Visit our SourceForge project page for technical information, or go directly to the download page to download the latest release! A live demo and a screen cast demo are now available.
Overview
UP2P is a project aiming to build a schema-based P2P infrastructure, exchanging files and their metadata described in XML.
This system relies on the idea that documents are described by specific metadata, formatted according to some metadata schema, and any peer can define arbitrary schemas to describe the documents it hosts. A peer may use several different schemas, to describe different types of documents (e.g. music files, research papers, photographs...); each schema is shared among a group of peers and defines a community, which will support specific searches. Searches written with the music schema will only be broadcast among peers who use this music schema.
The overlapping communities of peers allow a variety of documents to be exchanged within interest communities on the same infrastructure, and as messages are only sent to relevant peers, the system will scale to larger numbers of peers.
The latest version of U-P2P supports links. This means that a document shared in the network may contain a link to another document, also shared in the network.
For example, a science paper shared in the community of science papers could contain a link towards a Bibtex entry shared in the bibtex community. The beauty of this is that the link will work -- that is, UP2P is capable of retrieving the linked file --- wherever this linked file is stored in the network, as long as it is available, a search will be sent out and the user will be prompted to download the document. If it's stored locally, it will be shown immediately.
Overview
UP2P is a project aiming to build a schema-based P2P infrastructure, exchanging files and their metadata described in XML.
This system relies on the idea that documents are described by specific metadata, formatted according to some metadata schema, and any peer can define arbitrary schemas to describe the documents it hosts. A peer may use several different schemas, to describe different types of documents (e.g. music files, research papers, photographs...); each schema is shared among a group of peers and defines a community, which will support specific searches. Searches written with the music schema will only be broadcast among peers who use this music schema.
The overlapping communities of peers allow a variety of documents to be exchanged within interest communities on the same infrastructure, and as messages are only sent to relevant peers, the system will scale to larger numbers of peers.
The latest version of U-P2P supports links. This means that a document shared in the network may contain a link to another document, also shared in the network.
For example, a science paper shared in the community of science papers could contain a link towards a Bibtex entry shared in the bibtex community. The beauty of this is that the link will work -- that is, UP2P is capable of retrieving the linked file --- wherever this linked file is stored in the network, as long as it is available, a search will be sent out and the user will be prompted to download the document. If it's stored locally, it will be shown immediately.
This system relies on the idea that documents are described by specific metadata, formatted according to some metadata schema, and any peer can define arbitrary schemas to describe the documents it hosts. A peer may use several different schemas, to describe different types of documents (e.g. music files, research papers, photographs...); each schema is shared among a group of peers and defines a community, which will support specific searches. Searches written with the music schema will only be broadcast among peers who use this music schema.
The overlapping communities of peers allow a variety of documents to be exchanged within interest communities on the same infrastructure, and as messages are only sent to relevant peers, the system will scale to larger numbers of peers.
The latest version of U-P2P supports links. This means that a document shared in the network may contain a link to another document, also shared in the network.
For example, a science paper shared in the community of science papers could contain a link towards a Bibtex entry shared in the bibtex community. The beauty of this is that the link will work -- that is, UP2P is capable of retrieving the linked file --- wherever this linked file is stored in the network, as long as it is available, a search will be sent out and the user will be prompted to download the document. If it's stored locally, it will be shown immediately.
Sceencast Demo:
Live Demo
What's available...U-P2P version 2.0 is currently available from our SourceForge download page, either as a binary release deployed in Apache Tomcat v.5.5 (copy it to your hard disk and run it!), or as Source Code (with all the required libraries and an Apache ant build file, which can take care of all the compiling and deployment for you). The Source code is available from a zip file or from the SVN. Details of how to set-up and run U-P2P from the zip releases is found is the User Guide, also downloadable from the SourceForge download page. |
Publications
|
Status of development work
- April 28th, 2009: UP2P v.3, supporting link navigation between resources, was presented at LDOW2009 last week. The source code and a binary deployment were released today on our SF.net download page.
- February 18th, 2009: A separate release of JTella has been made available for download on our sourceforge page.
The main interest is that the implementation is fully documented, which can be useful for a developer who wants to reuse or modify this part.
- January 28th, 2009: UP2P v2.0.3 released with some major bug fixes. JTella seems to be working quite well, and the ditrect file uploads work. Likely the last release before introduction of the tuple-space architecture.
- October 15th, 2008: The SVN repository is up to date with U-P2P v.2.0.2
- October 9th, 2008: Another small bug fixed. U-P2P v.2.0.2 available on SourceForge.net
- October 8th, 2008: Two major bugs have been fixed, and today U-P2P v.2.0.1 is available for download.
- September 29th, 2008: release of U-P2P v.2.0 source code, user guide, and update of U-P2P sourceforge home page to point back here... The beginning of a new era for U-P2P?
Coming soon: a developer's guide for the refactored source code!
- September 11th, 2008: release of U-P2P v2.0 alpha (binary).
This version has reverted to the old functionalities of U-P2P (support of multiple communities without cross-community searches), but a number of problems have been fixed, and the general architecture of U-P2P has been overhauled to serve as basis for the future integration of a tuple-space core.
- May 2008 : after some big construction the refreshed UP2P is wobbly on its legs. The different parts of the applications are decoupled.
- summer 2007: we start a plan to refactor the application to allow further research and development.
- 2006 : a prototype of the system works, based on a web application architecture, rebuilt by Neal Arthorne. A paper is published at CASCON'06 based on this new prototype.
All functional parts of the application are heavily coupled and the beast is difficult to modify. - 2002: the first version of UP2P sees the light, created by Alokhe Mukherjee and Babak Esfandiari
Status of development work
- April 28th, 2009: UP2P v.3, supporting link navigation between resources, was presented at LDOW2009 last week. The source code and a binary deployment were released today on our SF.net download page.
- February 18th, 2009: A separate release of JTella has been made available for download on our sourceforge page.
The main interest is that the implementation is fully documented, which can be useful for a developer who wants to reuse or modify this part. - January 28th, 2009: UP2P v2.0.3 released with some major bug fixes. JTella seems to be working quite well, and the ditrect file uploads work. Likely the last release before introduction of the tuple-space architecture.
- October 15th, 2008: The SVN repository is up to date with U-P2P v.2.0.2
- October 9th, 2008: Another small bug fixed. U-P2P v.2.0.2 available on SourceForge.net
- October 8th, 2008: Two major bugs have been fixed, and today U-P2P v.2.0.1 is available for download.
- September 29th, 2008: release of U-P2P v.2.0 source code, user guide, and update of U-P2P sourceforge home page to point back here... The beginning of a new era for U-P2P?
Coming soon: a developer's guide for the refactored source code! - September 11th, 2008: release of U-P2P v2.0 alpha (binary).
This version has reverted to the old functionalities of U-P2P (support of multiple communities without cross-community searches), but a number of problems have been fixed, and the general architecture of U-P2P has been overhauled to serve as basis for the future integration of a tuple-space core. - May 2008 : after some big construction the refreshed UP2P is wobbly on its legs. The different parts of the applications are decoupled.
- summer 2007: we start a plan to refactor the application to allow further research and development.
- 2006 : a prototype of the system works, based on a web application architecture, rebuilt by Neal Arthorne. A paper is published at CASCON'06 based on this new prototype.
All functional parts of the application are heavily coupled and the beast is difficult to modify. - 2002: the first version of UP2P sees the light, created by Alokhe Mukherjee and Babak Esfandiari
