IMPLEMENTING JOINS OVER HBASE ON CLOUD PLATFORM

Amount of data and number of database accesses has increased enormously. Traditional databases are unable to fulfil these requirements. Along with the increased amount of data and number of accesses, data is becoming more unstructured. Relational database could not only serve these purposes efficiently but also they add a limitation on the size of data; upcoming cloud databases overcome these limitations. Cloud database can handle very huge amount of data and large number of database accesses. Also cloud databases support semi-structured and unstructured data along with the structured data. HBase is a cloud database which is an open source, non-relational, distributed database. HBase does not support SQL queries. HBase provides its own APIs to access data. Moreover HBase does not provide support for Joins and nested queries. The Joins were dropped from NoSQL databases because they add processing overhead which in case of huge amount of data becomes significantly large. But join is the way to combine results from one or more tables with very less code. In this paper we propose a layer over HBase which will support Joins over HBase. This layer will work and interact between user and HBase and will make use of HBase APIs for accessing the data, which is stored in the underlined HDFS. Developers will be able to use this layer as API in their program by just including the layer libraries in the program.