di serta tie 2005

8/13/2019 Di Serta Tie 2005

1/120

DEVELOPMENT OF MPLS TEST-BED FOR NETWORK TRAFFIC

ENGINEERING

MOHD TAUFIK BIN JUSOH @ TAJUDIN

A thesis submitted in fulfilment of the requirements for the award of the Degree of

Master of Engineering (Electrical-Electronic & Telecommunication)

Faculty of Electrical Engineering

Universiti Teknologi Malaysia

30 NOVEMBER 2005

8/13/2019 Di Serta Tie 2005

2/120

PSZ 19 :16 (Pind. 1/97)

UNIVERSITI TEKNOLOGI MALAYSIA

BORANG PENGESAHAN STATUS TESIS

JUDUL: DEVELOPMENT OF MPLS TEST-BED FOR NETWORK TRAFFICENGINEERING

SESI PENGAJIAN : 2005/2006

Saya MOHD TAUFIK BIN JUSOH @ TAJUDIN__________________( HURUF BESAR )

mengaku membenarkan tesis *( PSM / Sarjana / Doktor Falsafah )* ini disimpan di PerpustakaanUniversiti Teknologi Malaysia dengan syarat-syarat kegunaannya seperti berikut:

1. Tesis adalah hak milik Universiti Teknologi Malaysia.

2. Perpustakaan Universiti Teknologi Malaysia dibenarkan membuat salinan untuk tujuan pengajiansahaja.

3. Perpustakaan dibenarkan membuat salinan tesis ini sebagai bahan pertukaran antara institusipengajian tinggi.

4. * * Sila tandakan ( )

Disahkan oleh

__________________________________ __________________________________

( TANDATANGAN PENULIS ) ( TANDATANGAN PENYELIA)

Alamat Tetap :

LOT 749 LORONG CHEMBONG KECIL SATU, PROF. DR. NORSHEILA BINTI FISAL

KAMPUNG CHEMBONG KECIL, Nama Penyelia

71300 REMBAU, NEGERI SEMBILAN.

Tarikh : 30 NOVEMBER 2005 Tarikh : 30 NOVEMBER 2005

SULIT

TERHAD

TIDAK TERHAD

(Mengandungi maklumat yang berdarjah keselamatanatau kepentingan Malaysia seperti yang termaktub di

dalam AKTA RAHSIA RASMI 1972)

(Mengandungi maklumat terhad yang telah ditentukan

oleh organisasi / badan di mana penyelidikan dijalankan)

CATATAN : * Potong yang tidak berkenaan** Jika tesis ini SULIT atau TERHAD, sila lampirkan surat daripada pihak

berkuasa/organisasi berkenaan dengan menyatakan sekali sebab dan tempohtesis ini perlu dikelaskan sebagai SULIT atau TERHAD.

*** Tesis dimaksudkan sebagai tesis bagi Ijazah Doktor Falsafah dan Sarjana secara

penyelidikan, atau disertai bagi pengajian secara kerja kursus ataupenyelidikan, atau Laporan Projek Sarjana Muda (PSM).

8/13/2019 Di Serta Tie 2005

3/120

I hereby declared that I have read this thesis and in my opinion it is sufficient in

term of scope and quality for the award of the degree of Master of Engineering

(Electrical-Electronic & Telecommunication)

Signature :

Supervisors Name : Prof. Dr. Norsheila binti Fisal

Date : 30 November 2005

8/13/2019 Di Serta Tie 2005

4/120

ii

I declare that this thesis entitled Development of MPLS Test-Bed for Network Traffic

Engineering is the result of my own research except as clearly cited in the references.

The thesis has not been accepted for any degree and is not concurrently submitted in

candidature of any other degree.

Signature : .

Name : Mohd Taufik Bin Jusoh @ Tajudin

Date : 30 November 2005

8/13/2019 Di Serta Tie 2005

5/120

iii

To

My beloved Parents, Sisters and Brothers

for Their

Love, Sacrifices and Best Wished

8/13/2019 Di Serta Tie 2005

6/120

iv

ACKNOWLEDGEMENT

In the name of ALLAH S.W.T, the Most Beneficent and the Merciful,

praise be to ALLAH S.W.T for the incredible gift endowed upon me and for the

health and strength given to me in order to finish the project and to prepare this

thesis.

First of all, I would like to thank my supervisor, Prof. Dr. Norsheila Binti

Fisal for her keen effort, interest, advice, consistent guidance and insightful

comments throughout the period of this study. I also would like to take this opportunityto thank Mr. Adel and all masters student at Telekom Lab, Faculty of Electrical

Engineering and also not to forget a lot of thanks to all technicians at Switching Lab

and Acoustic Lab. Without their helps, this project could not be done.

Special words of gratitude are dedicated to my classmate batch 04/05 especially

to Miss Anis Shahida Niza Binti Mohktar and Mr. Kamaru for their help and advice. A

lot of thank to my best buddy, Mr. Ismail Ibrahim and Mr. Jack for his moral support on

me. To my entire clique, I deeply appreciate all your helps in finishing this project.

Finally, extraordinary thanks and love to my beloved parents, siblings and

relatives for their full support, encouragement and love throughout my studies in UTM.

With them, this life is very meaningful.

8/13/2019 Di Serta Tie 2005

7/120

v

ABSTRACT

Providing Quality of Service (QoS) and Traffic Engineering (TE) capabilities in

the Internet is essential, especially in supporting the requirement of real-time traffic, as

well as mission critical applications. For that reason, the current Internet must be

enhanced with new technology that enables it to offer capabilities for controlling its

behaviour as needed. Multiprotocol Label Switching (MPLS) is an emerging technology

that provides QoS and traffic engineering features in IP network. This study is mainly

concerned on how to develop MPLS test-bed in order to assess MPLS functionalities.

The goal is to develop MPLS test-bed using Linux operating system with kernel version

2.6.5 as the platform. In this work the MPLS test-bed consists of four MPLS routers andtwo host terminals. MPLS software package version 1.946 has been used to build up

routers similar to existing routers in MPLS domain called as Label Switched Router

(LSR). Once the routers are well configured, the connection between two routers is

established to create Label Switched Path (LSP). This LSP connection is also used to

create new LSP from ingress router to egress router. IP packets are sent from ingress

router to host terminals to validate the test-bed. These packets were encapsulated with

MPLS header and they are examined by using tcpdump command in Linux terminal

which shows that the test-bed is successfully developed. In order to enhance test-bed

functionalities, packet generator can be added to the test-bed so that UDP and TCP

throughput measurement can be done. Besides, RSVP and Diffserv can be integrated

into the test-bed for future study.

8/13/2019 Di Serta Tie 2005

8/120

vi

ABSTRAK

Menyediakan keupayaan Kualiti Perkhidmatan (QoS) dan Kejuruteraan Trafik

(TE) dalam Internet amatlah penting, lebih-lebih lagi bagi menyokong keperluan

aplikasi masa nyata seperti aplikasi-aplikasi yang lebih kritikal. Oleh itu, teknologi

Internet masa kini perlulah diperluaskan dengan teknologi terbaru untuk membolehkan

ia menyediakan keupayaan yang diperlukan bagi mengawal tingkahlakunya sendiri

seperti yang diinginkan. Pensuisan Label Pelbagai Protokol (MPLS) adalah satu

teknologi yang sedang berkembang dimana ia menyediakan ciri-ciri QoS dan TE dalam

rangkaian Protokol Internet (IP). Kajian tesis ini tertumpu bagaimana membangunkan

rangkaian-uji MPLS (MPLS test-bed) bagi menilai fungsi-fungsinya. Dengan objektifmembangunkan rangkaian-uji MPLS, sistem operasi Linux dengan kernel versi 2.6.5

diperlukan sebagai asas pembangunan. Dalam kajian ini pakej perisian MPLS versi

1.946 telah digunakan untuk membangunkan routeryang berfungsi sama seperti router

yang terdapat dalam domain MPLS dan ia dipanggil sebagaiRouterPensuisan Label

(LSR). Setelah routersiap dibina, penyambungan antara dua routerdilakukan untuk

menyediakan Laluan Label Tersuis (LSP). Penyambung LSP ini juga turut digunakan

untuk membina satu LSP baru yang menghubungkan ingress router danegress router.

Paket-paket IP dihantar dari ingress router ke terminal hostbagi mengesahkan

rangkaian-uji MPLS. Paket-paket IP yang telah digabungkan dengan kepala MPLS ini

akan diuji menggunakan arahan tcpdump yang terdapat dalam terminal Linux untuk

menunjukkan bahawa rangkaian-uji ini telah siap dibangunkan. Sebagai langkah untuk

memperluaskan fungsi rangkaian-uji ini, penjana paket boleh diganding bersama.

Dengan itu, kualiti penghantaran TCP dan UDP boleh ditentukan. Selain itu, RSVP dan

Diffserv boleh juga digabungkan bersama untuk kajian dimasa akan datang.

8/13/2019 Di Serta Tie 2005

9/120

vii

TABLE OF CONTENTS

CHAPTER TITLE PAGE

TITLE OF PROJECT i

DECLARATION ii

DEDICATION iii

ACKNOWLEDGEMENT iv

ABSTRACT v

ABSTRAK vi

TABLE OF CONTENTS viiLIST OF FIGURES xi

LIST OF ABBREVIATIONS xiv

LIST OF APPENDICES xvi

CHAPTER I INTRODUCTION

1.1 Overview of Internet Challenges 1

1.2 Project Background 3

1.3 Objectives of Research 4

1.4 Research Scopes 4

1.5 Thesis Outline 5

1.6 Project Flow Chart 6

8/13/2019 Di Serta Tie 2005

10/120

viii

CHAPTER II LITERATURE REVIEW

2.1 Introduction 8

2.2 MPLS Basics 9

2.2.1 MPLS Architecture 11

2.3 MPLS Applications 15

2.3.1 Virtual Private Network (VPN) 16

2.3.2 IP and ATM Integration 16

2.4 Traffic Engineering with MPLS 17

2.5 MPLS QoS 202.6 MPLS Future Technology 22

2.7 Linux as the Operating System (OS) 25

2.7.1 Linux Differences 26

2.8 Details of MPLS Forwarding Basics

Implemented in Linux OS 27

2.8.1 MPLS Label 28

2.8.2 MPLS Label Stacking 29

2.8.3 MPLS Data Structures 30

2.8.3.1 Incoming Label Map (ILM) 30

2.8.3.2 The Next Hop Label

Forwarding Entry (NHLFE) 31

2.8.3.3 FEC-to-NHLFE Map (FTN) 31

2.9 MPLS Details 32

2.9.1 MPLS ILM Details 34

2.9.2 MPLS Receive Processing 35

2.9.3 MPLS FTN Details 36

2.9.4 MPLS NHLFE Details 37

2.9.5 MPLS Transmit Processing 37

2.9.10 MPLS Commands in Linux OS 38

8/13/2019 Di Serta Tie 2005

11/120

ix

CHAPTER III DESIGN AND PROCEDURE

3.1 Introduction 40

3.2 Kernel Recompilation 40

3.2.1 Non-RPM System 41

3.2.2 RPM System 47

3.3 Kernel IP Forwarding 48

3.4 Ethernet Card Drivers 49

3.5 Assigning Ethernet Device with IP Address 50

3.5.1 ifconfig 513.5.2 ip route 51

3.5.3 Assigning IP Address in the

Test-bed 52

3.6 Enabling MPLS 54

3.7 Label Switching Path Set Up 55

3.7.1 LSP R-4 to R-1 57

3.7.2 LSP R-1 to R-4 59

3.7.3 LSP R-1 to R-2 60

3.7.4 LSP R-2 to R-1 62

3.7.5 LSP R-1 to R-3 62

3.7.6 LSP R-3 to R-1 64

3.7.7 LSP R-2 to R-3 64

3.7.8 LSP R-3 to R-2 66

3.8 Label Switched Path 1 68


CHAPTER IV RESULT AND ANALYSIS

4.1 Introduction 71

4.2 Result of LSP Set Up Between Two Routers 71

4.2.1 LSP between R-4 to R-1 73

8/13/2019 Di Serta Tie 2005

12/120

x

4.2.2 LSP between R-1 and R-2 74



4.3 Result of LSP 1 77

4.4 Result of LSP 2 81

4.5 Round Trip Time (RTT) 86

CHAPTER V CONCLUSIONS AND FUTURE WORKS

5.1 Conclusions 895.2 Proposed Future Works 90

REFERENCES 92

APPENDIX A 94

APPENDIX B 96

APPENDIX C 99

APPENDIX D 101

8/13/2019 Di Serta Tie 2005

13/120

xi

LIST OF FIGURES

FIGURE NO TITLE PAGE

2.1 MPLS in OSI Model 12

2.2 MPLS Architecture 13

2.3 The format of LDP messages 14

2.4 Shim Label 14

2.5 MPLS Domain 19

2.6 MPLS main structure 332.7 ILM structure 34

2.8 FTN structure 36

2.9 NHLFE structure 37

2.10 MPLS commands in Linux terminal 39

3.1 Kernel option before compilation; no MPLS option 43

3.2 Kernel option after compilation; MPLS is selected 44

3.3 Kernel options before kernel compilation; no Special

Next Hop option 45

3.4 Kernel option after compilation; Special Next Hop is

selected 46

3.5 Test-bed 52

3.6 The connection between R-4 and R-1 57

3.7 Label 36 is assigned for network 10.0.5.0/24 58

8/13/2019 Di Serta Tie 2005

14/120

xii









4.1 LSP between R-4 and R-1 73

4.2 LSP between R-1 and R-2 744.3 LSP between R-1 and R-3 75

4.4 LSP between R-2 and R-3 76

4.5 Pinging 10.0.5.2 (sending IP packets into the MPLS

domain) 77

4.6 Packets addressed to 10.0.5.2 are attached with label

36 at Ingress LSR 78

4.7 Packets labeled with 36 are received by eth0 at R-1 78

4.8 Packets labeled with 36 are swapped with label 26 at

R-1 79


4.10 Label 26 is removed at R-3, IP packets are

forwarded to destination 80

4.11 IP packets are received at destination without MPLS

header 80

4.12 Pinging 10.0.6.2 (sending IP packets into the MPLS

domain) 81

4.13 Packets addressed to 10.0.6.2 are attached with label

56 at Ingress LSR 82

4.14 Packets with label 56 are received by eth0 at R-1 82

4.15 Label 56 is swapped with label 18 at R-1 83

4.16 Packets with label 18 are received by eth0 at R-2 84

8/13/2019 Di Serta Tie 2005

15/120

xiii

4.17 Packets with label 18 are swapped with label 20 at R-2 84


4.19 Label 20 is removed at R-3, IP packets are

forwarded to destination 85

4.20 IP packets are received at destination without MPLS

header 86

4.21 RTT for LSP 1 87

4.22 RTT for LSP 2 87

4.23 Comparison of RTT between LSP2 and LSP1 88

8/13/2019 Di Serta Tie 2005

16/120

xiv

LIST OF ABBREVIATIONS

ATM - Asynchronous Transfer Mode

CoS - Class of Service

CR-LDP - Constraint-based Routing Label Distribution

Protocol

DLCI - Data Link Connection Identifier

DWDM - Dense Wavelength Division Multiplexing

FEC - Forwarding Equivalent Class

FSF - Free Software Foundation

GMPLS - Generalized Multiprotocol Label SwitchingGPL - General Public License

IETF - Internet Engineering Task Force

IGP - Interior Gateway Protocol

IntServ - Integrated Services

IP - Internet Protocol

IS-IS - Intermediate System-Intermediate System

LDP - Label Distribution Protocol

LER - Label Edge Router

L-LSP - Label inferred Label Switched Path

LMP - Link Management Protocol

LSR - Label Switching Router

LSP - Label Switched Path

MEM - Micro-Electric Mechanical Systems

8/13/2019 Di Serta Tie 2005

17/120

xv

MPLS - Multiprotocol Label Switching

MPS - Multiprotocol Lambda Switching

OADM - Optical Add-Drop Multiplexer

OS - Operating System

OSI - Open System Interconnection

OSPF - Open Shortest Path First

OXC - Optical Cross-Connects

PXC - Photonic Cross-Connect

PVC - Permanent Virtual Circuits

PC - Personal Computer QoS - Quality of Service

RSVP - Resource Reservation Protocol

RSVP-TE - Reservation Protocol with Traffic Engineering

RTT - Round Trip Time

SONET - Synchronous Optical Network

TDM - Time Division Multiplexing

TE - Traffic Engineering

TLV - Type Length Value

TTL - Time to Live

VCI - Virtual Circuit Identifier

VoIP - Voice over IP

VPI - Virtual Path Identifier

VPN - Virtual Private Network

WAN - Wide Area Network

WWW World Wide Web

8/13/2019 Di Serta Tie 2005

18/120

xvi

LIST OF APPENDICES

APPENDIX TITLE PAGE

A Kernel Script for R-4 94

B Kernel Script for R-1 96

C Kernel Script for R-2 99

D Kernel Script for R-3 101

8/13/2019 Di Serta Tie 2005

19/120

CHAPTER I

1.1 Overview of Internet Challenges

The overwhelming growth of the Internet and the growing popularity of real-time

applications set new challenges to the Internet community. Big Internet service

providers are growing ever larger and supporting increasingly a variety of services, with

different requirements both from different applications and their customers. Service

providers need for a commercially viable and scalable tools to make the most of their

networks in order to increase their revenues by supporting the needs of time or andmission critical applications [24]. This is because different applications have varying

needs for delay, delay variation (jitters), bandwidth, and packet loss.

For example, real-time applications such as Voice over IP (VoIP) and video

conferencing are extremely latency-dependent. Here, the timeliness of data delivery is an

issue of utmost importance. But in the Internet, where there is no predictable traffic

control, these applications do not run effectively. Service differentiation for traffic flows

and performance optimization of the operational networks are very critical for the

Internet to remain as successful as before. However, the Internet, particularly its core

protocol IP (Internet Protocol) was never designed with Quality of Service (QoS) in

mind. Instead it was originally designed as a research and educational resource, and thus

the underlying technology that forms the backbone of today's Internet is largely based on

8/13/2019 Di Serta Tie 2005

20/120

2

that philosophy [24]. But times have changed a lot since, and service differentiation

using QoS mechanisms while optimizing the operational network has become quite an

important issue in the Internet for these applications to run effectively and efficiently.

On the other hand, as the Internet is required to support different types of

services, effective and efficient bandwidth management tools in IP networks becomes

increasingly important, especially when dealing with how to allocate the available

network resources in order to optimize the overall performance of the networks. And

yet, when the network has to sustain heavy traffic load, and has limited resources, thesituation of having some congested links, while others remain underutilized is almost an

inevitable phenomenon [24]. One of the main reasons to cause such congestion events in

IP networks is that of the destination based forwarding paradigm.

In IP networks, the Interior Gateway Protocols (IGPs), such as Open Shortest

Path First (OSPF), and Intermediate System-Intermediate System (IS-IS) routing

protocols use destination-based forwarding algorithm, without considering other

network parameters, such as the available bandwidth. In effect, all traffic between any

two nodes traverses across the IGP shortest path. Hence, it is obvious that such situation

can create hot spots on the shortest distance between two points, while other alternative

routes may still be underutilized. As a result, degradation of throughput, and long delay,

and packet losses can be noticed. In such situation, minimizing the effects of congestion

by optimizing the performance of the operational networks becomes more critical.

Traffic Engineering (TE) is very important in this regard [11], and plays a key

role in that it offers service providers a means for performance optimization and

bandwidth provisioning. In fact, without TE, it is also difficult to support QoS on a large

scale and at reasonable cost [12]. Therefore, the key to address the problem of traffic

engineering is to have the ability to place the traffic onto the network as flexibly as

8/13/2019 Di Serta Tie 2005

21/120

3

needed, so that the congestion in the network can be minimized before it leads to poor

network performance. Because minimizing congestion by optimizing the distribution of

traffic on a given network is the central goal of TE [13].

In order to address the QoS issue, however, the ability to introduce connection-

oriented forwarding techniques to connectionless IP networks becomes necessary. In

effect, this allows IP networks to reserve resources, such as bandwidth over

predetermined paths for service differentiation in order to provide QoS guarantees.

1.2 Project Background

Multiprotocol Label Switching (MPLS) is an emerging technology, which plays

a key role in IP networks by delivering QoS, as well as traffic engineering features.

MPLS has been developed and standardized by the Internet Engineering Task Force

(IETF) to address these issues, in a more scalable and cost effective way.

A lot of research has been done on MPLS performance study, and MLPS does

provide QoS and TE. In University Teknologi Malaysia, MPLS test-bed not yet been

developed and used. Hence, this project is prominent since it would be a pioneer in

MPLS study in our university.

8/13/2019 Di Serta Tie 2005

22/120

4

1.3 Objectives of Research

Internet nowadays is facing a lot of challenges due to the gigantic numbers of

users and fast growing of Internet applications. One way to overcome this problem is to

provide QoS. However QoS cannot be achieved without having TE along. Due to these

challenges, this project aim is to develop MPLS test-bed for network traffic engineering

and to preserve QoS through explicit routing in Internet using MPLS.

1.4 Research Scopes

In order to develop MPLS test-bed, the components in MPLS domain itself need

to be set up previously. For this project, four routers are needed to provide alternative

routes between routers in MPLS network. Routers in MPLS test-bed which are able to

forward IP packets in MPLS domain are called Label Switching Router (LSR). The

routers basically based on Linux operating system since this operating system is open-

source system, and the software to turn ordinary Linux PC into LSR are freely available

in the Internet. Other scopes of this project are to enable the connection between routers

and to create Label Switched Path (LSP). Finally IP packets passing through routers

using LSP are validated using Linux command.

8/13/2019 Di Serta Tie 2005

23/120

5

1.5 Thesis Outline

The contents of the project are further subdivided into chapters that described in

details. Chapter I discussed about the overviews of current challenges in Internet

technology. The factors why core network must provide new technology and must be

improved are also specified in details. This chapter also introduced project background,

objective of the research and scopes that are involved in order to finish the project. The

overall projects process flow chart also shown.

The Chapter II discussed literature studies and related works. The basic

fundamentals and architecture of MPLS are explained in details. Other MPLS

applications such as Virtual Private Network (VPN) and Asynchronous Transfer Mode

(ATM) with IP integration also discussed. How MPLS provides Traffic Engineering

(TE), Quality of service (QoS) and Class of Service (CoS) are elaborated in this chapter.

Moreover, operating system used in this project is explained briefly. Finally the

implementations of MPLS in Linux are explained in details such as its data structures

and the processes happened when Linux machines run MPLS.

Chapter III explained about design and procedures done for this project. The

steps to configure Linux machine into MPLS router are also explained. Label Switched

Path (LSP) establishment between routers are described step by step for easy

understanding.

In Chapter IV, results are shown and discussed. This chapter covers the results of

LSP set up between two routers and the results of LSP between ingress and ingress

router. Finally in this chapter, the results of Round Trip Time (RTT) for the test-bed are

shown.

8/13/2019 Di Serta Tie 2005

24/120

6

Lastly in Chapter V, the conclusions of this project are described. Some futureworks that can be done in order to enhance are explained.

1.6 Project Flow Chart

Literature reviews on MPLS

Upgrade knowledge in using LINUX

Progress of this study can be summarized into flow chart shown above. First of

all, literature reviews on MPLS need to be done to acquire basic knowledge about

MPLS. Journals and papers were referred to enhance the understanding and to get the

overview of the project. Since this project requires Linux operating system as the

MPLS software

Recompile Kernel

Install Binary Files

Enable Kernel IP Forwarding

Build Ethernet Driver

Assign IP Address

Develop Test-bed

Turn On MPLS

LSP set up between 2 PCs

Create LSP1 & LSP 2

Test-bed Testing

8/13/2019 Di Serta Tie 2005

25/120

7

platform, the knowledge using Linux must be enhanced. The understanding of Linux

critical scripts that exists in Linux kernel is also very important. MPLS software is taken

from Sourceforge homepage. This software package is still under development thus

MPLS package version 1.946 is chosen since it is the latest up date package. Before

Linux PC can be change into MPLS router, the kernel need to be compiled. The

appropriate binary files are installed in order to make MPLS router to function.

Furthermore, kernel IP forwarding in Linux machine has to be change into

enable state since it is disabled by default. Linux machines are connected together withappropriate IP addresses. But this can only be done when the ethernet cards driver is

installed in every Linux machine. Before LSP between two routers can be established,

each router must be MPLS enabled by turning on MPLS. The connection between two

routers is used again to create new LSP but now it connects ingress router to egress

router.

In this study, two LSP that are LSP 1 and LSP 2 which connects ingress to egress

are created. LSP 2 connects four routers while LSP 1 connects only three routers. The

test-bed is tested by sending IP packet into it. When the packets captured were

encapsulated with MPLS header it shown that the MPLS test-bed was well developed.

8/13/2019 Di Serta Tie 2005

26/120

CHAPTER II

LITERATURE REVIEW

2.1 Introduction

Due to the increasingly numbers of personal computers that are connected to the

Internet and the popularity of the World Wide Web (WWW), Internet use has gone

beyond all expectations. The adoption rate of the Internet has surpassed all other

preceding technologies such as telephone, radio, television, and computer. This

phenomenal growth has made the Internet Protocol (IP) suite the most predominant

networking technology. With the rapid growth of Internet services and the recent

advances in Dense Wavelength Division Multiplexing (DWDM) technology, an

alternative to Asynchronous Transfer Mode (ATM) for multiplexing multiple services

over individual circuits is needed. The once fast and high-bandwidth ATM switches are

not good enough because Internet backbone routers are outperforming them [10].

Multiprotocol Label Switching (MPLS) offers a simpler mechanism for packet-

oriented traffic engineering. The term multiprotocol implies that its techniques are

8/13/2019 Di Serta Tie 2005

27/120

9

applied to any network layer protocol; however, in this discussion IP is regard as the

network layer protocol.

MPLS is an extension to the existing IP architecture. It is the latest step in the

evolution of routing and forwarding technology for the Internet core. It is a new

technology being standardized by the Internet Engineering Task Force (IETF) designed

to enhance the speed, scalability, and service provisioning capabilities in the Internet. As

a technology for backbone networks, MPLS can be used for IP as well as other network-

layer protocols. It has become the prime candidate for IP-over-ATM backbone networks[10].

The next part of this thesis presents the basic features and architecture of MPLS. It

also discussed its popular applications such as Virtual Private Network (VPN), Traffic

Engineering (TE), IP and ATM integration. It introduced the next technology of existing

MPLS that is Multiprotocol Lambda Switching or MPS. Finally, the implementations of

MPLS in Linux operating system that covers from all perspectives are discussed.

2.2 MPLS Basics

Cisco has been working within the IETF to develop a standard for invented tag

switching since they first shipped it in 1998 [10]. That standard is MPLS. Thus, tag

switching is a pre-standard implementation of MPLS. MPLS provides a solid

framework that supports the deployment of advanced routing services because it solves

a number of complex problems. It addresses the scalability issues associated with the

8/13/2019 Di Serta Tie 2005

28/120

10

currently deployed IP-over-ATM overlay model and significantly reduces the

complexity of network operation [10].

MPLS is a means of forwarding information at a very high rate. It combines the

speed and performance of Layer 2 with the scalability and IP intelligence of Layer 3. It

separates an IP router's function into two parts; forwarding and control. It is also a

collection of distributed control protocols used in setting up paths in IP networks.

MPLS provides a mean to map IP addresses to simple, fixed-length, protocol-

specific identifiers known as label (i.e., it separates forwarding information [label]from the content of the IP header). MPLS also uses a single-forwarding paradigm

(label swapping) at the data plane to support multiple-routing paradigms at the

control plane. Furthermore, it eliminates the need for interrogating the IP header of

every packet at every intermediate router [10].

MPLS enables constraint-based routing, which allows the distribution of the

traffic load on alternate routes when the shortest path is over loaded. It is an

integration of Layer 2 and Layer 3 technologies. By making conventional Layer 2

features available to Layer 3, MPLS enables traffic engineering and independent of the

underlying link layer technology. It uses different technologies and link layer

mechanisms to realize the label swapping forwarding paradigm. In addition, MPLS

supports the IP, ATM, and frame relay Layer 2 protocols [10].

Moreover, MPLS provides flexibility in the delivery of new routing services.

It supports the delivery of services with QoS guarantees and thereby facilitates voice,

video, and data service integration. It facilitates scalability through traffic aggregation

and allows the construction of VPN in IP networks. Furthermore, MPLS offers a

standards-based solution that promotes multi-vendor interoperability [10].

8/13/2019 Di Serta Tie 2005

29/120

11

The aim of MPLS is to improve the scalability and performance of the prevalent

hop-by-hop routine and forwarding across packet networks. Its primary goal is to

standardize a technology that integrates the label switching forwarding paradigm with

network layer routing.

2.2.1 MPLS Architecture

MPLS is a connection-oriented protocol that appears between the link layer and

the network layer in Open System Interconnection (OSI) model, as ATM or frame relay,

but without a dependence on the link layer. Figure 2.1 shows the MPLS in OSI model.

The concept of label switching is keys to MPLS. The key function of MPLS is to perform

traffic aggregation.

In a conventional IP network, streams of traffic between two points in the

network are divided into many IP packets with each packet having its own header; the

source address and destination address. Along the way, each router inspects every

packet header in order to route it. The less time routers spend inspecting packets, the

more time they have to forward them. It is a waste of resources to inspect every IP

packet header when transferring a large number of packets destined for the same

destination from the same source. MPLS recognizes such packets and adds a special

label identifying them as a unified flow of information. In other words, a label is a short,

fixed length, locally significant identifier used to identify a stream. Intermediate routers

would then see the labels and forward them rapidly toward their destination [10].

8/13/2019 Di Serta Tie 2005

30/120

12

TCP

Physical (Optical - Electrical) 1

2

IP 3

4

Applications7to

5

UDP

PPP FR ATM (*)

MPLS

TCP

Physical (Optical - Electrical) 1

2

IP 3

4

Applications7to

5

UDP

PPP FR ATM (*)

MPLS

Physical (Optical - Electrical) 1Physical (Optical - Electrical) 1

2

IP 3IP 3

4

Applications7to

5

Applications7to

5

UDP

PPP FR ATM (*)PPP FR ATM (*)

MPLS

Figure 2.1: MPLS in OSI Model

At the ingress point of an MPLS network, router called as Label Edge Router

(LER) adds a label to each incoming IP packet according to its destination and the state

of the network. The label is embedded in the Layer 2 protocol header of the packet. The

basic task of the routers is to forward packets as efficiently as possible from source to

destination. To achieve this, each router needs information on the topology and status of

the network, and where the egress of the respective destination is located. Label

switching relies on the setup of switched paths through the network known as Label

Switching Paths (LSPs). In other words, LSP is an established logical MPLS connection,

which links an LER via a LSR to another LER, as illustrated in Figure 2.2. A key feature of

MPLS is that one or more flows can be assigned to the same label or LSP. When a packet is

sent on an LSP, a label is applied to the packet. On ATM links, for example, the label

may be carried as the Virtual Circuit Identifier (VCI) or Virtual Path Identifier (VPI)

applied to each ATM cell. A similar scheme has been proposed for Synchronous Optical

Network (SONET).

In an MPLS network, an LSP is set up for each path through the network. The

router examines the header to determine which LSP to use before adds the appropriate

LSP identifier (a Layer 2 label) to the packet and forwards it to the next hop. All the

subsequent nodes forward the packet along the LSP identified by the label. LSPs

8/13/2019 Di Serta Tie 2005

31/120

13

typically follow the shortest path from source to destination. Once the traffic is within

the MPLS network, only the label is used to make decisions on the next hop where the

packet is to be sent, in contrast to traditional destination-based hop-by-hop forwarding

used in IP networks. Each label has only hop-by-hop relevance. This feature allows for

an efficient resource usage and provides high-speed forwarding.

Figure 2.2: MPLS architecture [10]

Hence, an MPLS network consists of MPLS nodes called LSRs connected by circuits

called LSPs. The main component of MPLS architecture is the Label Distribution Protocol

(LDP), which is a set of protocols by which an LSR communicates with an adjacent peer by

means of exchanging labels. To establish an LSP, a signaling protocol such as LDP is

needed. Each LSP has a Forwarding Equivalent Class (FEC) associated with it. The

classification of packets into specific FEC identifies the set of packets that will be mapped

to a path through the network. At the ingress of an MPLS domain, all incoming packets

will be assigned to one of these FECs. Once a packet gets to the end of the last egress edge

router, the labels are popped and the packet is routed according to the conventional or

traditional IP forwarding rules.

8/13/2019 Di Serta Tie 2005

32/120

14

As mentioned before, label distribution among the various LSRs is done by the

LDP. The LDP protocol has four types of messages serving different purposes [10];

i. Discovery messages: to advertise the presence of LSRs

ii. Session messages : to establish and maintain LDP sessions

iii. Advertisement messages : to create, change, and delete label mappings for FECs

iv. Notification messages : to carry advisory and error information

U

1 bits

Message

Type

15 bits

Message

Length

16 bits

Message ID

32 bits

Mandatory

Parameters

Variable bit

Optional

Parameters

Variable bit

Figure 2.3: The format of LDP messages

The format of each LDP message is presented in Figure 2.3, where U (1 bit)

denotes unknown messages and is used to notify the sender in case an LSR fails to

recognize a message. The 15-bit message-type field is used to identify an LSP as one of

the 10 defined types. The 16-bit message-length field specifies the total length of the

message in bytes. The 32-bit message ID uniquely identifies the message. Mandatory

and optional parameters employ a special coding called Type-Length-Value (TLV).

Within the IETF, two candidate protocols for the LDP are Resource Reservation Protocol

(RSVP) and Constraint-based Routing Label Distribution Protocol (CR-LDP).

Labels exist within some transport media that MPLS nodes can use in making

forwarding decisions. Examples include ATM's Virtual Path Identifier/Virtual Circuit

Identifier (VPI/VCI) field and frame relay's Data Link Connection Identifier (DLCI).

Other transport media, such as Ethernet and point-to-point links, must use a shim label.

8/13/2019 Di Serta Tie 2005

33/120

15

The MPLS generic shim label format is illustrated in Figure 2.4. It consists of the

following fields.

Data Link layer

Layer 2 Header

MPLS Shim Label IP Header

Layer 3 Header

Packet Data

Label

20 bits

EXP

3 bits

S

1 bits

TTL

8 bits

Figure 2.4: Shim Label

The 32 bits shim label is used to identify an FEC and the details of this shim

label are discussed in section 2.8.1.

2.3 MPLS Applications

Multiprotocol Label Switching can be used for the Internet Protocol (IP) or any

other network-layer protocols. It can be deployed in corporate networks as well as in

backbone networks. Its greatest strength is perhaps its seamless coexistence with IP

traffic and its reuse of proven IP routing protocols. It addresses today's network

backbone requirements by providing a standards-based solution that supports network

scalability, builds interoperable networks, and integrates IP and ATM.

8/13/2019 Di Serta Tie 2005

34/120

16

MPLS may be regarded as a general-purpose tunneling protocol that enables

several network services, which are distinguished by LDPs they used to build tunnels and

what kind of traffic flows through the tunnels. Popular applications of MPLS are

discussed next.

2.3.1 Virtual Private Network (VPN)

MPLS supports Virtual Private Network (VPN) services [10]. VPNs share a

single infrastructure of routers or switches between multiple independent networks.

Using MPLS for VPNs is an attractive alternative to building VPNs using either ATM or

frame relay Permanent Virtual Circuits (PVCs), or various forms of tunnels, to

interconnect routers at customer premises. MPLS based VPNs have some advantages

over a PVC based model. They allow customers to choose their own addressing plans,

and encryption is usually not required. They also enable the development of extranets

using flexible policies. An MPLS based VPN employs LSPs to provide tunnel-like

isolation. The key benefits of MPLS based VPNs include increased scalability, QoS

support, and implicit data security mechanism.

2.3.2 IP and ATM Integration

MPLS enables IP and ATM integration [10]. In fact, MPLS is primarily a solution

for IP-over-ATM backbone. MPLS enables an ATM switch to perform virtually all of the

functions of an IP router. This is made possible by the fact that MPLS label swapping is

8/13/2019 Di Serta Tie 2005

35/120

17

the same as the forwarding paradigm provided by ATM switch hardware. MPLS is the

Internet's best long term solution to efficient, high performance forwarding and traffic

differentiation. Although MPLS can operate over any data link layer, ATM is by far the

most suitable technology due to its ability to transport data, voice, and video with the offer

of a defined QoS. MPLS has some advantages over a typical IP-over-ATM overlay

model. First, it offers simpler integration of control information. It also avoids complex

mechanisms to provide mapping from IP multicast addresses to multipoint virtual

circuits. Second, the scalability of IP routing is improved relative to an overlay model.

Third, it makes routing at the IP layer more likely to be optimal.

The selling point for MPLS is its ability to support LSPs from edge to edge. This

allows sophisticated load balancing, QoS, and MPLS based VPNs to be deployed by

service providers. Although MPLS is a technology designed for IP based terrestrial

networks, its capabilities can be made for satellite constellations where extensive routing is

required. It should be mentioned that, as great as MPLS is, not every network needs it.

Only large networks such as Wide Area Networks (WANs) with complex mesh

topology must engineer traffic.

2.4 Traffic Engineering with MPLS

The challenge of traffic engineering is how to make the most effective use of the

available bandwidth in a large IP backbone networks, or place the traffic over them so

that we can achieve the best performance characteristics from the same IP networks even

in the event of congestion. MPLS addresses the traffic engineering issue by setting up

explicit paths through the network using constraint based routing.

8/13/2019 Di Serta Tie 2005

36/120

18

MPLS traffic engineering dynamically establishes and maintains an LSP tunnel

across the MPLS domain using signaling protocols. The two signaling mechanisms used

for distributing labels across an MPLS domain, in the context of traffic engineering and

QoS, are Constraint-based Routing Label Distribution Protocol (CR-LDP) [15], and

Resource Reservation Protocol with Traffic Engineering extension (RSVP-TE) [16].

Explicit routing or constraint-based routing is particularly interesting for traffic

engineering purpose.

The label distribution protocols (both CR-LDP and RSVP) determine the pathacross which the LSP tunnel is established based on its resource requirements and

available network resources, such as bandwidth. Then, they move the label binding

information along a predefined route. At the ingress, LSRs assign label to packets as

they enter the network. Here, the label binding to packets is done based on FEC

membership. This feature of MPLS allows scalable aggregation of flows into a single

FEC based on their requirements. In fact, it also makes the task very simple for MPLS

traffic engineering to route traffic flows across an LSP tunnel by associating the

resources required by a given FEC (or LSP) with actual backbone capacity and

topology.

To illustrate how explicit LSPs are used to solve the traffic engineering problem,

let consider the example shown in Figure 2.5. In this figure, LSR 1 through LSR 3 are

label switching routers, while Host A through Host D are traffic sources and sinks.

Assume 100 Mbps PVC connections among all routers. Further assume the traffic from

Host A to Host C is 100 Mbps and that of the traffic from Host B to Host D is 100

Mbps.

8/13/2019 Di Serta Tie 2005

37/120

19

Figure 2.5: MPLS Domain [24]

With MPLS explicit routing capability, the traffic from Host B to Host D can be

assigned to LSP 2, which in turn traverses the path LSR 1-LSR 2-LSR 3, while the

traffic from Host A to Host C can be made to travel across LSP 1, which consists of path

LSR 1-LSR 3. As a result, the traffic flows between hosts one side (in this case, Host A

and Host B) and hosts on the other side (Host C and Host D) can be distributed over the

network according to their demands, such as bandwidth guarantees, by just establishing

different LSPs. This enables to attain efficient utilization of bandwidth, as well as a

significant performance gain.

8/13/2019 Di Serta Tie 2005

38/120

20

2.5 MPLS QoS

The important issue to overcome here is the need for QoS capability in the

Internet, which arises from the fact that the current Internet offers just best effort packet

delivery service, and thus all packets are treated equally, regardless of the needs of

applications for some levels of resource assurance. This is mainly because IP on which

the Internet is based, was not typically designed as a connection oriented protocol.

Instead IP is providing a connectionless or datagram service, thus no end to end

guarantees of packet delivery can be provided. But in order to provide QoS guarantees inthe Internet, all packets of a flow must traverse the same path, and some means for

reserving resources along that path must exist. At least, in this way, the Internet can have

the ability to differentiate between high priority and lower priority network traffic in

order to ensure guaranteed quality of service.

MPLS provides a robust QoS control feature in the Internet. In addition, MPLS

Class of Service (CoS) feature can work in conjunction with other QoS architectures for

IP networks defined by the IETF. Integrated Services (IntServ) [17] using Resource

Reservation Protocol (RSVP) [18] and Differentiated Services (DiffServ) [19] are the

two models defined by the IETF for providing QoS in IP networks. Combining MPLS

with DiffServ is particularly interesting, and provides the required levels of end to end

QoS management in a scalable way. MPLS does support for DiffServ [20]. In fact, both

MPLS and DiffServ are the real enhancements to IP networks. However, they do not

make any requirements about each other, so they can work independently.

The QoS feature of MPLS represents the capability to provide differentiated

levels of service and resource assurance across an MPLS network. This capability

typically includes a set of techniques necessary to manage network bandwidth, delay,

jitter, and packet loss. For example, the ability to mark packets with a certain priority

8/13/2019 Di Serta Tie 2005

39/120

21

combined with buffer management and queuing schemes ensures that voice traffic

remains within the acceptable bounds for packet loss, delay, and jitter. So, in an MPLS

network, when a packet arrives at a LSR, the label is used to determine outbound

interface and new outbound label, but the CoS (EXP) field value is used to determine the

type of treatments, such as queuing and scheduling.

With MPLS QoS, there are two approaches to mark traffic for controlling QoS

within an MPLS network. That means, when IP traffic enters an LSP tunnel, the CoS

bits in the MPLS header are set in one of two ways. In the first way, queuinginformation is encoded into the experimental (EXP) field of the MPLS shim header.

Since the EXP field allows eight different CoS markings, the marking is used as packet's

CoS value. Here, different packets might receive different markings depending on their

requirements, so that they can receive different treatments along the path. This approach

is referred to as Experimental bit inferred Label Switched Paths (E-LSPs), to indicate

that QoS information is inferred from the EXP field.

In the second method, the label associated with MPLS packet specifies how a

packet should be treated, and all packets entering the LSP will be marked with a fixed

CoS value. This means that all packets entering the LSP receive the same Class of

Service. This approach is known as Label inferred Label Switched Paths (L-LSPs), to

indicate that QoS information is inferred from the MPLS label.

Either way, MPLS offers an effective way to allocate network resources to traffic

according to their requirements with different granularity. Since MPLS also allows

dedicated paths to be set up and bandwidth reservation across the same path, QoS can be

guaranteed.

8/13/2019 Di Serta Tie 2005

40/120

22

2.6 MPLS Future Technology

The next logical step toward all optical networks is to employ a variation of

MPLS that supports different wavelengths for different data flows on the transport layer.

This technique is already under investigation and it is called Generalized Multiprotocol

Label Switching (GMPLS) [10], which is also an enhancement of Multiprotocol Lambda

Switching (MPS). It is a suite of protocol extensions that support multiple switching

types such as packet switching, Time-Division Multiplexing (TDM) switching, lambda

switching, and fiber switching. The basic features of GMPLS include the following:

i. It is a set of traffic engineering and optical extensions to existing MPLS routing

and signaling protocols.

ii. It is an optical networking standard that represents an integral part of next-

generation data and optical networks.

iii. It provides the necessary bridge between IP and optical layers, therefore enabling

networks to be both interoperable and scalable.

iv. It brings more intelligence to optical networks; more intelligent optical devices

help network operators do better restoration and better utilize bandwidth.

v. It is designed to allow edge networking devices such as routers and

switches to request bandwidth from the optical layer.

vi. It supports hierarchical LSPs and optical LSPs (or lightpaths).

vii. It is designed to handle multiple traffic types simultaneously. It creates a control

plane to support multiple switching layers:

a. Packet switching for forwarding based upon packet or cell headers

b. Time-division switching for forwarding based upon the data's

time slot in a repeating cycle (e.g., SONET/SDH, PDH)

c. Wavelength (lambda) switching for forwarding data based on the

wavelength on which it was received

8/13/2019 Di Serta Tie 2005

41/120

23

d. Spatial switching for forwarding data based on a position of the

data in real-world physical spaces (e.g., incoming port or fiber to

outgoing port or fiber)

The main focus of GMPLS is on the control plane of these various layers

because each of them can use physically diverse data or forwarding planes. The goal is

to cover both the routing and signaling portions of the control plane. GMPLS extends

MPLS to encompass time division, wavelength, and spatial switching. It extends the

proven MPLS LSP mechanisms to create generalized labels and generalized LSPs.These extensions of MPLS include [10]:

i. Enhancements to signaling protocols - The RSVP or CR-LDP are used as

signaling protocol. GMPLS requires that an LSP start and end on similar types of

devices. Although LPSs are unidirectional in MPLS architecture, they could be

unidirectional or bidirectional light paths in GMPLS. GMPLS signaling enables

an upstream node to suggest a label, although the suggestion may be overridden

by a downstreams node.

ii. Enhancements of routing - GMPLS enables many improvements to routing. With

GMPLS, explicit paths can be established across network layers. Open Shortest

Path First (OSPF) or Intermediate System to Intermediate System (IS-IS) is used

as Interior Gateway Routing Protocol (IGP).

iii. Introduction of Link Management Protocol (LMP) - This is a new protocol

designed to resolve issues related to link management of optical networks. LMP

is what a router uses to decipher the availability of a link between itself and

another router. It provides four basic functions for a node pair; control channel

management, link connectivity verification, link property correlation, and fault

isolation.

8/13/2019 Di Serta Tie 2005

42/120

24

iv. Additional functionality to address issues related to MPLS control plane -

GMPLS allows the control to be physically diverse from the associated data

plane.

These extensions affect routing and signaling protocols for activities such as

label distribution, traffic engineering, protection, and restoration. They also enable rapid

provisioning and management of network services.

Although MPLS was designed exclusively for packet services, the primary goal

of GMPLS is to provide a single suite of protocols that would be applicable to all kinds

of traffic. By consolidating different traffic types, GMPLS permits simplification of

networks and improves their scalability in ways unheard of until now. It offers the

means by which networks can be scaled and simplified by deployment of a new class of

network element designed to handle multiple traffic types simultaneously.

GMPLS network architecture offers a common mechanism for routing, data

forwarding, and traffic engineering on transport networks with dense wavelength

division multiplexing (DWDM). Such a future network will consist of elements such as

routers, switches, DWDM systems, Optical Add-Drop Multiplexers (OADMs), Optical

Cross-Connects (OXCs), Photonic Cross-Connects (PXCs) so forth that will employ

GMPLS to dynamically provision resources and to provide network survivability using

protection and restoration techniques. These optical network elements will have full

control of the wavelengths and could create self connecting and self regulating networks.

The technology behind GMPLS is Micro-Electric Mechanical Systems (MEMs), which

can use micro-mirrors to redirect lambdas. This has opened the doors to a bandwidth

explosion. This will eventually enable networks that require less management and

overhead.

8/13/2019 Di Serta Tie 2005

43/120

25

GMPLS offers several key benefits to service providers to migrate their services

from their current model to the GMPLS unified control plane. First, GMPLS gives

network operators the freedom to design their networks to best meet their specific

objectives. Second, GMPLS provides accelerated service provisioning of any type of

service, at any time, with any level of QoS, and to any destination. Third, GMPLS

enables greater service intelligence and efficiency because it allows the network system

to view the entire network topology, network node switching capability, and network

resource status across all packet, TDM, and wavelength services. GMPLS will play a

major role in the deployment of next generation networks; it will provide a common

control mechanism across both the packet-switched and transport domain forestablishing label switched paths.

2.7 Linux as the Operating System (OS)

Linux is a free operating system that was created by Linus Torvalds when he was

a student at the University of Helsinki in 1991 [3]. It is a version of the UNIX operating

system that has become very popular over the last several years.

Linux has been copyrighted under the terms of the GNU General Public License

(GPL). This is a license written by the Free Software Foundation (FSF) that is designed

to prevent people from restricting the distribution of software. In brief, it says that

although money can be charged for a copy, the person who received the copy can not be

prevented from giving it away for free. It also means that the source code must be

available. This is useful for programmers. Anybody can modify Linux and even

distribute his/her modifications, provided that they keep the code under the same

copyright.

8/13/2019 Di Serta Tie 2005

44/120

26

In this project Linux Fedora Core 2 has been used since it has kernel of version

2.6.5 that is needed to support MPLS forwarding mechanism. Fedora Core 2 is operating

system that developed under Red Hat project and it is freely available. When this project

is done, Fedora has just released it Fedora Core version 4.

2.7.1 Linux Differences

Linux is generally cheaper than other operating systems and is frequently less

problematic than many commercial systems. But what makes Linux different is not its

price but its outstanding capabilities which are:

i. Linux is a true 32-bit multitasking operating system, robust and capable enough

to be used in organizations ranging from universities to large corporations.

ii. It runs on hardware ranging from low-end 386 boxes to massive ultra-parallel

machines in research centers.

iii. Out-of-the-box versions are available for Intel, Sparc, and Alpha architectures,

and experimental support exists for Power PC and embedded systems, among

others such as SGI, Ultra Sparc, AP1000+, Strong ARM, and MIPS

R3000/R4000.

iv. When it comes to networking, Linux is choice. Not only because networking is

tightly integrated with the Operating System (OS) itself and a plethora of

applications is freely available, but for the robustness under heavy loads that can

only be achieved after years of debugging and testing in an Open Source project

8/13/2019 Di Serta Tie 2005

45/120

27

v. Finally, Linux kernel has built-in support for routing functions. The kernel can

be configured to support many type of routing mechanism such as MPLS, RSVP

and other. A Linux box can acts either as an IP or IPX router for a fraction of the

cost of a commercial router. Recent kernels include special options for machines

acting primarily as routers.

2.8 Details of MPLS Forwarding Basics Implemented in Linux OS

The goal of Multiprotocol Label Switching (MPLS) is to remove the process of

layer three lookups at each hop of a network. This has many direct affects. MPLS

forwarding makes traffic engineering a more viable solution for network congestion.

MPLS forwarding may enable wire speed routing (switching). MPLS forward could

allow network hardware vendors to make cheaper trunk cards that only understand

labeled packets. MPLS abstracts the media specific qualities away from more complex

network engineering (QoS, CoS, DiffServ). MPLS forwarding can be used to make VPN

service more flexible and deployable without compromising the performance of edge

routers (switches). In addition to the above, the concepts used to setup MPLS forwarding

are the first step towards creating completely optical networks.

As a packet traverses an MPLS enabled network it must make three transitions.

First, is that it must go from its native Layer 3 forwarding into labeled MPLS

forwarding. This process entails the adding of a label to the head of the packet. Second,

a (now labeled) packet must be able to traverse an MPLS path. This path consists of all

the devices that know how this particular packet (and packets like it) needs to traverse a

network. This path is called a LSP. It is a connection oriented path that is setup ahead of

the forwarding of any packets. Finally a packet must make its way back into Layer 3

8/13/2019 Di Serta Tie 2005

46/120

28

forwarding. This process consists of removing the label from the head of the packet and

then sending it to the appropriate Layer 3 protocol for additional handling. In an MPLS

enabled network, Layer 3 forwarding is used by the edges of the network, and MPLS

forwarding is used in the core of the network.

2.8.1 MPLS Label

When a label is added to a packet this means that at minimum a 4 bytes "shim"

has been added to the packet. This shim is added between the Layer 3 header and Layer

2 header. Therefore an IP packet on Ethernet would add the shim before the IP header

but after the Ethernet header. MPLS forwarding is currently defined for the following

Layer 2 implementation: Ethernet, packet over SONET, ATM, frame-relay. MPLS has

also been defined for any medium that PPP runs on top of. On most of this Layer 2

implementation a label consists of a 20 bits number. The shim that is added to the packet

contains more then just a label.

As illustrated by Figure 2.4 in section 2.2.1, the label is 20 bits. This value is

used to determine how a packet will be label switched. The next 3 bits are called the

EXP bits. They are currently reserved for experimental purposes (although DiffServ

over MPLS has claimed these 3 bits for its use). The next bit is referred to as the

"bottom of stack bit" (S bit). Due to the fact that MPLS adds a shim to the packet, a LSR

needs to know if what follows this top shim is the Layer 3 header or another shim.

(Multiple shims are called a label stack and the purpose of a label stack will be

explained later). The S bit signifies that what follows this shim is the Layer 3 header.

For typical single shim MPLS forwarding the S bit is on. Finally the shim contains the

Time to Live (TTL) counter. This is used to allow current Layer 3 functions to occur

8/13/2019 Di Serta Tie 2005

47/120

29

even though an LSR cannot use the Layer 3 header. Some examples of these are

traceroute, loop detection, and multicast domains.

2.8.2 MPLS Label Stacking

When an LER adds a shim to a packet, it is feasible that is can add more thanone shim. This concept is called Label Stacking. The stack of shims is treated just as its

name sake data structure. A POP means that the top shim is removed, exposing either

another shim or the Layer 3 header (determined by the S bit). A PUSH adds a new

shim to the top of the stack or on top of the Layer 3 header. So standard label swapping

is defined as a POP followed by a PUSH.

In some cases a labeled packet may need to be tunneled across another MPLS

network. In the case the labeled packet gets another shim pushed on top without POPing

the original shim off. This results in a label stack of size 2. This operation can occur

multiple times by separate LSRs or a single LSR could add more then one shim. In

general any labeled packet has a label stack, although most have a label stack of size 1.

As one may expect a packet with a label stack greater then 1 may shrink in size as well.

This operation would be defined by a POP without a PUSH. The exact details of how

this occurs are explained in the next section.

8/13/2019 Di Serta Tie 2005

48/120

30

2.8.3 MPLS Data Structures

For MPLS forwarding to occur correctly some data structures need to exist to

assist in the interpretation of labels and the processing of them. In general there exists

three data structure. One data structure to help interpret incoming labels (labels from

packets that are entering an LSR or egress LER). A second data structure to assist in

adding out going labels (labels on packets that are leaving a LSR or ingress LER). A

third data structure is used by ingress LERs to figure out what label to add to a packet.

2.8.3.1 Incoming Label Map (ILM)

The common name for the data structure used to interpret incoming labels is

called the Incoming Label Map (ILM). The ILM table consists of all the incoming labels

that an LSR or egress LER will recognize. The contents of each ILM entry are: label,

opcode, FEC, and an optional link to an outgoing data structure. The general order of

operation for incoming label processing are:

i. extract label from top shim

ii. lookup the label in the ILM table

iii. further processing on the packet based on the opcode stored with the label

Details of ILM processing are explained in the section 2.9.1.

8/13/2019 Di Serta Tie 2005

49/120

31

2.3.8.2 The Next Hop Label Forwarding Entry (NHLFE)

As stated earlier there also exists a data structure to assist with outgoing labeling.

The common name for this data structure is the Next Hop Label Entry (NHLFE). The

NHLFE table consists of all the labels which can be push onto packets. Each NHLFE

entry contains: label, outgoing interface, and nexthop information. The general

processing that occurs when a packet reaches the NHLFE table are:

i. form new shim that contains the label

ii. push the shim onto the packet

iii. forward the packet to the nexthop via the outgoing interface

Details of NHLFE processing are discussed in section 2.9.4.

2.8.3.3 FEC-to-NHLFE Map (FTN)

A third data structure exists for the sole purpose of helping ingress LERs decide

what labels to add to a packet. For the explanation of this data structure we need to take

one step back and introduce the term Forwarding Equivalence Class (FEC). Up until

now we have referred to individual packet in terms of adding and removing labels. How

do we know what label to add to a packet? How do we know what type of packet I have

after removing a label?

8/13/2019 Di Serta Tie 2005

50/120

32

In general packets are labeled according to which FEC they belong to. For

example FEC A is defined to be the class of packets that are heading to the host with IP

address 1.1.1.1. Therefore all packets that have a destination of IP address 1.1.1.1 belong

to FEC A. In MPLS each FEC is assign a label and each label refers to a FEC (1:1

mapping). The definitions of FECs may change but the 1:1 mapping should always stay

constant. The data structure used to map FECs to labels is called a FEC to NHLFE or

FTN. An FTN table consists of all FECs which we know how to add labels too. An FTN

entry consists of FEC and NHLFE entry. The general orders of operation for FTN

processing are:

i. decide what FEC a packet belongs to

ii. find the FEC in the FTN table

iii. forward the packet to he NHLFE entry that corresponds to the FTN

Details FTN processing are explained in section 2.9.3.

2.9 MPLS Details

In this section the details about MPLS are explained in the sub-sections. Data

structures such as ILM and NHLFE are also explained. The process of receiving and

transmitting MPLS are also shown in the easy way to understand. Figure 2.6 shows the

main MPLS structure.

8/13/2019 Di Serta Tie 2005

51/120

33

struct mpls_label {

u32 label_res:1,

label_value:28,

label_type:3:

#define MPLS_LABEL_VPI ((label_value>>16)&0xFFF)

#define MPLS_LABEL_VCI (label_value&0xFFFF)

#define MPLS_LABEL_GEN (label_value&0xFFFFF)

#define MPLS_LABEL_DLCI_10 (label_value&0x3FF)

#define MPLS_LABEL_DLCI_17 (label_value&0x1FFFF)

#define MPLS_LABEL_DLCI_23 (label_value&0x7FFFFF)

};

#define MPLS_GEN_LABEL 0x01

#define MPLS_VPIVCI_LABEL 0x02

#define MPLS_VPI_LABEL 0x03

#define MPLS_VCI_LABEL 0x04

#define MPLS_FR10_LABEL 0x05



Figure 2.6: MPLS main structure

8/13/2019 Di Serta Tie 2005

52/120

34

2.9.1 MPLS ILM Details

Figure 2.7 shows the ILM structure.

struct ilm_ent {

struct mpls_label label;

struct route_ent* outgoing_rt;

u16 protocol;

u8 opcode;

};

Figure 2.7: ILM structure

Each logical interface needs to stores its own ILM table. MPLS packets that

arrive via that interface will do label lookups into that interfaces ILM table. The full lists

of opcodes that can be stored in a ILM entry are represented like below:

POP_AND_LOOKUP

Ifthe top shim has the S bit on:

Extract the protocol type from the ILM

POP the top shim

Copy the TTL to the layer 3 header

Using the protocol type, do a lookup on the layer 3 header that is exposed

Else

POP the top shim

Extract the label from the shim that is exposed

8/13/2019 Di Serta Tie 2005

53/120

35

Extract the S bit

Extract the EXP

Extract label and create ILM Index

Using the ILM Index Lookup the ILM Entry

Execute the opcode in the ILM Entry

End

POP_AND_FORWARD

Extract the outgoing route entry from the ILM

POP the top shim

If the outgoing route entry is a layer 3 route entry copy TTL to layer 3

header

Using the outgoing route entry forward the packet to the outgoing

interface

NO_POP_AND_FORWARD

Extract the outgoing route entry from the ILM Using the outgoing route entry forward the packet to the outgoing

interface

SEND_TO_RP

Send the entire packet to the Route Processor

2.9.2 MPLS Receive Processing

Ifinterface is of type ATM and VCC is of type MPLS_ENC

Use VPI, VCI to create ILM Index

8/13/2019 Di Serta Tie 2005

54/120

36

Else

Extract generic label from shim and create ILM Index

End

Extract TTL from top shim

Extract EXP from top shim

Extract S bit from top shim

Decrement TTL

If TTL

8/13/2019 Di Serta Tie 2005

55/120

37

2.9.4 MPLS NHLFE Details

The NHLFE is located on the transmission interface such as ethernet card.

Therefore the NHLFE entry does not need to store the outgoing interface. NHLFE

structure is shown in Figure 2.9.

struct mpls_nhlfe {

struct mpls_label **label;

u8 number_of_label;

struct sockaddr_un next_hop;

u8 exp;

};

Figure 2.9: NHLFE structure

2.9.5 MPLS Transmit Processing

If the route entry that the FTN point to is a NHLFE or the ILM refers to a

NHLFE then the following processing pertains.

Ifthe shim with an S bit on was POPed on ingress then make sure to add an S bit to the

bottom shim

Foreach label in the NHLFE

8/13/2019 Di Serta Tie 2005

56/120

38

Ifmore labels can be added

If the label is of type ATM then create a NULL shim and

add it to the packet make sure the shim with the S bit is on

the bottom of the stack extract VCC info from the ATM

label mark this packet will leave via this ATM VCC mark

that no more labels can be added.

Else

Create a shim from the generic label and add it to the

packet make sure the shim with the S bit is on the bottom

of the stackEnd

End

End

Forward the packet to the next hop stored in the NHLFE.

2.9.10 MPLS Commands in Linux OS

In order to display MPLS commands used in Linux, users have to type mpls in

Linux terminal. The resultant output is shown in Figure 2.10. There are many commands

can be used together with mpls to create Label Switched Path (LSP), assigning MPLS

packet receiver and transmitter device and also to create incoming label map (ILM).

Details of the functions are discussed in Chapter III.

8/13/2019 Di Serta Tie 2005

58/120

CHAPTER III

DESIGN AND PROCEDURE

3.1 Introduction

In order to enable MPLS in Linux operating system the right version of kernelmust be installed earlier. In this project, kernel of version 2.6.9 is required. Since this

project used Fedora Core 2 by Red Hat that has kernel of version 2.6.5, the Linux kernel

must be upgraded to version 2.6.9 or above. The next part discussed on how to compile

Linux kernel and so forth.

3.2 Kernel Recompilation

The prefer method of installing a MPLS enabled kernel is to use the RPMs file.

But if the system runs a non-RPM system developers might have to compile the kernel

by hand.

8/13/2019 Di Serta Tie 2005

59/120

41

Section 3.2.1 shows some quick steps how to recompile the kernel whereas in

section 3.2.2 the installation using RPM file is discussed.

3.2.1 Non-RPM System

Below are the steps to recompile kernel. But before the compilation can be done,developers are required to remember what his/her previous Linux config options.

Figure 3.1, Figure 3.2, Figure 3.3 and Figure 3.4 show the kernel options selected after

kernel is patched and before kernel compilation is done.

1) Get Linux kernel 2.6.9

2) Change directory to /usr/src

# cd /usr/src

3) Uncompress/untar linux-2.6.9.tar.gz

# tar -zxvf linux-2.6.9.tar.gz

4) Uncompress or untar mpls-linux-1.946.tgz

# tar -zxvf mpls-linux-1.946.tgz

5) Change directory to linux

8/13/2019 Di Serta Tie 2005

60/120

42

# cd linux

6) Patch Linux kernel for MPLS for Linux

# patch -p1 < ../mpls-linux/patches/linux-kernel.diff

7) Configure Linux kernel to turn on MPLS

# make menuconfig

These are the standard configurations of Linux option. Standard config:

Code maturity level options --->

[*]Prompt for development and/or incomplete code/drivers

Networking options --->

[*]Multi Protocol Label Switching - MPLS

For more advanced traffic mappings (in addition to Standard config):

[*]Network packet filtering (replaces ipchains)

IP: Netfilter Configuration --->

IP tables support (required for filtering/masq/NAT)

8/13/2019 Di Serta Tie 2005

61/120

43

spec_nh target support

8) Compile and install kernel

# make clean; make vmlinux

9) Reboot

# reboot

Figure 3.1: Kernel option before compilation; no MPLS option

8/13/2019 Di Serta Tie 2005

62/120

44

Figure 3.2: Kernel option after compilation; MPLS is selected

8/13/2019 Di Serta Tie 2005

63/120

45

Figure 3.3: Kernel options before kernel compilation; no Special Next Hop option

8/13/2019 Di Serta Tie 2005

64/120

46

Figure 3.4: Kernel option after compilation; Special Next Hop is selected

8/13/2019 Di Serta Tie 2005

65/120

47

3.2.2 RPM System

For RPM supported system, there is only one step must be taken before direct

installation of RPM files. The step is used to test the particular RPM file dependencies.

This is required to check either the system is already has the previous version of desired

RPM file or not. If yes, the upgrade command must be used instead of install command.

To check the file dependencies, test command must be used.

Command usage:

# rpm Uvh - - test

# rpm Uvh

Here are the steps to install RPM files.

1) # rpm Uvh - - test kernel-2.6.9-1.6_FC2mpls_1_946a.i686.rpm

# rpm Uvh kernel-2.6.9-1.6_FC2mpls_1_946a.i686.rpm

2) # rpm Uvh - - test quagga-0.97.3-1.FC3mpls1_946b.i386.rpm

# rpm Uvh quagga-0.97.3-1.FC3mpls1_946b.i386.rpm

3) # rpm Uvh - - test iptables-1.2.11-3.1.FC3mpls1_946.i386.rpm

# rpm Uvh iptables-1.2.11-3.1.FC3mpls1_946.i386.rpm

4) # rpm Uvh - - test iproute-2.6.9-3mpls1_946.i386.rpm

# rpm Uvh iproute-2.6.9-3mpls1_946.i386.rpm

8/13/2019 Di Serta Tie 2005

66/120

48

These RPM files are taken from MPLS package of version 1.946. These files are

prominent to enable MPLS on Linux machine. For example, the first file is needed to

adds MPLS functionality into Linux machine, the third file is used to add support for

SPEC_NH marking while last file is needed to add spec_nh that is a next hop option.

3.3 Kernel IP Forwarding

Before IP packets can be sent into the network, Linux router must be able to

forward IP packets. In Linux kernel, the option of IP packet forwarding is set to 0 by

default. So that, this set up need to be changed to enable IP packet forwarding in the

Linux kernel. There are two ways to set up IP forwarding capability. Initially this

capability can be set up for temporally. That means when the Linux machine restarted

the capability must be set up again. But for the second choice, it can be set up for

permanent. Below are the commands to enable IP packet forwarding.

1) For temporally

# echo 1 > /proc/sys/net/ipv4/ip_forward

2) For permanently

Edit sysctl.config script with this command:

# gedit /etc/sysctl.config

- change ipv4 forwarding to 1 and save the script

8/13/2019 Di Serta Tie 2005

67/120

49

Temporally command also can be used to make it permanently. Simply type this

command to rc.local script and this command will be run by Linux kernel just after

booted. The script can be found at this directory;/etc/rc.d/rc.local.

3.4 Ethernet Card Driver

This project required four routers to be connected to each order. Therefore at

each router more than one ethernet cards have to be assigned. Linux is a different

operating system environment compared to Windows operating system where in

Windows users can plug-and-play the ethernet cards. Current Windows has been

designed to find the ethernet driver within itself without requires users to install it.

Different with Linux since users are needed to compile his/her own driver by

him/herself.

One problem that always occurs in Linux; there is no driver source code for

particular network card. Hence for advanced Linux users, they have to design and

program their own source code.

This project used source code that is available in Internet. The source code has

been designed for Broadcom ethernet card version 5700. Below are the steps to compile

source code from RPM file.

a. Install the source RPM package:

i. # rpm -ivh bcm5700-.src.rpm

8/13/2019 Di Serta Tie 2005

68/120

50

b. Change directory to the RPM path and build the binary driver for Linux

kernel:

i. # cd /usr/src/{redhat,OpenLinux,turbo,packages,rpm ..}

ii. # rpmbuild -bb SPECS/bcm5700.spec

c. Install the newly built package (driver and man page):

i. rpm -ivh RPMS/i386/bcm5700-.i386.rpm

d. The driver will be installed in the following directory:i. /lib/modules//kernel/drivers/net/bcm5700.ko

e. Load the driver

i. # modprobe bcm5700

The version of Broadcom ethernet card driver used is bcm5700-8.1.55-1.

3.5 Assigning Ethernet Device with IP Address

There are two methods available to assign ethernet card with IP address in Linux

operating system. Users can either use ifconfig command or ip route command.

These two commands have to be typed in Linux terminal window that is similar with dos

prompt in Windows operating systems. The next sections discussed on these two

methods.

8/13/2019 Di Serta Tie 2005

69/120

51

3.5.1 ifconfig

This command requires inputs such as device name, IP address, broadcast

address and network mask.

1) Usage :

a. # ifconfig broadcast

netmask

2) Activate device :

a. # ifconfig dev up

3) Example :

a. # ifconfig eth0 10.0.1.1 broadcast 10.0.1.255 netmask 255.255.255.0

b. # ifconfig dev eth0 up

3.5.2 ip route

Differ with ifconfig, this command requires only IP address with prefix length

and device name. It does not require broadcast address.

1) Usage:

a. # ip route add / dev

8/13/2019 Di Serta Tie 2005

70/120

52

2) Example:

a. # ip route add 10.0.1.1/32 dev eth0

To activate device for ip route command, the same method with ifconfig is used.

3.5.3 Assigning IP Address in the Test-bed

The ip route command has been used to assign Linux PC with IP address. The

connection between Linux PCs for the test-bed with ethernet cards assigned to them is

shown in Figure 3.5. R-1 and R-2 functioned as intermediate routers while R-4 and R-3

functioned as edge routers. However R-5 and R-6 are the hosts attached to the test-bed.

Figure 3.5: Test-bed

8/13/2019 Di Serta Tie 2005

71/120

53

Below are IP addresses assigned to Linux PCs. Number of IP addresses assigned

to Linux PCs depends on how many connection they are connected. For example from

Figure 3.5, R-3 is connected to two routers (R-1 and R-2) and two hosts (R-5 and R-6).

So that it requires four ethernet cards and four IP addresses.

1) R-4


2) R-1a. # ip route add 10.0.1.2/32 dev eth0

b. # ip route add 10.0.3.1/32 dev eth1

c. # ip route add 10.0.2.1/32 dev eth2

3) R-2



4) R-3



c. # ip route add 10.0.3.2/32 dev eth2

d. # ip route add 10.0.6.1/32 dev eth3

5) R-5


6) R-6


8/13/2019 Di Serta Tie 2005

72/120

54

3.6 Enabling MPLS

When all Linux PCs are connected with proper IP addresses, then only the MPLS

can be enabled. The command shown is used to enable MPLS. This command has to run

at each Linux PC to change it to be MPLS router.

# echo 1 > /sys/mpls/debug

This command will produce MPLS debug equal to 1 that means MPLS is

already enabled. How the process looked like on Linux terminal is shown below.

[root@TJ-4 root]# echo "1" > /sys/mpls/debug

[root@TJ-4 root]# dmesg

ivering

MPLS DEBUG net/mpls/mpls_output.c:174:mpls_output2: enter

MPLS DEBUG net/mpls/mpls_output.c:192:mpls_output2: opcode PUSH

MPLS DEBUG net/mpls/mpls_opcode.c:117:mpls_push: enter

MPLS DEBUG net/mpls/mpls_opcode.c:131:mpls_push: using gap

MPLS DEBUG net/mpls/mpls_opcode.c:187:mpls_push: exit

MPLS DEBUG net/mpls/mpls_output.c:192:mpls_output2: opcode SET

MPLS DEBUG net/mpls/mpls_output.c:56:mpls_send: output device = eth0

MPLS mpls_skb_dump: from eth0 with len 88 (304)headroom=12 tailroom=28

6e02920088b84d0a01100000{|0003814045000054cd740000400192320a000101

#0a00060200001

dab5a103276114c3643233c000008090a0b0c0d0e0f101112131415161718191a1

b1c1d1e1f20212 2232425262728292a2b2c2d2e2f303132333435363700}

8/13/2019 Di Serta Tie 2005

73/120

55

MPLS DEBUG net/mpls/mpls_output.c:116:mpls_send: mpls_send result 0

MPLS DEBUG net/mpls/mpls_output.c:226:mpls_output2: exit

MPLS DEBUG net/mpls/mpls_input.c:274:mpls_skb_recv: exit(0)

MPLS DEBUG net/mpls/mpls_input.c:228:mpls_skb_recv: enterMPLS DEBUG

net/mpls/mpls_netlink.c:197:mpls_dump_ilm: Dump: idx 1

MPLS DEBUG net/mpls/mpls_netlink.c:197:mpls_dump_ilm: Dump: idx 2

MPLS DEBUG net/mpls/mpls_netlink.c:209:mpls_dump_ilm: Exit: idx 3

MPLS debug = 1

[root@TJ-4 root]#

3.7 Label Switching Path Set Up

After compiling and installing the MPLS enabled version of iproute, there will be

a new command in Linux terminal that is mpls and a new next-hop option for the ip

route command. This part described some examples on how to use mpls and the new

next-hop option available for ip route.

1) # mpls nhlfe add key 0 instrcutions push gen 16 nexthop eth0 ipv4

10.0.1.2

key: 0x00000002 (produced by kernel after upper line is typed)

This line notifies that entry (IP packet) for next hop will be assigned as

label 16 represented by key 0x2 in Linux kernel. This packet will be forwarded

to the next hop with address of 10.0.1.2 through device eth0.

8/13/2019 Di Serta Tie 2005

74/120

56

2) # ip route add 10.0.6.2/32 via 10.0.1.2 spec_nh 0x8847 0x2

This line adds IP routing table in Linux kernel. This route is assigned as

Label Switching Path (LSP) that means all packets that addressed with 10.0.6.2

will follow this route and will be encapsulated with MPLS header label 16.

0x8847 is the ethernet identification of unicast MPLS packets.

3) # mpls nhlfe show

NHLFE entry key 0x00000002 mtu 1496 propagate_ttlpush gen 16 set eth0 ipv4 10.0.6.2 (0 bytes, 0 pkts, 0 dropped)

This command produced the output that

di serta tie 2005

Documents