Brief Announcement: Convergence Analysis of Scalable Gossip Protocols

TitleBrief Announcement: Convergence Analysis of Scalable Gossip Protocols
Publication TypeConference Paper
Year of Publication2006
AuthorsPatterson, S, Bamieh B, El Abbadi A
Conference Name20th Int. Symp. on Distributed Computing
Pagination540–542
Abstract

We present a simple, deterministic gossip protocol for solving the distributed averaging problem. Each node has an initial value and the objective is for all nodes to reach consensus on the average of these values using only
communication between neighbors in the network. We first give an analysis of the protocol in structured networks, namely d-dimensional discrete tori and lattices, and show that in an n node network, the number of rounds required for the protocol to converge to within c of the average is O(|log(c)| n^{2/d}). We then extend our results to derive upper and lower bounds on convergence for arbitrary graphs based on the dimensions of spanning supergraphs and subgraphs.