Logical Design of Graph Databases from an Entity-Relationship Conceptual Model

NoSQL database models and data store systems have emerged in response to the needs of new big data applications. A common feature of these logical models is the schemaless property that promotes flexibility in data organization. Because of that flexibility, significant decisions are taken regarding database design, often based on best practice recommendations, which can affect the evolution of the database requirements of integrity, interoperability, and scalability. The traditional database design process transforms an abstract conceptual schema into a logical schema based on a target database model. In this paper, we propose a process to transform a conceptual modeling into a representation of NoSQL graph data model. In our approach, entity-relationship model is used for conceptual modeling, and an extended graph database model is proposed for logical modeling. We define a set of mapping rules for conceptual-to-logical scheme conversion based on integrity constraints. Some extensions of the Cypher query language for the Neo4j graph database management system are also proposed.