Performance Analysis Of Zone Based Efficient Multicast Routing Protocol For Multicasting Over MANET

Performance Analysis Of Zone Based Efficient Multicast Routing Protocol For Multicasting Over MANET

Savitha H. P1 Vinodha K2

1PG Student, The Oxford College of Engineering, India

2Asst Professor, The Oxford College of Engineering, India

ABSTRACT

Group communication in Mobile Adhoc Network is achieved using an efficient method called multicasting. Efficient, secured and scalable multicasting in MANET is a challenging approach, due to dynamic topology of the MANET. So here we propose a novel Scalable and Efficient Multicast Routing Protocol (SEMRP).It is a zone based multicast routing protocol, where a zone based bidirectional tree is constructed to achieve efficient group membership management and multicast packet delivery. The position information of nodes is used in zone construction, multicast tree creation and multicast data delivery which reduce the route searching overhead and tree structure maintenance. The scalability and the efficiency of SEMRP are evaluated through simulations and quantitative analysis. Our simulation results show that SCMRP achieves better throughput and less packet loss as compared to DSDV routing protocol. And SCMRP is proved Scalable in both network size and group size.

Keywords

Mobile Adhoc Networks, Multicast, Routing, Cluster, Protocol.

1. INTRODUCTION

   Mobile Adhoc Network (MANET) is a network of mobile nodes with dynamic topology, no proper infrastructure, absence of Centralized administration and communication is multihop wireless communication. Each node acts as host as well as router to assist in forwarding of data to other nodes which may not be related to it. Two broad classifications of routing protocols namely proactive and reactive routing protocols exist[2]. In proactive routing protocols each node maintains list of all other nodes and their routes by some sort of pre- processing function such as broadcasting a hello message, hence these protocols respond quickly to the network changes. When a packet arrives at a node, node verifies its routing table, and accordingly forwards the packet. In

   reactive routing protocols nodes do not maintain details of other nodes in the beginning, but the nodes build the routing table on demand, if a session has to be started then the path is setup between the source and the destination and then the packet is forwarded. Here the path-setup process is called as Route Discovery.

   In some of the applications data need to be transmitted from one to many destination nodes. The service which allows this type of data delivery is called multicasting, but providing multicast service in MANET is a challenging approach as MANET has dynamic topology.

   Table 1.1 characteristics and applications of

   MANET

   Characteristics
   Applications

   • Less bandwidth

   • Dynamic topology

   • Limited energy

   • Lack of Security

   • Military applications – Battlefields

   • Search and rescue operations

   • Multimedia applications

   • Teleconferences

   In this work, proactive geographic routing protocol called a Scalable and Efficient Multicast Routing Protocol (SEMRP) is proposed. SEMRP uses virtual Zone based structure to provide efficient group membership management. This protocol scales to a larger group size as well as network size. Zone is formed around the zone leader with all other member nodes of the nodes one hop or N-hops away from zone leader.

   The paper is organized into following sections, section 2 deals with some of the related work, section 3 gives the detailed study of proposed protocol, Simulation

   results in section 4, result analysis in section 5 and finally conclusion.

2. RELATED WORK

   In this section, we introduce some of the conventional multicast protocols and zone supported forwarding approaches. Multicast Routing protocols can be categorized as shown in fig 2.1, based on the routes constructed for the members of the multicast group (Network topology) as in [4].

   Fig. 2.1. Multicast Routing Protocols

   Tree based multicasting protocols are both power and capacity efficient. In these types of protocols, single path is established between source and destination pair. Tree based approaches are further divided into two kinds, Source tree based and Shared tree based. In source tree based approach (MZRP [5], ODOMP [6], ADMR [7]]), each source constructs individual multicast tree, where the path between source and member of the multicast tree is not shortest. In shared tree based approach (MAODV [8], AMRIS [9]) only one multicast tree is constructed which include all the source nodes and members of all multicast groups. Shared tree based approach is less efficient, each node maintains more routing information, single point of failure cause major problems, since traffic is aggregated in the shared tree ,which gives it a low throughput. Whereas source tree based protocols provide better throughput since each source is responsible for single multicast tree, but has scalability problem. SEMRP is a source tree based protocol.

   small. In an environment with frequent mobility, tree and mesh construction and maintenance introduce more overhead, the stateless approach is introduced to minimise this effect as in [17].

   In zone based approaches, clusters of nodes are formed which make multicasting routing mechanism simple and efficient [18].Every cluster have cluster head or leader to which all other cluster nodes are connected either by single hop or multiple hops. In zone based approaches the routing strategies can be either intra- cluster (within a cluster) or inter-cluster (among the clusters).

3. SCALBLE AND EFFICIENT MULTICAST ROUTING PROTOCOL

   1. Protocol overview

      In SEMRP the reliable membership management and multicast data delivery is performed through two tier virtual zone based structure, the lower layer nodes in the network organize themselves into set of zones in reference to fixed origin as shown in fig.3.1. Among the member nodes of the zone, one node is elected as zone leader. At the upper layer zone leader acts as representative and lets the nodes of the zone to join and leave the multicast group. For Proper zone construction and group membership management, bidirectional tree construction and maintenance, multicast data delivery, the location information is integrated with design. Bidirectional tree allows bidirectional packet forwarding to reduce delay and forwarding overhead. The sources forwards the packets directly along the tree to both upstream root zone and downstream leaf zones instead of only forwarding it first to the root of the tree.

      In mesh based protocols (FGMP [10], ODMRP [11], CAMP [12]), single mesh is constructed connecting the source to all the multicast group members, redundant paths in mesh structure provide more robustness and higher packet delivery ratio. But then introduce power inefficiency, capacity wastage and more overhead caused due to data packet duplication. In hybrid approach (AMRoute [13], ASTM [14], and OGHAM [15]) attempt to achieve both robustness and efficiency by combining tree based and mesh based approaches. Stateless multicast approach (DDM [16]) is useful if the multicast group is

      Fig .3.1 Zone Structure with bidirectional multicast tree and Multicast Group

   2. Neighbour table generation and Zone leader Election

      The election of zone leader with less overhead is mandatory for efficient zone stucture maintenance. Location service is assumed to be part of the protocol design, every node is aware of its own position without the need of any external location service. This is useful in leader election. Every node constructs the neighbour table with no extra signalling. Position, Node ID, Flag information of the neighbouring node is recorded into every other nodes neighbouring table upon receiving the Hello message for every minute. Zone Leaders Neighbour table include a extra field ZoneID, as it records all zone leaders information along with nodes information of same zone .

      Node ID	Position	Flag	Zone ID
      9	(x9y9)	1	Zone0
      18	(x18 y18 )	1	Zone1
      20	(x20 y20 )	—	Zone0

      Node ID	Position	Flag	Zone ID
      9	(x9y9)	1	Zone0
      18	(x18 y18 )	1	Zone1
      20	(x20 y20 )	—	Zone0

      Table 3.2.1 Neighbour table of leader node Zone0

      Table 3.2.2 Neighbour table of ordinary node of Zone0

      Node ID	Position	Flag
      9	(x9y9)	1
      16	(x16 y16 )	—
      20	(x20 y20 )	—

      Zone leader is elected based on following criteria

      1) if no other node in the same zone then the node announces itself as the leader. 2) If there are many nodes then, one which is near to the centre of the zone is elected as leader. 3) If more than one node is at the centre of the zone, then one which higher NodeID is elected as leader.

   3. Zone Supported Geographic Forwarding

      Communication process includes intrazone transmission and interzone transmission. As nodes in a zone are within each others transmission range and each node knows every other node position in the zone, intrazone communication is possible. As source and destinations may be multiple hops away, geographic unicasting is used to ensure the reliable transmission, with packet forwarding guided by the position information of the destinations. To avoid the overhead incurred in tracking the exact location of the large number of

      destinations, location service used as part of Zone based membership management. Only NodeID is needed at the network tier, to forward the packet to the destination zone. At the destination zone the packet is forwarded to the specific node or broadcast based on type of the message.

   4. Multicast tree construction and multicast session life cycle

      Multicast session GN is initiated by the source node X by flooding whole network with a message NEWSESSION (GN , ZoneIDX ), this message carries group name, and zone id of source node. When a node Y receives this message and is interested in this session, it will join the group. If Y did not receive NewSession message, then Y can query its neighbours to search the available groups. To join the group node Y send

      JOINREQ(Y, PosY , GN ) to respective zone leader. And node maintains Membership table, with the entry

      (GN , rootzone ID, isAck) , here isAck field confirm that the node is on the multicast tree. Zone leader will send updated NewSession message to the upstream zone

      towards the root zone. JOINREPLY message is sent to source of the JOINREQ. Then the isAck field of the membership table is set.

      Every zone leader maintains Multicast Table to keep track of all groups and group members of that zone. This table has an entry for each group with the group ID, root zone ID, upstream zone ID, downstream zone list, downstream nodes list. From source to all Group members multicast tree is constructed. To end the session the source node floods ENDSESSION (GN ) message. Upon

      receiving this message all the nodes will remove an entry

      from their Membership tables and Multicast table.When a member node Y wants to leave the group GN ,it sends the LEAVE(Y, GN ) message to its zone leader. If zone leader

      wants to leave the group it will send LEAVE (ZoneID,GN ) message to its upstream zone. All the entries related to

      these nodes are removed from multicast table. Source can forward the packets after the multicast tree is constructed. SEMRP uses the Bidirectional tree based strategy to forward the packets to upstream zone and downstream zones simultaneously to reduce the delay incurred in tree- based routing protocols.

   5. Multicast data forwarding

      Source and destinations may be far away from each other, group members may not be reached in one hop from source. When a source X wants to forward multicast packet to list of destinations (D1, D2, D3), it finds next hop node for each destination. X now inserts the next hop node and list of destinations associated with next hop node in the packet header and broadcasts the message. When a node receives this packet, keeps the packet if it is one among the next hop node, otherwise drops the packet.

      This process continues until packet reaches to all the destinations.

      For reliability, the packets are sent to the center of the destination zone although there may be no node located at the center position. There is no guarantee that the packet will be sent to the destination node correctly since the destination is a virtual reference point. This problem is neglected in previous geographic protocols that use region as destination. To avoid this problem zone forwarding mode is introduced in SEMRP protocol. In zone mode, the node selects as its next hop the neighbouring node whose zone is closer to the destination zone than its own zone. For this the distances of different zones to the destination zone are compared.

4. SIMULATION RESULT

   SEMRP protocol is simulated using NS2 in an area of 1250m×1250m. The area is further divided into equal sized zones. Inter-zone communication is performed within zone with no extra overhead, Intra-zone communication is through the zone leaders. The empty zone scenario is also considered , where nodes of one zone are given more mobility and zone is made empty, so the further intra zone communication for such nodes will be through leader of new zone to which the node has moved.

   PARAMETERS	VALUES
   Network Simulator	NS 2.34
   Channel type	Wireless channel
   Radio-propagation model	Two Ray Ground
   Antenna type	Omni Antenna
   Routing protocol	AODV, DSDV
   MAC type	802.11
   Traffic Type	CBR
   Packet size	512
   Max packet in Queue	50
   Number of packets	Upto 100
   Area covered	1250 X 1250

   PARAMETERS	VALUES
   Network Simulator	NS 2.34
   Channel type	Wireless channel
   Radio-propagation model	Two Ray Ground
   Antenna type	Omni Antenna
   Routing protocol	AODV, DSDV
   MAC type	802.11
   Traffic Type	CBR/p>
   Packet size	512
   Max packet in Queue	50
   Number of packets	Upto 100
   Area covered	1250 X 1250

   Table 4 : configurations made for simulation set up

5. RESULT ANALYSIS

   We compare the performance of SEMRP and DSDV with the variation on node density and multicast group size. The main focus is on the scalability and efficiency of the protocols, under the dynamic environment. And the performance metrics used for multicast performance evaluation are throughput and packet loss.

   1. Effect of Node Density

      Geographic routing protocols perform better in dense networks. The simulation is run with fixed group size of 10 and by varying the node density to 30, 40, 50,

      60 and 70 nodes.

      The SEMRP has better throughput when the network becomes more dense as compared to the DSDV as shown in Fig.4.1.1 Packet is dropped by the forwarding node if it doesnt find the neighbour node or the destination zone. And also packet drop increases with increase in the node density, compared to DSDV, SCMRP has less packet drop, as shown in the Fig.4.1.2.

      Table 5.1.1 Throughput performance of SEMRP and DSDV by varying the node density

      Node density	Number of Packets Received (kbps) In SEMRP	Number of Packets Received (kbps) In DSDV
      30	101.72	53.73
      40	98.46	54.05
      50	108.41	52.10
      60	109.41	54.83
      70	109.70	60.02

      Fig 5.1.1 Throughput comparison by varying node density

      Table 5.1.2 Packet drop in SEMRP and DSDV by varying the node density

      Node Density	Number of Packets Dropped in SEMRP	Number of Packets Dropped in DSDV
      30	23	111
      40	46	105
      50	37	112
      60	46	126
      70	50	145

      Fig. 5.1.2 Comparison of packet drop by varying the node density

   2. Effect of Group Size

      SEMRP can scale to a larger group size and performs well with different group sizes. The simulation is run with 50 nodes by varying the group size to 5,10,15,20 and 25.SEMRP has better throughput even with varying group sizes, where as DSDV performs poor even with smaller group sizes as shown in Fig.4.2.1. In SEMRP, larger group size results in temporary disconnections in the tree resulting into more packet loss. Packet drop in SEMRP is less as compared to DSDV as shown in Fig.4.2.2

      Table 5.2.1 Throughput performance of SEMRP and DSDV by varying the group size

      Group Size	Number of Packets Received (kbps) In SEMRP	Number of Packets Received (kbps) In DSDV
      5	71.75	8.38
      10	112.03	19.24
      15	143.29	33.03
      20	152.61	28.39
      25	169.55	39.30

      Fig. 5.2.1 Throughput comparison by varying the group size

      Table 5.2.2 Packet drop in SCMRP and DSDV by varying the Group Size

      Group Size	Number of packets dropped in SEMRP	Number of packets dropped in DSDV
      5	21	84
      10	57	99
      15	62	122
      20	83	149
      25	111	194

      Fig. 5.2.2 Comparison of packet drop by varying the group size

6. CONCLUSION

   Designing more scalable and efficient multicasting protocol over dynamic environment (MANET) is a challenging approach. In this paper Zone based Scalable and Efficient Multicast Routing Protocol (SEMRP) is proposed. Efficiency and scalability are achieved by constructing virtual zone based structure, position information is used for zone structure construction, packet forwarding, multicast tree construction and maintenance. A zone based bidirectional multicast tree is constructed for efficient multicast membership management and data delivery. SEMRP makes use of geographic forwarding for reliable multicast packet delivery.

   Simulation results prove that compared to DSDV the SCMRP achieves better throughput and less packet loss and is scalable to both network size and group size. SEMRP can further be enhanced to provide security by placing an agent in each host to detect the intrusion by analyzing the incoming request and response messages and neighbour node activities.

7. REFERENCES

1. X. Xiang, Member, IEEE, X. Wang, Member, IEEE, and Y. Yang, Fellow, IEEE, Supporting Efficient and Scalable Multicasting over Mobile Ad Hoc Networks, IEEE transactions on mobile computing, vol. 10, no. 5, April 2011.

2. J. Broch, D. A. Maltz, D. B. Johnson, Y. C. Hu and J.Jetcheva. A performance comparison of multi-hop wireless ad hoc network routing protocols. In Proceedings of the 4th annual ACM/IEEE international conference on Mobile computing and networking, pages 8597, New York, NY, USA, October 1998.

3. C. Wu, Y. Tay, and C.-K. To, Ad Hoc Multicast Routing Protocol Utilizing Increasing Id-Numbers (AMRIS) Functional Specification, Internet draft, Nov. 1998.

4. Osamah Badarneh and Michel Kadoch, Multicast Routing Protocols in Mobile Ad Hoc Networks: A Comparative Surey and Taxonomy, This research was supported by grant from NSERC.

5. X. F. Zhang and L. Jacob, "MZRP: An Extension of the Zone Routing Protocol for Multicasting in MANETs,"Journal of Information Science and Engineering, vol. 20, pp.535-551, 2004

6. O. Stanze and M. Zitterbart, "On-demand overlay multicast in mobile ad hoc networks," Proc. IEEE Wireless Communications and Networking Conference, vol. 4, pp. 2155-2161, 2005.

7. J. Jetcheva and D. B. Johnson, "Adaptive Demand- Driven Multicast Routing in Multi-Hop Wireless Ad Hoc Networks," Proceedings of the 2nd ACM International Symposium on Mobile Ad Hoc Networking & Computing (MobiHoc), pp. 33- 44, 2001.

8. E. M. Royer and C. E. Perkins, "Multicast ad hoc on demand distance vector (maodv) routing," Internet-Draft, draft-ietf-draft-maodv-00.txt, 2000.

9. C. W. Wu, Y. C. Tay, and C.-K. Toh, "Ad hoc Multicast Routing protocol utilizing Increasing id- numberS (AMRIS),"draft-ietf-manet-amris-spec-00.txt, 2000.

10. M. G. Ching-Chuan Chiang, and Lixia Zhang, "Forwarding Group Multicast Protocol (FGMP) for Multihop, Mobile Wireless Networks," ACM-Baltzer Journal of Cluster Computing: Special Issue on Mobile Computing,vol. 1, pp. 187-196, 1998.

11. S. J. Lee, W. Su, and M. Gerla, "On-demand multicast routing protocol in multihop wireless mobile n

    etworks," ACM/Baltzer Mobile Networks and Applications ,special issue on Multipoint Communications in Wireless Mobile Networks, vol. 7, pp. 441-453, 2002.

12. J. J. Garcia-Luna-Aceves and E. L. Madruga, "The Core Assisted Mesh Protocol," IEEE Journal on Selected Areas in Communications, Specia Issue on Ad-Hoc Networks, vol. 17, pp. 1380-1394, 1999.

13. J. Xie, R. R. Talpade, A. McAuley, and M. Liu, "AMRoute: Adhoc Multicast Routing Protocol," Mobile

    AUTHORS

    Mrs Savitha H.P received her Bachelor of Engineering in Computer Science and Engineering in 2007. Currently She is a M.Tech student in Computer Networking Engineering from Visvesvaraya Technological University at The Oxford Collage of Engineering,

    Networks and Applications, vol. 7, pp. 429-439, 2002.

14. Ching-Chuan Chiang, Mario Gerla, and L. Zhang, "Adaptive Shared Tree Multicast in Mobile Wireless Networks, Proceedings of GLOBECOM, pp. 193-207, 1998.

15. C.-C. Hu, E. H.-K. Wu, and G.-H. Chena, "OGHAM: On- demand global hosts for mobile ad hoc multicast services, Ad Hoc Networks, vol. 3, pp. 709-723, 2006.

16. J. Lusheng, and M. S Corson, "Differential destination multicast: A MANET multicast routing protocol for small groups," INFOCOM, vol. 2, pp. 1192- 1201, 2001.

1. Rajashekhar C.Biradar, SunilkumarS.Manvi, Review of multicast routing mechanisms in mobile ad hoc networks Journal of Network and Computer Applications 35 (2012) 221239

2. Son J, Cho Y, Chung YD. Direction-based clustering for efficient multicasting in wireless ad-hoc networks. In: Proceedings of the international workshop on data engineering for wireless and mobile access, Vancouver, Canada, 2008. p. 5562.

3. Zhou J, Li J, Li X, Cao F. Timer forecasting zone based on-demand multicast routing protocol in mobile ad- hoc networks. In: Proceedings of the IEEE international conference on communication systems (ICC), Guangzhou, China, 2008. p. 17125.

Bangalore. Her research interests are Computer Networks and Network Security.

Mrs Vinodha K received her Bachelor of Engineering in Electronics and Communication in 1998. She received her ME in Computer Science and Engineering from Anna University in 2008.She is persuing PhD in Computer Science and Engineering at VTU, Belgaum.

Currently she also holds a faculty position as Assistant Professor, Department of ISE, The Oxford College of Engineering. Her main research interests are Computer Networks and Distributed Networks.