4 ms·
The whitepaper is an interesting read. The "large data throughput" bullet-point is justified by the claim that no intermediary forwarding node has any idea whet
by pshc 15y ago
The whitepaper is an interesting read. The "large data throughput" bullet-point is justified by the claim that no intermediary forwarding node has any idea whether it's talking to another intermediary node or the one of the actual endpoints. So once you have a buzzing network of anonymized nodes with large volumes of anonymous traffic, you can open up a direct connection to your buddy and transfer at high speed. Later, you can claim with "reasonable doubt" that you couldn't have possibly known whether or not they were a forwarding node or an endpoint in that conversation. (You can also claim that you didn't know whether or not YOU were a forwarding node!) This seems reasonable to me so long as you're transferring at a rate comparable to the average throughput, though I'm not sure how you could determine that.
Well, I'm not really familiar with this sort of distributed scheme, so that might be old hat, insanely insecure, or what have you.
- sp332 15y agoThat doesn't quite work, because if you open direct links to that friend on a regular basis, someone's going to notice that it's not really random.
- mambodog 15y agoThe "reasonable doubt" part sounds very much like Freenet[1], though IIRC Freenet typically has fairly poor data throughput. [1] http://en.wikipedia.org/wiki/Freenet#Technical_design http://en.wikipedia.org/wiki/Freenet#Technical_design
- Groxx 15y agoThe reasonable doubt would be easily almost-certainly disproved by an ISP watching both nodes, though - the high-throughput is partially achieved by not delaying data / sending junk data, so it's reasonably certain that a high degree of data transfer between two nodes without any high transfer out of either to unknown nodes represents a direct connection. If they're both actively sending data out to other nodes, you're more protected, but odds are you won't be giving the network as much bandwidth as your buddy, much less significantly enough to hide your communication in the noise. If your messages are small, yeah, you're lost in the noise and completely deniable. Which is an interesting advantage.
- chopsueyar 15y agoI think it would work to saturate your allocated upload bandwidth and any data sent would be a combination of sources to constantly have a bunch of anonymous random source data. Anytime you are not actively sending something yourself, it would still saturate the uploads with other people's data. Then the question remains, "What did you send to that guy?" I do not think it would ever allow a direct full transfer. But hopping between a dozen or so users with each serving as dual node types would be enough, particularly if all are within the same ISP network.