Semester of Graduation
First Major Professor
Master of Science (MS)
Relational databases have been popular since a very long time. They store data in a structured way providing optimisation and simplicity. Although the strict structure does not provide flexibility to the developer. It works on having primary key and foreign keys. Joins are created at runtime which eat a lot of memory and time. Graph databases involve nodes and edges. Each node represents an entity and each edge represents a relationship. So when an equivalent JOIN operation is run in graph databases, it saves the time from doing extensive searching. Although graph databases have started gaining a lot of popularity recently, everyone is interested in comparing these popular database management systems. We propose CyDIW (Cyclone Database Implementation Workbench) benchmarking, for measuring performance of MySQL (Relational Database Management System) and Neo4j(Graph Database Management System) in various aspects. CyDIW provides a one-click system to perform an extensive experiment to compare query runtimes. This benchmarking involves a graph model with multiple node types and CRUD operations. We discuss in details the process of creating Neo4j adapter for CyDIW system, which was successfully completed. After extensive study, we see that MySQL is faster than Neo4j for most of the CRUD operations.
Prakash, Ishita, "A platform for benchmarking Database Management Systems: CyDIW" (2019). Creative Components. 234.