PROCESSING SUB-GRAPH QUERIES ON GRAPH DATABASES EFFICIENTLY

Most of the modern applications depend on data modeled by graphs. Processing queries efficiently over graph databases serves these applications and are currently a performance bottleneck Given a graph query q, sub-graph query processing finds all the graphs g in a database of graphs D where q is contained in g. The graph databases are assumed to contain a lot of graphs and it is known that sub-graph isomorphic tests are NP-complete. So, an indexing-based technique should be adopted for an efficient solution. In this paper we propose an efficient technique for processing sub-graph queries. The technique consists of an index called eSuhlndex and a query processing algorithm. Having a query graph, the database is filtered to generate candidate graphs. The proposed technique takes into consideration the full structure of the database graphs besides considering the frequent fragments while filtering. The proposed technique reduces the sub-graph isomorphism tests required for query processing. This is achieved through polynomial time algorithms and hence the total processing time is reduced.