Re: retro crypto
John Denker via cryptography <[email protected]> Sat, 11 Jan 2020 23:32:56 -0700
| Newsgroups | gmane.comp.encryption.general |
|---|---|
| Message-ID | <[email protected]> |
On 1/11/20 10:44 PM, Ben Laurie wrote: > Are you assuming the contents of the EPROM is secret? If so, why not use it > as a OTP? It's not a one-time pad. It's meant to be a few-time pad. A true OTP requires that no byte of the pad ever gets reused. In contrast, in a rotor-like machine, the various bytes do get reused, but in a key-dependent pseudo-random order that the adversary cannot easily figure out. This means the combined length of all messages sent by all users of this system can greatly exceed the amount of randomness stored in the EPROMs, while still maintaining some decent security. > If not, then surely this construction is trivially insecure? Not quite so trivial. The S-boxes in DES are not secret, yet triple-DES is not trivial to break, because there is still some security that comes from the key. You need the key-space to be big enough to foil brute-force searching. I still strongly recommend that the EPROMs be kept secret. However, the key space is large enough to make life unpleasant for the attacker, even if one or two of the EPROMs were compromised. Add a few more rounds it you think it would help. _______________________________________________ The cryptography mailing list [email protected] https://www.metzdowd.com/mailman/listinfo/cryptography