Toward lightweight transparent data middleware in support of document stores

With the advent of rapid increase in the size of data in legacy applications, it is urgent to build efficient and flexible data middleware that supports SQL to MapReduce transformation. In this paper, we propose a data middleware to translate the SQLs to the operations that the NoSQL can understand using the MapReduce framework we design. This middleware transforms the read and write SQL statements into MapReduce jobs. In addition, a set of transformation rules is discussed in detail. This middleware can significantly reduce redundant computations and improve the performance of I/O operations.