Nalanda

Citation: Broch, J., Maltz, D. A., Johnson, D. B., Hu, Y., and Jetcheva, J. 1998. 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 (Dallas, Texas, United States, October 25 - 30, 1998). W. P. Osborne and D. Moghe, Eds. MobiCom ‘98. ACM, New York, NY, 85-97. DOI= http://doi.acm.org/10.1145/288235.288256

This paper presents a comparison of four wireless routing protocols used in ad hoc networks - DSDV, TORA, DSR and AODV - with tweaks to DSDV and AODV. Tests are run in a simulator, with a number of scenarios involving nodes moving at varying speeds. The metrics used for comparison are the packet delivery ratio, routing overhead and path optimality.

The most surprising of the four, to me, was TORA, for its in-order delivery requirement, which seems an impractical choice, especially for a wireless environment. Unsurprisingly, TORA fares the worst in the performance tests, with wildly varying performance under similar conditions, and undergoing congestion collapse beyond a certain number of sources in the network.

DSR provides a consistently high packet delivery ratio, closely followed by AODV, albeit both with increasing routing overhead as the number of sources increase (since they are on-demand protocols), though the incremental cost of adding a source does decrease with the number of sources. Even so, the routing overhead for both these protocols is lower than either DSDV or TORA. DSDV also exhibits little variation in the packet delivery ratio, though it does function best under lower degrees of mobility; at higher mobility, it can completely fail to converge. As might be expected, DSDV’s routing overhead remains largely constant.

The discussion of the effect of packet size was interesting: DSR is more expensive than AODV when routing overhead is measured in bytes. However, the cost to acquire the medium for transmission of a packet is much higher than the incremental cost of adding these extra bytes to a packet.