A Framework for Enabling Privacy Preserving Analysis of Graph Properties in Distributed Graphs

In the real world, many phenomena can be naturally modeled as a graph whose nodes represent entities and whose edges represent interactions or relationships between the entities. Past and ongoing research on graphs has developed concepts and theories that may deepen the understanding of the graph data and facilitate solving many problems of practical interest represented by graphs. However, little of this work takes privacy concerns into account. This paper contributes to privacy preserving graph analysis research by proposing a framework for enabling privacy preserving analysis of graph properties in distributed graphs. The framework is composed of three modules. We discuss the functionality of each module and describe how the modules together ensure the privacy protection while retaining graph properties and answer users’ queries pertaining to graph properties.