Re: Online Certificate Revocation Protocol

[Tony Bartoletti <azb@xxxxxxxx>]

At 03:11 PM 6/8/01 -0700, you wrote:
> Using POP for revocation causes problems in two scenarios:
>
> - I have deleted my private key and discover later that it was stolen but 
> not used before I deleted it
>
> - I have physically lost my private key to an attacker (such as my 
> computer was stolen)
>
> The latter is probably much more likely. Given that private keys are often 
> (usually?) protected with crackable passwords, the loss of a computer to 
> an attacker can be pretty disastrous. I assume that many CAs have 
> out-of-band revocation mechanisms for this case, but they certainly would 
> take a long time, and are probably difficult for a typical end user to 
> find out about.
>
> --Paul Hoffman, Director
> --Internet Mail Consortium

It is interesting to take each "misfortune" that might befall a key, and 
consider the likelihood and timeliness of the discovery.  It my PC is 
stolen, I'll notice it pretty quick.  If someone gets inside, either 
remotely or by physical access, and intends to steal the key (to decrypt at 
some later time), they are not likely to destroy my copy (tends to defeat 
the purpose of stealing it, I would assume.)  In such a case, discovery 
depends upon observing activities that may or may not be obvious, even when 
the key is used.  If the key is used to access sensitive information, how 
would I know, unless there is a policy of the information agency to send 
the legitimate requesting party a confirmation/record of accesses.

Of your two examples, the latter is certainly "more likely discovered 
soon", but I don't know if it is a more likely occurrence, overall.  I 
suppose it depends upon how long it is before you notice the theft, or the 
illicit use of the key.

The problem in considering these threats is that it is precisely the 
"stealthy professional" who will trigger the most insidious 
consequences.  The risk of damage is perhaps greater, and yet we tend to 
discount the likelihood.

If CAs may "take a long time" to effect the out-of-band revocation, I have 
to assume the issue is really some combination of "staffing" and "low 
priority", rather than a concern about DoS.

Maybe I'm wrong.

___tony___



Tony Bartoletti 925-422-3881 <azb@xxxxxxxx>
Information Operations, Warfare and Assurance Center
Lawrence Livermore National Laboratory
Livermore, CA 94551-9900