I. Introduction
The steady growth of data makes it increasingly important to control and reduce the communication of algorithms. The CONGEST model [1] is a widely studied model for low communication algorithms on large graphs and sparse matrices. In this model, each vertex/variable occupies a separate machine, and communicates in synchronous rounds by sending messages of length to its neighbors given by the edges of the underlying graph. This bandwidth restriction implies a polynomial lower bound in the round complexity for many fundamental graph problems [2]–[4]. While early work on efficient algorithms in this model has focused on the minimum spanning tree problem [5]–[7], extensive work over the past few years has led to efficient algorithms for several more fundamental graph problems, such as approximate and exact single-source shortest paths [8]–[13], approximate and exact all-pairs shortest paths [14]–[21], approximate and exact minimum cut [22]–[26], approximate maximum flow [27], bipartite maximum matching [28], triangle counting [29]–[31], and single-source reachability [32], [33].