An Improvement On Load-Balancing Cluster For Virtual And Realitv Combined Internet-based Lab

This thesis is an extensional project of a national science and technology support program titled "Research and Application of the Key Technologies of Virtual Experimental Teaching Environment" (No.2008BAH29B00). Currently physical experimental system and virtual simulation experimental system has been combined as a whole system called Virtual and Reality Combination Experimental Platform(VRCP), which provides 5 classes, more than 80 physical experiments and virtual simulation experiments. In order to make the performance of Electrical Engineering Virtual Lab meet growing customer demand, this paper makes an improvement on the architecture of load-balancing cluster on Linux Virtual Server, and proposes a genetic optimized BP neural network algorithm to improve the performance and stability of the cluster.In the arrangement of the content, this thesis firstly discusses the background and significance of the research, and describes the existing status of load balance technology. Then in chapter 2, this thesis focuses on the overall structure of the VRCP system. The third chapter briefly introduces the principle of LVS and implementation of the original system. Then chapter 4, after analyzing the shortcomings of the original system, designs a new architecture and an improved load-balancing algorithm. At the end, the fifth chapter explores the effects of improved algorithm and architecture by three groups of tests respectively, and the results indicate the superiority of the new structure and algorithm.