Let's Encrypt certificate lifetime incident: is there any security risk?



  • A few days ago, Let's Encrypt discovered that they misinterpreted RFC 5280, thus making every certificate they issued valid for one second longer than expected.

    The associated issue on Mozilla bug tracker definitively looks serious, and some people are suggesting revoking all current Let's Encrypt certificates (about 183 millions).

    However, I fail to see any tangible security issue associated with this. Is there any, or are they just following standard procedures, whatever the impact?



  • The related task that they've opened explains why they do not plan to revoke certificates, and offers the following reasons for not doing so:

    • "we do not believe that revoking certificates already issued as part of our response would benefit the Web PKI."
    • Since LE certificates are 90 days, 90 days + 1 second is well below maximum allowable lifetime.
    • They claim Chrome and NSS treat their certificates as 90 days, not 90 days + 1 second
    • "...it is not clear... that this constitutes misissuance"

    The 2nd and 3rd claims represent technical reasons why there is no tangible security impact. So, to answer your question, I think those people calling for revocation - which appear to be a minority, based on a casual review of comments to that issue - are doing so out of a letter-of-the-law viewpoint. The letter of the law says misissuance calls for revocation, but the spirit of the law says that the severity of the issue may be taken into account.

    If it aids the letter of the law, then consider that LE has a plan to replace the bad certificates - over a 90 day period.



Suggested Topics

  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2
  • 2