Graph-based data management system for efficient information storage, retrieval and processing

Abstract

Data management systems rely on a correct design of data representation and software components. The data representation scheme plays a vital role in how the data are stored, which influences the efficiency of its processing and retrieval. The system components design realizes software engineering concepts to enable performance metrics such as scalability, efficiency, flexibility, maintainability, and extendibility. This paper presents a data management system that uses a graph-based data representation scheme to achieve an efficient data retrieval when using graph-based databases. Input data are transformed into vertices, edges, and labels while inserting them into the database. The proposed system consists of three layers which are: system beans layer, data access layer, and the database engine. Healthcare data are used to evaluate the system in comparison with resource description framework (RDF) semantics. Extensive experiments are conducted to compare different scenarios of data storage and retrieval using Neo4J, OrientDB, and RDF4J. Experimental results show that the performance of the proposed graph-based approach outperforms RDF4J framework in terms of insertion and retrieval time.