k2-Trees for Compact Web Graph Representation

The directed graph representation of the World Wide Web has been extensively used to analyze the Web structure, behavior and evolution. However, those graphs are huge and do not fit in main memory, whereas the required graph algorithms are inefficient in secondary memory. Compressed graph representations reduce their space while allowing efficient navigation in compressed form. As such, they allow running main-memory graph algorithms on much larger Web subgraphs. In this paper we present a Web graph representation based on a very compact tree structure that takes advantage of large empty areas of the adjacency matrix of the graph. Our results show that our method is competitive with the best alternatives in the literature, offering an interesting space/time tradeoff. It gives a very good compression ratio (3.3–5.3 bits per link) while permitting fast navigation on the graph to obtain direct as well as reverse neighbors (2– 15 microseconds per neighbor delivered). Moreover, we ...

• Type: Conference
• Year: 2009
• Where: SPIRE
• Authors: Nieves R. Brisaboa, Susana Ladra, Gonzalo Navarro
• Keywords: Directed Graph Representation, Graph Algorithms, Graph Representations, Information Retrieval, SPIRE 2009