Research On Application Of Relational And Non-Relational Database

The database technology has made great progress from the emergence to today, and the relational database has always been the mainstream products. But in recent years, the data of Internet shows a variability and explosive growth trend with the rapid development of Web2.0. Faced with huge amounts of data, the management and maintenance of the traditional relational database fall across unprecedented challenges. Therefore, NoSQL database management technology has attracted the attention of researchers.The relational database supports complex query work, has a strong mathematical theory basis and the data model of it is simple, but it does not support complex nested data and complex data types. The data model of NoSQL is plentiful and easy to expand, but it is short of support of unified technology and the application of it is not mature enough. It is hard to say which type of database is better, so providing a scientific selection scheme is imperative. The subject is put forward in this background.Firstly, this thesis makes comparisons of the features and application scenarios between the relational database and NoSQL, and analyses their advantages and disadvantages, which provides theoretical basis for the selection of database.Secondly, this thesis designs and brings out the flower basket system based on MySQL and MongoDB. This thesis puts forward a design scheme of MongoDB, which provides a technical support for the developers.Then, this thesis designs details experiment to compare the operating performance of inserting, deleting, updating, selecting of the two kinds of relational database and NoSQL based on the designed model of the flower basket system. This thesis analyses the experimental results from the consumption of time, disk, CPU, I/O and so on.Finally, this thesis puts forward a selection model for database based on the comprehensive evaluation method, which proposes a reasonable database selection scheme for the database administrators. The performance of increasing, deleting, modifying, querying are the primary indicator of this model, and the consumption of time, disk, CPU, I/O are the second indicator. This model can be further improved in the future research work.