Nelson B wrote:
> Ben,  (and others who think SSH is the answer):

Nelson, I wonder if you misunderstand what people
are saying.  I know this is a tough ask, most people
who have spent much of the last decade working on
SSL security systems also take any criticism of the
whole SSL thing as an unwarranted attack.

Nobody is saying that "SSH is *the* answer."

And, nobody is saying that the CA system be disabled,
or dismantled or disposed of.  Nobody is suggesting
that the lines of code that people have slaved
over are to be wasted or thrown away.  Also, nobody
is saying "we should get rid of all the CAs."

There is no change to the crypto protocols mooted,
no reduction in the information available to the
user, and no reduction in their choices.

What instead people *are* saying is that CAs aren't
the only way to do security, and that SSH offers a
successful example of another way.  What is being
proposed is not less CA certs, but more CA certs.
CAs are needed by some people, but not by all people.
There shouldbe more information for users describing
the real details on the cert, rather than a bland
and tiny lock symbol that we all agree is ignored.


> Live with what you propose.  No cheating.  Then after 90 days, come back
> and tell me you never needed any CAs to do anything.  Tell us how many
> calls you had to make to get fingerprints.  Tell us how many times you
> wanted to visit a web site but found that the cert was untrusted.
> Tell us how much better that was than with CAs.


People do this all the time over HTTP.


...
> When you and numerous others can honestly tell us that's better, THEN
> maybe mozilla should start to consider that approach.


This is why I keep my ear to the ground for any data
about MITMs.  There is very little.  There is the
one story I heard on this group, relating to a
credit card on a student campus, and then there
a few stories from other protocol areas (one email
story, and one other, can't recall right now).

This allows me to claim - honestly - that MITM is
not the threat that you think it is.  I can't prove
it because there is an absence of information.  But,
I can show - honestly - that credit cards do not in
general get stolen over HTTP(S), because the card
issuers know of few or no cases, even though there
are large numbers of merchants that use HTTP and
email for transmission of credit card info.  They
get stolen by hackers and inside thefts, because
crooks can count and they understand the risks.

Now, can you honestly tell us that phishing is not
a threat, and Mozilla shouldn't respond to it?

Phishing is a direct attack on the browser's security
model, and it breaches it very nicely.  Your favourite
bank has been attacked by phishing (if it is of any
size and is in the USA, that is, you can look it up
on the antiphishing.org site).  Your favourite bank
is suffering because what was supposed to be secure,
isn't.

Phishing is addressed by the sort of measures we
have proposed, which you think to be bad.  So,
honestly, what do have to do to get Mozilla to
start considering improving security to cope with
the threats that are today breaching the security
model?


> In the meantime, the issue before mozilla foundation is to choose new
> CAs for admission to the list of trusted CAs.


> I'm all for open discussions, in the newsgroups, where they belong.


We're ok, then.  The point of discussing the CA
admission policy within the realms of a wider
security discussion is that without an understanding
of the "one size fits all" CA bug, any policy is
... as good as the next.

If however, the "one size fits all" issue can
included in the discussion, then a policy can
be created that a) makes sense in the face of
the bug, and b) aims at a suitable point when
the bug can be addressed in the security model.

(All of which can be done without disabling certs.)


iang
_______________________________________________
mozilla-crypto mailing list
[EMAIL PROTECTED]
http://mail.mozilla.org/listinfo/mozilla-crypto

Reply via email to