Research And Implementation Of ERP Caching Mechanism

The ERP system can mix enterprise resources properly, save enterprise management cost to the maximum and increase their profits. With the promotion and development of the ERP system in China, the business module is divided more and more specific and particular. Meanwhile, the cost of sub-modules in ERP server grows larger and larger and optimization design task of system becomes more and more difficult. According to the research, the lower efficiency of system is caused by the frequent connection with database server by the ERP system. Therefore, proper caching algorithm is the basic solution of this problem. So the research of ERP cache is the important way to change the low efficiency of ERP.The paper mainly discusses the cache replacement strategy, cache consistency policies, semantic caching. According to different ERP businesses characteristic, this paper presents an Index Caching Replacement Strategy. The Index Caching Replacement Strategy has designed three kinds of index tables, Least Frequently Used Index, First in First Out Index and Timestamp Index. The Index Caching Replacement Strategy connects the different index tables with different businesses characteristics in order to replace the caching based on the Index Caching table when the system needs to replace the caching. It query caching using semantic analyze so that the utilization of caching item can more high. The paper designs VersionSensitiveMap class and ElementVersionSensitiveMap class to control the caching replacement grain. When the data changed in the database, the caching system promise that the data of cache is right and promises minimum cost of the replacement price.The experiment uses IBM Rational Functional Tester to design the automatic test script. The test script imitates the company daily operation to test the node of Committee Gathering Account on 50 times of operations, and test the hit rate and average response time of LRU, FIFO and Index Replacement Strategy. The result reveals a positive data by using Semantic Index Replacement Strategy.