4 ms·
Serious question: does the same apply to encryption between routers? Like large PiB/day loads. Between the old 'bandwidth of a 747 full of tapes' exercise and t
by samatman 12y ago
Serious question: does the same apply to encryption between routers? Like large PiB/day loads. Between the old 'bandwidth of a 747 full of tapes' exercise and the speed of XOR, I've wondered for awhile now why it isn't standard. Specialized entropy sources can make a lot of the stuff rapidly, and it seems like a case where reducing that layer of security to a a couple consumable tapes that get swapped out like batteries when they're dead would be practical, and would have the advantage that seizing the cipher clearly involves having several tens of grammes of tightly-encoded data.
- tptacek 12y agoIt doesn't matter how many AES ciphertexts you collect. No number of them that physics suggests could ever be feasible will allow you to break the cipher. That's part of what makes one time pads so silly. The gap between the security margin of AES and the "infinite" security margin of a one time pad is totally meaningless in reality. Meanwhile, attempting to use a one time pad has huge collateral costs throughout the rest of the system.