A robust, worst-case design framework is developed for a tandem router network based on rival queueing models. Worst-case design is concerned with determining the best design or strategy in the most hostile operating environment, i.e. simultaneously with worst-case model parameters. In the router network, end-to-end performance is computed as the mean transmission delay subject to an upper limit on the rate of losses and finite capacity queueing and recovery buffers. The worst-case optimal ratio of arrival-buffer size to recovery-buffer size is determined, which is a critical quantity that affects both loss rate and transmission time. The impact of the retransmission probability is also investigated.
Information from pubs.doc.ic.ac.uk/worst-case-routers.