一种基于Bloom filter的加强队列公平性改进算法  

Improved algorithm enhancing queue fairness based on Bloom filter

作　　者：何建新[1,2] 王建新[2] 

机构地区：[1]湖南城市学院计算机科学系,湖南益阳413000 [2]中南大学信息科学与工程学院,长沙410075

出　　处：《计算机应用研究》2010年第8期3100-3103,共4页Application Research of Computers

基　　金：湖南省教育厅2009年度教学改革研究资助项目;国家自然科学基金资助项目(90304010);湖南省杰出青年基金资助项目(06JJ10009)

摘　　要：随着Internet基于非TCP的应用不断涌现,基于异质流网络拥塞控制公平性研究越来越重要。针对流与流之间传输的公平性问题,基于BLUE算法,结合Bloomfilter,提出了一种改进的AQM算法EFBLUE。通过仿真实验对新算法从分组丢失率、吞吐量、延时等方面的性能进行了测试并与RED算法进行了性能对比。NS2仿真实验结果表明,该算法只需使用极少量的状态位和很小的缓存空间就能较好地鉴别出非响应流,并限制其速率,保护TCP流免受非响应流影响,实现了流量传输的公平性。最后对EFBLUE的性能优化问题作了进一步的分析。With the teeming of application based on the non TCP in the Internet,fairness research based on heterogeneous flow network congestion control becomes more and more important. Combining with Bloom filter,this paper put forward an improved AQM algorithm based on the BLUE algorithm to solve fairness between the stream and stream. Tested the new algorithm through network simulation,had done simulation analysis and comparison with the RED algorithm from aspects of packet loss rate,throughput and delay by using NS2. The NS simulation results show that the algorithm can distinguish the unresponsive stream and restrict the velocity,protect the TCP stream from being affected by the unresponsive stream so as to actualize the fairness of stream transmission by little status bit and small buffers. At last,analyzed the performance optimization to EFBLUE.

关 键 词：拥塞控制 公平性 异质流网络 主动队列管理 布鲁姆过滤器 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]