Gossip-based Computer Networking

This workshop is part of a “Series of Events” in the Marie Curie Action SCF (Conferences and Training Courses) and co-sponsored by the EU’s Sixth Framework Programme.

This workshop concentrates on the construction of overlay networks that can link together thousands to millions of devices. A typical example is formed by the infamous peer-to-peer file-sharing networks. The combination of size, decentralized management, and instability of these networks has brought researchers to look for new protocols that can easily scale, support

self-management, and are robust. One such class is formed by gossip-based protocols leading to what are now known as gossip-based networks. In such a network each node has a small collection of neighbors. All decisions are taken based on the information that a node can collect from these neighbors. To this

end, a node periodically contacts one of its neighbors at random to exchange (application dependent) information.

Gossip-based networks have shown to be extremely robust, self-managing and above all, simple. Strangely enough, our understanding of these networks is poor: in many cases we cannot say why a certain property emerges, let alone predict it. Much more research is needed.

The aim of this workshop is to bring together some of the leaders in the field of gossip-based networks along with experts in related fields, to identify the important areas of research and potential solutions. Problem areas we propose to concentrate on include:

Applications

Security

Gossiping in wireless networks

Mathematical modeling and analysis

Experimental evaluations

The format of the workshop is to let small groups work on a specific topic relevant to gossip-based networks, and to eventually produce a position paper that is to be published in the October issue of ACM's Operating System Review  http://www.sigops.org/osr.html. Based on these papers, there will also be an open call for a special issue in Elsevier's Computer Networks on gossip-based networks http://www.elsevier.com/wps/find/journaldescription.cws_home/505606/description#description