r4 - 05 Apr 2007 - 15:13:50 - RobinGareusYou are here: TWiki >  Blog Web > BlogEntry3
13 Mar 2007 - 10:54 tagged , , by RobinGareus?

Besides our enthusiastic kick-off, we slowly gain a more realistic but not pessimistic view on peer-to-peer web-indexing and searching: The upcoming year(s) will bring various PhD theses on that subject, addressing shortcomings of pioneer work, and providing a fundament for next-generation P2P applications.

open-search.net is aiming towards a tightrope walk between academic and real-world appliance. [NPO, open-source, free -> distinct from software-industry !] – For the time being, we feel that we can best provide valuable solutions for issues that are not addressed at University, while refraining from actual research on Peer-to-Peer and resort to existing P2P-implementations.

open-search.net will wrap one or more p2p lib/app and provide a open-source framework to aid development, prototyping, testing and distribution a p2p search-engine.

We split the project into 2 major task-forces:

  • (A) API + back-end devel
  • (B) Community + front-end

(A)

It is still one of our goals to povide the core in POSIX-C but it feels much more realistic to glue P2P libraries and search cores via local sockets and provide an API to bind crawlers, P2P-storage and data-processing to other programming languages.

(using java itself introduces latencies which are an order of magnitude larger that a local-socket API. Python, perl or haskell bindings might come in handy.)

The framework comprises a simple HTTP server, open-search configuration and session management, and API abstration to manage search-requests/results wrapped in a installable package.

The database to store the index, will evolve from static-file, via a dynamic query interface (forward queries to other search-engines) to a P2P example implementation (eg. using gnutella & existing torrents). well use the opensearch and other XML based import/export mechanisms.

p2p is not yet ready for a commercial grade search engine: thus flexibility and rapid prototyping are to be focused on!

(B)

different audiences:
  • devels (core + framework)
  • devels (serach + p2p)
  • desigeners + front-end devels
  • testers and users
  • privacy – part of the open-search frameworks API.
  • anonymity – separate layer (eg. TOR) – check p2p lib. support


Leave a Reply

<input type="hidden" name="Nr" value="1 *~~" />
You may have to login or register to comment if you haven't already.

Edit | WYSIWYG | Attach | Printable | Raw View | Backlinks: Web, All Webs | History: r4 < r3 < r2 < r1 | More topic actions
 
Copyright © 1999-2009 by the contributing authors. All material on this collaboration platform is the property of the contributing authors. Ideas, requests, problems regarding TWiki? Send feedback.