Searching Across Peers

In the current P2P world, systems have provided for simple keyword searches but not much more. Gnutella's Infrasearch, for example, enables multiple term definition and processing. Thus, you can search for multiple terms concurrently across a space to locate a match. The Infrasearch model enables simple searches to be executed without significant capabilities to place constraints on the search field. Although Freenet has the capability to couple a metadata file to a piece of content within its P2P network, it does not currently enable searching across a document. The structure of the information in the document is also left up to the user, which often means reverting to the <META> tag issue in HTML.

P2P applications have successfully defined search mechanisms for special cases such as file exchange, which we have discussed using Napster as an example. However, retrieving "all songs by Sting" does not require complex query languages or complex metadata, so special purpose formats for these P2P applications have been sufficient. By concentrating on domain-specific formats, current P2P implementations appear to be fragmenting into niche markets instead of developing unifying mechanisms for future P2P applications.

Edutella

The Edutella open source project (http://edutella.jxta.org/) is attempting to address some of the shortcomings of current P2P applications by building on the W3C RDF metadata standard. The project is an effort to deploy an RDF-based metadata infrastructure for P2P-networks based on the recently announced JXTA framework. The initial services that this framework will be providing include the following:

• Query service Standardized query and retrieval of RDF metadata.

• Replication service Provide data persistence/availability and workload balancing while maintaining data integrity and consistency.

• Mapping service Translate between different metadata vocabularies to enable interoperability between different peers.

• Annotation service Annotate materials stored anywhere in the Edutella Network.

The first application is a P2P network for the exchange of educational resources between German universities, Swedish universities, Stanford University, and others. By linking heterogeneous educational resources across this academic framework, the project aims to enable a user to easily discover content simply and efficiently.

Edutella's generic framework of unique ontologies for different domains that still interrelate will certainly be the winning solution to the metadata challenge in the P2P world.