Internet performance and control of network systems

This Special Issue contains selected papers that were presented at the Internet Performance and Control of Network Systems conference that was held in Boston, MA, 6–7 November 2000. The conference was part of the International Symposium on Voice, Video and Data Communications, sponsored by SPIE— The International Society for Optical Engineering. The purpose of the conference was to promote the discussion on the development of performance evaluation techniques, traffic control principles and traffic engineering methods and practices. The selected papers are all extended versions of papers presented at the conference and included in [1]. The papers have been reviewed by the following subject-matter experts, whose efforts are reflected in the high quality of the papers presented here: E. van den Berg (Telcordia Technologies), J.L. van den Berg (KPN Research), A. Brandt (Humboldt University of Berlin), K. Christensen (University of South Florida), H.R. Gail (IBM), B. Haverkort (Technical University of Aachen), Y. Kogan (AT&T Labs), S. Low (California Institute of Technology), M.R.H. Mandjes (CWI/Lucent Technologies/Twente University), P.K. Reeser (AT&T Labs), T. Strayer (BBN Technologies). The authors wish to thank W. Bux, Editor-in-Chief, for his interest in the conference and his efforts in publishing this Special Issue. They also like to thank W.S. Lai (AT&T Labs) for his advice in organizing the conference and preparing this Special Issue. The selected papers cover various topics within the field of performance and control of network systems. The contributions of the selected papers are outlined below. Squillante, Xia and Zhang study the performance of high-volume commercial Web sites. In particular, they study the control policy for each server in a multiclass queuing network that maximizes a particular profit function, or minimizes a particular cost function, across the different classes of Internet services. To this end, they formulate a simple stochastic tandem queuing network with feedback that captures a fundamental tradeoff between the diverse set of Internet services supporting customer needs and the different importance levels of these services to both the customer and the e-commerce merchant. For this model, they derive the optimal Brownian deviation control for the corresponding diffusion control problem, which can be shown to be asymptotically optimal for the original stochastic network. Simulation is then used to compare the effectiveness of this control policy with other scheduling disciplines. Houck and Meempat study the performance of voice-over-IP (VoIP) and present a new methodology for call admission control (CAC) and load balancing (LB) for VoIP services. Their approach employs a network function referred to as call admission manager (CAM) that keeps track of the voice occupancy levels across the IP network. Based on this model, an exact algorithm is presented to perform control functions. Additionally, an approximate variant is presented that allows scalability for large-scale networks. This algorithm employs a measurement-based strategy, where IP routers directly monitor the voice traffic volumes on the links to assist in inferring network congestion. The framework presented is suitable