FastTrack

I hate to ask anything about FastCrap but how does it work?

Is it fully distributed like Gnutella or does it connect to indexing servers like eDonkey? Or none of those? Thanks.

FastTrack is a so-called second generation P2P protocol. It uses supernodes to improve scalability. The supernode functionality is built into the client; if a powerful computer with a fast network connection runs the client software, it will automatically become a supernode, effectively acting as a temporary indexing server for other, slower clients.

In order to be able to initially connect to the network, a list of supernode IP numbers is stored in the program. The client attempts to contact these, and as soon as it finds a working supernode, it requests a list of currently active supernodes, to be used for future connection attempts. The client picks one supernode as its "upstream" and uploads a list of files it intends to share to that supernode. It also sends search requests to this supernode. The supernode communicates with other supernodes in order to satisfy search requests. The client then connects directly to a peer to download the file; this transfer is done using HTTP.

To allow downloading from multiple sources, FastTrack employs the UUHash hashing algorithm. While UUHash allows very large files to be checksummed in a short time, even on slow computers, it also allows for massive corruption of a file to go unnoticed. Many people, as well as the RIAA, have exploited this vulnerability to spread corrupt and fake files on the network.