Using p2p to share the news
p2pnet.net News:- “9th April 2004 – fame at last!” – writes Cambridge professor Ross Anderson on his web page here. “A project of ours is the lead story in the technology section of the BBC website. We are adapting some of the ideas I floated in my paper on the Eternity Service to making usenet news distribution more efficient. Watch this space…”
He’s referring to a BBC story which leads off, “By the year 2010, file-sharers could be swapping news rather than music, eliminating censorship of any kind.
“This is the view of the man who helped kickstart the concept of peer-to-peer (P2P) file-sharing, Cambridge University’s Professor Ross Anderson. In his vision, people around the world would post stories via anonymous P2P services like those used to swap songs. They would cover issues currently ignored by the major news services, said Prof Anderson.”
Currently, “only news that’s reckoned to be of interest to Americans and Western Europeans will be syndicated because that’s where the money is,” he told the Beeb.
“But if something happens in Peru that’s of interest to viewers in China and Japan, it won’t get anything like the priority for syndication. If you can break the grip of the news syndication services and allow the news collector to talk to the radio station or local newspaper then you can have much more efficient communications.”
And to allow this to come about, Anderson is proposing “a new and improved version of Usenet, the internet news service,” says the BBC item.
We’ll drink to that.
April 20th, 2004 at 12:14 am
P2P doesn’t make sense for news.
P2P is an architecture that only makes sense for certain kinds of problems. For normal problems pure p2p isn’t a great option, because it’s very inefficient.
For example, compare the original Napster to modern, pure p2p networks. It was far more efficient to have a single, central database of everything because a search could be performed easily across all available data. In a p2p network you need immense volumes of communication to very slowly give you a search result across a small percentage of all of the nodes in the network. However, it appears that it’s more important not to have a single, central database of everything in the network from a legal perspective, even if the result is slow and inaccurate.
Aside from trying to duck lawsuits, P2P architectures work well if you have to do something that scales infinitely (like TCP/IP networking) or where the information being exchanged is naturally local, not global. For example, Groove’s groupware makes sense for p2p, because once your group assembles, all interactions are within the group, and there’s an advantage in not having to set up a server.
For news, it’s very easy to put news up on a web site somewhere so that people can read it, and perhaps implement some variant of RSS in order to let people who want it know when you have new news. No one news site needs to scale infinitely, so p2p’s scalability isn’t important. And if you can host your site at an anonymous ISP somewhere (Sealand?) you don’t need to worry about hiding yourself too much.
I’ll differentiate between “news” and what FreeNet does, for example, You could put “news” into FreeNet, but nobody would know unless you send emails or put up a web site (or whatever) to tell people that it’s in the network.