Fixing the download mirrors @ GCW...

[PrinceGaz]

The emule network does use servers but the vast majority of them have no restrictions on what is shared on them (except perhaps a max number of files per client). A server just keeps a record of what each client connected to them currently has shared. When you want to download a file, your client first queries the server its connected to to find out if any other clients connected to it have the file you want. If it has less than about a hundred sources for the file, after a few minutes delay your client starts asking other servers if they know anyone with the file. You don't need to change servers to do this or to download from people on other servers.

Servers are just a centralised database of what everyone connected to it has shared. It therefore doesn't matter which server you connect to either for uploading or downloading (but it is more efficient when files have few sources for everyone to be on one server to help them find each other quickly).

Taking your points one at a time:

• You won't need a "GCW Server", it would be nice to have one but totally unnecessary
• Most people don't join a specific server regardless of what they're downloading, they just use the first one they are able to connect to.
• Yes, many emule users do d/l and share illegal files. As most people don't care which server they connect to, servers usually have people sharing all types of files. The server itself isn't involved in any uploading or downloading of files however, only keeping a database of what everyone on it is sharing. It might be possible to restrict the files the server keeps track of (by size, type etc), I really don't know. If thats a problem, the solution is simple-- don't have a GCW server
• While servers do usually have a limit on how many files a person can share, that limit is per person, not per server. In theory everyone connected to a server could be sharing a different 500 files.

I do like the queue idea, waiting in a queue for an hour or two knowing you will get a file at the end of it is better than attempting to get something from an overloaded server which keeps stalling.