Re: [therightkey] DNSNMC deprecates Certificate Authorities and fixes HTTPS security

Rob Stradling <> Tue, 31 December 2013 14:31 UTC

Return-Path: <>
Received: from localhost ( []) by (Postfix) with ESMTP id B09701AE28E for <>; Tue, 31 Dec 2013 06:31:55 -0800 (PST)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: 1.41
X-Spam-Level: *
X-Spam-Status: No, score=1.41 tagged_above=-999 required=5 tests=[BAYES_50=0.8, HELO_MISMATCH_NET=0.611, SPF_PASS=-0.001] autolearn=no
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id x7CLl-liabDV for <>; Tue, 31 Dec 2013 06:31:52 -0800 (PST)
Received: from ( []) by (Postfix) with ESMTP id 5FA121AE027 for <>; Tue, 31 Dec 2013 06:31:51 -0800 (PST)
Received: (qmail 7616 invoked by uid 1000); 31 Dec 2013 14:31:43 -0000
Received: from (HELO []) ( (smtp-auth username rob, mechanism plain) by (qpsmtpd/0.40) with (CAMELLIA256-SHA encrypted) ESMTPSA; Tue, 31 Dec 2013 14:31:43 +0000
Message-ID: <>
Date: Tue, 31 Dec 2013 14:31:43 +0000
From: Rob Stradling <>
User-Agent: Mozilla/5.0 (X11; Linux i686; rv:24.0) Gecko/20100101 Thunderbird/24.1.1
MIME-Version: 1.0
To: Jacob Appelbaum <>,, Seth David Schoen <>
References: <> <> <> <> <>
In-Reply-To: <>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [therightkey] DNSNMC deprecates Certificate Authorities and fixes HTTPS security
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Tue, 31 Dec 2013 14:31:55 -0000

On 23/12/13 18:29, Jacob Appelbaum wrote:
> Phillip Hallam-Baker:
>> You can't calculate the number of CAs the way the EFF tried to. An
>> intermediate certificate does not equate to a CA. Pretending it does to
>> peddle an alternative PKI scheme calls into question their veracity.
> I disagree strongly. I have an intermediate certificate. I am as
> powerful CA as a result.

Jake, you're only that powerful if you control the intermediate private key.

> Other estimates appear to be much higher than the EFF count. What is
> your qualification for what counts as a CA? For example - Debian
> GNU/Linux ships with one set of ca-certificates, Chrome on Windows ships
> with another, heck Microsoft even adds new CA certs dynamically, right?
> So what is your metric exactly?

I would prefer to count the number of distinct organizations that 
control at least 1 private key that is associated with at least 1 
non-Name-Constrained root or intermediate certificate that chains to (or 
is) a root in the Microsoft, Mozilla and/or Apple root store and which 
can issue certs that are trusted for Server Authentication.

It's not possible to measure this purely by examining the body of 
root/intermediate certificates that are known to exist (although this 
body of certificates is of course useful for cross-referencing).

>> 2) Continuing to count the DFN as 300 CAs when they know it is one.
> The number matters because it isn't just an issue of control over a
> single signing key. I'd be interested to hear how many of those
> CAs/sub-CAs are able to sign leaf certificates.

All of the DFN Sub-CAs are able to sign leaf certificates, but it is 
_only_ DFN that controls the private keys that would be used to sign 
these leaf certificates.  The various German universities are 
essentially only RAs, even though they are named as the Subjects of the 
intermediate certificates.

Many Sub-CA certificates issued by major commercial Root CAs exist 
purely for branding reasons.  i.e. the Subject is at most an RA, and 
sometimes only a Reseller.

On the other hand, if there are still any RAs/Resellers that control 
root or intermediate private keys, then by my metric they should be 
counted as CAs.

My gut feeling is that the real number (by my metric) is likely to be a 
lot nearer to 60 than to 600.

Rob Stradling
Senior Research & Development Scientist
COMODO - Creating Trust Online