An Analysis of a Simple P2P Protocol for Flash Crowd Document Retrieval

Today’s Internet periodically suffers from hot spots, a.k.a., flash crowds. A hot spot is typically triggered by some unanticipated news event that causes an unanticipated surge of users to request data objects from a particular site, temporarily overwhelming the site’s delivery capabilities. During this time, the large majority of users that attempt to get these objects face the frustrating experience of not being able to retrieve the content they want while still being able to communicate effectively with all other parts of the network. In this paper, we mathematically model and simulate a simple, distributed, randomized peer-to-peer (P2P) protocol. This protocol runs atop a network overlay during a hot spot to quickly deliver the “hot” object from a small set of users who were able to obtain the object to a much larger set of users that desire it. Our results based on this evaluation show that this type of protocol can scale effectively to millions of users, demonstrating that peer-to-peer approaches can be an effective means to alleviate problems commonly associated with network hot spots.