Information systems have the requirement of a large amount of data’s collectionand analysis, which need the database system having the ability of large capacitystorage, real-time processing and scalability. The traditional relational databasesexpose a lot of problems when it process a large amount of data, which can notprocess data efficiently. Therefore, the traditional relational database can not solve thenew demands of the information system.Non-relational database compensate the shortage of a relational database.Existing Key-Value main memory database system have good ability of processing asmall amount of data, but not good at processing a large amount of data, which cannot meet the new demand of the current system.Therefore, we analyze user requirement, the shortage of current Key-Value mainmemory database and the technology of distributed system, implement a kind ofdistributed main memory database. Firstly, we design and describe the systemarchitecture and functions. Secondly, we focus on the detailed design andimplementation of the distribution array module, the message transfer module, mainmemory resource management modules and business processing module. In order tosolve the problems of data distribution of distributed systems, we put forward adistribution array method with hash algorithm to solve the problems. Thirdly, test thefunctionality and performance of the system, and give the test results. Finally,summarize the shortage of this paper and the improvement aspect. |