You are here

A Lightweight Network Proximity Service Based On Neighborhood Models

TitleA Lightweight Network Proximity Service Based On Neighborhood Models
Publication TypeConference Paper
Year of Publication2015
AuthorsLiao, Y., W. Du, and G. Leduc
Conference Name22nd IEEE Symposium on Communications and Vehicular Technology in the Benelux (SCVT)
Date Published11/2015
Conference LocationLuxembourg

This paper proposes a network proximity service based on the neighborhood models used in recommender systems. Unlike previous approaches, our service infers network proximity without trying to recover the latency between network nodes. By asking each node to probe a number of landmark nodes which can be servers at Google, Yahoo and Facebook, etc., a simple proximity measure is computed and allows the direct ranking and rating of network nodes by their proximity to a target node. The service is thus lightweight and can be easily deployed in e.g. P2P and CDN applications. Simulations on existing datasets and experiments with a deployment over PlanetLab showed that our service achieves an accurate proximity inference that is comparable to state-of-the-art latency prediction approaches, while being much simpler.

Citation KeyYon2015
Refereed DesignationRefereed
Project year: 
Third year
WP(s) associated with the paper: 
WP4 - mPlane Supervisor: Iterative and Adaptive Analysis
Partner(s) associated with the paper's author(s): 
Universite de Liege
Is this an OFFICIALLY supported mPlane paper?: