Data Modelling and Database Migration for NOSQL Database

Abstract

With the exponential growth of data in the digital age, traditional relational databases (RDB) are no longer sufficient for managing big data applications. To overcome the limitations of RDB, Not Only SQL or NoSQL databases have emerged as a promising solution. NoSQL databases offer flexible schemas, horizontal scalability, and support for diverse data models, making them well-suited for the dynamic and heterogeneous nature of modern data. Decades of research have focused on developing tools and techniques for relational databases, enabling non-expert users to utilize them effectively. However, the emergence of the NoSQL database has introduced new research challenges. This thesis addresses two critical challenges in the NoSQL database: a) data modelling in the NoSQL database because it often follows a trial-and-error or ad-hoc rule-based approach, lacking a formal process like normalization in relational databases. Unfortunately, these approaches can result in suboptimal database designs, leading to increased costs in query processing, data storage, and the need for redesigning, and b) managing database migration from relational to NoSQL database because it is a complicated process as it involves the architecture changes of stored data which involve challenges like the difference in data storage approach between RDB and NoSQL, with RDB favouring normalized data and NoSQL preferring denormalized data. These challenges are critical within the context of the NoSQL database due to their inherent complexity and highlight the need for automated methodologies to streamline the data modelling and migration process in the NoSQL database. The challenges outlined above lead to identifying specific research problems (RPs). These research problems are presented below, accompanied by four innovative solutions for each: RP1: To study the influence of schema alternatives for data modelling in NoSQL databases and understand their impact on query performance, data redundancy, and overall database efficiency.