Data Structures for Implementation

Multicast group members in HOMRP need to maintain the following data structures:

z Multicast Connection Table (MCT): Each multicast group member stores the relationship of its parent node, children nodes and relay nodes in this table.

z Multicast Relay Routing Table: All multicast group members need to maintain this table. This table is expended from MCT. This table records if an MTN needs to forward multicast packets by multicasting or not when it receives multicast packets. If an MTN only has a parent node, or has one child node without a parent node, it will not forward the multicast packets when it receives multicast packets. If an LCN has more than two children nodes, it needs to forward multicast packets.

z Unicast Relay Routing Table: This table is expended from MCT. LCNs and MRNs maintain this table for recording those group members who need to be

forwarded multicast packets by unicast tunnels. An LCN records its Parent_Node in this table. An MRN records the Relay_Node_List in this table.

z Message Cache: This is maintained by each multicast group member to detect duplicated packets.

Chapter 4

SIMULATION AND DISCUSSION

4.1 Simulation Environment

HOMRP was simulated using the Global Mobile Simulation (GloMoSim) [15].

The GloMoSim is a scalable simulation environment for wireless network systems using the parallel discrete-event simulation capability provided by PARSEC [16].

Our simulation models a wireless network with 50 hosts placed randomly within a 1000 m × 1000 m area. The movements of hosts is based on the random-waypoint model. The MAC protocol is the IEEE 802.11 with 2 Mbits/sec channel capacity.

The simulation time is 600 seconds. The pause time is 10 seconds. Each node moves with speed of 0, 10, 20, 30, 40 or 50 meters per second. The constant bit rate (CBR) is assigned for data flows and the payload is 512 bytes. The adopted unicast routing protocol is dynamic source routing (DSR) [17]. The simulation parameters are listed in Table 8.

Table 8：Simulation Parameters

4.2 Evaluation of Different Group Sizes

In this setion, we simulated the performances of HOMRP with group sizes of 5, 10, 20 and 30. In this experiment, there were 5 sources. Figure 8 illustrates the packet delivery ratio of HOMRP under different group sizes. All cases of different group sizes perform well without mobility. As the mobility speed increases, the multicast group with smaller group sizes, e.g. 5 and 10, performed worse, because the multicast packets were transmitted by using unicast tunnels. The performance of larger group sizes, e.g. 20 and 30, were better than that of the smaller ones under high mobility. This is because the larger the group size is, the denser the placement of the group members is, and more local multicast trees can be created. It is efficient to deliver multicast packets by multicasting.

The average number of control packets transmitted per data packet delivered (control overhead) versus mobility speed is shown in Figure 9. The control overhead increases with the mobility speed, because the probability of sending TREE_CONSOLIDATION becomes high at high mobility speeds. In addition, a

×

smaller multicast group had a higher control overhead.

Figure 10 shows the average end-to-end delay per data packet delivered. The multicast group with smallrt group sizes, i.e. 5 and 10, had higher end-to-end delay, because a large percentages of packet delivery were transmitted by unicast tunnels.

Using unicast tunnels increase end-to-end delay. On the other hand, the multicast group with larger group sizes had lower end-to-end delay. This is because a large percentage of packet deliveries were transmitted by multicasting.

Figure 8：Packet delivery ratio as a function of mobility speed.

0.9

Figure9：Control overhead as a function of mobility speed.

Figure 10：End-to-end delay as a function of mobility speed.

0.1

4.3 Comparison with Other Approaches

In this section, we compare HOMRP to an efficient overlay protocol, AMRoute [11], and a high performance meshed-based protocol, ODMRP [5]. Both AMRoute and ODMRP are IETF Internet Draft. There are 5 sources and each node’s mobility speed is 20 meters/second. The group size is 20. Figure 11 shows the packet delivery ratio as a function of mobility speed among different approaches. ODMRP is shown to have the highest packet delivery ratio even in high mobility, because ODMRP uses the mesh topology. If one route is fail, there is possible another route to delivery packet successfully. However, ODMRP used more system resources such as bandwidth than the other two approaches. AMRoute is good with no mobility, but it has very low packet delivery ratio in high mobility. Its packet delivery ratio is getting worse as the mobility speed increases. One reason is that AMRoute may build inefficient share trees and has the problem of loop paths. Route maintenance in HOMRP is initiated by each group member. Each member only maintains those members that are close to it, so each member knows available links to other members clearly. Therefore, compared to AMRoute, HOMRP increases 50% packet delivery ratio. It has slightly lower packet delivery ratio at high mobility speeds compared to ODMRP, because the performance of the unicast protocol decreases at high mobility speeds.

Figure 12 illustrates control overhead as a function of mobility speed among different approaches. ODMRP has the highest control overhead, because it needs to maintain the forwarding group members and transmit JOIN_TABLE during the member discovery period [5]. Since AMRoute only creates one multicast tree per multicast group for multiple sources, AMRoute is more efficient than ODMRP. We

can see HOMRP has the lowest control overhead. This is only multicast group members need to maintain multicast routes in HOMRP. A multicast member only needs to maintain those members that are close to it, but not the whole members.

Therefore, the sizes of exchanged control packets can be reduced. The two-level flooding scheme also reduces the range of flooding to avoid unnecessarily control packets. Those contribute to low control overhead. In sum, HOMRP reduces 18%

control overhead compared to ODMRP. As to AMRoute, HOMRP reduce 12%

control overhead than AMRoute.

Figure 13 illustrates the end-to-end delay per data packet delivered. All packets transmissions in AMRoute are by unicast tunnels, so that AMRoute has the worst end-to-end delay. ODMRP maintains forwarding group members for packet forwarding. This increases the speed of packet delivery. HOMRP has higher end-to-end delay than ODMRP, because some parts of transmissions in HOMRP are by unicast tunnels. Since HOMRP transmits packets by multicasting in local trees, it reduces 9% end-to-end delay compared to AMRoute.

Figure 11：Packet delivery ratio as a function of mobility speed among different approaches.

Figure 12：Control overhead as a function of mobility speed among differnent approaches.

Packet delivery ratio (%) ^ODMRP

HOMRP

Figure 13：End-to-end delay as a function of mobility speed among different approaches.

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5

1 2 3 4 5 6

Mobility speed (m/sec)

End-to-end delay (sec)

AMROUTE AOMRP ODMRP

Chapter 5

CONCLUSIONS AND FUTURE WORK

5.1 Concluding Remarks

The proposed hybrid overlay multicast routing protocol (HOMRP) has been presented in this thesis. It integrates multicasting and unicast tunnels for packet transmissions. It creates multiple local multicast trees where each pair of parent node and child node is at a distance of one-hop. It uses multicasting to deliver multicast packets in local multicast trees to provide efficient data forwarding.

Unicast tunnels are used for transmitting packets between local multicast trees. It avoids looping routes between local multicast trees by assigning each local tree a tree ID. Two-level flooding is used for member discovery, which limits the value of TTL in a packet to reduce the control overhead of flooding. Due to the dynamic network topology in MANETs, HOMRP uses a tree consolidation scheme for highly efficient multicast delivery. In addition, HOMRP does not require to use any specific unicast routing protocol; hence it can operate with any unicast routing protocols. In addition, each multicast group only needs to maintain a route structure for multiple sources; this is efficient for route maintenance.

Simulation results have shown that HOMRP have a higher packet delivery ratio, lower control overhead and lower packet delay as the group size become larger. In addition, it also performs well with small group sizes, because of the high performance of the unicast routing protocol. In sum, HOMRP improves 50% packet delivery ratio and reduces 9% end-to-end packet delay compared to AMRoute. In addition, HOMRP provides high packet delivery ratio which is close to ODMRP and reduces 18% control overhead compared to ODMRP. Therefore, our HOMRP is an efficient multicast routing protocol, which is very suitable for a large multicast group with a large number of multiple sources (senders).

5.2 Future Work

HOMRP is a hybrid overlay multicast routing protocol based on a unicast routing protocol. We adopt DSR as the underlying unicast protocol for HOMRP.

There are other unicast routing protocols that have been proposed for MANETs, such as AODV, DSDV and OLSR. We will evaluate the performance of HOMRP with different unicast routing protocols. By this evaluation, we can select a more suitable unicast routing protocols to further improve the performance of HOMRP.

Since HOMRP is efficient for a large multicast group with a large number of multicast sources (senders), we can use HOMRP to implement multicast applications with multiple sources, such as real time games with multiple players or instant video stream exchanging with multiple senders. This deserves for further study.

Bibliography

[1] Carlos de Morais Cordeiro, Hrishikesh Gossain, and Dharma Agrawal,

“Multicast over Wireless Mobile Ad Hoc Networks: Present and Future Directions,” IEEE Network, Volume: 17 Issue: 1, pp. 52 -59, Jan.-Feb. 2003.

[2] Li Xiao, A. Patil, Yunhao Liu and L.M. Ni, “Prioritized Overlay Multicast in Mobile Ad Hoc Environments,” IEEE Computer, Volume: 37 Issue: 2, pp. 67 -74, Feb. 2004.

[3] E. M. Royer and C. E. Perkins, “Multicast Operation of the Ad Hoc On-Demand Distance Vector Routing Protocol,” in Proc. of ACM MOBICOM, Aug. 1999, pp.

207 - 218

[4] C.W. Wu, Y.C. Tay, and C.-K. Toh, “Ad Hoc Multicast Routing Protocol Utilizing Increasing id-numberS (AMRIS) Functional Specification,” Internet Draft, Nov. 1998.

[5] Sung Ju Lee, William Su, and Mario Gerla, “On-Demand Multicast Routing Protocol,” in Proc. of IEEE WCNC, Sept. 1999, pp. 1298 - 1302.

[6] C.-C. Chiang, M. Gerla, and L. Zhang, “Forwarding Group Multicast Protocol (FGMP) for Multihop, Mobile Wireless Networks,” ACM/Baltzer Journal of Cluster Computing: Special Issue on Mobile Computing, Volume: 37, pp.

187–96, 1998.

[7] Subir Kumar Das, B. S. Manoj, and C. Siva Ram Murthy, “A Dynamic Core based Multicast Routing Protocol for Ad hoc Wireless Networks,” in Proc. of ACM MobiHoc, June 2002.

[8] Yao Zhao, Leirning Xu, and Meilin Shi, “On-Demand Multicast Routing Protocol with Multipoint Relay (ODMRP-MPR) in Mobile Ad-Hoc Network,”

in Proc. of International Conference on ICCT , Volume: 2, April 2003, pp. 1295

- 1300.

[9] Thomas Heide Clausen and Philippe Jacquet, “Optimized Link State Routing Protocol,” in Proc. of International Conference on INMIC, Dec. 2001, pp.

62-68.

[10] J.J. Garcia-Luna-Aceves and E.L. Madruga, “The Core-Assisted Mesh Protocol,” IEEE Journal on Selected Areas in Communications, vol. 17, no. 8, pp. 1380 -1394, Aug. 1999.

[11] E. Bommaiah, M. Lui, A. McAuley, and R. Talpade, “AMRoute: Adhoc Multicast Routing Protocol,” Internet Draft, Aug. 1998.

[12] Chao Gui and Prasant Mohapatra, “Efficient Overlay Multicast for Mobile Ad Hoc Networks,” in Proc. of IEEE WCNC, March 2003, pp. 1118 - 1123

[13] Guangyu Pei, M. Gerla, and Tsu-Wei Chen, “Fisheye State Routing: A Routing Scheme for Ad Hoc Wireless Networks,” in Proc. of ICC, Volume: 1, June 2000, pp. 70 -74.

[14] A. Laouiti, A. Qayyum and L. Viennot, “Multipoint Relaying: An Efficient Technique for Flooding in Mobile Wireless Networks,” in Proc. of 35th Annual Hawaii International Conference on System Sciences (HICSS'2002), April 2003,

pp. 1295 - 1300.

[15] UCLA Parallel Computing Laboratory and Wireless Adaptive Mobility Laboratory, “GloMoSim: A Scalable Simulation Environment for Wireless and

Wired Network Systems”, [online]. Available:

http://pcl.cs.ucla.edu/projects/domains/glomosim.html.

[16] R. Bagrodia, R. Meyer, M. Takai, Y. Chen, X. Zeng, J. Martin, and H. Y. Song,

“PARSEC: A Parallel Simulation Environment for Complex Systems,” IEEE Computer, Volime. 31, October 1998, pp. 77 - 85.

[17] J.Broch, D. Johnson, and D. Maltz. “The Dynamic Source Routing Protocol for

Mobile Ad Hoc Networks,”

http://www.ietf.org/internet-drafts/draft-ietfmanet-dsr-03.txt, IETF Internet Draft, Oct. 1999.

[18] Sung-Ju Lee, William Su, Julian Hsu, Mario Gerla, and Rajive Bagrodia, “A Performance Comparison Study of Ad Hoc Wireless Multicast Protocols,” in Proc. of INFOCOM, Mar. 2000, pp. 565-74.

[19] Josh Broch, David A. Maltz, David B. Johnson, Yih-Chun Hu, and Jorjeta Jetcheva, “A Performance Comparison of Multihop Wireless Ad Hoc Network Routing Protocols,“ in Proc. of IEEE/ACM MOBICOM ’98, Oct. 1998, pp.

85-97.

[20] IEEE Standards Department, Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications, Standard 802.11, 1999.

[21] Larry L. Peterson and Bruce S. David, “Computer Networks: A systems Approach, Second Edition,” Morgan Kaufmann Publishers, 2000.

[22] T. Ohta, S. Inoue and Y. Kakuda, “An Adaptive Multihop Clustering Scheme for Highly Mobile Ad Hoc Networks,” Autonomous Decentralized Systems, pp.

293 – 300, April 2003.

[23] V. Devarapalli and D. Sidhu, "MZR: A Multicast Protocol for Mobile Ad Hoc Networks," in Proc. of IEEE ICC, Jun. 2001, pp. 886 - 891.

[24] Xiaofeng Zhang and L. Jacob, “Multicast Zone Routing Protocol in Mobile Ad Hoc Wireless Networks,” in Proc. of IEEE International Conference on Local Computer Networks, Oct. 2003, pp. 150 – 159.

[25] L. Ji, and M. S. Corson, “Differential Destination Multicast — A MANET Multicast Routing Protocol for Small Groups,” in Proc. of INFOCOM, 2001, pp.

1192 – 02.

[26] Chun-Chieh Huang, Ruay-Shiung Chang and Ming-Huang Guo,

“Weight-based Clustering Multicast Routing Protocol for Mobile Ad Hoc Networks,” in Proc. of IEEE WCNC, Volume: 2 , March 2003, pp. 1112 - 1117.