摘要
针对互联网现有服务质量(QoS:quality of service)模型的扩展性差、端到端延时高等问题,提出了基于域内全局流标签的QoS保障模型。该模型以WFQ队列调度模型为基础,采用域内全局流标签进行资源预留和路由转发,能够实现灵活的服务质量保证,具有更强的可扩展性。在此基础上,利用网络微积分理论对所提模型进行了端到端延时分析,给出了该模型端到端服务曲线和延时上界的表示式。理论分析和数值实验表明基于域内全局流标签的QoS保障模型具有比现有模型更低的端到端延时上界。
Each of the existing Intemet Quality of Service （QoS） models has some difficult problems to be solved, including weak scalability, low end-to-end （E2E） QoS guarantee, etc. Aiming at this problem, this paper proposes a novel QoS model on the basis of Intra- domain Global Flow Label （IGFL）, which adopts the Weighted Fair Queuing （WFQ） mechanism to serve the flows. Our model uses the IGFL to do resource reservation and fast forwarding, so itcan achieve flexible QoS provisioning and good scalability. Moreover, we ana- lyze the E2E delay bound of the model by network calculus. We give the formulations of E2E service curve and the E2E delay bound of the IDFL-based QoS model. Theoretical analysis and numerical experiments show that our model has lower E2E delay bound than existing ones.
出处
《软件》
2012年第8期14-19,共6页
Software
基金
国家973重点基础研究发展规划(2007CB307101)和(2007CB307106)
关键词
服务质量
端到端延时
标签转发
网络微积分
quality of service
end to end delay
label switching
network calculus
