WF2Q: Worst-Case Fair Weighted Fair Queueing

The Generalized Processor Sharing (GPS) discipline is proven to have two desirable properties: (a) it can provide an end-to-end bounded-delay service to a session whose tra c is constrained by a leaky bucket; (b) it can ensure fair allocation of bandwidth among all backlogged sessions regardless of whether or not their tra c is constrained. The former property is the basis for supporting guaranteed service tra c while the later property is important for supporting best-e ort service tra c. Since GPS uses an idealized uid model which cannot be realized in the real world, various packet approximation algorithms of GPS have been proposed. Among these, Weighted Fair Queueing (WFQ) also known as Packet Generalized Processor Sharing (PGPS) has been considered to be the best one in terms of accuracy. In particular, it has been proven that the delay bound provided by WFQ is within one packet transmission time of that provided by GPS. In this paper, we will show that, contrary to popular belief, there could be large discrepancies between the services provided by the packet WFQ system and the uid GPS system. We argue that such a discrepancy will adversely e ect many congestion control algorithms that rely on services similar to those provided by GPS. A new packet approximation algorithm of GPS called Worst-case Fair Weighted Fair Queueing (WFQ) is proposed. The service provided by WFQ is almost identical to that of GPS, di ering by no more than one maximum size packet.