Distributed Big Graph Caching

Zichao Qi, Yanghua Xiao, Bin Shao, and Haixun Wang

Abstract

Distributed systems have the potential to support big graphs. A

distributed system can scale out to hold graphs of any

size. Distributed systems also have the parallelized computing power

to boost the performance of big graph computation. To deploy a big

graph in a distributed system, we must first partition the big graph

into different parts (typically by random hashing) each of which is

loaded to a single machine's memory. Then, all succeeding computations

will run in distributed memories.

Details

Publication typeTechReport
NumberMSR-TR-2013-18
> Publications > Distributed Big Graph Caching