An investigation of multilevel service provision for voice over IP under catastrophic congestion

IEEE Communications Magazine(2004)

引用 12|浏览0
暂无评分
摘要
QoS has been the dominating issue faced in VoIP development today. In this article we investigate and empirically evaluate three DiffServ (Kilkki 1999, Blake et al. 1998) design strategies, WRED Dropper (Cisco Systems), PQ meter, and multiple differentiated meter (MDM), and study how multilevel communication services can be guaranteed for multiple VoIP classes. The novel aspect of our work is that we are trying to provide solutions with the use of features that are not currently available in real routers (Westhead, 2002). By taking advantage of network simulation techniques, a series of experiments have been designed and carried out under different network conditions by using the intersim simulation tool (Westhead, 2002). Our simulation results effectively reflect that: 1) all three designs can provide service differentiation for multiple VoIP classes; 2) the highest-priority VoIP class can always be protected under catastrophic congestion; and 3) each design has some pros and cons in terms of performance stability under varied traffic distribution and changed packet size mix. Based on this investigation, we conclude that MDM is comparatively better at providing even service distribution and staying robust against traffic distribution variation and packet size changes.
更多
查看译文
关键词
catastrophic congestion,intersim simulation tool,multilevel communication services,network simulation technique,service distribution,multiple voip class,network simulation techniques,quality of service,multiple differentiated meter,varied traffic distribution,internet telephony,traffic distribution,qos,multilevel service provision,multiple voip classes,voip development,telecommunication traffic,telecommunication network routing,simulation result,voice over ip,highest-priority voip class,traffic distribution variation,network simulator,stability,convergence,jitter,robustness
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要