TBSGM: A fast subgraph matching method on large scale graphs

Subgraph matching, which belongs to NP-hard, faces significant challenges on a large scale graph with billions of nodes, and existing methods are usually confronted with greater challenges from both stability and efficiency. In this article, a subgraph matching method in a distributed system, tree model-based subgraph matching method (TBSGM) is proposed. The authors provide a transformed efficient query tree as a replacement for a query graph. In order to get the tree, they present a cost evaluation model which may help to generate the efficient query tree according to network communication-cost and calculation-cost evaluation. Also, a key set based indexing strategy for intermediate results is given to simplify the matching results during network communication. Extensive experiments with real-world datasets show that TBSGM significantly outperforms other methods in the aspects of scalability and efficiency.