Julien Pierre wrote:

I have left a lot of your posts unanswered as I could have gone on for ages, but I couldn't let this one slip.

What is your criteria for comparing the respective success of SSH and SSL ? I bet the later gets used by a heck of a lot more people every day, and it is supported by many more programs.


This is a very good question.  My original
stab was here:

http://iang.org/ssl/how_effective.html

which suggested that HTTPS was not successful,
at least by the one single measure of how
many HTTP servers sought to use SSL (about
1%).  This is one area where SSH, for example,
shines, as it dominates the alternate of
telnet within the Unix world.

Of course, many other measures are possible.
Design wins, mindspace, etc, are areas where
SSL is successful over SSH.  Ease of use at
the implementation and setup level is one area
where SSH dominates over SSL.

A particularly interesting area is to measure
how successful these protocols are against
real attacks, which of course is fraught
with complications, as real attacks are hard
to research.

Then, as you suggest, how many usages (per day)
and how many users - that I simply don't know,
and it would take some ISP or router stats to
work that out.  E.g., what's the relative
count of sessions on ports 22, 443, although
there are many complications in those sorts
of simple counts.


You are free to use any proprietary security protocol you want, including a home built-one or double ROT13 . Nobody is forcing you to use SSL/TLS for your own applications.


As of a year ago, a lot of people would have
disagreed with that.  In fact, in mid 2003, a
group of allegedly respectable crypto people
hounded Tom (of libtomcrypt fame) for presenting
an alternate to SSL, on more or less the grounds
that it wasn't SSL.

Eventually, to quell the complaints, Tom withdrew
the protocol.  The crypto field is the poorer for
this, as it is a fairly consistent observation
that SSL/TLS libraries are not the easiest things
to use.

However, Tom's bruising may have had benefits, as
those same people are now rethinking the notions
of "if it's not SSL, it's evil" [1].  Especially as
the whole phishing thing arises, and as the stats
on low HTTPS cert usage sink in.


Obviously most businesses and financial institutions don't agree with your approach, as they prefer to use the open and very popular SSL/TLS protocol, which is supported by a host of programs from many sources, commercial and free, rather than reinvent the wheel and start from scratch as you advocate.


Like I say, there's no fun in singing with
the choir :-)  Frankly, those businesses
and financial institutions either don't know
whether they are secure or not, and when they
find out that they are insecure, they are
indeed forced to reinvent the wheel.  There
is an awful lot of circular engineering
going on, it's just not going on in terms
of "replace SSL with LSS."  More, it's
going on in terms of examining the entire
security and threat space.

One of the often stated goals of the NSS library is to support this SSL protocol, for the now-defunct Netscape client, the Mozilla client, some other AOL clients, the Sun/iPlanet servers, and incidently anyone else who wants a free source implementation of that protocol.

Supporting proprietary security protocols is not one of the stated goals of NSS, or of Mozilla .


Perhaps I should have clarified - this entire
thread was not necessarily oriented to the
NSS library, which I agree has more strict
goals that might be closer to "just SSL
support." [2]

I also am mostly interested in improving the
security of Mozilla in the face of threats
that are documented, costly, and continuous,
but are otherwise ignored by the crypto people
everywhere.

I.e., phishing.  To that end, with some minor
tweaks of HTTPS and the way Mozilla deals
with users, I think it is possible to reduce
phishing to a dull roar.  I don't know any
other way to address it though, without
looking squarely at HTTPS and the browser.

iang



[1]  There is a wider debate going on within
the crypto field as to the way in which crypto
people - cryptographers and cryptoplumbers -
should interface with the apps world.  As there
is a wider debate as to whether we as a security
field know what we are doing...

In some senses, there are claims of success in
the past;  I find those claims to be without
much founding, and where the claims are examined,
they tend to be based on the hype and press of
corporate campaigns, and some expert's opinion,
as taught by another expert, which latter expert
go from a book, which has since been recanted.

Rather than hard evidence and/or measurable
protection.

So, the debate is also trying to address what
it was that really worked in the past.  And,
it's a surprisingly difficult question, and
also an uncomfortable one, as more and more,
it seems to suggest that SSH worked, and HTTPS
did not.

[2]  Hence, the answer to the original question
in the Subject line is probably "not here,
sorry".


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

Reply via email to