Research on multi-virtual queue based on flow estimation
1Department of Computer and Information Engineering, Heze University, Heze 274015, Shandong, China
2Key Laboratory of computer Information Processing, Heze University, Heze 274015, Shandong, China
To solve the equity issues in the network bandwidth competition of TCP/UDP mixed flow, we have raised an AQM algorithm framework based on “granularity”, which is aimed to realize the divide-and-rule goal of the conservative TCP and the greedy UDP so as to alleviate the equity issues of the network resources caused by different protocols by designing a common framework. Furthermore, under the AQM framework of “granularity”, we have brought forth Multi-Virtual Queue based on Flow Estimation (VFQ) and we have also given the detailed implementation schemes for different modules of VFQ algorithm. In order to further verify the feasibility and validity of VFQ algorithm, we have designed 5 computer network congestion scenarios in AQM network simulation platform and made simulation comparisons with the classical ARED, PI and Blue algorithms. Numerous simulation results demonstrate that in contrast with other AQM algorithms, VFQ can better adapt to various sudden network congestion scenarios and it has better response speed and queue management performance.