Model based development of data integration in graph databases using triple graph grammars

Abstract

Graph databases such as neo4j are designed to handle and integrate big data from heterogeneous sources. For flexibility and performance they do not ensure data quality through schemata but leave it to the application level. In this paper, we present a model-driven approach for data integration through graph databases with data sources in relational databases. We model query and update operations in neo4j by triple graph grammars and map these to Gremlin code for execution. In this way we provide a model-based approach to data integration that is both visual and formal while providing the data quality assurances of a schema-based solution.