The Designand Implementation Of Distributed Database Middleware Server Layer

With the continuous development of the network, various new technique andnew business spring up unceasingly, the number of users is growing far beyondpeople’s imagination. And large sites and applications require uninterrupted service7*24hours, the database processing huge amounts of data and the number ofconcurrent faces enormous challenges. It will cause very serious consequences ifnot handled properly. In such a situation, and constantly improve the performanceof a single database to spend more and more, so that the burden on operators toincrease. In this context, database clustering technology is required for people.This paper describes the design and implementation of a distributed databaseserver layer middleware. First, determine the technical solutions based onapplication scenarios and requirements definition system, using the horizontalsegmentation to improve the processing speed for massive data clusters, using aconnection pool of connections to achieve the integration of connectionmultiplexing, while supporting distributed transactions to ensure that transactionsatomicity, isolation, durability, consistency, using data space, data sources, dataserver, three-tier structure to ensure that when the back-end server problems can bepromptly switched to the backup machine and continues to provide highly availableservices. Support for load balancing with separate read and write differentstatements to further improve the overall system performance.According to requirements definition, the system is divided as an optioncontroller module, connection pool management module, backend managementmodule, the user request processing module and the implementation plan buildingblock. Each parameter options to ensure the system can be configured throughoption controller module. Users can configure the file system functions for cutting,to achieve functional application needs. Connection pool management module tomanage the back-end connections, the connections will be dynamically created andcleaning according to the current cluster load, ensure the rational use of resources.Backend management module to initialize the entire cluster to provide support fordata space, data sources, data server. User request processing module is used foruser authentication, data to optimize the user’s request, including literacy separationand load balancing. Under the premise of guaranteed features, providehigh-performance services. The implementation plan building block builds a user requests the execution plan and executes, exchange data with the back-end server,and return data to the user.Finally, perform module testing and system testing. Perform functional testingfor each module. Perform performance testing for the entire system. Test resultsshow that the modules meet the requirements, system responsiveness and operatingresults meet design goals.