3

I'm phasing out my old OpenPGP key and am creating a new one. I plan to lock away the main key and do daily work only with subkeys.

The subkeys will be valid for a year and I'll add new ones after that time.

Should I make one subkey each year that is used for both signing and encryption (RSA keys support that)? The alternative would be to have two keys, one for each task.


There is a question about that in general, but I don't think it applies to me:

  • I'm specifically talking about OpenPGP/GnuPG
  • Key escrow at work is no problem since I only use the key for private communication

If the answer is "potentially insecure" - does this hold for RSA OpenPGP keys? Why?

1

1 Answer 1

2

You're right, a single RSA key pair supports both signing and encryption; all restrictions are imposed by key flags configured during key creation time. Other public/private cryptography algorithms only allow one or the other action because of technical reasons (for example, DSA for signature and el-gamal for encryption).

The reasons in the linked Q&A apply also to OpenPGP. On one hand, you have key management advantages like individual key escrow. Generally, by "reusing" the same key, you might enable (known or yet unknown) attacks on the key that would not have been possible otherwise.

All in all, you have some benefits of using distinct keys for encryption and signature, at the same time OpenPGP takes all the hassles in maintaining multiple subkeys. The costs mostly boil down to some additional kilobytes added to your key and a negligible amount of computation time for creating the keys. Apart from running addkey twice, there is no additional administrative effort in creating a second key pair.

Considering there are some advantages (or at least might prevent possible attacks) for no costs, I do not see any reason not to create a second subkey pair.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .