SGVCut: A Vertex-Cut Partitioning Tool for RandomWalks-based Computations over Social Network graphs
Abstract
Several distributed frameworks have recently emerged to perform computations on large-scale graphs. However somerecent studies have highlighted that vertex-partitioning approaches, e.g. Giraph, failed to achieve workload-balancedpartitioning for skewed graphs, typically having a heavy-tail degree distribution. While edge-partitioning approachessuch as PowerGraph and GraphX provide better balancing and performances for graph computation, they supply ageneric framework, independent from the computation. Thisdemonstration presents SGVCut to display our edge partitions designed for random walks-based computation, whichis the foundation of many graph algorithms, on skewed graphs.The demonstration scenario introduces SGVCut interface andillustrates the benefits of our approach compare to other partitioning strategies for different settings and algorithms.