Improving Gossip Dynamics Through Overlapping Replicates

Abstract : 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.
Complete list of metadatas

Cited literature [24 references]  Display  Hide  Download

https://hal.inria.fr/hal-01631721
Contributor : Hal Ifip <>
Submitted on : Thursday, November 9, 2017 - 4:14:02 PM
Last modification on : Thursday, June 21, 2018 - 4:38:01 PM
Long-term archiving on : Saturday, February 10, 2018 - 2:40:03 PM

File

416253_1_En_12_Chapter.pdf
Files produced by the author(s)

Licence


Distributed under a Creative Commons Attribution 4.0 International License

Identifiers

Citation

Danilo Pianini, Jacob Beal, Mirko Viroli. Improving Gossip Dynamics Through Overlapping Replicates. 18th International Conference on Coordination Languages and Models (COORDINATION), Jun 2016, Heraklion, Greece. pp.192-207, ⟨10.1007/978-3-319-39519-7_12⟩. ⟨hal-01631721⟩

Share

Metrics

Record views

67

Files downloads

60