On the Tradeoff Between Effectiveness and Scalability of Measurement-based Admission Control

One of the key problems to be solved in Mesurement-Based Admission Control is to efficiently utilise the information provided by the sources and measured by the network. There is a direct relationship between the amount of information available and the resulting effectiveness of the admission algorithm. In this paper we analyse the tradeoff between complexity and effectiveness using a practial class of AC methods in the context of Differentiated Services Internet.