[p2p-hackers] Stanford Distributed Systems Seminar, Wed. 05/15, Edith Cohen

[Laurence Melloul]

            Stanford Distributed Systems Research Seminar

Title :  Associative Search in Peer to Peer Networks: Harnessing Latent
Semantics

Speaker :  Edith Cohen, AT&T Labs-Research

When:      12:45PM, Wednesday, May 15th, 2002

Where:     Mc Cullough 115, Stanford University
URL:       http://cs548.stanford.edu/schedule.shtml
Map:
http://www.stanford.edu/home/map/search_map.cgi?keyword=&ACADEMIC=McCullough+Electrical+Engineering

Abstract:

The success of a P2P file-sharing network highly depends on the
scalability and versatility of its search mechanism.  Two particularly
desirable search features are scope (ability to find infrequent items)
and support for partial-match queries (queries that contain typos or
include a subset of keywords). While centralized-index architectures
(such as Napster) can support both these features, existing
decentralized architectures seem to support at most one: Popular
unstructured P2P networks (such as Gnutella and FastTrack-based
Morpheus) deploy a "blind" search mechanism where the set of peers
probed is unrelated to the query; thus they support partial-match
queries but have limited scope.  On the other extreme, distributed
hash tables (DHTs) such as CAN and CHORD, which were recently
proposed in the research community, couple index location with the
item's hash value, and thus have good scope but can not effectively
support partial-match queries.  Another drawback of DHTs stems from
their tight control of the overlay structure: the set of neighbors of
each peer and the information (part of the index) each peer maintains.
This tight control makes them more sensitive to peer and network
failures and frequent joins and disconnects.

We develop a new class of decentralized P2P architectures.  Our
design is based on unstructured architectures such as gnutella and
FastTrack, and retains many of their appealing properties including
support for partial match queries, and relative resilience to peer
failures.  Yet, we obtain orders of magnitude improvement in the
efficiency of locating rare items.  Our approach exploits associations
inherent in human selections to steer the search process to peers that
are more likely to have an answer to the query.

This work is joint with Amos Fiat and Haim Kaplan.

Bio:

Edith Cohen is a researcher at AT&T Labs-Research, which she joined in
1991 (then Bell Labs), after recieving a PhD in Computer Science from
Stanford University. She spent part of 1997 as a visiting professor in
UC Berkeley.  Her research interests include the design and analysis
of algorithms, combinatorial optimization, Web and Internet
performance, networking, and data mining.





+----------------------------------------------------------------------------+
| This message was sent via the Stanford Computer Science Department         |
| colloquium mailing list.  To be added to this list send an arbitrary       |
| message to colloq-subscribe at cs.stanford.edu.  To be removed from this list,|
| send a message to colloq-unsubscribe at cs.stanford.edu. For more information,|
| send an arbitrary message to colloq-request at cs.stanford.edu. For directions|
| to Stanford, check out http://www-forum.stanford.edu                       |
+-------------------------------------------------------------------------xcl+