Congestion Control in Integrated Services Networks

Abstract

Since feedback control is developed to be able to handle significant modelling errors and inaccuracies it is not surprising that preliminary results based on these models were very successful and encouraging for pursuing this approach further. It is safe to state that models which use non-linear control theory can be suitable for congestion control and that this seems to be the area we must concentrate our efforts. Early results given by the IDCC model are very promising and encourage us to investigate this model further to see if it can be proposed as a more general congestion control scheme suitable for ISN networks. Further analysis and formal evaluation of the model is required. This means that the following steps must be done: • formal analysis and evaluation of the model • proof of fairness • design of general and comprehensive scenarios • implementation in multiple suite of networks (e.g. ATM, IP Diff-Serv) • extension to large scale • extensive simulations and evaluation of results IDCC is a good starting point for further research.