Design And Implementation Of Database Access Middleware On Massive Data Management

Databases stored in Operator’s Data Center always have different types, which brings theproblem that the application system can not use the unified interface and have a difficulty to accesstransparently, with the promotion of3G network, mobile phone network produces hundreds TB dataevery month, the huge amount of data makes the queries inefficient. In order to resolve theseproblems, Data Center needs a data access service which can handle massive data and providepublic interface for application system. The application system will use the data access service tointeract with databases in Data Center. Database Access Middleware technology used in data accessservice not only realize query request of massive data, but provide unified interface, shielding thedifferent structure and achieve transparent access to distributed heterogeneous databases.This paper discusses the design model of Data Access Middleware based on EJB3.0technology,realizes the remote data transfer in each layer by encapsulating the result set to custom object; forthe process of massive data, this paper discusses a method based on data partition, reduces the sizeof data and optimizes the query performance by partitioning the data into many tables or views; inorder to solve the problem of uniform access to heterogeneous databases, this paper discusses SQLstatements translation mechanism to shield the syntax between different databases.Finally, this paper discusses the deployment and application of middleware, verifies thefeasibility and efficiency of it, test shows that the database access middleware can has a unifiedaccess in good performance.