Slashdot Mirror
P2P Streaming Radio
sonicsft writes "RIAA, CARP, and streaming internet radio, oh my. Well these guys may have found a solution. With the tag line, pirate radio for the digital age, they've released a peer to peer streaming radio solution and claim that it is untracable/closable by the RIAA."
The first obvious problem is that by connecting to a number of servers, I don't even need a spanning tree attack to trace the source of the broadcast; I can do it all with the latency differential, to find the "root" node.
The second obvious problem is that you can't find a broadcast source unless it's advertised, and once it's advertised, it can be found.
The third obvious problem is that, even if you solve the second obvious problem using a distribguted naming service and... for lack of a better name for it... "AntiBGP", you still have the problem of being able to use differential to find the source of the inital advertisement.
The fourth obvious problem is that you can find the source through traffic analysis: it's the one without an equal number of packets in and out.
Something like this can only ever be effective with a distributed flood-fill model, where you can trust your nearest neighbor, or your nearest neighbor doesn't actually know what he has. Effectively, this means that you have to go to a store-and-forward model, use hard crypto on the interconnects, and then generate bogus traffic to avoid analysis.
At that point, you would have to find a legitimate an legal use for the network before deploying it, or you are minimally an accessory before the fact and/or involved in a conspiracy to commit. If they can point at your node in the network and prove intent, then you are screwed.
"BlackNet" is really unsuitable for streaming data.
-- Terry