RésuméThis thesis draws a panorama of main technologies related to databases, from relational to non-relational, including NoSQL technologies but also NewSQL. They are compared and a critical eye is brought on advantages and disadvantages of each of them. Moreover, this thesis describes two approaches that complete each other to apprehend database migration: the transformational and co-transformational approaches. After providing that global view about databases and database migration, this thesis covers two distinct aspects.
In a first time, it aims at bringing a method to perform a complete migration from a relational database towards a non-relational database with a particular focus on Datomic as target technology. To do so, the method addresses three crucial elements.
The first element addressed is the structure. By combining two different processes that are the database reverse engineering and the database forward engineering, the migration method has for a goal to match the relational model and Datomic’s data model in order to ensure that there is no loss in the information capacity. The second element addressed is the data. The method describes the complete data conversion process that consists of inserting the existing data within the target database while transforming it according to the structure changes and preventing information loss. The last element addressed is the application program(s). The method aims at providing a systematic approach that allows the existing application program(s) to communicate with the target database while balancing aspects such as the possibilities offered by the new technology, the security and the necessary effort to convert the program.
In a second time, this thesis focuses on a sample of NoSQL technologies which support Datomic. An evaluation of the performance of these databases has been made in order to choose the most appropriate database for the OSCAR system, in a future work. The benchmark “Yahoo! Cloud Serving Benchmark” has been used to evaluation the sample of NoSQL technologies.
|la date de réponse||1 sept. 2015|
|Superviseur||Anthony Cleve (Promoteur)|