Performance Modeling of Computation and Communication Tradeoffs in Vertex-Centric Graph Processing Clusters

Amirreza Abdolrashidi; Lakshmish Ramaswamy; David Narron

10th IEEE International Conference on Collaborative Computing: Networking, Applications and Worksharing

Research Article

Performance Modeling of Computation and Communication Tradeoffs in Vertex-Centric Graph Processing Clusters

Download544 downloads

Cite: BibTeX Plain Text

@INPROCEEDINGS{10.4108/icst.collaboratecom.2014.257474,
    author={Amirreza Abdolrashidi and Lakshmish Ramaswamy and David Narron},
    title={Performance Modeling of Computation and Communication Tradeoffs in Vertex-Centric Graph Processing Clusters},
    proceedings={10th IEEE International Conference on Collaborative Computing: Networking, Applications and Worksharing},
    publisher={IEEE},
    proceedings_a={COLLABORATECOM},
    year={2014},
    month={11},
    keywords={distributed vertex-centric graph processing performance modeling parallel processing graph partitioning},
    doi={10.4108/icst.collaboratecom.2014.257474}
}

Amirreza Abdolrashidi
Lakshmish Ramaswamy
David Narron
Year: 2014
Performance Modeling of Computation and Communication Tradeoffs in Vertex-Centric Graph Processing Clusters
COLLABORATECOM
IEEE
DOI: 10.4108/icst.collaboratecom.2014.257474

Amirreza Abdolrashidi¹^,*, Lakshmish Ramaswamy¹, David Narron¹

1: University of Georgia

*Contact email: ara@cs.uga.edu

Abstract

Distributed vertex-centric graph processing systems have been recently proposed to perform different types of analytics on large graphs. These systems utilize the parallelism of shared nothing clusters. In this work we propose a novel model for the performance cost of such clusters. We also define novel metrics related to the workload balance and network communication cost of clusters processing massive real graph datasets. We empirically investigate the effects of different graph partitioning mechanisms and their tradeoff for two different categories of graph processing algorithms.

Keywords: distributed vertex-centric graph processing performance modeling parallel processing graph partitioning

Published: 2014-11-11
Publisher: IEEE

: http://dx.doi.org/10.4108/icst.collaboratecom.2014.257474