Towards Scalable Graph Analytics on Time Dependent Graphs
Authors: Suraj Poudel (University of Alabama at Birmingham), Roger Pearce (Lawrence Livermore National Laboratory), Maya Gokhale (Lawrence Livermore National Laboratory)
Abstract: The objective of this study is to annotate temporal metadata into the partitioned graph topology of a scale-free graph. Our time dependent graph has been modeled using multiple metadata per edge, where each edge metadata field encodes temporal information. We performed experiments with large scale CAIDA datasets, anonymized internet traces, on a Linux HPC cluster (Catalyst) by deriving flow between two IP’s and annotating that flow information as metadata into the graph topology using the asynchronous visitor computation model of HavoqGT.
With the dataset, graph-representation, and HavoqGT, we implemented and evaluated time dependent Single Source Shortest Path (TD-SSSP) and Temporal Betweenness Centrality.
Two-page extended abstract: pdf