MODIFICATION OF QUERY-ORIENTED DATA MODELING (QODM) APPROACH TO MODEL DOCUMENT-ORIENTED NOSQL DATABASE

• Main Author: Sean Putra, Rafael
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/78306
• Institution: Institut Teknologi Bandung

Li et al. (2014) proposed the Query-Oriented Data Modeling (QODM) approach for modeling document-oriented NoSQL databases. This approach uses UML class diagram and queries that are used to the database system as a basis to create logical model of databases. But, the QODM approach has flaws that makes the created logical model to not be able to store all the data that the used conceptual model has. This project aims to modify the QODM approach so that it can solve the problem that has been identified. The first problem is regarding two different queries that has the same representative form. This problem is solved by adding relationships to the representative form. The second problem is regarding the logical model created that is unable to store all the information regarding the class and relationship that the used UML class diagram has. This problem is solved by adding a checking stage to make sure that every class and relationship in the used conceptual model is completely stored in the logical model created. The third problem is regarding the process on how to create the logical model that is not provided. This problem is solved by creating steps to create the logical model using embedding concept that is used in the database modeling method using Peter Chen Framework (Shin et al., 2017). The modified QODM approach is then tested using two test cases. On the first test case, the modified QODM approach is compared to the original QODM approach to check the completeness of the data stored. Based on this test, it can be concluded that the modified approach can completely store all the data. On the second test case, the database model created using modified QODM approach is compared to the database model created using the Peter Chen framework (Shin et al., 2017) to assess the performance of the database model. From the test, it can be concluded that the modified QODM approach creates database model that is better in handling read operation, especially the operation that used the queries that are used to model the database. But the created database model is worse in handling write and update operation.