BitPath – Label Order Constrained Reachability in Large Graphs

With ever expanding size of the web, sizes of graphs with semantic relationships are growing very rapidly. The semantic relationships are typically presented as edge-labeled graphs where nodes are the entities and the edge label signifies the relationships between the two entities. RDF graphs are an example of such graphs. Pure reachability queries on these graphs do not give any information about the semantic association between two remote entities. On the other hand, the typical size of semantic graphs present on the web has made general regular path query processing infeasible. In this paper we focus on the following graph path query problem – given a source node s, a destination node d and a sequence of ordered edge labels “label seq”, does there exist any path between the two nodes which has the edges labels in “label seq” (in the given order) on the path? We solve this problem by a combination of graph indexing, and a query processing algorithm based on a divide-and-conquer procedure and greedy pruning of the search space. We have evaluated our technique on graphs with more than 22 million edges and 6 million nodes – much larger compared to the datasets used in published work on path queries. We compare our approach with optimized-DFS, optimized-focusedDFS, and bidirectional-BFS methods.