Re: [TLS] Eleven out of every ten SSL certs aren't valid

["Steingruebl, Andy" <[e-mail]>]

Peter said: (combining two different posts)

> his is the same thing, the figure is very meaningful
> because it supports the analysis of the effects of externalities on web site
> security done by some guys at Microsoft Research a few months ago which
> pointed out that since certificate warnings are close to one hundred percent
> false positives (they were aware of no known cases in which someone had been
> saved from being phished by a cert warning),

>  "So Long, And No Thanks for the Externalities: The Rational Rejection of
>  Security Advice by Users", Cormac Herley, Proceedings of the 2009 New
>  Security Paradigms Workshop (NSPW.09), September 2009, p.133.

The problem of course is that this is purely anecdotal.  Cormac didn't cite any research pointers for that, and it is at the heart of the current debate.   If we can actually check for users receiving self-signed certificate warnings when browsing the ordinary web (and perhaps not their corporate intranets) and get some real data about how frequently users actually get these warnings, and/or get MITM'd, then we'd really be getting somewhere.

Frankly, if you think most certificate warnings are false positives, and you also believe that they will forever be such, then why do any certificate verification at all?   This line of reasoning essentially says "there aren't really any active MITM attackers, and won't ever be" so let's just do away with the whole mess of ever checking certificates.

If you want to stop not just attacks today, but attacks that active attackers move to, you need to actually worry about the real numbers here, and the potential for abuse. This does argue for a change in client behavior, but not perhaps in the way some people have suggested - removing all warnings about self-signed certificates.  I don't think that helps us long term.

--
Andy