Using Dense Subgraphs to Optimize Ego-centric Aggregate Queries in Graph Databases

In this paper, we present an approach to optimize ego-centric aggregate queries, in graph databases, by precomputing (materializing) some of their results. Ego-centric aggregate queries allow to graph nodes, called consumers, to aggregate events from others nodes, called producers. Our contribution consists of discovering the densest subgraph, that represents the tightly coupled nodes. We consider that these nodes are the most actives i.e. they have two main features: the high access frequency and the strong correlation between them. Then, we precompute the results of the ego-centric aggregate queries that are implemented on the active nodes. Our experimentation has shown that, when the graph database size is less voluminous, our approach is able to improve the response time of the ego-centric aggregate queries. However, for large graph databases, we should adjust the selection way of active nodes to improve the management load of ego-centric aggregate queries.