Group Meeting Minutes

Date: June 30^th, 2005

Place: SB232C

Attendees: Wai Gen, Linh, Dongmei

² Linh’s presentation on the technical report “Analyzing the Search Characteristics of Space Filling Curve-based Indexing within the Squid P2P Data Discovery System”

Ø Squid DHT-based system: keyword searching including wildcards, partial keyword and ranges; information guaranteed to be found; preserve locality of data.

Ø Indexing schema: Hilbert Space Filling Curve (SFC).

Ø Query processing and query optimization: reduce the number of generated clusters, the number of messages.

Ø Load balancing: two load-balancing algorithms used – one at node join, another at run-time. Details are in a reference paper.

Ø Simulations were done using both synthetic and real CiteSeer data.

Ø Their experiments were only based on a 3-dimentional keyword space. However, the number of dimensions could be pretty high, which might affect system performance much.

² Linh:

Ø Keep on implementing B+ tree to speed up simulator.

Ø Try both client and server centric masking and compare results.

² Dongmei:

Ø Pick a paper to present for next week.

An architecture for information retrieval over semi-collaborating peer-
to-peer networks. I. A. Klampanos and J. M. Jose. In Proceedings of
the 2004 ACM Symposium on Applied Computing, volume 2, pages 1078–1083, Nicosia, Cyprus, March 14–17 2004.

　　http://www.dcs.gla.ac.uk/~iraklis/publications/klampanos2004p2parch.pdf

Ø Run simulation on the small dataset using uniform distribution and compare it with the results of the skewed data.

Ø Keep on working on limewire prototype. First goal: figure out how metadata got stored and implement hashing and our grouping techniques.