Search Tools Product Report

JXTA Search

(formerly InfraSearch)

Product Information
FAQ

JXTA Search is a distributed search system, allowing clients to query multiple search engines -- it's "peer-to-peer" because it's part of Sun's JXTA system which allows peer publishing and other P2P operations. JXTA Search defines a hub system for routing queries to the best information provider, a Query Routing Protocol (QRP) for sending the queries and responses, and "Queryspaces": meta-data for search engine identification, coverage and scope information. Hubs can forward queries to other hubs as well. This is more standard than a metaseach query and allows access control. All messages are encoded in XML, although the response format is not defined by JXTA Search.

Platform: Java 1.1.4 (Windows 98/ME/2000, Solaris, Linux and Macintosh)
Price: free, Open Source (Apache Software license)

• Takes advantage of existing search engines and database search functions.
• No central index.
• Standards-based, XML and HTTP or JXTA messages.
• Common structure for sending and accepting queries.
• Searches can be distributed among multiple hubs.
• Designed to be scalable to billions of queries per day.

Articles & Reviews

• JXTA Search: A look at the future of searching O'Reilly P2P, June 6, 2001 by Rael Dornfest and Kelly Truelove.
  Interview with Gene Kan (a designer of Gnutella) and Steve Waterhouse.
  
  
• White Paper [in PDF] Steve Waterhouse, Sun Microsystems, early 2001
  Describes the origins and structure of the JXTA peer-to-peer search hub system.