Triangle enumeration on massive graphs using AWS lambda functions

Abstract

Triangle enumeration is a fundamental task in graph data analysis with many applications. Recently, Park et al. proposed a distributed algorithm, PTE (Pre-partitioned Triangle Enumeration), that, unlike previous works, scales well using multiple high end machines and can handle very large real-world networks. This work presents a serverless implementation of the PTE algorithm using the AWS Lambda platform. Our experiments take advantage of the high concurrency of the lambda instances to compete with the expensive server-based experiments of Park et al. Our analysis shows the trade-off between the time and cost of triangle enumeration and the numbers of tasks generated by the distributed algorithm. Our results reveal the importance of using a higher number of tasks in order to improve the efficiency of PTE. Such an analysis can only be performed using a large number of workers which is indeed possible using AWS Lambda but not easy to achieve using few servers as in the case of Park et al.