[net.crypt] Cracking RSA

sibley@psuvax.UUCP (06/25/83)

Sam Wagstaff was here at Penn State yesterday and gave a very interesting
seminar on how to break the RSA cryptosystem.  I suppose people who really
follow these things already know about this, but I hadn't realized just
how much progress had been made on this problem.

He described some (not all) of the algorithms involved and then described
a machine that he and others had built which factors 50 digit integers in
about an hour.  It can factor 70 digit integers in a month -- apparently
they actually did it.  Most surprising is that the machine only cost
$6000.  They have funds for an even better second generation version.
He claimed that it would be possible to factor 100 digit numbers in a
month with such a machine.  However, they won't build a machine with quite
that capacity since some 100-digit RSA systems are actually being used.
Even so, the second version will be a lot more expensive.

Does this mean the end of RSA as a serious system?  He didn't say, but he
did claim that keys much larger than 100 digits entail encryption and
decryption rates which might be too slow for a high-volume operation.

You could, of course, keep both the encryption and decryption processes
secret.  You wouldn't have a public-key system any more, but you wouldn't
have to worry about anyone factoring your key either.

Dave Sibley
...psuvax!sibley