A performance comparison of Multi-Hop Wireless Ad Hoc Network Routing Protocols

This paper compares different network routing protocols including the Destination-sequenced distance vector, Temporally-Ordered routing Algorithm, Dynamic source routing, and Ad Hoc On-Demand Distance Vector. The paper gives a brief summary of the different routing protocols and their implementation of each one. This provides a good amount of background for the paper. The paper takes into account the lower level layer characteristics into their simulation which they claim has not previously been done.

The way they test these protocols is that they setup the nodes to move around. The algorithm used to move the nodes cases each node to move and remain stationary for a pause time. When pause time is up, then it picks a destination within a 1500m x 300m space, and then moves to that location in a speed uniformly distributed between 0 and a max speed. Then it waits for another pause time (also random), and then moves again. Different packet sizes and source communicators were used to observe the different effects of the variants. The metrics used were: Packet delivery ratio, routing overhead, and path optimality.

The results showed that DSDV delivered all packets when mobility rate and movement speed are low, but does not perform well when mobility increses. TORA was a little better than DSDV, but introduced a high overhead in the routing protocol. DSR was good at all mobility rates, but the paths need to be determined ahead of time, and introduces more overhead at the source. The AODV performs as well as DSR at all mobility rates, but is more expensive to implement than DSR. This paper however doesn't really consider the effects of congestion on the protocols. In fact, it seems that most wireless protocols focus on the link loss and link quality of the physical layer, while wired protocols focus on congestion control. It makes sense, but i think wireless protocols would still need to consider congestion, and there probably needs to be a mechanism separating congestion loss and quality loss.