Design And Implementation Of Load Balancing Service In Massive Request Data

With the development of computer networks and network applications, information technology has entered a period of rapid growth and overall consolidation. Because the network load has a significant growth recent years, many web servers have weighted too heavy. This situlation making the network service response time longer, and making the service quality has a significant reduction. In view of this situlation, we try to design and develop a wserver system based on load balancing algorithm. Using wserver system, we optimize the quality of network service, improve user experience and reduce the cost of network server.This thesis proposed a way to design and development a wserver system based on epoll model and load balancing algorithm, and using this system to provide a reverse proxy service mode and high-performance web service. In general the server design, we use a connection blocking method for one thread. This method has been widely used because of the simple design and development. However, this method will appear many problems when need to face large number of user requests. Reference operating systems asynchronous IO design, we propose the same work scheme using non-blocking operation. Using this way, we can handle all blocking operation with epoll model and handle all non-blocking operation with master process.Firstly, this thesis introduces the research background and the current problems. Secondly, this thesis doses the analysis and modeling of the system requirements and briefly introduces the basic architecture and model design of the system. Thirdly this thesis describes the design of the system in each functional module units in detail, describes the realization of each functional module in detail and also dose the analysis of the realization codes of key processes. Finally, the system was used in the "web of things" project. And this thesis analyzes the deficiencies of current works and plans the future works. At last, summarizes the works and the results of my-self during the master-degree study.