> Ian Grigg wrote:
>> 1.  The reason there is a strong dominating player at
>> the moment is because there is no way to compete.
>
> But the reason there's no way to compete is due to whose root certs are
> in the main browsers, not any other reason like branding or lack of it.

Yes, the reason there is no way to compete is because
the root certs are a fixed and hidden bucket of certs.
The solution is to unfix it and to unhide it.  One way
of unhiding it is to brand the CAs.  Another way is to
petname the certs.  A third way is to put the counts
in, and a fourth way is to treat self-signed certs as
simply a cert with a less popular brand.  A fifth way
is to have the user sign off on each cert and attach
logos to them.

All of these things are useful for competition, and for
that reason, the CAs generally actually want these
things (because only competition in a public space
will let them grow the market) ... they actually want
to compete against alternates like self-signed certs
that are not based on a fixed/hidden bucket of root
certs.

But, the real reason for them is security.  Currently,
the reason the SSL browser security model is
breached by phishing is because the cert is hidden
from the user, and not surfaced (as per brand and
other activities).  Competition and monopolies
are side issues, really.

>> 3.  Incumbents don't currently do anything to justify
>> their brands, basically because they don't have to.  If
>> they had to, there would be a shakeup in the marketplace.
>> Those that are small, lean and mean would be much better
>> placed to deal with the new branding issues than those
>> that are large, and encumbered with the baggage of past
>> mistakes.
>
> But there can never be a proper market in certs, can there?

So, why should we keep the current un-proper one?

> If Amazon is
> secured by Verisign, I can say "Well, I don't trust Verisign", but if I
> want to buy from Amazon, I don't have much choice, do I?

I'm not sure where to start!  You have plenty of choice.

If you don't "trust Verisign" you examine the cert that
is presented and you check it is the one that Amazon
uses.  Hopefully your browser will then allow some
relationship building to assist in this.  Bear in mind that
you almost certainly won't be phished the *first* time,
only later on when you have a relationship established
(when it is a valuable relationship...).

The reason you might think you have no choice is
because the browser doesn't give you any choice.
Hence.... phishing.  Solution:  have the browser give
you a choice!

(But, you are right in one facet:  x.509 doesn't support
trust as we humans know it, which is with multiple vectors.
But, the browser doesn't have to follow that.)

>> I don't see the problem.  The user sees that they
>> are different sites.  She analyses the sites, and
>> if they are bona fide, enters different pet names
>> and carries on.  When she gets phished over to
>> https://www.barclaysbank.com/ and finds that the
>> site might not be right, she doesn't petname it.
>
> But surely the point about phishing is that https://www.barclaysbank.com
> looks totally genuine even if it's not.

RIght.  So the user has to determine whether it is.

Verisign isn't going to help, because there is no
cert in use.  In fact the only thing that will help is
that there is no cert in use;   but right now, the
brower hides the presence or otherwise of the
cert, down in the lower left corner where nobody
notices.

(Note that this prescription of fixes also will lead
to much more cert usage ... which will allow Verisign
to help, *iff* it can rely on its brand being obvious.)

> I assumed the point about petnames is that the user "goes to their
> bank", but the petname doesn't appear, and they go "Huh?". Was I wrong?

Yes, that is the first defence for the immediate
problem.  But what we have to do is to migrate
all important sites across to SSL, and to place
the browser in control of the relationship, so
that the information tied to Verisign and the cert
and all the other personal information can be
integrated.

>> to hide all the real information.  So the problem
>> with the above is to stop assuming that they are
>> valid, and to insist that the user authenticates
>> them in some fashion or other.
>
> Using what evidence? Inspecting the website? The only sensible thing I
> can think of is comparing the URL with one printed in e.g. a magazine.

If you open a bank account online then generally,
the bank takes some care to establish the
relationship.  Maybe it's the URL, maybe it is a
token that is sent in the mail, or text'd to their cell.
Personally, I'd rather see the fingerprint printed
on statements, but that might be a ways in the
future.

If you are just shopping at Amazon, then there
would be less of a burden, because it matters less,
but Amazon still has a way of reaching people,
and generally, they have no difficulty in finding
excuses to try.

Consumers and merchants do use a little bit of
common sense when establishing a relationship;
they just need to be given the tools to do that.

> But who would bother doing that?

Anybody who's been phished!

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

Reply via email to