[cryptography] current limits of proving MITM (Re: Gmail and SSL)

Kevin W. Wall kevin.w.wall at gmail.com
Mon Dec 17 18:18:39 EST 2012


[A bit OT. Sorry]

On Sun, Dec 16, 2012 at 5:51 PM, Jeffrey Walton <noloader at gmail.com> wrote:
> On Sun, Dec 16, 2012 at 4:48 AM, ianG <iang at iang.org> wrote:
>> On 16/12/12 11:47 AM, Adam Back wrote:
[snip]
>>> On Sun, Dec 16, 2012 at 10:52:37AM +0300, ianG wrote:
>>>>
>>>> [...] we want to prove that a certificate found in an MITM was in the
>>>> chain
>>>> or not.
>>>>
>>>> But (4) we already have that, in a non-cryptographic way.  If we find
>>>> a certificate that is apparently signed by say VeriSign root and was
>>>> found in an MITM, we can simply publish it with the facts.  Verisign
>>>> are then encouraged to disclose (a) it was ours, (b) it wasn't ours,
>>>> or (c) mmmmummm...
>>>
>>> Verisign cant claim it wasnt theirs because the signing CA it will be
>>> signed
>>> by one of their roots, or a sub-CA thereof.
>>
>> Just to nitpick on this point, a CA certainly can claim that they or an
>> agent did not sign a certificate.  And, they can provide the evidence, and
>> should have the ability to do this:  CAs internally have logs as to what
>> they did or did not sign, and this is part of their internal process.
> That brings up a good point: the CA should be responsible for their
> reseller's or agent's actions. The CA entered into the relationship,
> and no one forced them into the partnering.
>
> I also envision a scenario where a CA sets up a subsidiary (that is, a
> distinct corporate entity) and then uses the new corporate entity to
> subvert the spirit and intentions of the system. Later, the CA claims
> "it was them, not us."
>
> Lack of responsibility and accountability are part of the problem. It
> needs to be addressed.

IANAL (thank God! ;), but I really don't see how this could work, at
least unless there were laws specific and restricted to only cases like
this narrow focus and I don't see that as likely. I think that this will
continue to be enforced by legal binding contractual agreements
rather than regulatory issues. There would be great resistance from
most businesses to have it otherwise.

What you propose all sounds good on the surface, specifically if
the intent of the CA is to create such a subsidiary for potentially nefarious
purposes, but intent is something is difficult to regulate as well as being
difficult to prove.

I'm sure if the scenario that you outline were to happen and a breach
resulted because of it, both the CA and their subsidiary could be sued
even without any specific existing laws governing this.  In many (most?)
states in the USA (well, at least in Ohio), one cannot completely waive
tort despite what the contract that sign says. (Or so my attorney informs
me.) If it can be shown that there is an intent to defraud or negligence
is involved (especially if it is intentional), the contract is thrown out the
window. (Of course, showing enough evidence in court is another
matter entirely.) In some such cases, even criminal charges are
possible.

But there is also an inherent risk in doing business and no business in their
right minds would ever sign an agreement where they are liable for some
other service provider's screw-ups. Usually, businesses want contractual
agreements with their service providers where the service providers agree to
liability in cases where service provider screws-up. (E.g., where their buggy
software causes an unexpected service outage.) In my experience, most service
providers--at least those with deep pockets--are reluctant to agree to
even that much. So it is highly unlikely of businesses are going to support
any type of legislation that makes them liable for what their service providers
do. They naturally want to shed liability, not take it on.

In the specific case that you mention, even if there were such specific laws
it would likely mean an end CAs creating such CA subsidiaries which
probably means higher prices for certificates all of us.

If you think about what you are asking for in the *general* sense, I think you
might reconsider. For example, consider a case where a merchant wants to
do a credit check so they send a credit bureau an SSN and DOB and get back
a credit rating. Let's suppose the merchant does all this securely and
doesn't even
permanently store the SSN / DOB, but only holds it long enough to get a credit
rating back from some credit bureau.  Surely, you would not hold that merchant
responsible for the credit bureau's lack of security, would you? Would you want
that merchant to be able to be (successfully) sued even though a security breach
of the credit bureau resulted in the identity fraud of all the
merchant's customers?
And of course a similar scenario could be possible with credit cards. Why should
the merchant be responsible in such cases when that merchant has pretty much
lost control of how someone downstream is handling sensitive data?

I realize that we are somewhat comparing apples and oranges here, but law
often tends to become become more encompassing than originally intended
because it gets used a precedent in similar cases that are brought to
trial. There
is enough similarity here between what I think you are suggesting and the
example that I've outlined and I'm fairly certain you would not approve of a
merchant being held liable for another service provider's lack of security.
So just be careful of what you wish for as there are likely to be unintended
consequences.

Regards,
-kevin
--
Blog: http://off-the-wall-security.blogspot.com/
"The most likely way for the world to be destroyed, most experts agree,
is by accident. That's where we come in; we're computer professionals.
We *cause* accidents."        -- Nathaniel Borenstein



More information about the cryptography mailing list