Summary: | 碩士 === 國立東華大學 === 電機工程學系 === 92 === Weighted Fair Queueing (WFQ) is a popular scheme because it guarantees bandwidth and bounded delay. However, the session delay bound is related to the allocated service sharing limit. Therefore, a high service sharing limit has lower delay boundary. To make WFQ more flexible, a priority system is added to give WFQ better bandwidth guarantee and bounded delay.
In this paper, we propose a Queueing algorithm that combines WFQ and a sliding window called Sliding Weighted Fair Queuing (SWFQ). In this sliding window, the session bandwidth guarantee and bounded delay are adjusted according to the session service sharing and priority. SWFQ decouples the effect of the service sharing limit in WFQ. In other words, if a session with low sharing but a high priority could still receive a small delay. This algorithm combines share-driven scheduling and priority-driven scheduling. This algorithm gives WFQ simultaneous share-driven and priority-driven characteristics. Therefore, the proposed method balances the share and priority-driven characteristics and allows WFQ application to various network environments. The proposed Queueing algorithm is applied to either IPv4 or IPv6 networks as long as they can provide adequate quality of service (Qos) support.
|