Picolo: Peer-to-peer Infrastructure For Content Delivery Overlays (completed)
About the project
The term content distribution networks is often related to the on demand delivery of video data to many clients via a fixed hierarchical infrastructure that is composed out of server(s) and proxy caches. It is obvious that in the future the content that is distributed will not only be encoded in video data, instead interactive multimedia documents will combine animations, graphics, audio, video, text, and provide a rich set of interaction features to the end-user. The application domains will be more diverse, including tele-teaching applications, scientific collaboration, etc. Finally, the infrastructure to be used for different applications will not only comprise a fixed hierarchy of servers and proxy caches. The peer-to-peer computing paradigm will enable new possibilities for distribution infrastructures. However, today it is not well understood what the impact of using dynamic peer-to-peer infrastructures (and their combination with fixed hierarchical infrastructures) will be onto Quality-of-Service, costs, scalability, and availability. Our two research groups have complementary expertise in the area of content distribution: Eurecom in simulation and Oslo in systems development and experimental evaluation. We intent to combine our expertise to understand the impact of dynamic peer-to-peer infrastructures on content distribution, both from a modelling and simulation perspective and from an experimental perspective. For the latter, we want to emphasize that the location of our research groups in Europe will enable Wide-Area-Networks experiments to observe effects that are not obvious in local laboratories. Our new understanding of this impact shall lead to first ideas for new algorithms and protocols in content distribution infrastructures.
01-jan-2002 - 01-jan-2003
Research Partners: EURECOM http://www.eurecom.fr/