网络端到端流量的QoS优化技术研究

发布时间：2018-05-12 16:14

本文选题：服务质量 + 综合服务　；参考：《电子科技大学》2014年硕士论文

【摘要】：在网络信息高速发展的时代,网络多媒体业务不断涌现,网络用户越来越多地提出包括多媒体业务通信在内的综合服务要求。传统分组交换网数据传输多面向对实时性要求不高的数据通信(如FTP和Email等),采用TCP/IP协议主要为了优化网络业务流吞吐量并保证其通信可靠性。而现今网络是处于多种业务流并存的状态,其中包括一些实时性较强的多媒体应用,例如VoIP等。这些应用不仅包括文本信息,而且还包括语音(Voice)、图形(Graphic)、视频(Video)等形式的信息。这些信息在传输时不仅要求网络具有较高的带宽保证,同时要求具有传输低延迟、低抖动和低错误率等性能。由此可见,在网络不断发展的情况下,新兴业务流特别是对实时性较高的多媒体应用对网络提出了更高的服务质量(Quality of Service,QoS)需求,需要网络提供更加有效的QoS控制和保证。有鉴于此,本文将对QoS优化技术进行研究。本文核心旨在对QoS网络技术中的关键技术——队列调度算法进行研究和优化。在文中通过对比研究分析常见队列调度算法,选择公平性较好的加权公平队列(Weighted Fair Queuing,WFQ)调度算法进行深入地研究,并在研究分析基础上根据其不支持实时性业务流的缺点提出改进方案。在实施改进方案中形成三种改进模型:基于权值改进的加权公平队列调度算法、基于优先级保证带宽的加权公平队列调度算法以及将两种相结合的支持实时业务的加权公平队列调度算法。最终经过NS-2网络仿真器在时延、吞吐量和丢包率QoS参数方面进行仿真,通过仿真分析表明:三种改进模型都能使加权公平队列调度算法很好地支持实时业务流。本论文是以先研究分析后提出改进方案为思路。在本文中首先对QoS网络应用背景进行阐述,然后对QoS网络技术发展过程中的主要模型作研究,这包括最大努力模型、综合服务模型、区分服务模型,对这三种模型先研究分析其原理与结构,分析优缺点,然后研究目前因特网所应用的主要服务模型及其关键技术。基于上述对目前应用较广泛的典型队列调度算法进行研究,并分析其优缺点。相继深入研究加权公平队列调度算法并提出改进方案,然后根据改进方案利用NS-2网络仿真软件进行模拟仿真,并对比分析改进算法与未改进算法仿真结果,最终通过仿真分析得到改进算法比原有算法在支持实时业务流方面得到了很大改善。
[Abstract]:In the era of rapid development of network information, network multimedia services continue to emerge, network users increasingly put forward comprehensive service requirements, including multimedia service communications. The traditional packet switching network data transmission is oriented to the data communication (such as FTP and Email, etc.) which does not require high real-time performance. The TCP/IP protocol is mainly used to optimize the throughput of network traffic and ensure its communication reliability. Nowadays, the network is in the state of multiple traffic, including some real-time multimedia applications, such as VoIP and so on. These applications include not only text information, but also voice, graphics, video and other forms of information. The transmission of these information requires not only high bandwidth guarantee, but also low transmission delay, low jitter and low error rate. It can be seen that with the continuous development of the network, the new traffic, especially for the multimedia applications with high real-time performance, puts forward a higher quality of service (QoS) requirement for the network, which requires the network to provide more effective QoS control and guarantee. In view of this, this paper will study the QoS optimization technology. The core of this paper is to research and optimize queue scheduling algorithm, which is the key technology of QoS network. In this paper, by comparing and analyzing common queue scheduling algorithms, we select weighted Fair queuing WFQs with better fairness to do in-depth research. On the basis of research and analysis, an improved scheme is put forward according to its shortcoming of not supporting real-time traffic. In the implementation of the improved scheme, three kinds of improved models are formed: weighted fair queue scheduling algorithm based on weight improvement, A weighted fair queue scheduling algorithm based on priority guaranteed bandwidth and a weighted fair queue scheduling algorithm which combines the two algorithms to support real-time traffic are proposed. Finally, the NS-2 network simulator is used to simulate the delay, throughput and packet loss rate QoS parameters. The simulation results show that the three improved models can make the weighted fair queue scheduling algorithm support real-time traffic very well. This paper is based on the research and analysis, and then put forward the improvement plan as the train of thought. In this paper, the application background of QoS network is first expounded, and then the main models in the development of QoS network technology are studied, which include the best effort model, the integrated service model, the differentiated service model. First, the principle and structure of the three models are analyzed, and then the main service models and their key technologies used in the Internet are studied. Based on the above, the typical queue scheduling algorithms, which are widely used at present, are studied, and their advantages and disadvantages are analyzed. The weighted fair queue scheduling algorithm is studied in succession and an improved scheme is put forward. According to the improved scheme, the NS-2 network simulation software is used to simulate and simulate, and the simulation results of the improved algorithm and the unimproved algorithm are compared and analyzed. Finally, the simulation results show that the improved algorithm is much better than the original algorithm in supporting real-time traffic.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.06

【参考文献】