The Case for a Hybrid P2P Search Infrastructure

The Case for a Hybrid P2P Search Infrastructure (2004)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Boon Thau Loo , Ryan Huebsch , Ion Stoica , Joseph M. Hellerstein

Citations:

75 - 1 self

BibTeX

@INPROCEEDINGS{Loo04thecase,
    author = {Boon Thau Loo and Ryan Huebsch and Ion Stoica and Joseph M. Hellerstein},
    title = {The Case for a Hybrid P2P Search Infrastructure},
    booktitle = {},
    year = {2004},
    pages = {141--150}
}

Years of Citing Articles

Bookmark

OpenURL

Abstract

Popular P2P file-sharing systems like Gnutella and Kazaa use unstructured network designs. These networks typically adopt flooding-based search techniques to locate files. While flooding-based techniques are effective for locating highly replicated items, they are poorly suited for locating rare items. As an alternative, a wide variety of structured P2P networks such as distributed hash tables (DHTs) have been recently proposed. Structured networks can efficiently locate rare items, but they incur significantly higher overheads than unstructured P2P networks for popular files. Through extensive measurements of the Gnutella network from multiple vantage points, we argue for a hybrid search solution, where structured search techniques are used to index and locate rare items, and flooding techniques are used for locating highly replicated content. To illustrate, we present experimental results of a prototype implementation that runs at multiple sites on PlanetLab and participates live on the Gnutella network.

Citations

423	Measurement, Modeling, and Analysis of a Peer-to-Peer File-Sharing Workload - Gummadi, Dunn, et al. - 2003 (Show Context) Citation Context ...tribution as well. While individual rare items in the tail may not be requested frequently, these queries represents a substantial fraction of the workload, and are worth optimizing. A separate study =-=[10]-=- has shown that the popularity distribution of a file-sharing workload is flatter than what we would expect from a Zipfian distribution. The most popular items were found to be significantly less popu...
375	Making gnutella-like p2p systems scalable - Chawathe, Ratnasamy, et al. - 2003 (Show Context) Citation Context ...nefits of reducing per-query bandwidth might outweigh the publishing overheads of storing the filename redundantly, which makes Join Indexes a more attractive option. 5 Related Studies A recent study =-=[9]-=- has shown that most file downloads are for highly replicated items. One might think that their findings contradict our analysis in Section 2.1 that shows that queries for rare items are substantial. ...
309	Querying the Internet with PIER - R, HELLERSTEIN, et al. - 2003 (Show Context) Citation Context ...t the tail of the distribution. To evaluate our proposal, we present experimental results of a hybrid file-sharing implementation that combines Gnutella with PIER, a DHT-based relational query engine =-=[11]-=-. Our prototype runs at multiple sites on the PlanetLab testbed, and participates live on the Gnutella network. 2 Setting and Methodology To analyze the Gnutella network, we have instrumented the Lime...
187	Looking up data in P2P systems - Balakrishnan, Kaashoek, et al. (Show Context) Citation Context ...works are effective for locating highly replicated items, they are less so for rare items 3 . As an alternative, there have been proposals for using inverted indexes on distributed hash tables (DHTs) =-=[8]-=-. In the absence of network failures, DHTs guarantee perfect recall, and are able to locate matches within a small number of hops (usually log(n) hops, where n is the number of nodes). However, DHTs m...
144	On the feasibility of peer-to-peer web indexing and search - Li, Loo, et al. - 2003 (Show Context) Citation Context ...ing the inverted indexes for these two keywords are high. A “Britney Spears” query also requires shipping large posting lists to perform the distributed join. Recent back-of-the-envelope calculati=-=ons [12]-=- suggest that shipping large posting lists over DHTs is bandwidth-expensive.sWhile compression techniques and Bloom filters would reduce the bandwidth requirements of publishing, a flooding scheme tha...