The Community for Technology Leaders
2013 IEEE International Conference on Cluster Computing (CLUSTER) (2013)
Indianapolis, IN, USA
Sept. 23, 2013 to Sept. 27, 2013
ISBN: 978-1-4799-0898-1
pp: 1-8
Ahmet Erdem Sariyuce , Depts. 1Biomedical Informatics, The Ohio State University, USA
Erik Saule , Depts. 1Biomedical Informatics, The Ohio State University, USA
Kamer Kaya , Depts. 1Biomedical Informatics, The Ohio State University, USA
Umit V. Catalyurek , Depts. 1Biomedical Informatics, The Ohio State University, USA
ABSTRACT
Networks are commonly used to model the traffic patterns, social interactions, or web pages. The nodes in a network do not possess the same characteristics: some nodes are naturally more connected and some nodes can be more important. Closeness centrality (CC) is a global metric that quantifies how important is a given node in the network. When the network is dynamic and keeps changing, the relative importance of the nodes also changes. The best known algorithm to compute the CC scores makes it impractical to recompute them from scratch after each modification. In this paper, we propose Streamer, a distributed memory framework for incrementally maintaining the closeness centrality scores of a network upon changes. It leverages pipelined and replicated parallelism and takes NUMA effects into account. It speeds up the maintenance of the CC of a real graph with 916K vertices and 4.3M edges by a factor of 497 using a 64 nodes cluster.
INDEX TERMS
CITATION

A. E. Sariyuce, E. Saule, K. Kaya and U. V. Catalyurek, "STREAMER: A distributed framework for incremental closeness centrality computation," 2013 IEEE International Conference on Cluster Computing (CLUSTER), Indianapolis, IN, USA USA, 2014, pp. 1-8.
doi:10.1109/CLUSTER.2013.6702680
88 ms
(Ver 3.3 (11022016))