[cabfpub] Proposal of a SHA-1 exception procedure

Dean Coclin Dean_Coclin at symantec.com
Fri Jun 17 22:35:45 UTC 2016


If you do, it would be beneficial if they read the proposed process and come with concrete concerns about why aspects of it would be problematic.

>>Yes, they have read it and do have comments they would like addressed. 


That said, a perhaps more useful and productive use of the discussion time would be to hear from those who examine the audit reports - that is, root store operators - to get a sense of what their expectations are if an organization issues a SHA-1 certificate and presents a qualified audit stating as much.

>>That would be good too.


For example, I think Jody's made it clear that Microsoft just wants someone (and it's unclear if it'll be Microsoft or if they'll just delegate that to the ecosystem) to run counter-cryptanalysis on the proposed tbsCertificate. Presumably, but not stated as such, if the probability of a collision is raised during that counter-cryptanalysis, then Microsoft wants to be able to 'veto' and say they won't accept such issuance.

>>I never heard him say it was going to be Microsoft. My understanding is that he was fine with what was proposed in Bilbao (i.e. the 3rd party approach).


For Google, the procedure we laid out is one that, so far, we think best represents the balance between the ecosystem participants. That includes the necessary disclosures and information so that we can gather information necessary to avoid such situations in the future, while having the necessary transparency for us effectively accepting, on behalf of the Internet trust ecosystem, the security risks.

>>I think this is the crux of the issue which will require dissecting this sentence. First, “That includes the necessary disclosures and information so that we can gather information necessary to avoid such situations in the future”: This is great and  I don’t think anyone has an issue with gathering this information so the CA/B Forum and root store operators can avoid future issues. The second part, “while having the necessary transparency for us effectively accepting, on behalf of the Internet trust ecosystem, the security risks” focuses on the security risks which I thought were ameliorated by the cryptanalysis. Is this not true?


It's useful to know what Apple/Mozilla/Opera/Qihoo360 think, as well as any other root store program that may be presented with such audits.



To reiterate, the goal is not to say that the procedure is the minimum (of what we can all agree on) - rather, it's to make sure that any procedures are not intrinsically in conflict - for example, Google saying you MUST NOT do X, while Microsoft saying you MUST do X.

>>So far, I haven’t heard any conflicting requirements.


Realize as well that it's not guaranteed that other programs even consider this at all something they'd consider. For example, if Mozilla were to say "We will remove any CA that issues SHA-1 certificates", then this whole discussion is a moot-point, is it not?

>>Perhaps, but as you say, it would be good to hear from all browsers first.


While we (Google) are interested in understanding the as-yet-unclarified concerns you/your customers have, to understand and inform our suggested procedure, it would seem that, as a CA and as the Forum, there's a need for broader involvement from root programs - of which you've only heard from two so far (Microsoft and Google) making any statements pro or con to the proposal.


To your implicit suggestion that the phone call is equivalent, I would point out that Interested Parties are not able to participate on such calls or ask questions, while they are permitted such on the mailing list. So while I appreciate the implication that it's post-hoc transparent (typically, 2 weeks after such calls), it's not open, nor in a discussion of this nature is it timely. We saw as much with WorldPay discussions, so I'm somewhat surprised to see it suggested as a viable solution again, when it's limitations were clear.

>>I’m only suggesting it as a more direct approach. I prefer, as an intermediary, not to be a roadblock. As you know, our bylaws provide the Chair with some latitude for inviting guests to meetings. Interested Parties that are fascinated by this discussion and express this to me could see an invitation coming their way for this portion of the call. Alternatively, I can promise to provide a meeting excerpt within 24 hrs to the public list of this particular discussion. 


On Fri, Jun 17, 2016 at 12:02 PM, Dean Coclin <Dean_Coclin at symantec.com <mailto:Dean_Coclin at symantec.com> > wrote:

Perhaps I should offer up again one of the representatives of these companies or trade associations the opportunity to present on our next call. This will give you and others the chance to ask your questions directly w/o this back and forth which is not productive. This dialog will be recorded in our minutes so there will be no opportunity for anything other than a transparent discussion. You can hear firsthand what I stated below (which I quoted directly from them).  


From: Ryan Sleevi [mailto:sleevi at google.com <mailto:sleevi at google.com> ] 
Sent: Friday, June 17, 2016 12:58 PM
To: Dean Coclin <Dean_Coclin at symantec.com <mailto:Dean_Coclin at symantec.com> >
Cc: Gervase Markham <gerv at mozilla.org <mailto:gerv at mozilla.org> >; CABFPub <public at cabforum.org <mailto:public at cabforum.org> >
Subject: Re: [cabfpub] Proposal of a SHA-1 exception procedure




On Fri, Jun 17, 2016 at 9:46 AM, Dean Coclin <Dean_Coclin at symantec.com <mailto:Dean_Coclin at symantec.com> > wrote:

No Processor's legal department will allow them to put out a public form saying, "We are using SHA-1". They don't understand why browsers think this is a good idea.




If you're going to make broad, sweeping, absolute statements, then it would help if you - or the customers you're claiming to represent - would explain why. If your goal is to suggest that Google reconsider the need for transparency, then you - and those customers - have an obligation to explain why that is. Statements like the above, and statements like you've made on the thread, objectively do not help further the discussion, and only serve to postpone and delay any further consideration of SHA-1 allowances.


If your goal is to support your customers, you're only hurting them with statements like this.


A useful furtherance of the discussion, rephrasing what yous aid, might be

"It's unlikely that payment processor's legal department will allow them to publicly admit "We are using SHA-1", because of [concerns X, Y, Z]."


Of course, to also reiterate the previous discussions, "because security and privacy" aren't really concrete or actionable concerns - they're opaque, vague, and broad. They don't help inform the discussion about the tradeoffs - about the need for ecosystem transparency.


If the proposition is that "Admitting you use SHA-1 is to put yourself at risk", then please consider what you're asking - that the entire Internet trust ecosystem accept the risk on behalf of that payment processor (and those like them), that need SHA-1 certificates. That's a completely unreasonable request, without further details.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cabforum.org/pipermail/public/attachments/20160617/07c91207/attachment-0003.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5723 bytes
Desc: not available
URL: <http://lists.cabforum.org/pipermail/public/attachments/20160617/07c91207/attachment-0001.p7s>

More information about the Public mailing list