Peer to Peer networks

There are essentially two kinds of peer to peer networks. There are the regular ones, as Gnutella originally was, and there are the kinds with Hubs. Hubbed Peer to Peer networks work because some user’s of the internet are insane, and buy really expensive, really overpowered internet connections and computers, perfect for use as miniature centralized servers. These kinds of networks are great for times when you need to be able to quickly search for something and find it, because the hubs can keep a list of all the things their Leaves have available. Most modern file sharing networks work on this principal as a result, though the number of ‘leaves’ to each ‘hub’ vary between different systems. If you don’t need to do fast searching on your network, you can probably go for the simpler normal peer to peer design, where all of the members are created equally, and connect to a few other people each.

One thing to consider with hubbed designs also is that depending on the content of your network, you might be more or less likely to have many members with overpowered systems. You’ll also need a way to figure out who has the higher speed connections.

Another thing about Peer to Peer networks is that you need to bootstrap them if you want them to work on the internet. This is because your computer won’t initially know who else is on the internet using the same software, or how to connect to them, so modern peer to peer networks use “Web Cache’s” run on regular web servers, which keep a list of people who recently joined the network, so people looking to join now, can look at the list, and connect to those previous people.

This pony has been thinking about making an extra doodad for Legs to make Peer to Peer networking as simple as Legs makes regular networking, but it’s such a complex thing to think about that it could take some time!