web集群负载均衡系统的设计与实现

发布时间：2018-06-06 07:08

本文选题：互联网 + WEB集群　；参考：《电子科技大学》2014年硕士论文

【摘要】：随着Internet的飞速发展,人们的工作生活越来越离不开互联网,截至2013年12月,中国网民规模达6.18亿,互联网普及率为45.8%。网民数量的急剧增加带来的庞大的访问流量给很多受人欢迎的网络站点带来了巨大的挑战。热门站点每天要处理数亿的请求,爆炸式的访问次数增长使得建立更强的网络服务系统成为迫在眉睫的问题。为了给网民提供优质的体验,解决网络访问量剧增对网络带宽和服务器性能提出的需求,绝大多数网站通过构建WEB服务器集群来提高系统的处理能力。WEB集群技术是构建大型网站的关键技术之一。本文以WEB服务器集群的负载均衡调度为研究课题,对WEB集群的一些关键技术进行了研究,重点研究了负载均衡调度策略以及负载均衡系统实现的关键技术,并取得了以下成果:1.设计了一种新的基于HTTP请求内容的调度策略,当请求到达时根据HTTP请求内容将请求划分为动态请求和静态请求两类,并针对两类请求做出不同的调度决策,对于动态请求主要考虑负载的均匀分配,对于静态请求主要考虑提高请求的Cache命中率。并在算法中设计了基于模糊聚类的动态反馈调节机制,周期性的对WEB集群的负载做出调整。2.设计实现了负载均衡系统。详细叙述了负载均衡系统的设计与实现的过程,使用epoll事件驱动模型、散列表、连接池等关键设计技术,实现了高性能的负载均衡系统。最后构建测试环境,模拟多用户测试,对负载均衡系统进行了测试,并对本文提出的策略、LARD及加权最小连接算法的性能进行了对比分析。
[Abstract]:With the rapid development of Internet, people's work and life are more and more inseparable from the Internet. As of December 2013, the number of Internet users in China reached 618 million, and the Internet penetration rate was 45.8%. The rapid increase in the number of Internet users brings huge traffic to many popular web sites. Hot sites handle hundreds of millions of requests a day, and the explosive number of visits makes it urgent to build stronger web services. In order to provide a high-quality experience for Internet users and address the demand for network bandwidth and server performance caused by the sharp increase in the number of visitors to the network, Web cluster technology is one of the key technologies to build large web sites, which can improve the processing ability of the system by building WEB server cluster. In this paper, the load balancing scheduling of WEB server cluster is taken as the research topic, and some key technologies of WEB cluster are studied, especially the load balancing scheduling strategy and the key technology of the implementation of load balancing system, and the following results are obtained: 1. A new scheduling strategy based on HTTP request content is designed. When the request arrives, the request is divided into dynamic request and static request according to the HTTP request content, and different scheduling decisions are made for the two types of requests. For dynamic requests, load distribution is mainly considered, and for static requests, the Cache hit ratio of requests is mainly improved. In the algorithm, the dynamic feedback adjustment mechanism based on fuzzy clustering is designed to adjust the load of WEB cluster periodically. A load balancing system is designed and implemented. The design and implementation process of load balancing system is described in detail. A high performance load balancing system is realized by using key design techniques such as epoll event-driven model, hash table, connection pool and so on. Finally, the test environment is constructed, the multi-user test is simulated, the load balancing system is tested, and the performance of the proposed strategy lard and the weighted minimum connection algorithm are compared and analyzed.
【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】：TP393.09

【参考文献】