Group priority scheduling

For many applications, the end-to-end delay of an application-specific data unit is a more important performance measure than the end-to-end delays of individual packets within a network. From this observation, we propose the idea of group scheduling. Specifically, consecutive packet arrivals in a flow are partitioned into groups, and the same deadline (called group priority) is assigned to every packet in a group. We first present an end-to-end delay guarantee theorem for a network of guaranteed-deadline (GD) servers. The theorem can be instantiated to obtain end-to-end delay bounds for a variety of source control mechanisms and GD servers. We then specialize the delay guarantee theorem to group scheduling for a subclass of GD servers. We work out a detailed example to demonstrate how to use group scheduling in a particular class of networks. The advantages of group scheduling are discussed and illustrated with empirical results from simulation experiments