SQL-to-MapReduce Translation for Efficient OLAP Query Processing with MapReduce

Abstract

Substantial research has addressed that frequent I/O required for scalability and faulttolerance sacrifices efficiency of MapReduce. Regarding this, our previous work discussed a method to reduce I/O cost when processing OLAP queries with MapReduce. The method can be implemented simply by providing an SQL-to-MapReduce translator on top of the MapReduce framework and needs not modify the underlying framework. In this paper, we present techniques to translate SQL queries into corresponding MapReduce programs which support the method discussed in our previous work for I/O cost reduction.