Automatic Migration of Data to NoSQL Databases Using Service Oriented Architecture

Abstract

For the past few years there has been an exponential rise in the use of databases which are not true relational databases. There is no correct definition of such databases but can only be described with a set of common characteristics such absence of a fixed schema, inherent scalability features, high performance, data etc. These databases have come to be known as NoSQL databases. Various companies are seeing the advantages of NoSQL and want to migrate to these databases. But they find it difficult to migrate their data as a lot of study and analysis is required. Each type of database have their own terminology and query language. We propose a novel automated migration model which utilizes the power of service oriented architecture to help these companies easily migrate to NoSQL databases of their choice. We utilize web services which encapsulates few of the most popular NoSQL databases such as MongoDB, Neo4j, Cassandra etc. so that inner details of these databases are hidden yet providing efficient migration of data with little or no knowledge of the inner working of these databases. As proof of concept relational data was migrated successfully from Apache Derby database to MongoDB, Cassandra, Neo4j and DynamoDB, each vendor representing a different type of NoSQL database.