Il 26/02/2016 16:02, Peter Lebbing ha scritto:
>> Rotating does only make sense if you take the old key soon offline.
> Why is this the case? I must admit I'm fairly comfortable not rotating
> my keys (which are on OpenPGP smartcards). But I can think of lines of
> reasoning where it makes sense to rotate, but still keep the old
> decryption key available.
In my case: every year will have its own PIN, different from the one
used for signing, and *really* different from the one for certification.
> Think: "There's a non-zero chance that someone
> got my private key material, but at least they can only decrypt stuff
> encrypted in 2011, all other years use a different key".
Extreme case: a judge orders to hand over the key to a set of messages
('cause they won't trust your decryption). Rotating keys minimizes
exposure of other material.
> Note in this scenario it is nice if I can still easily access my
> 2011 material as well.
Exactly.
> I'm not saying this is a solid line of reasoning. I'm just curious why
> limiting access to the decryption key is the only thing that makes sense.
Well, everybody can have his own perfectly valid reasons... Why limit
keys on smartcards more than technically necessary? Years ago cards had
space only for 3 keys, but a 144K Javacard can handle many more!
And if PKCS#11 was useable, one could use as many keys as needed by his
policy.
Note that I really don't like PKCS#11, but it's the de-facto standard to
access nearly every crypto-capable device.
BYtE,
Diego
_______________________________________________
Gnupg-users mailing list
Gnupg-users@gnupg.org
http://lists.gnupg.org/mailman/listinfo/gnupg-users
