Design And Implementation Of Large-scale Distributed Virtual Reality Platform For Server Load Balancing

The big scale and distributed virtual reality platform is the key technology in the virtual reality field. It has a great application in web games, city planning, national defense and military affairs, engineering design and so on. The server structure of the platform is distributed server cluster. The whole system has a coordinator server and several data servers, so how to keep the servers going perfectly to meet the need of the big scale clients and at the same time assign the load of the whole system to every data server equally are the questions that the technology of load balance should answer. Yuanqing Wu has read a lot of papers and literatures and chose the load balance technology of big scale and distributed virtual reality platform as the study task of his master paper according to the goal of the lab's 863 project. The task has an important theory value and good prospect. The main jobs of the paper are below:1. Provide a load balance algorithm that a data server stores several unit data. In the algorithm, the first allocation should meet the CCR principle. For the more, it gives the measurement of the load of data server and how to bring in and out the data server. The algorithm has the advantages: the unit data are stored compactly and the measurement of load is easy. But it asks the good net bandwidth and the equal division of graph scene, which is also its localization.2. Provide a load balance algorithm that a data server stores a unit data. The load is decided by the using rates of CPU, disk, memory and bandwidth. It also uses the ARMA model to forecast the load of the data server. The algorithm could meet the requirement of load balance of the big scale and distributed data.The paper makes use of a simulation program to validate the above algorithms. The result shows that the algorithms are effective.