Tighter end-to-end delay bound for aggregate traffic scheduling

Aggregate scheduling techniques have been developed and deployed to avoid complexities of Integrated Services model. A major breakthrough in this regard is in terms of scheduling with Packet Scale Rate Guarantee. However, fundamental limitation of the method lies in unbounded growth of end-to-end delay, whenever the strict relation between utilization and number of hops traversed, fails. A tighter end-to-end delay bound is derived in this paper for highest priority packet, called EF packet. The paper also presents an analytical expression for quantifying maximum possible increase in delay of an EF packet. Using these results, a modified relation between utilization and number of hops traversed by a packet is shown. The modified relation is shown to allow higher utilization compared to earlier result.