Approximation of Distances and Shortest Paths in the Broadcast Congest Clique

Abstract

We study the broadcast version of the CONGEST-CLIQUE model of distributed computing. This model operates in synchronized rounds; in each round, any node in a network of size n can send the same message (i.e. broadcast a message) of limited size to every other node in the network. Nanongkai presented in [STOC'14] a randomized (2+o(1))-approximation algorithm to compute all pairs shortest paths (APSP) in time ~{O}(sqrt{n}) on weighted graphs. We complement this result by proving that any randomized (2-o(1))-approximation of APSP and (2-o(1))-approximation of the diameter of a graph takes ~Omega(n) time in the worst case. This demonstrates that getting a negligible improvement in the approximation factor requires significantly more time. Furthermore this bound implies that already computing a (2-o(1))-approximation of all pairs shortest paths is among the hardest graph-problems in the broadcast-version of the CONGEST-CLIQUE model, as any graph-problem where each node receives a linear amount of input can be solved trivially in linear time in this model. This contrasts a recent (1+o(1))-approximation for APSP that runs in time O(n^{0.15715}) and an exact algorithm for APSP that runs in time ~O(n^{1/3}) in the unicast version of the CONGEST-CLIQUE model, a more powerful variant of the broadcast version. This lower bound in the broadcast CONGEST-CLIQUE model is derived by first establishing a new lower bound for (2-o(1))-approximating the diameter in weighted graphs in the CONGEST model, which is of independent interest. This lower bound is then transferred to the CONGEST-CLIQUE model. On the positive side we provide a deterministic version of Nanongkai's (2+o(1))-approximation algorithm for APSP. To do so we present a fast deterministic construction of small hitting sets. We also show how to replace another randomized part within Nanongkai's algorithm with a deterministic source-detection algorithm designed for the CONGEST model.