I read
Swarming: Scalable Content Delivery for the Masses by
Dr. Daniel Zappala on my way home yesterday. It was a fascinating article and a peek at content delivery of the future. Dr. Zappala has put a serious content delivery framework on the table and demonstrates the value of peer-to-peer content delivery beyond file sharing. There are a number of hurdles to overcome, such as dynamic server initiation and peer selection processes. He openly recognizes these as challenges and is working to address them.
One of the things that I am hoping to get out of the future research is an understanding of the system’s equilibrium. As described in the paper, I imagine equilibrium being the point at which download performance from the root server equals the download performance from the peers. This is also the point at which a dynamic server initiated component need to be working. Actually, you would want the server to initiate the dynamic component before this equilibrium point is reached based on the amount of time it takes to construct a minimal mesh to balance the rate of arrival of new clients versus the rate of dispersal and normalize download performance realized by the peers. I believe that based on this formula -- however it is represented mathematically -- and some predefined metrics for performance you could create a steady state system whose participants are constantly changing.
