ABSTRACT
Emerging high speed broadband networks, such as those based on the Asynchronous Transfer Mode, will carry a variety of traffic such as video, voice and data. The problem of determining feasible sets of guarantees has attracted a lot of attention. This chapter focuses on delay guarantees. Many approaches to establishing hard delay guarantees follow the seminal work of R. L. Cruz. There are two main aspects to determining delay bounds: describing the traffic, and specifying a scheduling rule at each node in the network. The traffic model generally includes a bound on the burstiness, and specification of the average and maximum rates at which the source generates traffic. There are several techniques available for approximating the mean delay in a variety of queueing systems. The chapter examines upper bounds on the mean delay in a tandem system of switches, and upper bounds for the first two moments of delay when service times are deterministic and equal for all packets.
