Nelson B wrote:
...
The user who receives the cert presentation relies as much on the
accuracy of the cert revocation information after the issuance as
on the accuracy of the information placed in the cert itself.

Hence both aspects of CA operation (pre and post issuance) are worthy
criteria for CA selection, IMO.

Nelson,


you are asking CAs to do all this work and cover all
these risks, but, have you shown that the risks are
worth covering?

I.e., how frequent are they?  Has anyone ever lost any
money because of a failure in the revocation process?

If this were a real-world problem, and losses could
be shown, we'd be able to say something like "as the
probability of loss is 1% and the likely loss is
$1000, it makes some sort of sense to spend up to
$10 on an insurance policy."

As the CAs are going to spend money on this (as
they are "forced to" by selection criteria) and as
that money then gets extracted from the purchasers
of CA-signed certs, it behoves to show that the
money is well spent.

For example, imagine that total losses for revocation
failures were $10,000 in any one year.  But, fixing
those losses cost $10 per cert.  As there are some
41,450 signed cert years every year, that means that
the user base has to pay $414,500 to cover $10,000
in losses.

That's not a good deal.  I would say that if that
was the best that could be done, then revocation
losses should *not* be fixed in the protocols and
in the CA mix.

Hence, how likely is all this and how much has been
lost?  Is the situation bad enough such that it is
worthwhile to beat up on the CAs and get them to lift
their game?

Crypto protocols are great things because they can
reduce and sometimes eliminate risks without cost.
But, the moment, we step outside the assumption of
no cost, and ask others to do stuff, at their cost,
we have to be very careful not to create an edifice
that loses more money than it saves.

iang

PS: http://www.securityspace.com/s_survey/sdata/200401/certca.html
_______________________________________________
mozilla-crypto mailing list
[EMAIL PROTECTED]
http://mail.mozilla.org/listinfo/mozilla-crypto

Reply via email to