Re: Why is the crlNumber an OCTET STRING?
Jeffrey Walton <[email protected]> Wed, 21 Apr 2021 23:28:43 -0400
| Newsgroups | gmane.ietf.x509 |
|---|---|
| Message-ID | <CAH8yC8nbBKd3_gSYqOUMkAsE4yZhcW7m-fvqt+4+PG5A8Qg-eQ@mail.gmail.com> |
On Wed, Apr 21, 2021 at 4:19 PM Stefan Santesson <[email protected]> wrote: > > Yes, I understand the logic. I just say that as implementer I never > found the reason to use this in practice. > > In practice I have a policy that test if my current cached CRL is valid > to use. If yes, then I use it. if no, then I discard it and attempts to > get a new CRL. > > The new CRL is treated the same way before being used. > > Perhaps my implementations are to simplistic, but I never found a > compelling reason to check and compare CRL number. It would make sense to use the existing CRL if the CRL has not changed if it is expensive to download a CRL. The use case would include a large PKI, like a federal PKI with a CRL that is hundreds of MB, and a mobile client over 3G on a restrictive data plan. Jeff