Design And Research Study On Data Model Of Document Database

After the relation model was proposed in 1970, the relational database began to appear. Based on the theory of relational algebra and the improvements of relational database, the relational database has the strong ability to store and manage the structured data. However, with the continuous development of Internet technology and the coming of the era of Big Data, the relational database has its disadvantages on processing different types of mass data, including poor scalability, flexibility etc. NoSQL databases, including MongoDB, HBase, Redis, because of their good expansibility, low cost, high flexibility, are gradually applied in each big enterprise and cloud storage system.After years of development, the relational database has formed a comparatively perfect theoretical basis to design the data model. By contrast, the NoSQL database has not yet formed a set of storage model design criteria. Although a characteristic of NoSQL is that it has a flexible data model, this doesn’t mean that it does not need to design the data model, using different data models of different scenarios, may have great influence on the performance of the storage system. Because of the difference between relational database and NoSQL in terms of data organization and storage model, it is not efficiently to design the data model of NoSQL with guidelines of relational database.This paper will take Document Database as the research object to discuss this issue. Fisrt of all we analyse the complexity of create, retrieve, update and delete operations in Document Database and using formula to quantity each operation. After the quantitative we build a model to present different relationships, then develop strategies for each relationship in designing data model. Furthermore, this paper also proposes an algorithm for designing data models called Model_Design. Using the storage strategies in the previous work, this algorithm is able to design a global data model for all relationships in the database, so as to improve the performance. In the end, we construct experiments to verify the study of this paper. The experiments results indicate that the design strategies and algorithm can improve the overall performance of storage system.