APICe » Publications » Improving Gossip Dynamics Through Overlapping Replicates

Improving Gossip Dynamics Through Overlapping Replicates

Danilo Pianini, Jacob Beal, Mirko Viroli
Gossip protocols are a fast and effective strategy for computing a wide class of aggregate functions involving coordination of large sets of nodes. The monotonic nature of gossip protocols, however, mean that they can typically only adjust their estimate in one direction unless restarted, which disrupts the values being returned. We propose to improve the dynamical performance of gossip by running multiple replicates of a gossip algorithm, overlapping in time. We find that this approach can significantly reduce the error of aggregate function estimates compared to both typical gossip implementations and tree-based estimation functions.
Keywords: gossip, distributed algorithms, aggregate computing
Coordination Models and Languages, Lecture Notes in Computer Science 9686, pages 192--207, 2016.
Alberto Lluch Lafuente, Jos{'{e}} Proen{c{c}}a (eds.), Springer Berlin Heidelberg
@inproceedings{GossipingCoordination2016,
  author    = {Pianini, Danilo and
               Beal, Jacob and
               Viroli, Mirko},
  title     = {Improving Gossip Dynamics Through Overlapping Replicates},
  booktitle = {Coordination Models and Languages - 18th {IFIP} {WG} 6.1 International
               Conference, {COORDINATION} 2016, Held as Part of the 11th International
               Federated Conference on Distributed Computing Techniques, DisCoTec
               2016, Heraklion, Crete, Greece, June 6-9, 2016, Proceedings},
  pages     = {192--207},
  year      = {2016},
  url       = {http://dx.doi.org/10.1007/978-3-319-39519-7_12},
  doi       = {10.1007/978-3-319-39519-7_12},
  editor    = {Alberto Lluch Lafuente and
               Jos{\'{e}} Proen{\c{c}}a},
  series    = {Lecture Notes in Computer Science},
  volume    = {9686},
  publisher = {Springer},
  year      = {2016},
  url       = {http://dx.doi.org/10.1007/978-3-319-39519-7},
  doi       = {10.1007/978-3-319-39519-7},
  isbn      = {978-3-319-39518-0}
}