Computing That Serves

Swarming: Peer-to-Peer Content Distribution for the Masses


Thursday, January 29, 2004 - 10:00am


Dr. Daniel Zappala, University of Oregon

Most web sites run by "ordinary" users are not able to serve high volumes of traffic, since doing so requires spending thousands of dollars, either for a faster Internet connection or for a Content Distribution Network. In this talk I will discuss swarming, a new content delivery system that is based on peer-to-peer file transfer. With swarming, any client that has downloaded even a portion of a file may serve that portion to other clients. This means that the capacity of a swarming system actually increases as more clients join the system. I will present recent results from our performance evaluation of swarming, illustrating the benefits and limitations of a basic swarming protocol and identifying several key opportunities for performance improvements. I will also use swarming as a motivating example of the power of peer-to-peer applications and outline future directions for my research.


Daniel Zappala ( is an Assistant Professor in the Department of Computer and Information Science at the University of Oregon. His research interests include multicast, peer-to-peer networking, and routing in ad-hoc wireless networks. Dr. Zappala started the networking program at the University of Oregon, founding the Network Research Group and establishing their networking curriculum. Prior to joining Oregon, Dr. Zappala was a research assistant at the USC Information Sciences Institute, an intern at Xerox PARC, and a member of the RSVP project. He is a coauthor of "RSVP: A New Resource ReSerVation Protocol", which was chosen as one of ten landmark articles for the 50th anniversary of the IEEE Communications Society. Dr. Zappala received his B.S. in Electrical Engineering from Stanford University in 1990 and his Ph.D. in Computer Science from the University of Southern California in 1997.