Transcript Slide 1
Performance Analysis of AODV and SAODV Routing Protocols in Ad-Hoc Mesh NetworksA Simulation Study Sangeeta Ghangam Division of Computing Studies, Arizona State University July 30, 2007 Advisor: Prof. Bruce Millard Introduction to Mesh Networks • Wireless Mesh Networks(WMNs) have a 3-tier architecture where different protocols work in tandem • Infrastructure, client and hybrid WMN’s • WMN’s have several applications Metro area coverage, Military applications Disaster recovery, Intelligent transportation systems • Examples: MIT’s RoofNet- An experimental WMN Tempe, Gilbert, Chandler, and Philadelphia – city WMN Wireless Mesh Network A Survey of Wireless Mesh Networks, IEEE Radio Communications, pp S23—S30, Sept 2005 WMN’s & Mobile Ad-hoc Networks (MANET’s) • WMN’s diversify the capabilities of MANET’s • IEEE 802.11’s Extend Service Set (ESS) mesh networking task group is working on a routing protocol for WMN based on AODV(Ad-hoc On demand Distance Vector) • BelAir, Tropos, Strix Systems, Cisco etc. have WMN products using routing protocols based on AODV Why AODV? • A reactive protocol which finds routes on demand • Low overhead and adjusts to dynamic environment • Provides loop free routes with the use of sequence #’s • Has both unicasting & multicasting capability • Works in 3 modes Route discovery Route maintenance Local connectivity management Reverse Path Setup Forward Path Setup C.E.Perkins, E.M.Royer, “Ad-hoc On-Demand Distance Vector Routing”, Proc. of ACM, 2002 AODV • AODV works in a implicit ‘trust your neighbor’ mode • Two main security requirements Node Authentication Message Integrity • Secure AODV (SAODV) adds an extension to different AODV packet formats to incorporate digital signature for protecting the non-mutable information and hash chains to protect the mutable information(hop count) OPNET: The Simulation Environment • Provides an OOP approach to network design • Has a hierarchical modeling architecture similar to the structure of communication networks • Provides ease of use via a GUI or the flexibility to create custom models via programming interface • Integrated post simulation analysis tools Designing the SAODV Model • SAODV model in OPNET follows from the pre-existing AODV model • Implements the hash chain functionality • New data structures for packet formats • New functions to implement the revised packet creation, memory allocation & de-allocation etc. has been created • A cryptoSys API toolkit, which implements a random number generator & MD5 hash algorithm, from DiManagement services (http://www.dimgt.com.au/crypto.html) is use The Hash Chain Mechanism Creation(RREQ (Route REQuest)and RREP(Route REPly)): Set Max-hop-count = TTL (time to live) Generate a Random Number ( seed) Calculate Top-hash = h max _hop_count ( seed) Set Hash = seed Send: Type, Length Hash- function, Max-hop-count, Tophash& Hash fields with the original AODV RREQ and RREP Verification: Top_hash== h (max_hop_count – hop_count) (Hash) Account for the additional hop h( Hash) Network Model for Comparisons in OPNET Server IP backbone Gateway1 R BSS1 Gateway2 R BSS2 R BSS3 R BSS4 Simulation Parameters Parameters Values Room Size Mobile Nodes 4000m x 4000m 2 in each BSS Random Waypoint Speed - 2m/s to 7 m/s Start: Uniform( 0, 100)s Pause: Uniform(50, 100)s Wired Links: 100 Base T Wireless Links: IEEE 802.11g @ 11Mbps Wired Links: OSPF Wireless Links: AODV/SAODV Bursty Uniform(100,110)s Poisson (9000) Poisson (4000) Poisson (8000) Poisson (80) 200 minutes Throughput (bits/sec) Route discovery time(sec) Delay (sec) Routing traffic sent ( bits/sec) Routing traffic received ( bits/sec) Load (bits/sec) Mobility Mac Layer Routing Protocols Application Traffic Start End Request Response Request/hr Simulation Time Metrics Throughput & Delay SAODV AODV Time Delay (sec) Throughput (bits/sec) SAODV AODV Time SAODV has a higher delay and throughput for a similar sized network even with less number of nodes creating application traffic. Load & Route Discovery Time Load (bits/sec) AODV Time Route discovery time (sec) SAODV AODV SAODV Time SAODV and AODV have comparable route discovery time. SAODV though imposes a higher load on the network than AODV. SAODV AODV Time Routing Traffic Sent (bits/sec) Routing Traffic Received (bits/sec) Routing Traffic Received & Sent SAODV AODV Time SAODV is bandwidth intensive compared to AODV as it has a higher amount of control traffic. Scalability of AODV and SAODV AODV SAODV 30 nodes Route Discovery Time (/sec) Route Discovery Time (/sec) 50 nodes 50 nodes 15 nodes 30 nodes 15 nodes Time Time The Route Discovery time more than doubles, as the number of nodes increases for AODV and SAODV. Summary & Conclusions Summary: SAODV model created in OPNET and compared against the existing AODV model. Different sized networks designed for comparing and analyzing the protocols. Conclusions: Both SAODV and AODV show comparable results up to a mid-sized network of about 30 nodes with some mobile nodes. AODV performs well over different network architectures and is well suited for mesh networks. SAODV only works well in a single tier and hence cannot be used without modifications for mesh networks.