A NEW APPROACH FOR MIGRATION OF A RELATIONAL DATABASE INTO COLUMN-ORIENTED NOSQL DATABASE ON HADOOP

摘要

This article presents a new approach based on the "Object" concept, to successfully migrate a relational MySQL database to a column oriented HBase NoSQL database. The purpose of this article is to provide a new model of migration process divided into three phases, the first of which allows to obtain a copy of these metadata using the principle of semantic enrichment, and this to extract the different principles of the objects, including aggregation, inheritance and composition, the second phase of the process concerns the automatic generation of a New Optimised Data Model 'NODM' containing all relational database information in a flattened way. The last phase serves for the migration of the existing relational database into column-oriented database in the Hadoop ecosystem. The whole approach proposes a migration solution from a relational database to a NoSQL column-oriented database, which exploits the fast extraction of data columns for several types of applications, thus generating a better factor for analytic query performance, minimizes the input / output load of the disk, and reduces the amount of data being addressed from the disk.