Hello forum:

I am looking for some MFC/Java Programers that would want to develop the backend for a gnutella-compatable network. This new network would provide true consistancy with the current gnutella protocalls (and allow connections between these clients), but would also incorporate some of the more advanced features of programs like Edonkey (with partial file sharing and creating a mesh of people sharing the same files, as well as some way of using the hash data to open from a download), OpenNAP (with the capabilities of connecting to a central server if one is available), FreeNet (with limited automated distribution of popular files and information), and of course Kazaa/Grokster (for the ease of use as well as the full metadata capabilties).

The ultimate aim is to have an open source p2p network coexisting with the Gnutella network, only this network would be focused on much more than mp3 and pirated movie distribution. The goal here is to have sites like Download.com or FilePlanet move their download servers from FTP and HTTP Server>Client transmission to a system that is compatable with this new network that ensures that a "mesh" of users dealing in the same file at the same time are sharing that file in an effort to distribute the file the fastest. All the while this network will ensure that smaller files (such as MP3s) are easily found (especially if there is intergration with existing networks such as OpenNap) and recieved.

As stated before, this will be a primarily open-source project. There will be 4 areas of development.

First off will be a backend system written in C++ and compatable with all the major platforms. This will be a program that will sit as a service on a user's machine that will allow multiple programs to access the same connection to the p2p infrastructure. This will ensure that multiple programs on a user's system can interface with the network. This means that a user can swarm from the p2p network a file that he/she is downloading using Internet Explorer or Opera without having to provide seprate connections. This also means that existing client's will have to do very little work to interface with the network, which will allow the client developer more time to worry about interface and less time on the core infrastructure. With a central service there is one more advantage, the network can change very rapidly. If a digitally-signed update is sent accross the network within hours all of the users on the network will be upgraded with new functionality. This is a major drawback with Gnutella (with older clients) and will ease the pains of keeping the network updated. This behind-the-scenes engine will also allow connections to Gnutella, OpenNAP, and allow plugins that will eventuially allow Kazaa, Edonkey, WinMX, Direct Download, and other closed-source networks to become part of this new network.

Secondly, there will be a demo-client written in either MFC/C++ or Java. This will be for general use only, mostly to test out the new platform and to serve as a reference to developers. This will be kept up to date as the network evolves.

The third development will be a server engine that will allow a site like FilePlanet.com or Download.com to track all of the users actively sharing any one file. This will be provided to a user who will then be able to keep track of every user on the network with a specific file that is in demand. With sharing of partial files this will create a mesh of users all with a common need for the same file, which will be extremely useful for larger files and files that are extremely popular (the release of a new Half Life Counterstrike version comes to mind). This servlet will allow the deployment of completely legitimate files accross the network, and will allow the network to expand into other services than the general file-swapping realm.

The foruth development will be a general indexing server, similar to that used in EDonkey, that will allow more powerful supernodes with large indexes to offload their index to a server and proform some queries on these servers. The focus of these servers is the expediate locating of files that would otherwise not be able to be found, as well as to reduce network traffic because the query responses are returned directly to the user. This fourth development will take some manpower to make effecient, but it will help the network proformance out tremendously.

There will also need to be a smaller fifth development of header files and other various interfaces between the network and a program. The idea here is that any program can interface with the network easily, and thus bring their service offering (and the files associated with that service offering) to the network.


Other things that will need to be looked at-
A) Encryption of Secure Files
B) Effecient Query Management Between Networks
C) Correct Placement of Nodes Within Network (Inteligent LoadBalancing)
D) Intergration with other Networks (to ensure benefit to both networks involved)
E) Support of Developers on this new Network



Note: This is not meant to be a competitor to Gnutella, rather an evolution of Gnutella. Gnutella is based on HTTP commands that work file for the simple propigation of files, but adding a layer of inteligence to the network would require too much work and upgrading all the nodes on the network is impossible. The development of this second network will (when done) provide an effecient structure to develop the next generation of peer to peer applications on, after Gnutella is deemed to be ineffecient in comparason. Remember, this is open source (as open as Gnutella is), and the intention of this project is not to make money nor cause the loss of money for those who allready have invested in Gnutella (if anything, Limewire, Bearshare, and ESPECIALLY Morpheus should be devoting some manpower to this new project).

2nd Note: This is a rough draft, the actual structure of the network is up for nogiation (although features wise I think I have a farily good idea of where everything fits and how it needs to be implimented..) Unforturnantly, my programing skills arn't up to a level that would make such a development possible, thus I am looking to developers who are looking to develop an open platform that is competitive with other systems. I can get 15-25 developers onboard this network could be in beta within 2-3 months (it's not like we are reinventing the wheel here, just taking knowledge of current technology (what works and what doesn't) and applying it to an open network design).

Think about it. Gnutella 2.0 may be here sooner than you think ;-)

__________________
- Nick Catalano