From: “Alan Pugh” <Alan Pugh@MAILSRV2.PCY.MCI.NET>
To: cypherpunks@toad.com
Message Hash: 08fa62b5ae6c0693f3e7e4369bee3b920d199ae48bb1847409c15e2e7196678d
Message ID: <01HV3KI1816A8ZH61F@MAILSRV1.PCY.MCI.NET>
Reply To: N/A
UTC Datetime: 1995-09-10 08:54:29 UTC
Raw Date: Sun, 10 Sep 95 01:54:29 PDT
From: "Alan Pugh" <Alan Pugh@MAILSRV2.PCY.MCI.NET>
Date: Sun, 10 Sep 95 01:54:29 PDT
To: cypherpunks@toad.com
Subject: Re: Certificates/Anonymity/Policy/True Names
Message-ID: <01HV3KI1816A8ZH61F@MAILSRV1.PCY.MCI.NET>
MIME-Version: 1.0
Content-Type: text/plain
> What about when the CA signing key is stolen, factored, or otherwise falls
> into the wrong hands, thereby possibly making every signature made by the CA
> worthless, or at least questionable?
>
> I assume liability will be based on the CA's efforts to ensure the integrity
> of the signatures it makes (and therefore the confidentiality of the secret
> key components), but what constitutes due diligence? As we all know,
> security measures cover a very wide range and can reach ridiculous
> proportions on both ends of the spectrum... How much security will be
> 'enough' from a legal standpoint...?
an excellent point, and one that i'd not seriously considered until i saw
your post. given today's legal climate, assuring the confidentiality of
a ca's key would be pretty expensive. i suppose burning cd-roms with all
transactions would help to document all transactions, but would not be
definitive as far as the signatures go. any document signed with the key
would really have to be considered valid if the signature itself is to
really mean anything in a legal sense. if one were to allege that a
signature is not valid, even though it checked out cryptologically, how
could one defend against a charge that the secret key had been
compromised. you cannot prove a negative. obviously, ianal, but i would
think it would be reasonably easy to convince at least one jury member
that there is a reasonable dought that the key had been compromised.
then again, it might be similar to a claim that a signature has been
forged.
i think there could be a danger of allowing the confidence in a given
piece of crypto to unduly influence a jury of a document's authenticity
when the key _had_ been compromised. the 'gee wiz' factor could be fairly
significant among juries. given statements like 'you could take all the
computers on the planet and let them crunch on it for a billion years
would be needed', a jury might miss issues of security.
fwiw, my 2 cents...
*********************************************
* / Only God can see the whole *
* O[%\%\%{<>===========================- *
* \ Mandlebrot Set at Once! *
* amp *
* <0003701548@mcimail.com> *
* <alan.pugh@internetmci.com> *
*********************************************
Key fingerprint = A7 97 70 0F E2 5B 95 7C DB 7C 2B BF 0F E1 69 1D
Return to September 1995
Return to ““Alan Pugh” <Alan Pugh@MAILSRV2.PCY.MCI.NET>”
1995-09-10 (Sun, 10 Sep 95 01:54:29 PDT) - Re: Certificates/Anonymity/Policy/True Names - “Alan Pugh” <Alan Pugh@MAILSRV2.PCY.MCI.NET>