Priorities and Scheduling

Priority arises specifically when a label is assigned to each customer, specifying a level of preference for service. Scheduling is usually referred to policies to serve concurrent flows of customers. Priority and scheduling serve two purposes: creating differentiated service; and protecting customers against excessive demands of other customers that could hog the service capacity thus degrading unacceptably the quality of service experienced by less overwhelming customers. This chapter provides an overview of priority queueing systems with Poisson input traffic. It discusses server capacity sharing by analyzing the processor sharing policy, its generalization and possible practical scheduling algorithms that implement that policy. The chapter outlines the generalized processor sharing model and discusses ways to realize theoretical sharing models in a packet multiplexer. It also explores the scheduling of radio resources in the downlink of a cellular access network, as well as job dispatching in a cluster of computing servers.