Investigation of graph modeling for data virtualization of SQL and NoSQL Databases

With the rise of data-centric applications, the diversity of data models has posed challenges in combining, interpreting, and retrieving data from disparate sources. As a response, data integration solutions have emerged, streamlining these intricate processes by automating the identification of entities spanning both structured and unstructured data sources. Among these endeavors, data virtualization, a prominent facet of data integration technology, emerges as a promising avenue. However, the successful implementation of data virtualization solutions mandates meticulous consideration of various aspects, most notably the intricate data heterogeneity stemming from syntactic, semantic, and structural disparities inherent in data sources. This paper is centered on tackling challenges related to syntax and structure, with a particular emphasis on investigating how flexible and feasible it is to use graph databases to enhance data virtualization. By examining findings from data integration efforts and identifying gaps in existing research, this study delves into the potential of using graph modeling to bridge these gaps and enhance data virtualization.