A Model for Relational to NoSQL database Migration: Snapshot-Live Stream Db Migration Model

Data migration is simply a process of transforming data from a source database to target database. Companies are moving their database from one database (e.g., RDBMS) to another new one (e.g., NoSQL) because of many reason such as new database can handle more data, that could be fast, scalable etc. There are various tools and methods such as, mongify [1], mongoimport [2], sqoop [3] etc. These are available for database migration from RDBMS to NoSQL databases. NoSQL uses document model, graph model, key-value etc., as compared to RDBMS databases. NoSQL was primarily designed for storing and retrieving a large amount of data. In this paper, we have proposed a database migration model, which can migrate real-time as well as old data in a parallel way. The proposed model is implemented for MySQL (RDBMS) to MongoDB document database in Java. The prototype migrates database instance at a particular time (Snapshot of database) as well as live data in a parallel way. Finally, our experimental evaluation of snapshot and live data migration shows the higher performance as compared to other methods.