基于区分服务中的队列调度算法研究

发布时间：2018-10-26 21:09

【摘要】：随着Internet网络技术的飞速发展,用户对网络服务质量(QoS)的要求越来越高。传统的IP网络主要针对非实时业务设计,不保证传输数据的实时性,只提供“尽力而为”的服务,因此难以满足复杂网络中不同业务的QoS要求。在此背景下,IETF联盟提出了区分服务模型DiffServ,该模型下的队列调度策略成为IP QoS核心技术。首先,论文详细介绍了综合服务IntServ和区分服务DiffServ两种模型的设计背景和实现原理,分析了两种模型的优缺点。网络的复杂性使得任何一种模型独立使用都不能实现网络端到端的QoS,因此将IntServ与DiffServ两种服务模型结合起来,相互协调补充。其次,论文分析了网络产生拥塞的原因以及拥塞控制的方式。队列调度是缓解网络拥塞最直接的方法,因此对队列调度机制进行了重点研究,作为本论文的核心内容。在轮询类DWRR算法的基础上提出改进型算法VDWRR,并插入了新的队列管理思想。用NS2网络仿真器验证了VDWRR算法的可行性和有效性。最后,为了弥补轮询类算法和优先级算法(PRI)各自缺陷,将PRI算法与VDWRR算法相结合,提出了PRI+VDWRR调度策略。用NS2网络仿真器对此调度策略的性能进行了评价,仿真结果表明,该调度策略在保证EF流实时性的同时,为非EF流量提供可靠的带宽保证。
[Abstract]:With the rapid development of Internet network technology, users are demanding more and more high quality of service (QoS). Traditional IP networks are mainly designed for non-real-time services, which do not guarantee the real-time transmission of data, and only provide "best effort" services, so it is difficult to meet the QoS requirements of different services in complex networks. In this context, the IETF Alliance proposed the differentiated Service Model (DiffServ,), the queue scheduling strategy under the model becomes the core technology of IP QoS. Firstly, the paper introduces the design background and implementation principle of integrated service IntServ and differentiated service DiffServ in detail, and analyzes the advantages and disadvantages of the two models. The complexity of the network makes it impossible for any model to implement end-to-end QoS, alone. Therefore, two service models, IntServ and DiffServ, are combined to complement each other. Secondly, the paper analyzes the causes of network congestion and congestion control methods. Queue scheduling is the most direct way to alleviate network congestion. Based on the polling class DWRR algorithm, an improved VDWRR, algorithm is proposed and a new queue management idea is inserted. The feasibility and validity of VDWRR algorithm are verified by NS2 network simulator. Finally, in order to make up for the shortcomings of polling class algorithm and priority algorithm (PRI), a PRI VDWRR scheduling strategy is proposed by combining PRI algorithm with VDWRR algorithm. The NS2 network simulator is used to evaluate the performance of the scheduling strategy. The simulation results show that the scheduling strategy not only guarantees the real-time performance of EF flows, but also provides reliable bandwidth guarantee for non-EF traffic.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP393.09;TP301.6

【相似文献】