Cluster Streaming Caching Proxy Server Research Of Cache Management System

With the explosive development of Internet, it becomes more and more easy for people to gather much more information than ever. However, with the rapid increasing of the user, currently Internet is suffering from problems such as server overload, increase of client perceived latency and backbone network congestion. Especially with the growing popularity of applications such as video-on-demand, remote education and electronic business,such problems become even worse. The traditional methods used are upgrading network servers and improving user-end network accessing speed, howerver, these can not radically resolve the problems: usually, the upgrading speed of servers can not catch up with the increasement of user numbers, further more, just upgrade the user-end accessing speed could not reduce the backbone network congestion, but to upgrade the whole backbone network is too costly to implement.While such problems can be effectively settled through adopting caching proxy technique. The purpose to use caching proxy is decreasing the frequency of accessing remote servers via keeping copys of data which is frequently accessed in the storage devices nearby users. So, it directly increases the responding speed of client-end, meanwhile, reduces the remote servers' loads and the backbone network congestion indirectly; on the other hand, thanks to the data copys reserved in caching proxy, users will not be disturbed by the crash of remote servers, which drastically enhances the quality of service(QoS).The architecture of the cluster-based streaming caching proxy server is proposed in the paper. Content aware scheduling mechanism is adopted to improve the scheduling performance of the cluster. The segment based prefix/suffix caching mechanism fit into streaming media object is put forward, which improves the efficiency of cache management effectively. Moreover, a global replacement algorithm adapting to cluster architecture is presented, which takes not only the long-time popularity but also short-time popularity into account when calculating the object's popularity. Based on the analysis of traditional performance evaluating index, a new index named segment hit rate is raised and the pertinent mathematic definition is given.The system is implemented by C and C++ language and tested through simulation method. The tested results confirm that the system achieves good robust and high scalability, moreover, the streaming media caching management system is well-designed, and the new replacement algorithms surpass the present ones on performace.