Research And Application On Buffered Asynchronous Processing Model In Large Concurrence Environment

Currently,according to the development of Internet technology and popularization of its application,it is the information times. Information systems come into people's lives and help people to complete a variety of day-to-day business.With more and more people joined to the information age,the workload which information systems undertaken becomes heavier and heavier. With the rapid increasement of user number, challenge the performance of information systems in the large concurrent processing environment.How to solve this problem becomes the very important of various Applications.In different applications,use different methods or technologies to solve the performance bottleneck.Many of which uesed the strategy of buffering mechanism that originated from the research of operation system. The implement of this mechanism in the early time is designed to ease and improve the dismatch between CPU and I/O device.With the development and maturity of various buffering mechanisms.Such solutions like web cache, streaming media cache and database cache solved the performance bottleneck problems in the specific area.In the large concurrent processing environment often accompanied by a large number of data-processing requests. In traditional processing model, the data-processing requests are directly submitted to database. Database system separately to deal with every transaction based on the FIFO principle. The heavy data-processing tasks consumed most of time and system resources, which led to a rapid decline of system performance in the large concurrent processing environment.To this situation, the author of this paper proposed and implemented the Buffered Asynchronous Processing Model(BAP model) based on the data-buffering mechanism. It uses event-driven mechanism for data communication inside the model, considered both the strengths and weaknesses of memory cache and disk cache, and finally submit all the original requests in a batch way of which greatly proved the efficiency. The whole system runs through thread way, Asynchronous completed the original data-processing requests. Users only need to call the entry function in the specific code point to start the BAP model which implements buffering function, it maximize the reduction and the impact of changes of the original system. At last of this paper, applied this model to a real project, through a large number of functional testing and performance testing verified the reliability and advantage of the BAP model, and gived a quantitative analysis to the performance of the system.