Minimizing End-to-End Delays in Linear Multihop Networks

Abstract

The optimum designs of linear multihop networks that can minimize the end-to-end transmission delays are studied in this paper. In multihop networks, more hops mean a shorter transmission distance per hop but more overheads; thus, it is critical to identify the optimum number of hops that can minimize the end-to-end delay. Two network configurations are considered: one with deterministic equidistant relays and the other with random relays uniformly distributed between the source and destination. The average end-to-end delays of both networks operating in the presence of noise and/or interference are expressed as functions of a number of system parameters, such as channel propagations, noise, interference, protocol overhead, and the geometric distributions of the relays. Based on the analytical results, the optimum number of hops for various system configurations is obtained. It is shown that networks with equidistant relays can achieve a shorter end-to-end delay than those with uniformly distributed relays.