Re: [Acme] Proposed ACME Charter Language

Daniel Kahn Gillmor <> Fri, 15 May 2015 15:11 UTC

Return-Path: <>
Received: from localhost ( []) by (Postfix) with ESMTP id B39141A1BC8 for <>; Fri, 15 May 2015 08:11:12 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -1.9
X-Spam-Status: No, score=-1.9 tagged_above=-999 required=5 tests=[BAYES_00=-1.9] autolearn=ham
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id PFk7bBmF52nM for <>; Fri, 15 May 2015 08:11:10 -0700 (PDT)
Received: from ( []) by (Postfix) with ESMTP id 6691D1A039D for <>; Fri, 15 May 2015 08:11:10 -0700 (PDT)
Received: from (unknown []) by (Postfix) with ESMTPSA id BBDA9F984; Fri, 15 May 2015 11:11:06 -0400 (EDT)
Received: by (Postfix, from userid 1000) id 0F9AF20040; Fri, 15 May 2015 11:10:43 -0400 (EDT)
From: Daniel Kahn Gillmor <>
To: Ted Hardie <>, Randy Bush <>
In-Reply-To: <>
References: <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <> <CA+9kkMAJ-925hQ+wawkLvEjTaf5f1JRHdrGMtCR>
User-Agent: Notmuch/0.20~rc1 ( Emacs/24.4.1 (x86_64-pc-linux-gnu)
Date: Fri, 15 May 2015 11:10:38 -0400
Message-ID: <>
MIME-Version: 1.0
Content-Type: multipart/signed; boundary="=-=-="; micalg=pgp-sha512; protocol="application/pgp-signature"
Archived-At: <>
Subject: Re: [Acme] Proposed ACME Charter Language
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Automated Certificate Management Environment <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Fri, 15 May 2015 15:11:14 -0000

On Wed 2015-05-13 19:36:51 -0400, Ted Hardie wrote:
> ​So, the point I'm getting at is that the system ought to provide
> an automated way to request revocation if the requester can meet
> the same bar as it would take to request or renew a certificate.  If 42
> is one of the ways to meet that bar, well and good.  If 42 is not one
> of the ways to meet the original bar, then putting effort to automating
> revocation on that basis seems off to me.

There is at least one scenario where automated revocation seems
legitimate but automated issuance does not: posession of the
end-entity's secret key.

That is: If i have the secret key, i *also* need to prove something else
(control of the domain) to get a certificate issued.  But if i have the
secret key to an already-issued certificate, i should not need to prove
control of the domain to get the certificate revoked.

If I compromise your secret key, the nicest possible thing i can do with
it is get it revoked.  There is no reason to prevent this action from
anyone who has access to the secret key.

I think i like Russ's first suggestion best:

>> ACME certificate management must, in an automated manner, allow an
>> authorized party to request revocation of a certificate.

because it lets the WG (and implementers/deployers) decide to include
types of authorization that might be relevant for revocation but not for
issuance, while still covering the general case.