A Bandwidth Bargain Model based on Adaptive Weighted Fair Queueing

Efficient usage of network bandwidth is a key factor of providing quality of service guarantees in the Internet. In this paper, a bandwidth bargain model is developed which aims to dynamically allocate network bandwidth based on the varying demand of packet flows. To achieve the design goal, an adaptive weighted fair queueing (AWFQ) algorithm is presented which is more flexible than the generic weighted fair queueing (WFQ). By using the estimation of flow arrival rate and weight adjustment approach, AWFQ has the ability to guarantee bandwidth requirements of active service within a pre-defined range, without compromising the guarantee to assured service. Furthermore, AWFQ can provide a specified minimum bandwidth to best-effort service, which cannot be offered by WFQ, when the amount of competing traffic exceeds the link capacity. The simulation results, which are obtained from different network topologies with various packet generating processes, verify that AWFQ is feasible and performs better than WFQ in the context of dynamic bandwidth allocation