Granular Security in a Graph Database

Abstract : With the growing use of data in all aspects of government and commerce, the need for that data to be both accessible and secure is also growing. One solution to this dual need is provided by Accumulo, a database that allows multiple users of various security levels to access one platform but receive authorization to view only portions of the database. Various databases, however, organize information differently. This thesis examines the possibility of implementing a granular security on a graph database. Using Neo4j as a reference implementation, graph theory concepts are used to find a method of allowing data access while retaining security in a data environment that emphasizes connectivity. Using adjacency matrix multiplication on bipartite graph slices of the network of security layers, a mathematical justification exists for locating two step connections that exit from and return to a security layer. These connections can be revealed to a user without granting access outside of the assigned security layer.