On Peer-to-peer — Content Distribution, Acyclic Preference Networks
FR

Université Pierre et Marie Curie

Habilitation to Supervise Research

On Peer-to-peer:
Content Distribution,
Acyclic Preference Networks

by Fabien Mathieu

Defended on February 11, 2009 before a committee of:

  • M. Pascal Felber, Professor, University of Neuchâtel — Reviewer
  • M. Pierre Fraigniaud, Research Director, CNRS — Reviewer
  • M. Jérôme Galtier, Orange Labs — Examiner
  • M. Laurent Massoulié, Thomson Technology Paris Laboratory — Examiner
  • M. Philippe Robert, Research Director, INRIA — Reviewer
  • M. Sébastien Tixeuil, Professor, Pierre et Marie Curie University — Chair
  • M. Laurent Viennot, Research Scientist, INRIA — Examiner
Peer-to-peer (P2P) recently emerged as a new paradigm in computer science. Due to major economic and social stakes, mostly related to content distribution, P2P has brought back to the forefront many existing research fields related to distributed systems, providing new incentives and goals. In this work, we give some keys to the understanding of the research fields related to P2P systems. After a brief survey of our work on content distribution, we consider a more theoretical subject: acyclic preference-based systems, which recently appeared as an elegant way to model many P2P unstructured or hybrid systems. The strength of these models is a self-stabilizing property that allows us to provide analytical results in addition to empirical validation.

📄 Download PDF

Esc