Migrating Relational Database To MongoDB

In recent years, the increasing amount of data challenges the limit of relational databases. As a contrast, NoSQL databases get more and more preference because of their scalability, high performance and high availability. Therefore, there comes the requirements of migration from relational databases to NoSQL databases.To meet the requirements mentioned above, we present a framework that can migrate from relational databases to NoSQL databases, which assists users to automatic/semi-automatic migrate applications in three key phases:schema conversion, data migration and code migration.1. In the schema conversion phase, we present an automatic/semi-automatic model conversion technique based on foreign keys of relational databases and relations customed by users.2. In the data migration phase, we present an automatic data migration technique based on the topological order, and save relevant data together based on embedded documents or document references, and then insert indecies automatically.3.In code migration phase, we propose an automatic code generation machnism to generate entity classes. We also present an automatic technique to generate some DAO methods, and also a technique that can generate MongoDB methods from SQL statements.Based on these proposed techniques, we design two experiments to compare our framework with the classic ETL tool Spoon in all three phases. The results show that our framework has a better performance than Spoon, particularly in migrating complex data. All above have shown the effectiveness of our framework.