Research And Practice Of Performance Optimization Of Large-scale Network Application Systems Based On Java EE

With the rapid development of Internet and E-commerce's rise, more and more people begin to use the network platform to implement their needs. Therefore, the average response time becomes a key factor. Traditionally, using database connection pool optimizes network system that presents low efficiency. In order to access database less and less, this thesis introduces the secondary cache, full-text search and other optimization techniques that can improve performance of the original project that is insufficient in present network system.In this thesis, Java Web application hierarchical management ideas is used to solve the existing development problems in project. The traditional single-layer structure results in team inconsistent development, redundant procedure code that is severely limited by system scalability. At the same time, there are too many bloated software problems, and this thesis uses annotation to reduce unnecessary use of configuration files, so as to achieve the purpose of optimizing the code. Generally speaking, application systems have a large number of users who frequently access to relatively stable, infrequently changed data so that results in too much time that seriously affects the system's performance and the user's experience. In this thesis, using a combination of Hibernate ORM and Lucene full-text indexes important data and then stores data in the local database files or memories. When user quering, data can be obtained directly through the index file to improve response time. For other users who frequently visit data that reading-writing ratio is greater than 2:1 use Hibernate to achieve second-level cache EHCache. In addition, in order to achieve the purpose of improve the response time of network applications, this thesis also attempts to integrate Lucene and EHCache through the Spring and then applies to system.Finally, using JMeter performance testing tool tests design of the system, and the system design does not use the above optimization methods comparing with optimized system. At last the result proves that system that has techniques is effectiveness and credibility.