A FLEXIBLE TRAFFIC SHAPER FOR HIGH-SPEED NETWORKS - DESIGN AND COMPARATIVE-STUDY WITH LEAKY BUCKET

Maximizing bandwidth utilization and providing performance guarantees, in the context of multimedia networking, are two incompatible goals. Heterogeneity of the multimedia sources calls for effective traffic co ntrol schemes to satisfy their diverse Quality of Service (QoS) requir ements. These include admission control at connection set up, traffic control at the source ends and efficient scheduling schemes at the swi tches. The emphasis in this paper is on traffic control at the source end. Most multimedia sources are bursty in nature. Traffic shapers hav e been mainly studied hitherto from the point of view of their effecti veness in smoothing the burstiness. The Leaky Bucket (LB) scheme, to c ite an example, is a mean rate policer smoothing at the token generati on rate. Studies on bursty sources show that burstiness promotes stati stical multiplexing at the cost of possible congestion. Smoothing, on the other hand, helps in providing guarantees at the cost of utilizati on. Thus need for a flexible scheme which can provide a reasonable com promise between utilization and performance is imminent. Recent studie s have also questioned the suitability of LB for policing real-time tr affic due to the excessive delays. We argue for a policy which is less stringent on short term burstiness than the LB. We propose a new traf fic shaper which can adjust the burstiness of the input traffic to obt ain reasonable bandwidth utilization while maintaining statistical ser vice guarantees. The performance study is conducted in two parts. In t he first part, we study the effect of varying the shaper parameters on the input characteristics. In the second part, we dimension our schem e and a LB equivalently and compare the mean and peak rate policing be havior with delay and loss as the performance parameters. Adopting a l ess stringent attitude towards short term burstiness is shown to resul t in considerable advantage while policing real-time traffic. Future r esearch possibilities in this topic are explored.