On Sat, 24 Mar 2001, Jim Choate wrote:
Cryptographically interesting. It looks like starting now, the highest-end threat facing a cryptosystem involves liters of fluid performing molecular computation. The kick is that these molecular computers aren't all that hard to design and they scale. Which means someone can have a 1000-liter computer with somewhere near the same ease they can set up a milliliter computer. So we need to revise the recommended key lengths for security purposes. There are about 2^167 atoms in planet earth, about 2^30 nanoseconds per second, and 2^39 seconds till the next ice age. So, if a tenth of that mass were made into ten-atom computing units that could complete a calculation every ten nanoseconds, I get about 2^226 operations before the next ice age. Assuming each is a brute-force check, we should probably be looking at symmetric ciphers with keys a minimum of 225 bits long right about now. Round up to the nearest handy power of 2 and make it 256 bit keys. Fortunately, the AES, and several other recent symmetric ciphers, have a use for a 256 bit key. Unless they find some radical flaw in AES & Co, I don't see a real problem posed by molecular computing; we just need to start taking it into account when we decide what key length to use. Bear