AMANDA: A Middleware for Automatic Migration between Different Database Paradigms

Applied Sciences

Flavia S. Santos¹  Tafarel Brayan F. Souza¹  Pedro Vitor V. P. Cleis¹  Fabrício L. Soares¹  Thiago A. Falcão ¹  Felipe T. Giuntini¹  Phillip M. Furtado¹  Jordan S. Queiroz¹ 

[1] Sidia R&D Institute, Manaus 69055-035, Brazil;

关键词: big data;    database paradigms;    automatic migration;    database performance;    middleware;    non-relational database;   

DOI  :  10.3390/app12126106

来源: DOAJ

【 摘 要 】

In a world rich in interconnected and complex data, the non-relational database paradigm can better handle large volumes of data at high speed with a scale-out architecture, which are two essential requirements for large industries and world-class applications. This article presents AMANDA, a flexible middleware for automatic migration between relational and non-relational databases based on a user-defined schema that offers support for multiple sources and target databases. We evaluate the performance of AMANDA by assessing the migration speed, query execution, query performance, and migration correctness, from two Relational Database Management Systems (RBMSs), i.e., Postgres and MySQL, to a non-relational database (NoSQL), i.e., DGpraph. The results show that AMANDA successfully migrates data 26 times faster than previous approaches, when considering Northwind. Regarding the IMDB database, it took 7 days to migrate 5.5 GB of data.

【 授权许可】

Unknown