Delay Differentiation by Reserving Space in Queue

Introduction: A common problem in packet-based communication networks such as IP (Internet Protocol) or ATM (Asynchronous Transfer Mode) is the provisioning of adequate QoS (Quality of Service) guarantees to the traffic flows in the network. To assess the scheduling performance of a node in such networks, we can regard the node as a queue to which two flows of packets arrive: the first flow (type 1) carries delaysensitive traffic while the second flow (type 2) represents best-effort traffic. To accommodate the more stringent delay requirements of the packets of type 1 (1-packets), the scheduling mechanism must give the 1-packets a certain priority over the 2-packets. Under FIFO (First-In First-Out), no special arrangements are made to prioritise the delaysensitive flow. On the other hand, the most extreme way of priority scheduling is AP (Absolute Priority) [1], for which the 1-packets will always be scheduled first. AP guarantees the lowest possible delay for the type 1 traffic, but at the cost of increasing the delay for the 2-packets which may result in packet starvation or time-out.