Collaborative Research: Algorithmic Problems in Next Generation Networks

University Of Pittsburgh, Pittsburgh PA

Investigators

Abstract

This research involves the study algorithmic problems that arise in the design of next generation networking technology. Rapid growth of the internet user base, coupled with rapid parallel growth in mobile subscriber numbers, is creating powerful pent-up demand for wireless access to the internet/intranets and other data networks. Communication in wireless networks differs in some fundamental ways from communication in wired networks. In particular, the basic form of communication in wireless networks is broadcasting, as opposed to point-to-point communications in most wired networks. This research investigates the effect of switching to broadcast communication on server strategies in the client-server computing model, one of the most common computing paradigms. A wide range of real-time applications, such as multimedia conferencing, computer supported cooperated workspaces, remote medical diagnosis etc., has become very popular in recent years. These applications have dynamically changing bandwidth requirements. It is clear that controlling and allocating bandwidth for individual users and web browsers is becoming very important for bursty traffic flows utilizing highly constrained communications channels. From the network management's perspective, it is advantageous to allow the network to dynamically change the tariff parameters of the charging scheme in response to demand. This research investigates dynamic renegotiation and pricing problems that arise in the design of networks that give quality of service guarantees.