There are a number of existing peer-to-peer (P2P) software applications. Most applications are targeted for file sharing that uses caching of popular content for efficiency. There are a few distributed hash table (DHT) based applications as well that are more suited for Internet telephony. However, many of these existing systems suffer from one or more of the following drawbacks:
- The implementation is proprietary even if the application executable is free to use. This means developers cannot use it for experimenting with new ideas.
- The implementation is too complex with too many files, modules and interaction among modules. Sometimes this is due to the choice of implementation language and other times due to complex design philosophies.
- The implementation is either not portable or requires significant effort to port on various platforms.
- The application does not use standard protocols such as IETF's SIP and RTP for real-time communication.
- The application does not do full peer-to-peer, i.e., only media path is end-to-end by solving the NAT traversal problem. A true peer-to-peer communication application should try to do both signaling and media as peer-to-peer. It should not rely on central authentication server for every login.
There are some remarkably great software applications that fill in part of the picture. For example Benny's pjsip.org provides a very fine high performance, small footprint library for SIP, RTP, media, NAT traversal, etc. JAIN NIST stack for SIP has been used extensively in several projects. Sean's Bamboo DHT provides a open source Java based DHT implementation. The project goals of these projects are different from ours. In particular, we advise you to not use our software if you are interested in high performance, embedded platform or well tested piece of engineering.
Goals of this project:
Our project is meant as experimental for developers (hint: project students) to quickly try out new ideas, without having to write thousands of lines of code or without spending a lot of time to see if a small variation in the existing algorithm works well.
Another main goal of the project is to help the developers understand the RFCs and drafts of various protocols and how they translate into source code. This helps developing more inter-operable software easily. Usually a developer has to translate the specification into object oriented design and pseudo-code, and then write the implementation. This project provides design of several specifications in Python which is a pseudo-code style language.
