P2P MetaData Search Layers

Distributed Hashtables (DHTs) provide a scalable method of associating file-hashes with a particular location in a distributed network environment. Modifying DHTs directly to support meta-data is difficult, and meta-data search systems such as flooding tend to scale poorly. However, a number of more scalable distributed meta-data search systems have recently been developed that could be deployed in tandem with DHTs, and several are discussed here along with some novel simulation results that concern the scalability and resource limitations of a meta-data search layer that employs semantic routing. Semantic routing is a method of pruning a flooding search such that queries are preferentially forwarded to nodes that can answer those queries. Previous simulations [9] showed that under certain circumstances semantic routing leads to a reduction in search path length. This paper presents further simulation results indicating that the scalability of this effect is a function of the query distribution of individual user search activity.