Performance Optimization And Application Of Storage Area Network Under The Online Transaction Processing

With the leap of the amount of the information and the sharp increased transactions, the storage of the related data has become the center of the industries, such as, the fields of bank, finance, government and education. It would cause immeasurable loss once these data was damaged.Storage area network, called SAN for short, is one of the three forms of network storage, and also a very popular mode of storage recently. SAN connects the storage devices and application servers through the network, and this network is specially used for access between servers and storage devices. The entire system environments become more complex because of the network factors. Especially now, there are more and more online transactions, so it requires a smaller response time to meet clients'demands. In order to support different kinds of businesses, the optimization of the performance is needed.This paper had done the research on the storage area network and online transaction processing, and then analyzed the key factors that affects their performance and put forward the methods of configuration optimization and solving the performance bottleneck. In this way, when a SAN was built, there would be a clear guidance so that the hardware environment can reach a reasonable and stable state. At the same time, it analyzed how to quickly and accurately find out the performance bottleneck when the performance problems occur. This will provide important help to solve practical problems.Since the importance and high availability of the stored data, SAN uses the redundant method, that is, the accessed data can arrive at the storage devices through different paths and network devices. Because of the different size of the application load, unbalanced condition occurs among the controllers. Therefore, this paper applied a strategy that logical disks do not belong to any controller, that is all controllers can access the data. Based on the contrast test of the traditional strategy and optimization strategy with the SimPy simulation package, the results showed, the optimization strategy can keep the load balanced among the controllers, improve the balance application of the controller, prevent the problem that some controller is continuously utilized but some is vacant.Finally, in a SAN system, using the tuning method given in the paper, can quickly indentified the performance bottleneck and gave a correct resolution, so that the work efficiency was improved.