Efficiency Evaluation of Routing Protocol in MANET

IJCST Vol. 2, Issue 2, June 2011

ISSN : 2229-4333(Print) | ISSN : 0976-8491(Online)

Efficiency Evaluation of Routing Protocol in MANET 1

Neeraj Kr. Shakya, 2Gopal Singh Kushwah, 3Sandeep Singh Sengar

Dept. of CSE, Motilal Nehru National Institute of Technology, Allahabad, India. Abstract A Mobile Ad-Hoc Network (MANET) is characterized by dynamic nature without any physical infrastructure and centralizes access point. Routing in Ad-hoc networks is a challenging due to mobility of nodes. In this paper, a detailed simulation based performance analysis has been carried out of Ad- Hoc On-Demand Distance Vector (AODV), Dynamic MANET On-Demand (DYMO), and Destination-Sequenced Distance Vector Routing (DSDV), Dynamic Source Routing (DSR) based on different network scenario of network size, node mobility and pause time. Performance matrix includes parameters like packet delivery fraction, throughput, average delay, routing overhead, and packet loss using ns-2.34 as network simulator. Keywords MANET, AODV, DYMO, DSDV, DSR, NS-2.34. I. Introduction MANET [1 - 4] is a collection of dynamic self-organized nodes which communicate among each other in multi-hop fashion. These set of nodes form a temporary network without any centralized authority. In the initial days of networking wired network was the only mean of communication. But due the limitation of portability and setup cost, focused has changed to wireless ad-hoc network. In MANET each node communicates via other nodes as illustrated in Fig. 1. Mobility nature of node makes topology of MANET highly dynamic. Each node participates in topology discovery and thus acts as a router for other nodes. In order to select optimum route for data transmission many routing algorithms have been proposed. Routing

Fig. 1: MANET Network algorithms are basically categorized into proactive and reactive category. Proactive routing protocols maintain a routing table and keep tracks of all the routes available. Reactive routing table determines route on demands and thus having less routing overhead. Performance comparisons of some routing protocols have been performed like between AODV and DSR [5], AODV and DSDV [6], AODV and TORA [7], AODV and DYMO [8] and between STAR, AODV and DSR [9]. In this paper we have evaluate the performance of AODV [10], DYMO [11], DSDV [12], DSR [13], in three different mobility model using ns-2.34 [14] as network simulator on various parameters. The goal of

318 International Journal of Computer Science and Technology

simulation is to study the working of routing protocol and evaluate the condition under which the protocols work better. The organization of this paper is as follows. In next section II, we will discuss the routing protocols used. Section III describes the simulation model used and performance metric of simulation. Section IV depicts the simulation result and analysis. Finally section V concludes the paper. II. Routing Protocols This section depicts the working of routing protocols used in simulation analysis. 1. AODV AODV is a reactive routing protocol. Reactive protocols are those which create route to destination only if the source is required to transmit data to destination. AODV avoids the counting-toinfinity problem unlike other distance vector protocols by using sequence number for each route. In AODV, all nodes maintain a routing table containing the entry for each destination node. Each entry includes the next hop, sequence number and number of hops requires for reaching destination node. Source node starts path discovery process when it has data to send and does not have route to destination. Path discovery is accomplished by flooding the Route Request (RREQ) packet, when RREQ arrive at the destination or the intermediate node having route to destination, it send off a Route Reply (RREP) packet in unicast mode. Since the nature of MANET is dynamic, link failure can be occurs due to movement in-between nodes. In such case link failure message is sent to the source node by the neighbor nodes affected by the link failure. On getting failure message source node re-initiate path discovery process. Finally, the source node receives the RREP packet and start sending the data. 2. DSDV DSDV Routing is a proactive routing protocol based on idea of Bellman-Ford algorithm. Each node maintain a routing table which contain all the possible destination reachable by node along with number of hop required to reach destination and sequence number assigned by destination. This sequence number helps in finding fresh route and prevents loop formation. A node always selects a route with higher sequence number which indicates the freshness of route. A response to topology changes occurs in two incremental updates. In full update the whole ways, first by sending full updates or routing table is send to neighbor whereas in incremental update only the information changed since the last update is send. When a network is more stable incremental update are occurs more frequently and when network is changing then full update occur. 3. DYMO DYMO is a reactive routing protocol and its working is similar to AODV but in more enhanced way. DYMO has implemented the concept of path accumulation, removes gratuitous RREP and determines routes in a unicast way among DYMO nodes. In addition, the Internet connectivity is also defined in the DYMO Internet-Draft [11]. Each node maintains a routing table with information about nodes. Each entry in the routing table consists of a destination address, next hop address, hop count, sequence w w w. i j c s t. c o m



number, valid timeout, delete timeout and gateway flag. Valid timeout indicates the time at which route entry is no longer, the role of sequence number is same as in AODV, delete timeout shows time after which the entry will be deleted and gateway field shows if the destination node is internet gateway or not. The DYMO protocol consists of two operation route discovery and route maintenance. In route discovery process, a source node broadcast a RREQ message in network. During this propagation process, each intermediate node records a route to the source node. Upon sending the RREQ, the originating node waits for a RREP message from the destination. In case no RREP is received within valid time the node may yet again strive to determine a route by issuing another RREQ. On getting RREP message source node starts sending message. Secondly, route maintenance is the process of responding to changes in topology due to node disassociation that happens after a route has been initially created. In such case a Route Error (RRER) message is sent to source node indicating the route is no longer valid. On receiving the RRER, source node re-initiate route discovery if still has packets to send. 4. DSR DSR is on demand routing protocol but the difference between DSR and other on demand routing protocols is that in DSR source node completely specify the route to be taken by packet between source and destination. Furthermore, DSR is beacon-less and thus does not require transmissions of periodic hello packet (beacon), which a node uses to inform its neighbors of its existence. DSR limit the congestion consumed via control packets in ad MANET by eliminating the periodic table-update messages necessary in the table-driven approach and thus increases throughput. The DSR protocol consists of the two main operation of route Discovery and Route Maintenance. A source node initiate a route discovery process by broadcast a RREQ message in its locality. Source node adds its address and target address in the RREQ packet. When an intermediate node receives the packet and not having route to destination in its cache, appends its own address in RREQ packet and broadcast. As soon as the RREQ packet arrives at the destination node, the destination node throw a RREP message back to the source node, reply containing the accumulated list of intermediate nodes address which RREQ has traversed. In route maintenance process, the link break and topology changes related things are being done. III. Simulation Model and Performance Metric The NS2 [14] network simulator with the CMU extensions by Johnson et al. [16] has been chosen to carryout the simulation process. We have used Random Waypoint (RWP) model with Continuous Bit Rate (CBR) traffic sources. NS-2 is an objectoriented simulator written in C++ and OTCL. Fig. 2 show a snapshot of NS2 containing routing nodes. Various parameters that are considered for simulation are listed given below in Table: 1. Table I: Simulation Parameters PARAMETER VALUE Simulator NS-2 (2.34) Protocols

AODV,DYMO,DSDV,DSR

Traffic Source

CBR

Channel Type

Wireless Channel

Grid Size

670 m X 670 m

Mobility Model

Random waypoint

w w w. i j c s t. c o m

Simulation Time

300 seconds

Packet Size

512 bytes

No. of Nodes

5 to 70

Minimum Velocity

0 meters/seconds

Maximum Velocity

25 meters/seconds

Pause Time

0 to 100 sec

Queue Length

50

Topology

Random

Source data pattern

4 packets / sec

The performance of algorithms has been evaluated based on the following metric [15]. • Throughput: it is define as the total of data transferred over the given period of time. It is measured in bytes/sec or bits/sec, efficient routing protocols must have greater throughput. • Relative Routing overhead: The routing overhead is the total amount of control data packets sent by the routing protocol throughout the duration of the simulation. If more control packets are sent by routing agents, then delay in network will increase. • End to end delay: it is the average delay experienced by the data packets throughout the simulation experiment. It is calculated as the time taken between generation of data packet and the arrival of last bit at destination. • Packet drop fraction: it is define as the no. of packet drop to the total no. of packet generated during the simulation time. Lower the packet drop, lower would be the delay in network. • Packet delivery ratio: The packet delivery ratio is expressed as the percentage of number of received packets by destination node to the number of packets sent by all the sources nodes within the period of simulation. It is an essential performance metric of routing protocols.

Fig. 2: NAM Window IV. Simulation Result and Analysis The simulation of algorithms has been performed on three mobility models. First model depends on the number of nodes, second one based on the pause time of mobile nodes and third one is based on speed of nodes.

International Journal of Computer Science and Technology 319


1. Varying the number of nodes The number of nodes in the simulation setup varies from 10 to 20. Fig. 3 show the effect of number of nodes on the routing overhead. As the number of nodes increase routing overhead is also increase. AODV and DSR having almost same overhead whereas the DYMO shows greater routing overhead then others when number of nodes reaches 70. Fig. 4 shows the throughput, throughput of all the protocols initially increases sharply, but after reaching 30 nodes it becomes constant and when the number of nodes reaches 60 throughputs of all protocol decreases. DSR and AODV shows almost same throughput whereas DYMO throughput decreases rapidly as the number reaches 30. Fig. 5 show the packet delivery fraction. All four except DSDV shows almost 100% delivery fraction till the number of nodes is 30. Packet delivery fraction reduces to half when the no. of node increases to double. DSDV shows better delivery fraction than others as the no. of nodes increases. The average delay from source to destination is shown in fig 6: DSDV always shows the delay of less than one seconds and DYMO shows an only fraction of delay. Fig. 7 show the packet drop percentage of routing protocols. When the numbers of nodes are less DSDV drops more packets but as the number increases, the drop percentage of AODV and DSR increases. 2. Varying the pause time The pause time of nodes varies from 10 seconds to 100 seconds. Fig. 8 shows effect of pause time on the throughput of protocols. AODV and DSR shows constant throughput. Fig. 9 show the overhead, it can be seen that other protocols. AODV and DSR show constant overhead irrespective of pause time. Fig. 10 illustrate the packet delivery ratio of routing protocols. DSR and AODV having almost 100% of packet deliver ratio whereas DSDV packet delivery fraction decreases as the pause time increases. It can be seen from Fig. 11 that AODV, DSR, and DYMO having much lower packet drop ratio. AODV shows greater delay as shown in Fig. 12 and DYMO is the most effective as having the least delay. 3. Varying the speed of node This mobility model depicts the effect of node speed on the performance of protocols. Fig. 13 shows that as the speed of nodes increases the packet delivery decreases in case of DSDV and DYMO but AODV and DSR shows no effect in the delivery ratio. Fig. 14 indicates that DSR having greater overhead than AODV, DYMO and DSR. From Fig. 15 it can be verified that throughput of DSDV decreases as the nodes speed increases whereas AODV and DSR give constant throughput. Fig. 16 give the details analysis of delay; initially DYMO shows the least delay than AODV and DSR but as the speed decreases DYMO delay increases. Fig. 17 show the packet drop in the network. DSDV shows the highest drop ratio and least effective in term of packet loss.


Fig. 4: Throughput Vs No. of nodes

Fig. 5: Packet delivery fraction Vs No. of nodes

Fig. 6: End to end delay Vs No. of nodes

Fig. 7: Packet drop fraction Vs No. of nodes

Fig 3: Overhead Vs No. of nodes





Fig. 8: Throughput Vs pause time

Fig. 13: Packet delivery fraction Vs Node speed

Fig. 9: Overhead Vs pause time

Fig. 14: Overhead Vs Node speed

Fig. 10: Packet delivery fraction Vs pause time

Fig. 15: Throughput Vs Node speed

Fig. 11: packet drop fraction Vs pause time

Fig. 16: End to end delay Vs Node speed

Fig. 12: End to end delay Vs pause time

Fig. 17: Packet drop fraction Vs Node speed V. Conclusion This paper carried out the detailed analysis of AODV, DYMO, DSDV and DSR. Simulation result shows that as the number of nodes increase, the performance of routing protocols decrease. DYMO performance improves as the pause time increase. Nodes mobility affects the performance of routing protocol most. As the mobility speed increases, delay and overhead of routing protocols also increase whereas throughput decreases. In future, we can simulate more routing protocol like ESAODV [17], WRP [18], and TORA [19] as well as establish the best condition under which a particular routing algorithm performs best.


International Journal of Computer Science and Technology 321


References [1] Ramanathan, S., Steenstrup, M. E. “A survey of routing techniques for mobile communications networks”. Journal of Mobile Networks and Applications, Vol. 1, Oct, 1996. [2] Blum, J.J.; Eskandarian, A.; Hoffman, L.J. (Dec. 2004). “Challenges of inter- vehicle Ad hoc Networks”, IEEE transactions on Intelligent Transportation Systems 2004, Vol. 5, No. 4. [3] Macker, Joseph P., Corson, M. Scott. Mobile ad hoc networks (MANET). [Online] Available : http://www.ietf. org/html.charters/manet-charter. html. IETF Working Group Charter. [4] Macker, Joseph P., Corson, M. Scott. “Mobile Ad-hoc Networking and the IETF”, ACM SIG MOBILE, Mobile Computing and Communications reviews 1999, vol. 2, No. 2, pp. 9-14. [5] Das, S. R.; Perkins, C.E., Royer, E. M. “Performance comparison of two on-demand routing protocols for ad hoc networks”. In Proc. of IEEE INFOCOM 2000 Vol. 3. [6] Wei, Qingting, Zou, Hong. “Efficiency Evaluation and Comparison of Routing Protocols in MANETs”. In Proc. of ISISE-2008 Vol. 2. [7] Broch, J.; Maltz, D. A.; Johnson, D. B.; Hu, Y-C, Jetcheva, J.. “A Performance Comparison of Multihop Wireless Ad Hoc Network Routing Protocols”. In Proceedings of MOBICOMM ’98, on Mobile computing and networking. [8] Kum, Dong-Won; Park, Jin-Su; Cho, You-Ze, Cheon, ByoungYoon (2010). “Performance Evaluation of AODV and DYMO Routing Protocols in MANET”. In Proc. of CCNC 2010 on Consumer Communication and Networking. [9] Jiang, H. “Performance Comparison of Three Routing Protocols for Ad Hoc Networks”. Proceedings IEEE Conference on Computer Communication and Network, 2001. [10] Perkins, C.; Belding-Royer, E.; Das, S. “Ad hoc On-Demand Distance Vector (AODV) Routing”. IETF RFC 3561. [11] Chakeres, I. , Perkins, C. “Dynamic MANET On-Demand (DYMO) Routing”, IETF Internet-Draft, draft-ietf-manetdymo-17.txt. [12] Perkins, Charles E., Bhagwat, Pravin. “Destination-Sequenced


[19] Park, V. D., Corson, M. Scott 1997). “A Highly Adaptive Distributed Routing Algorithm for Mobile Wireless Networks”. IEEE Conference on Computer Communications, INFOCOM'97, IEEE, Japan, April 1997, pp. 1405-1413. Neeraj Kumar Shakya received B.Tech degree in computer science and engineering from Uttar Pradesh Technical University, in 2008. He is now pursuing the M.Tech degree in computer science from Motilal Nehru Nation Institute of Technology, Allahabad. His research interests are in the area of ad hoc networking, specifically routing in interconnected wireless/ wired network. Gopal Singh Kushwah received B.Tech degree in computer science and engineering from Madhav Institute of Technology and Science, 2008. He is now pursuing the M.Tech degree in computer science from motilal nehru national Institute of Technology, Allahabad. His research interests are in the area of ad hoc networking, and Wi-Fi networks. Sandeep Singh Sengar received the B.Sc. degree in Mathematics from Dr. B. R. Ambedkar University Agra, UP in 2005 and MCA degree from UP Technical University, Lucknow in 2009. He is now pursuing the M.Tech. degree in Information Security at Motilal Nehru National Institute of Technology, Allahabad. He is currently interested in Telecommunication Technology Standards. His research efforts are focused primarily on the WiMAX and LTE Networks.

Distance-Vector Routing (DSDV) Routing”, Proc. of the SIGCOMM 1994 Conference on Communications Architectures, Protocols and Applications, 1994, pp. 234-244.

[13] Johnson, D.; Hu, Y., Maltz, D. “The Dynamic Source Routing (DSR)”. IETF RFC 4728. [14] NS-2. [Online] Available: http://www.isi.edu/nsnam/ns/ tutorial/. [15] Corson, S., Macker, J. “Mobile Ad hoc Networking (MANET): Routing Protocol Performance Issues and Evaluation Considerations”. RFC 2501, IETF MANET Working Group. [16] CMU Monarch Group. “CMU Monarch extensions to the NS-2 simulator”. Available: http://monarch.cs.cmu.edu/cmuns. html. [17] Mandala, S.; Ngadi, M. A.; Abdullah, A.H., Ismail, A.S (2010). “A Variant of Merkle Signature Scheme to Protect AODV Routing Protocol, Recent Trends in Wireless and Mobile Networks". In Communications in Computer and Information Science Springer, 2010, Volume 84, Part 1, 87-98.

[18] MURTHY, S., GARCIA-LUNA-AVECES, J.J. “A Routing Protocol for Radio Packet Networks”. Proc. ACM International Conference on Mobile Computing and Networking, pp. 8695, November, 1995.

