Parallel Processing of Large Graphs

TL;DR

This paper compares three parallel graph processing techniques—MapReduce, map-side join, and BSP—on large datasets from various domains, demonstrating BSP's superior efficiency especially for iterative algorithms.

Contribution

It provides a comparative analysis of parallel graph processing methods on real-world large datasets, highlighting the efficiency advantages of BSP over MapReduce.

Findings

BSP significantly outperforms MapReduce in iterative graph algorithms.

MapReduce extension with map-side join improves efficiency over standard MapReduce.

MapReduce remains viable for extremely large networks that do not fit in memory.

Abstract

More and more large data collections are gathered worldwide in various IT systems. Many of them possess the networked nature and need to be processed and analysed as graph structures. Due to their size they require very often usage of parallel paradigm for efficient computation. Three parallel techniques have been compared in the paper: MapReduce, its map-side join extension and Bulk Synchronous Parallel (BSP). They are implemented for two different graph problems: calculation of single source shortest paths (SSSP) and collective classification of graph nodes by means of relational influence propagation (RIP). The methods and algorithms are applied to several network datasets differing in size and structural profile, originating from three domains: telecommunication, multimedia and microblog. The results revealed that iterative graph processing with the BSP implementation always and significantly, even up to 10 times outperforms MapReduce, especially for algorithms with many iterations and sparse communication. Also MapReduce extension based on map-side join usually noticeably presents better efficiency, although not as much as BSP. Nevertheless, MapReduce still remains the good alternative for enormous networks, whose data structures do not fit in local memories.