My research interests lie in the area of Peer-to-Peer systems. More
generally, I'm interested in distributed and networked systems and
related fields.
Stanford Peer-to-Peer
Group
Stanford Databases
Group
Stanford Computer Science
Department
Current peer-to-peer file sharing networks generally use a broadcast
mechanism to propagate queries, in some cases leading to
congestion collapse of the network. As an alternative, several
groups have proposed an iterative, crawl-like search mechanism
where nodes are queried in turn.
I presented an overview of these new mechanisms to the P2P
group. [PowerPoint Slides].
Beverly Yang
and I developed a simulator to analyze the performance of such
systems. The results of our study have been accepted to
ICDCS 2004.
Algorithms exist to facilitate communication in mobile and
ad-hoc networks; additionally, publish-subscribe systems have
been extended to enable mobility. We study the possibility of
using publish-subscribe communication semantics in an
unstructured, highly mobile ad-hoc network, such as a network of
moving automobiles. In this setting, the establishment of
communications hierarchies and long-lived routes is infeasible,
so we examine extensions of epidemic algorithms in order to
communicate effectively.
|