Distributed resource discovery using a content-sensitive infrastructure

Distributed Resource Discovery in a World Wide Web environment using full-text indices will never scale. The distinct properties of WWW information (volume, rate of change, topical diversity) limits the scaleability of traditional approaches to distributed Resource Discovery. An approach combining metadata clustering and query routing can, on the other hand, be proven to scale much better. This paper presents the Content-Sensitive Infrastructure, which is a design building on these results. We also present an analytical framework for comparing scaleability of different distribution strategies.