On Path Metrics and Routing Protocols for Wireless Networks

De Couto’s et. al. on the expected transmission count (ETX) metric works on the premise that the minimum hop-count metric does not necessarily provide the best performance. One direct contribution of this paper is their examination of the effects (or side-effects) of solely relying on the minimum distance. They also contribute the design and implementation of ETX in DSDV and DSR routing protocols and showed that it yields better performance on multi-hop wireless networks than a scheme based solely on minimum hop counts. Using bi-directional per-link packet loss measurements, ETX is a scheme that finds high quality paths to reduce transmissions. ETX incorporates in its formulation the wide range of link loss ratios, the asymmetry of these loss ratios and the interference between hops. One thing that stands out is the overhead of computing for the ETX would prevent it from being used when the network changes either through mobility or changes in the signal strength of the nodes.

–

Broch et al. compare the peformance of four ad-hoc network protocols namely DSDV, TORA, DSR and AODV.  They provide extensions to the ns-2 simulator to model IEEE 802.11 MAC and physical layers, node mobility and radio network interfaces. The study was very detailed with the assumption clearly explained. I also appreciate the discussion of the effects and results of the simulation which showed that depending on the need and cost (overhead), the nature of the application and problem must be taken into account when selecting an appropriate protocol to use.