Tamer Dag

Proactive/Reactive Congestion Control Schemes for High Speed Networks

Thursday, December 16, 1999
11:00 AM
406 Egan

Abstract

The next generation high speed communication networks which are characterized by large bandwidth propagation delay products are expected to support applications with diverse traffic characteristics and Quality of Service (QoS) requirements. Due to the statistical multiplexing employed in these networks, congestion control mechanisms are necessary for the delivery of the QoS required by the applications.

Various reactive flow control schemes have been developed in the past for congestion control of the non-real time (best effort) applications. It has been well established that these schemes are not effective for networks with large bandwidth propagation delay products, as the feedback information generated in the network can be outdated by the time it reaches the sources. In this thesis, this problem (referred to as the bandwidth mismatch problem) is investigated and a proactive feedback based flow control scheme which attempts to eliminate the bandwidth mismatch problem by generating and transmitting early feedbacks is introduced. Either by prediction or cooperation, the proactive feedback based flow control scheme is able to improve the delivered QoS.

Various congestion avoidance schemes have also been developed in the past, attempting to prevent congestion through traffic smoothing. Such schemes have been extensively studied for non-real-time applications. For real-time applications which are delay sensitive, the smoothing operation can reduce the deadline margin of the transmitted cells and may lead to increased cell losses due to deadline violations. In this thesis, the impact of various traffic shaping schemes on the buffer overflows and deadline violations is investigated and some simple metrics that capture the potential for buffer overflows and deadline violations are introduced. Such metrics can be useful for the design of effective traffic shaping schemes that balance the adverse effects of buffer overflows and deadline violations.

Thesis Committee:
Prof. I. Stavrakakis (advisor, external member)
Prof. E.S. Manolakos (co-advisor)
Prof. M. Salehi