
acl number 2001 acl number 2001
rule 5 permit source 10.0.20.0 0.0.0.255 rule 5 permit source 10.0.40.0 0.0.0.255
# #
traffic classifier pe operator or traffic classifier pe operator or
if-match any if-match any
# #
traffic behavior pe traffic behavior pe
car cir 15000 pir 20000 cbs 300000 pbs 500000 car cir 15000 pir 20000 cbs 300000 pbs 500000
green pass yellow pass red discard green pass yellow pass red discard
90
# #
traffic policy pe traffic policy pe
classifier pe behavior pe classifier pe behavior pe
# #
aaa aaa
authentication-scheme default authentication-scheme default
authorization-scheme default authorization-scheme default
accounting-scheme default accounting-scheme default
domain default domain default
domain default_admin domain default_admin
local-user admin password cipher %$% local-user admin password cipher %$%
$K8m.Nt84DZ}e#<0`8bmE3Uw}%$%$ $K8m.Nt84DZ}e#<0`8bmE3Uw}%$%$
local-user admin service-type http local-user admin service-type http
# #
firewall zone Local firewall zone Local
priority 15 priority 15
# #
interface GigabitEthernet0/0/0 interface GigabitEthernet0/0/0
ip address 10.0.11.2 255.255.255.0 ip address 10.0.15.2 255.255.255.0
mpls traffic-policy pe inbound
mpls te mpls
mpls te bandwidth max-reservable-bandwidth mpls te
100000 mpls te bandwidth max-reservable-bandwidth
mpls te bandwidth bc0 100000 100000
mpls rsvp-te mpls te bandwidth bc0 100000
mpls ldp mpls rsvp-te
# mpls ldp
interface GigabitEthernet0/0/1 #
ip address 10.0.10.1 255.255.255.0 interface GigabitEthernet0/0/1
mpls ip address 10.0.17.1 255.255.255.0
mpls te mpls
mpls ldp mpls rsvp-te
# mpls ldp
ip binding vpn-instance vpnaa interface GigabitEthernet0/0/2
ip address 10.0.20.1 255.255.255.0 ip binding vpn-instance vpnaa
traffic-policy pe inbound ip address 10.0.20.1 255.255.255.0
# #
ip binding vpn-instance vpnbb ip binding vpn-instance vpnbb
# #
interface LoopBack1 interface LoopBack1
# #
interface Tunnel0/0/0 interface Tunnel0/0/0
91
ip address unnumbered interface LoopBack1 ip address unnumbered interface LoopBack1

tunnel-protocol mpls te tunnel-protocol mpls te
destination 10.0.5.5 destination 10.0.3.3
mpls te tunnel-id 100 mpls te tunnel-id 100
mpls te bandwidth ct0 200000 mpls te bandwidth ct0 200000
mpls te commit mpls te commit
# #
bgp 100 bgp 100
peer 10.0.4.4 as-number 100 peer 10.0.3.3 as-number 100
peer 10.0.4.4 connect-interface LoopBack1 peer 10.0.3.3 connect-interface LoopBack1
# #
ipv4-family unicast ipv4-family unicast
undo synchronization undo synchronization
peer 10.0.4.4 enable peer 10.0.3.3 enable
# #
ipv4-family vpnv4 ipv4-family vpnv4
policy vpn-target policy vpn-target
# #
ipv4-family vpn-instance vpnaa ipv4-family vpn-instance vpnaa
import-route direct import-route direct
# #
ipv4-family vpn-instance vpnbb ipv4-family vpn-instance vpnbb
# #
ospf 1 ospf 1
opaque-capability enable opaque-capability enable
area 0.0.0.0 area 0.0.0.0
network 10.0.3.3 0.0.0.0 network 10.0.5.5 0.0.0.0
network 10.0.10.0 0.0.0.255 network 10.0.15.0 0.0.0.255
network 10.0.11.0 0.0.0.255 network 10.0.17.0 0.0.0.255
mpls-te enable mpls-te enable
# #
return return
<P1>display current-configuration <P2>display current-configuration
[V200R003C00] [V200R003C00]
# #
sysname P1 sysname P2
# #
snmp-agent local-engineid snmp-agent local-engineid
92
800007DB03000000000000 800007DB03000000000000
snmp-agent snmp-agent
# #
# #
# #
# #
# #
# #
mpls mpls
mpls te mpls te
# mpls te cspf
mpls ldp #
# mpls ldp
# #
aaa #
authentication-scheme default aaa
authorization-scheme default authentication-scheme default
accounting-scheme default authorization-scheme default
domain default accounting-scheme default
domain default_admin domain default
local-user admin password cipher %$% domain default_admin
$K8m.Nt84DZ}e#<0`8bmE3Uw}%$%$ local-user admin password cipher %$%
local-user admin service-type http $K8m.Nt84DZ}e#<0`8bmE3Uw}%$%$
# local-user admin service-type http
firewall zone Local #
priority 15 firewall zone Local
# priority 15
mpls ip address 10.0.13.2 255.255.255.0
mpls te mpls
# mpls rsvp-te
mpls ip address 10.0.16.2 255.255.255.0
mpls te mpls
93

# mpls rsvp-te
mpls ip address 10.0.12.1 255.255.255.0
mpls te mpls
# mpls rsvp-te
interface NULL0 #
# interface NULL0
interface LoopBack1 #
ip address 10.0.1.1 255.255.255.255 interface LoopBack1
# ip address 10.0.2.2 255.255.255.255
ospf 1 #
opaque-capability enable ospf 1
area 0.0.0.0 opaque-capability enable
network 10.0.1.1 0.0.0.0 area 0.0.0.0
network 10.0.11.0 0.0.0.255 network 10.0.2.2 0.0.0.0
network 10.0.13.0 0.0.0.255 network 10.0.12.0 0.0.0.255
network 10.0.15.0 0.0.0.255 network 10.0.13.0 0.0.0.255
mpls-te enable network 10.0.16.0 0.0.0.255
# #
Return return
<PE2>display current-configuration <PE4>display current-configuration

[V200R003C00] [V200R003C00]
# #
sysname PE2 sysname PE4
# #
800007DB03000000000000 800007DB03000000000000
# #
# #
# #
# #
# #
# #
mpls mpls
mpls te mpls te
94
mpls te cspf mpls te cspf

# #
mpls ldp mpls ldp
# #
# #
aaa aaa
# #
# #
mpls mpls
mpls te mpls te
mpls te bandwidth max-reservable-bandwidth mpls te bandwidth max-reservable-bandwidth
100000 100000
mpls te bandwidth bc0 100000 mpls te bandwidth bc0 100000
# #
mpls mpls
mpls te mpls te
mpls te bandwidth max-reservable-bandwidth mpls te bandwidth max-reservable-bandwidth
100000 100000
mpls te bandwidth bc0 100000 mpls te bandwidth bc0 100000
# #
# #
interface NULL0 interface NULL0
# #
interface LoopBack1 interface LoopBack1
# #
bgp 100 bgp 100
95
# #
# #
ipv4-family vpnv4 ipv4-family vpnv4
policy vpn-target policy vpn-target
# #
ospf 1 ospf 1
opaque-capability enable opaque-capability enable
area 0.0.0.0 area 0.0.0.0
network 10.0.4.4 0.0.0.0 network 10.0.4.4 0.0.0.0
network 10.0.10.0 0.0.0.255 network 10.0.10.0 0.0.0.255
network 10.0.12.0 0.0.0.255 network 10.0.12.0 0.0.0.255
# #
Return return
<CE1>display current-configuration <CE3>display current-configuration

[V200R003C00] [V200R003C00]
# #
sysname CE1 sysname CE3
# #
board add 0/10 4GET board add 0/10 4GET
# #
800007DB03000000000000 800007DB03000000000000
# #
# #
# #
# #
# #
# #
# #
96

rule 0 permit udp destination-port eq dns rule 0 permit udp destination-port eq dns
rule 1 permit udp destination-port eq snmp rule 1 permit udp destination-port eq snmp
rule 2 permit udp destination-port eq snmptrap rule 2 permit udp destination-port eq snmptrap
rule 3 permit udp destination-port eq syslog rule 3 permit udp destination-port eq syslog
rule 4 permit udp rule 4 permit udp
# #
traffic classifier a operator or traffic classifier udp-limit operator or
if-match acl 2001 if-match acl 3001
traffic classifier udplimit operator or traffic classifier udp-limit1 operator or
traffic classifier udplimit1 operator or traffic classifier a operator or
traffic classifier c operator or traffic classifier c operator or
traffic classifier b operator or traffic classifier b operator or
# #
traffic behavior e traffic behavior udp-limit
remark dscp cs5 traffic behavior udp-limit1
car cir 10000 cbs 150000 pbs 500000 green pass car cir 5000 cbs 100000 pbs 150000 green pass
yellow pass red discard yellow discard red discard
traffic behavior udplimit traffic behavior e
traffic behavior udplimit1 remark dscp cs5
car cir 5000 cbs 100000 pbs 150000 green pass car cir 10000 cbs 150000 pbs 300000 green pass
yellow discard red discard yellow pass red discard
traffic behavior g traffic behavior g
car cir 2000 cbs 100000 pbs 500000 green pass remark dscp default
yellow pass red discard car cir 2000 cbs 100000 pbs 300000 green pass
remark dscp default yellow pass red discard
traffic behavior f traffic behavior f
car cir 5000 cbs 100000 pbs 500000 green pass remark dscp af31
yellow pass red discard car cir 5000 cbs 100000 pbs 300000 green pass
remark dscp af31 yellow pass red discard
# #
traffic policy udplimit traffic policy udp-limit
classifier udplimit behavior udplimit classifier udp-limit behavior udp-limit
classifier udplimit1 behavior udplimit1 classifier udp-limit1 behavior udp-limit1
traffic policy 3 traffic policy 3
classifier c behavior g classifier c behavior g
classifier b behavior f classifier b behavior f
classifier a behavior e classifier a behavior e
# #
aaa aaa
97

# #
# #
traffic-policy udplimit outbound traffic-policy udp-limit outbound
# #
traffic-policy 1 inbound traffic-policy 1 inbound
# #
# #
traffic-policy udplimit outbound #
#
bgp 65410 bgp 65430
# #
# #
Return return
<CE2>display current-configuration <CE4>display current-configuration

[V200R003C00] [V200R003C00]
# #
sysname CE2 sysname CE4
# #
800007DB03000000000000 800007DB03000000000000
# #
# #
# #
98
# #
# #
# #
aaa aaa
# #
# #
# #
bgp 65420 bgp 65440
# #
# #
return return
BGP MPLS VPN service Customer satisfaction questionnaires
From: Anwar Kade (Federal TVET Institute) To: ____________________
99
Please fill out the provided questionnaire so that we may better understand the VPN service
that you are using. The more information you are able to provide about your VPN services
and the services that Ethio Telecom offered, the better understanding about the quality of
your VPN performance will be. Please answer all questions as completely as possible then
submit the completed form to us directly.
Company Name
Contact information.
Job
Name Title
Phone Fax Email
Address
Q1. How satisfied or dissatisfied are you the overall quality of VPN service?
Very satisfied Nether satisfied nor dissatisfied Dissatisfied Vary dissatisfied
Q2. How satisfied or dissatisfied are you when you download file or video /music?
Q3. How satisfied or dissatisfied are you when you upload file or video/music?
Q4. The browsing delay when you access website is:
Very low Low Moderate High Very high
100
Q5. How satisfied or dissatisfied are you with the quality of video?(in YouTube or
video chatting)?
Q6. Choose two locations where you can get the best VPN connection speed
Inside office outside office when walking inside bus/taxi/car
Q7. Choose two locations where you can get the worst VPN connection speed
Inside office outside office when walking inside bus/taxi/car
Q8. Choose two time period in which you can get the best VPN connection speed
Morning Afternoon Evening Lunchtime Night
Q9. Choose two time period in which you can get the worst VPN connection speed
Morning Afternoon Evening Lunchtime Night
Q10. If your answer is yes for the above question, why do you choose it?
To get faster download for file/videos To get faster upload for file /videos
To save money To get better quality
101
Q11. If a new service provider or operator launches a new VPN service in Ethiopia,
will you switch to the new service provider?
Very likely Likely Unlikely Very unlikely
Q12. If you have ever called ethio telecom contact center for VPN service quality
problem, how do you rate the response?
Excellent Good Fair Ppoor Bad
Q13. How do you access VPN service?(you can choose more than one )
ADSL/VDSL EPON GPON DIRECT FIBER AIRONET VSAT
Q14. Your gender is:
Male Female
Q15. Your age range:
Under 18 18-24 25-34 35- 54 55+
Q16. Your educational background is:
Master’s degree and above Bachelor’s degree University or college
student Diploma or certificate Grade 11 or12 Below grade 10
Q17. Any other comment
102
103

Last Submission

Uploaded by

Document Information

Original Title

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

Last Submission

Uploaded by

Copyright:

Available Formats

Federal Technical and Vocational Education and Training Institute

Faculty of Electrical Electronics Department of Information and

Communication Technology Management

Enhancing end-to- end Quality of Service of Border Gateway Protocol Multiprotocol

Label Switching Virtual Private Network of EthioTelecom Service Level Agreements

Dr. Sreenivas Nune

Thesis Submitted to the Faculty of Electrical Electronics in Partial Fulfillment of the

Requirements for the Degree of Master of Science in Information and Communication

Enhancing end- to- end Quality of Service of Border Gateway Protocol

Multiprotocol Label Switching Virtual Private Network of EthioTelecom

Service Level Agreements

of Master of Science in Information and Communication Technology Management

Faculty of Electrical Electronics Department of Information and Communication

Thesis Examination Committee:

Full Name Signature Date

Addis Ababa, Ethiopia

work have been duly acknowledged.

Anwar Kade Hassen

Full Name of Student

Dr. Sreenivas Nune

Full Name of Advisor

Addis Ababa Ethiopia

in this thesis would be impossible to be achieved.

am especially indebted to his support.

1.2 Statement of the Problem....................................................................................20

1.3.1 General Objective....................................................................................22

1.3.2 Specific Objectives...................................................................................22

1.4.1 General Approach and specific research method.................................22

1.4.2 Data type and source...............................................................................23

1.4.3 Sample size and sampling Techniques...................................................23

1.4.4 Instruments and procedures, Process, and Analysis............................24

1.4.5 Design and Evolution Procedures..........................................................25

1.5 The significance of the study...............................................................................25

1.7 Scope and Limitations.........................................................................................26

1.7.1 The scope of the Study.............................................................................26

1.7.2 Limitations of the Study..........................................................................26

1.8 Organization of the Thesis..................................................................................27

2 Review of Literature and Related Works.................................................................28

2.1 Review of Literature............................................................................................28

2.1.1 Border Gateway Protocol (BGP)............................................................29

2.1.2 Multi-Protocol Label Switching.............................................................30

2.1.3 Multiprotocol Label Switching and Virtual Private Network.............33

2.1.4 MP BGP MPLS VPN...............................................................................35

2.1.5 Quality of Service (QoS)..........................................................................36

2.1.6 QoS Models...............................................................................................42

2.1.7 DiffServ QoS Implementation over MPLS VPN..................................44

2.1.8 Congestion Management Mechanisms..................................................48

2.1.9 Congestion Avoidance Mechanisms.......................................................50

2.2 Related Works......................................................................................................51

3 Proposed Network Architectures..............................................................................54

3.2 Designed BGP MPLS VPN.................................................................................55

3.2.1 Network IP Address................................................................................55

3.2.3 Interior Gateway Protocol (IGP) Interconnection...............................57

3.2.4 MPLS and MP BGP Interconnection....................................................58

3.2.5 Configuring MPLS RSVP-TE................................................................58

3.2.6 Configure VPN instances on PEs...........................................................60

3.3 Designed QoS of Proposed network architectures...........................................61

3.3.1 Define Access Control List rules............................................................62

3.3.2 Define traffic classifiers...........................................................................63

3.3.3 Define traffic behavior............................................................................64

3.3.4 Define and apply traffic policies.............................................................65

3.4 Applying the predefined policies to the inbound interfaces routers...............65

3.5 Experimental Results of Proposed Architecture..............................................66