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 can
support both these features, existing decentralized
architectures seem to support at most onne.  Associative overlays are a
new class of decentralized P2P architectures which met both of these
desirable features. In this talk, I'll first talk about the main idea of
the associative overlays and the challenges to design such an
overlay. Then decribe a specific search strategy used to exploit
associations inherent in human selections, and finally present some
experimental results.