Our present model of distributed computing is built around network-based services. However, as yet, we do not have good analytical models for such systems. One way of analysing such systems is by using gate-limited queuing models. However, practical solutions for gate-limited models have not been readily available.
This talk looks at development of a new Markov Model for gated-limited service. An approximate solution is presented by making use of the Standard Partial Batch model. The solution is then applied to a system designed for video-on-demand applications which are served over the network using pre-fetching techniques. Finally the effects of different network types as well as different network loads are also examined.
Contact: Richard G. Clegg (richard@richardclegg.org) or Keith Briggs ()