Evolution of Databases: A Comparative Study of Relational Databases and Vector Databases
Authors: Okonta Samson Iwenyan, Abogunrin Oluwatimilehin Abayomi, Adegbite Gbenga Aduagbemi and Odim Mba ObasiThe architecture of data management systems is undergoing a profound transformation, shifting focus from symbolic computation to semantic intelligence to meet the demands of modern artificial intelligence (AI). This paper analyzes this evolution, which culminates in the emergence of Vector Databases (VDBs), systems purpose-built for the efficient storage and retrieval of high-dimensional vectors (embeddings) that encapsulate semantic meaning. VDBs employ specialized Approximate Nearest Neighbour (ANN) search algorithms (such as HNSW and IVF) to enable semantic search, multimodal processing, and Retrieval-Augmented Generation (RAG). This study presents a focused comparative evaluation of the two dominant data paradigms, traditional Relational Database Management Systems (RDBMS) and modern VDBs, examining their performance, scalability, and suitability for AI-driven applications. We establish that while RDBMS remain essential for structured data integrity and transactional workloads, VDBs are fundamentally optimized for high-dimensional similarity search and semantic retrieval efficiency, confirming their role as the indispensable infrastructure for the next generation of intelligent data systems

