Design On Streaming Server With LVS In SAN-based Environment

VOD is a client-oriented video system, in which users can actively choose any video program. The most attractive feature of VOD is that it's a way through which information-receivers can get self-development and self-perfection according to their own demands. In recent years, with the overall development of multimedia streaming technology, the network products and terminal facilities, streaming media became more and more prevalent and popular. In consideration of this very background, we carry on our research work of SAN and Cluster through the construction and operation of VOD system.In virtue of Tsinghua University's Advanced Network Storage Lab., the dissertation brings forward a LVS method to build a high performance, high scalable VOD system with several PC servers to handle high concurrent multimedia stream. What's more, combining with SAN technology, it can resolve the problem of huge storage as well as shared storage, and thus greatly improve the system-performance. To satisfy the system's demand of high reliability, high scalability and excellent performance, open resource LVS cluster is introduced in to construct VOD system. The system uses shared disc system structure, which helps to take SAN as subsystem, so that to reduce redundant copies and network bandwidth. Moreover, the size of multimedia objects, which are already huge, are getting even bigger as the resolution of output devices improve. As a result, we use hierarchical storage model to off-load the less popular objects to tertiary storage in order to keep costs down. Now, the system works well.Which combines the LVS cluster and SAN is a 3-Layer typical architecture. Consequently, this very system can be used in other applications after a few modifications. The other important part of the dissertation is carrying out a dynamic feedback scheduling algorithm as an extension for LVS. At present, the LVS software has eight scheduling algorithms, but all of them belong to static scheduling algorithm. As we all know, due to the variation of the tasks' attributes and the dynamic environment, the static scheduling algorithms, which depends greatly on the tasks' or servers' static attributes, can't distribute a workload evenly amongst a cluster of back-end servers. In order to improve the request allocation decisions made, we presents the dynamic feedback scheduling algorithm for flexible workload. Dynamic feedback scheduling requires mechanisms for monitoring and evaluating the current load on each server, gathering the results, combining them, and taking real-time decisions. Because evaluating the load of server has a characteristic of uncertainty, so we choose the theory of fuzzy optimal selection for multistage and multicriteria decision making system to solve this problem.In order to test the effectiveness of the dynamic feedback system, we use the VOD server mentioned above as test bench and bench marked in various configurations, with and without dynamic feedback in place. To simulate an unbalanced server load, one of the servers was given an additional network-intensive task. The results of the experimentation show that the DF scheduling does optimise the request allocation; network load of the servers is brought back to an even distribution with the dynamic feedback process running.Especially, the system we designed is fault tolerant. Since load balancer gathers the information of back-end servers regularly, if one node failed, all the new tasks can be handled by remaining intact nodes. So that, as far as the total performance of the system is concerned, single server failure can hardly lead to termination of the whole system, but only a little decrease of its performance.