Gervase Markham wrote:
Ian G wrote:
Good point. So all ISPs can sniff on traffic. Now,
the question is, why have ISPs had a very low incidence
of snooping and eavesdropping? You'd think that by now
there would have been dozens even hundreds of cases of
such? After all, we know there is a non-trivial amount
of credit card traffic going over HTTP, and ISPs are
ideally placed to do perfect DNS attacks.
I've heard of about one, maybe two if we push it. I
think the reason is that your average ISP is staffed
with the wrong sort of person to do insider attacks,
whereas banks, telcos, and other places have no such
good luck.
It's interesting to contrast this view of ISPs with your view of CAs,
which is almost entirely the opposite... :-)
LOL... I guess it may look that way, but the process is
the same: both are participants in the game of players,
and both should be modelled as economic agents with risks
and desires for dosh.
Statistically, we should have seen evidence of massive
fraud by now inside ISPs. Statistically, we should have
seen the same thing in CAs.
But in actual observable reality, we've only seen isolated
events that show a) it's easy to do and b) they aren't
doing it.
What that leads us to conclude (and I'm not the only one
that is thinking this) is that in neither case is there an
economic model for doing these things. Neither for the
CAs nor for the ISPs is there money in them breaking the
rules and futzing with the customers by spying, MITMing,
spoofing, identity theft, etc etc.
Now herein lies a key difference: CAs can issue false
certs which permit undetectable MITMs. ISPs are still
limited to crunching 40 bit crypto (rare, and not so
much of a danger but an interesting thought experiment
hence my claim of it being good enough for banking) or
by eavesdropping on sexchat on AIM or whatever takes
their fancy. Harder to make money there.... But ISPs
have got access to the traffic, which CAs don't really
have.
CAs combined with ISPs would be much more dangerous.
Having them separated means that any CA or ISP that
were to go rogue would still need a partner, and one of
our favourite anti-fraud techniques is to force frauds
into multi-player space. Once we have a conspiracy,
we know that it will unravel soon enough...
But what happens when a CA becomes an ISP? What happens
when a CA indicates it is in the business of helping
people to track, trace, spy, eavesdrop, spoof and MITM?
Now that's crossing a line called _conflict of interest_,
another one of our favourite anti-fraud techniques...
http://www.financialcryptography.com/mt/archives/000206.html
iang
--
News and views on what matters in finance+crypto:
http://financialcryptography.com/
_______________________________________________
mozilla-crypto mailing list
[email protected]
http://mail.mozilla.org/listinfo/mozilla-crypto
