Protocol Comparison

This paper was an analysis of various routing protocols on ad-hoc wireless networks. There wasn't anything particularly interesting. First, they create a simulator that will allow for accurate emulation of wireless links, and mobility on these links. As the work focuses on ad-hoc networks, it's mostly relevant to sensornets. Then they implement and run a variety of protocols with some reasonable workload.

Good result, the DSR algorithm is pretty much winner. The most interesting aspect of this paper was it's existence. The paper pretty much writes itself. Half of it is descriptions of the routing algorithms. The second half is all results. There is no future work. There's a good amount of work in getting it all set up, but it's just so unimaginative. Still, it's totally useful. I wouldn't have expected a paper like this out of CMU.

I'm interested in doing such work for delay tolerant networks. I don't think any particular motivating use case (such as ad-hoc) has emerged, which means that it maybe more difficult to set up a reasonable simulation.

It's surprisingly soft on results too. the DSR algorithm beats all others in almost all ways, losing only on bytes sent in overhead. This isn't emphasised in the conclusion. There must be a reason why, is it politics? It is bad form for a survey to make these judgements? Why be so generic?