John Smith, Dilip Kotak, and William A. GruverSchool of Engineering Science • Simon Fraser University

Burnaby, BC Canada

Dorian SabazIntelligent Robotics CorporationNorth Vancouver, BC Canada

Comparative Analysis of AODV, DSDV and DSR in ns-2

2004 IEEE International Conference on Systems, Man, and CyberneticsThe Hague, The Netherlands

October 11, 2004

ENSC 835: HIGH PERFORMANCE NETWORKSCMPT 885: SPECIAL TOPICS: HIGH PERFORMANCE NETWORKS

FINAL PROJECT PRESENTATION

SPRING 2006

COMPARATIVE ANALYSIS OF WIRELESS ROUTING PROTOCOLS

USING NS-2

Edward Chen (ekchen@sfu.ca)

Colin Ng (cnge@sfu.ca)

http://www.sfu.ca/~ekchen

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 2/26

Presentation Overview

IntroductionIntroductionMotivationMotivationRouting Protocol OverviewRouting Protocol Overview

Project OverviewProject OverviewProcess FlowProcess Flow

Project SimulationProject SimulationSimulation parametersSimulation parametersSimulation metricsSimulation metrics

AnalysisAnalysisComparative AnalysisComparative AnalysisIndividual AnalysisIndividual Analysis

Conclusion/QuestionsConclusion/Questions

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 3/26

Introduction

Traditional Centralized TopologyTraditional Centralized Topology

Advantages:Advantages:SimplisticSimplisticSecureSecure

Disadvantages:Disadvantages:ScalabilityScalabilityFaultFault--tolerancetolerance

Central Server

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 4/26

Introduction

Distributed TopologyDistributed TopologyLack of central server for storage/routingLack of central server for storage/routingEach node is both a Each node is both a serverserver and a and a clientclientMessages routed by intermediary nodesMessages routed by intermediary nodes

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 5/26

Introduction

Routing extremely importantRouting extremely importantMany types depending on user criteriaMany types depending on user criteria

Simplicity, low overhead, minimize dropped packet Simplicity, low overhead, minimize dropped packet …… etcetcAODV, DSDV and DSRAODV, DSDV and DSR

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 6/26

Protocols Overview - DSDV

DestinationDestination--Sequenced Distance Vector RoutingSequenced Distance Vector RoutingExtension of BellmanExtension of Bellman--Ford (shortest path between two points)Ford (shortest path between two points)Routing table list all available destinations, hops and sequenceRouting table list all available destinations, hops and sequence numbersnumbersSeq. # avoids loopsSeq. # avoids loopsNode periodically send out routing tablesNode periodically send out routing tables

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 7/26

Dynamic Source RoutingDynamic Source RoutingComplete hopComplete hop--byby--hop route to destinationhop route to destinationMultiple routes for each destinationMultiple routes for each destinationAggressive use of source routing and route cachingAggressive use of source routing and route cachingRouteRoute--discovery and routediscovery and route--maintenancemaintenance

Protocols Overview - DSR

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 8/26

AdAd--Hoc OnHoc On--Demand Distance Vector RoutingDemand Distance Vector RoutingCombination of DSR and DSDVCombination of DSR and DSDVDSDV DSDV NextNext--hop routing tablehop routing tableDSR DSR OnOn--demand route discoverydemand route discovery

Protocols Overview - AODV

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 9/26

Protocols Overview - Summary

HighHighMediumMediumMediumMediumNode OverheadNode Overhead

OnOn--DemandDemand

Route Table Route Table with next hopwith next hop

MediumMedium

AODVAODV

OnOn--DemandDemandPeriodicPeriodicRoute Route DiscoveryDiscovery

Complete Complete routes cachedroutes cached

Route Table Route Table with next hopwith next hop

Route Route MechanismMechanism

LowLowHighHighNetwork Network OverheadOverhead

DSRDSRDSDVDSDV

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 10/26

Project Overview

Implemented in nsImplemented in ns--2 2 Simulation of Wireless Distributed System (WDS)Simulation of Wireless Distributed System (WDS)Wireless package developed by CMUWireless package developed by CMU

VariablesVariablesRouting Protocols x 3Routing Protocols x 3Number of Nodes x 3Number of Nodes x 3Pause time (mobility) x 3Pause time (mobility) x 3

Initially wanted to simulate larger network Initially wanted to simulate larger network > 1000 nodes> 1000 nodes27 trace files 27 trace files > 1.5 > 1.5 GbGbProcessed with Pentium IV 2 GHz Processed with Pentium IV 2 GHz > 72 hours> 72 hoursMemory issue Memory issue aborted prematurelyaborted prematurely

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 11/26

Project Overview

To graph processed dataGraphing data

To combine the processed data of each trace-file piece

Combining processed data

To process each individual trace-file piece

Processing split trace-file

To divide the trace-file into smaller pieces

Splitting trace-file

To generate trace-fileTrace-file generation

Purpose of StateState

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 12/26

Project Simulation

Variable ParametersVariable Parameters Fixed ParametersFixed Parameters

5 kbpsRate

½ of number of nodesMaximum Connections

Constant Bit Rate (CBR)Traffic Type

Traffic Generation

5 m/sMaximum Speed

Node Movement

150 secondsSimulation Time

1000 metersY-Boundary

1000 metersX-Boundary

General Topology

1, 50, 10020, 60, 100DSR

1, 50, 10020, 60, 100DSDV

1, 50, 10020, 60, 100AODV

Pause Time (sec)Number of Nodes

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 13/26

Application LoadApplication LoadThe total number of sent messages and forwarded messages The total number of sent messages and forwarded messages (application(application--related)related)

Dropped LoadDropped LoadThe total number of dropped messages (applicationThe total number of dropped messages (application--related)related)

Received LoadReceived LoadThe total number of received messages (applicationThe total number of received messages (application--related)related)

Routing LoadRouting LoadThe total number of sent messages and forwarded messages The total number of sent messages and forwarded messages (routing(routing--related)related)

Project Simulation - Metrics

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 14/26

Analysis – Application Load

DSDVDSDVAODV AODV / DSR/ DSR

Low Low (P)(P)

DSDVDSDVAODV AODV / DSR/ DSR

High High (P)(P)

Low Low (N)(N)

High High (N)(N)

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 15/26

Analysis – Dropped Load

AODV AODV / DSR/ DSR

AODV AODV / DSR/ DSR

Low Low (P)(P)

AODV AODV / DSR/ DSR

AODV AODV / DSR/ DSR

High High (P)(P)

Low Low (N)(N)

High High (N)(N)

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 16/26

Analysis – Routing Load

DSDVDSDVDSDVDSDVLow Low (P)(P)

DSDVDSDVDSDVDSDVHigh High (P)(P)

Low Low (N)(N)

High High (N)(N)

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 17/26

Analysis – Received Load

Node VarianceNode Variance

Pause VariancePause Variance

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 18/26

Analysis - DSR

Routing LoadRouting Load Dropped LoadDropped Load

Node VarianceNode Variance

Pause VariancePause Variance

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 19/26

Analysis - DSDV

Routing LoadRouting Load Dropped LoadDropped Load

Node VarianceNode Variance

Pause VariancePause Variance

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 20/26

Analysis - AODV

Routing LoadRouting Load Dropped LoadDropped Load

Node VarianceNode Variance

Pause VariancePause Variance

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 21/26

Analysis – Throughput

Node VarianceNode Variance

Pause VariancePause Variance

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 22/26

Analysis – End-To-End Delay

Node VarianceNode Variance

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 23/26

Summary

MotivationMotivationDecentralized framework is better than a centralized frameworkDecentralized framework is better than a centralized frameworkEfficient routing is requiredEfficient routing is requiredCompared AODV, DSDV, DSR in nsCompared AODV, DSDV, DSR in ns--22

Simulation parametersSimulation parametersVarying nodes, pause time, and routing protocolsVarying nodes, pause time, and routing protocols

Performance metricsPerformance metricsApplication load, dropped load, received load, routing loadApplication load, dropped load, received load, routing load

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 24/26

Summary

Best Case / Worst Case:Best Case / Worst Case:Application loadApplication loadDropped loadDropped loadRouting loadRouting load

DSDVDSDVAODV / AODV / DSRDSR

Low (P)Low (P)

DSDVDSDVAODV / AODV / DSRDSR

High (P)High (P)

Low (N)Low (N)High (N)High (N)

AODV / AODV / DSRDSR

AODV / AODV / DSRDSR

Low (P)Low (P)

AODV / AODV / DSRDSR

AODV / AODV / DSRDSR

High (P)High (P)

Low (N)Low (N)High (N)High (N)

DSDVDSDVDSDVDSDVLow (P)Low (P)

DSDVDSDVDSDVDSDVHigh (P)High (P)

Low (N)Low (N)High (N)High (N)

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 25/26

References

[1] Agent Development Kit, http://www.madkit.org/[2] E. Cortese, F. Ouarta, and G. Vitaglione, “Scalability and Performance of the JADE

Message Transport System,” Proc. Of the AAMAS Workshop on AgentCities, Bologna, Italy, July 2002

[3] S.I. Kumaran, JINI Technology, An Overview, Upper Saddle River, NJ, USA, 2002[4] J. F. Kurose and K. W. Ross, Computer Networking, AW Education Group, USA, 2002.[5] E. Chen, D. Sabaz, and W.A. Gruver, “JADE and wireless distributed environments,”

IEEE International Conference on Systems, Man, and Cybernetics, The Hague, Netherlands, 2004.

[6] E. Chen, Jade and JXTA Extensions for Implementing a Better Distributed Systen, Master’s Thesis, School of Engineering Science, Simon Fraser University, Canada, 2005

[7] AODV, http://moment.cs.ucsb.edu/AODV/aodv.html, accessed March, 2006[8] DSDV, http://www.cs.virginia.edu/~c17v/cs851-papers/dsdv-sigcomm94.pdf, accessed

March, 2006[9] DSR, http://www.cs.cmu.edu/~dmaltz/internet-drafts/draft-ietf-manet-dsr-09.txt, accessed

March, 2006[10] New Wireless Trace-File Format, http://k-lug.org/~griswold/NS2/ns2-trace-

formats.html#wireless:new, accessed March 2006[11] The Network Simulator ns-2: Documentation, http://www.isi.edu/nsnam/ns/ns-

documentation.html, accessed April 2006

Comparative Analysis of AODV, DSDV and DSR using ns-2School of Engineering Science - Simon Fraser University – Slide 26/26

Q & A?Q & A?