NSA and the NCSA/Apache web servers

Note: We were informed by NCSA that the NSA (The US National Security
   Agency - yes, the folks who in 1994 said "we're only 10 years
behind schedule")  considered the hooks to encryption in NCSA's httpd 
   to be in violation of the munitions export law, thereby making its 
   distribution to foreign sites illegal. For various reasons, we decided 
   to remove the -DPEM_AUTH code completely.

This was followed by a pointer to http://www.apache.org/nopgp.html from 
which the following text was taken:

   On May 17th, 1995, we were asked by a representative of NCSA to remove 
   any copies of NCSA httpd prior to 1.4.1 from our web site. They were 
   mandated by the NSA to inform us that redistribution of pre-1.4.1 code 
   violated the same laws that make distributing Phill Zimmerman's PGP 
   package to other countries illegal. There was no encryption in NCSA's
   httpd, only hooks to publicly available libraries of PEM code. By the 
   NSA's rules, even hooks to this type of application is illegal. 

Wow -- hooks to encryption are unexportable -- now THAT's bullshit.  Sheesh.

	-Amir

       /\     Set the controls for the heart of the sun.    -Pink Floyd    
______/  \    ___________ __ __  _  _  _  _   .   .   . axon@neuron.net
          \  / 
           \/    For PGP 2.6 key send mail with subject: SEND PGPKEY

"Amir Y. Rosenblatt" <axon@neuron.net> wrote:
  Wow -- hooks to encryption are unexportable -- now THAT's bullshit.  Sheesh.

A few yuears ago I asked Matt Blaze if he would
publish CFS with the sryptography removed, and he
told me that AT&T's lawyers also believed this to
be true. (So, of course, his answer was "No".)
The hooks are as important as the crypto code.

Interestingly though, Kerberos made it to
Australia (Bond University I think) legally.

Greg Rose               INTERNET: greg_rose@sydney.sterling.com  
Sterling Software       VOICE:  +61-2-9975 4777    FAX:  +61-2-9975 2921
28 Rodborough Rd.       35 0A 79 7D 5E 21 8D 47  E3 53 75 66 AC FB D9 45
French's Forest         co-mod sci.crypt.research
NSW 2086 Australia.     USENIX Director.

> 
>   
> "Amir Y. Rosenblatt" <axon@neuron.net> wrote:
>   Wow -- hooks to encryption are unexportable -- now THAT's bullshit.  Sheesh.
> 
> A few yuears ago I asked Matt Blaze if he would
> publish CFS with the sryptography removed, and he
> told me that AT&T's lawyers also believed this to
> be true. (So, of course, his answer was "No".)
> The hooks are as important as the crypto code.
> 
> Interestingly though, Kerberos made it to
> Australia (Bond University I think) legally.

Actually, neither hooks nor encryption are unexportable, you just need
a license to export them.  I got a license to export an RSA encryption
scheme and a general purpose hook into encryption for integrity toolkit.
It took a few months and was not very difficult, but you have to apply.

-- 
-> See: Info-Sec Heaven at URL http://all.net
Management Analytics - 216-686-0090 - PO Box 1480, Hudson, OH 44236

On Jul 28, 12:18pm, Dr. Frederick B. Cohen wrote:
> Subject: Re: NSA and the NCSA/Apache web servers

> Actually, neither hooks nor encryption are unexportable, you
> just need a license to export them.

>-- End of excerpt from Dr. Frederick B. Cohen

I hope I'm not alone in wondering why on earth this is the case.

Ok, exporting cryptography from the USA is restricted, and highly
controversial. I think there has been something on this one already.

But what is it, in the legal wibble, that make _hooks_ to
cryptography restricted. How have they worded things to make this the
case.

The hooks are of course completely useless in and of themselves. You
can only do anything useful with them if you have the matching crypto
package.

Yours a confused Brit ... who doesn't have this problem ... yet!!

--
___________________________________________________________________

Andrew Meredith
 Senior Systems Engineer           Tel: (direct) +44(0) 1793 545377
  Network Engineering Tools Group  Tel: (main)   +44(0) 1793 541541
   Motorola ECID                   Fax:          +44(0) 1793 420915
    16, Euroway, Blagrove
     Swindon SN5 8YQ, UK       email: Andrew_Meredith@email.mot.com
___________________________________________________________________

...
> > Actually, neither hooks nor encryption are unexportable, you
> > just need a license to export them.
...

> I hope I'm not alone in wondering why on earth this is the case.

They don't want to encourage encryption if they can avoid it.  It
impairs their ability to gather intelligence.

> Ok, exporting cryptography from the USA is restricted, and highly
> controversial. I think there has been something on this one already.
>
> But what is it, in the legal wibble, that make _hooks_ to
> cryptography restricted. How have they worded things to make this the
> case.

Legal? What makes you think so? It hasn't made it to the courts yet
because people in the US aren't willing to risk jail for over their
right to do it.  The only court case I am aware of was the RSA case and
in that one, the courts ruled against the NSA - but in today's political
and economic environment, people who do cryptography don't want to risk
it.

> The hooks are of course completely useless in and of themselves. You
> can only do anything useful with them if you have the matching crypto
> package.

Not really right.  It's very easy to change a compression hook into an
encryption hook using standard off-the-shelf shareware, public domain
software, or commercial products.

> Yours a confused Brit ... who doesn't have this problem ... yet!!

Don't bet on it.  If you really try to export top-flight encryption
technology in a big way, you may find that your government can be
just as opressive as mine.

-- 
-> See:  Info-Sec Heaven using our New Super Secure World-Wide-Web Server
-> Free: Test your system's security (scans deeper than SATAN or ISS!)
---------------------- both at URL: http://all.net ----------------------
-> Read: "Protection and Security on the Information Superhighway"
	 John Wiley and Sons, 1995 ISBN 0-471-11389-1, 320 pp, $24.95
-------------------------------------------------------------------------
   Management Analytics - 216-686-0090 - PO Box 1480, Hudson, OH 44236

On Jul 28,  2:09pm, Dr. Frederick B. Cohen wrote:
> Subject: Re: Hooks to Crypto>
> ...
> > > Actually, neither hooks nor encryption are unexportable, you
> > > just need a license to export them.
> ...
>
> > I hope I'm not alone in wondering why on earth this is the case.
>
> They don't want to encourage encryption if they can avoid it.  It
> impairs their ability to gather intelligence.

I kind of meant "why they think they can" rather than "why they want
to". One can't really help being aware these days what the US
government (and indeed many others) are trying to do.

> Legal? What makes you think so? It hasn't made it to the courts yet

Oh ... I see ... just like the rest of this stuff, only more so.

> because people in the US aren't willing to risk jail for over their
> right to do it.  The only court case I am aware of was the RSA case
> and in that one, the courts ruled against the NSA - but in today's
> political and economic environment, people who do cryptography
> don't want to risk it.

Judging by the PZ case, I can't say as I can really blame them. it
would be better of course if they would go for it, but ...

> > The hooks are of course completely useless in and of themselves.
> > You can only do anything useful with them if you have the
> > matching crypto package.
>
> Not really right.  It's very easy to change a compression hook
> into an encryption hook using standard off-the-shelf shareware,
> public domain software, or commercial products.

I probably should have written "a crypto package" instead of "the
matching crypto package".

This would infer that anything that can cause information to be piped
out to a package and then the result sucked back in would fall into
this category.

Hmmm

map ^Xe :,$! /bin/sh -c 'pgp -feast 2>/dev/tty^V|^V|sleep 4'^M^L

Everyone DELETE VI NOW !!! >;)

> > Yours a confused Brit ... who doesn't have this problem ... yet!!
>
> Don't bet on it.  If you really try to export top-flight encryption
> technology in a big way, you may find that your government can be
> just as opressive as mine.

Our lot tend to work in a different way to yours. Similar end result
of course, but different approach.

So when I said "yet". I was meaning that, at the moment, we have no
laws specifically refering to Cryptography and it's export, but the
approach I can see being taken would be:

1 - Find yourself a tenuous link with some Psycho-Baby-Killer group.
2 - Start a "This must be stopped" campain.
3 - Propose the "Internet Pornography Act"
4 - Shove it through before anyone can get together enough
    opposition to get it squashed. (and that would have to be a
    GREAT DEAL of opposition).

That's how they did the "Criminal Justice Act" which breaks both
European and International law in a great many places. That's how
they'll do the "Internet Pornography Act". It'll be just loose enough
to include just about anything they want it to.

>-- End of excerpt from Dr. Frederick B. Cohen


Andy M

...
> > because people in the US aren't willing to risk jail for over their
> > right to do it.  The only court case I am aware of was the RSA case
> > and in that one, the courts ruled against the NSA - but in today's
> > political and economic environment, people who do cryptography
> > don't want to risk it.
> 
> Judging by the PZ case, I can't say as I can really blame them. it
> would be better of course if they would go for it, but ...

In my case, I just opted to moving my crypto business outside the US. 
This is the real result of the crypto policy.  The US is falling behind
the rest of thew world in crypto R+D.  For example, two good crypto
packages for the Internet have been released in the last few months.  I
was engaged in a similar project in the late 80s but abandoned it
because I couldn't export, so the market would not justify the work. 
Now it is owned by people in EC and Australia who are generous enough to
allow those of us in the US to use them.

Of course, I can't post them in info-sec heaven because even imported
crypto software may not be exported, and I cannot adequately detect the
difference between a foreign person using a US site to get the
information and a legitimate US site getting the information for itself. 
In other words, the policy prevents US firms from having better Internet
resources in the info-sec arena.

...
> This would infer that anything that can cause information to be piped
> out to a package and then the result sucked back in would fall into
> this category.

Right - in other words, nothing can be exported if it produces output
and takes input.  The point is, they want a way to arrest people who are
doing something they don't like.  Philo Zimmerman would almost certainly
win if they ever took him to court, but by harassing him in this more
subtle way, they destroy the impact of PGP in the marketplace, get MIT
to support an official (and perhaps customized for the NSA to have weak
keys) version, and prevent others from following in Phil's footsteps. 
So the strategy works until some brave person risks enough top get past
it.

...
> 1 - Find yourself a tenuous link with some Psycho-Baby-Killer group.
> 2 - Start a "This must be stopped" campain.
> 3 - Propose the "Internet Pornography Act"
> 4 - Shove it through before anyone can get together enough
>     opposition to get it squashed. (and that would have to be a
>     GREAT DEAL of opposition).
> 
> That's how they did the "Criminal Justice Act" which breaks both
> European and International law in a great many places. That's how
> they'll do the "Internet Pornography Act". It'll be just loose enough
> to include just about anything they want it to.

All true, but as the saying goes:

Representative democracy is a terrible form of government,
but every other form of government we know of is even worse.

-- 
-> See:  Info-Sec Heaven using our New Super Secure World-Wide-Web Server
-> Free: Test your system's security (scans deeper than SATAN or ISS!)
---------------------- both at URL: http://all.net ----------------------
-> Read: "Protection and Security on the Information Superhighway"
	 John Wiley and Sons, 1995 ISBN 0-471-11389-1, 320 pp, $24.95
-------------------------------------------------------------------------
   Management Analytics - 216-686-0090 - PO Box 1480, Hudson, OH 44236

On Fri, 28 Jul 1995, Dr. Frederick B. Cohen wrote:

> Philo Zimmerman would almost certainly
> win if they ever took him to court, but by harassing him in this more
> subtle way, they destroy the impact of PGP in the marketplace, get MIT
> to support an official (and perhaps customized for the NSA to have weak
> keys) version,

I've personally pulled apart the innards of both MIT pgp 2.6.2 and the 
non-MIT pgp 2.6.2i in order to generate large primes and full RSA keys.  

There are no hacks in MIT pgp that cause it to generate weak keys.


- Andy

+-------------------------------------------------------------------------+
| Andrew Brown  Internet <asb@nexor.co.uk>  Telephone +44 115 952 0585    |
| PGP (2048/9611055D): 69 AA EF 72 80 7A 63 3A  C0 1F 9F 66 64 02 4C 88   |
+-------------------------------------------------------------------------+

On Fri, 28 Jul 1995, Andrew D Meredith wrote:

> On Jul 28, 12:18pm, Dr. Frederick B. Cohen wrote:
> > Subject: Re: NSA and the NCSA/Apache web servers
> 
> > Actually, neither hooks nor encryption are unexportable, you
> > just need a license to export them.

The answer is to have some non-USA entity build shareable full fledged 
full powered crypto libraries and provide them for free for the rest of 
the world and for all machines.

On Windoze DLL's or WIN32's would be needed, on PPC Mac's shared Lib's, 
on 68K Macs, an INIT that hooks itself via Gestalt, on Unix, shared 
LIB's, etc.  I would also include routines for asking the user for a 
passphrase to prevent the running application from grabbing that password 
and providing weak security.  Or a program running in the background that 
handles all the calls via IAC's or whatever... (like AppleEvents to PGP)

All with full free source, etc and PGP compatibility, etc.  Then the rest 
of us could write code that uses that library.  Whoever wants crypto just 
downloads the library from whereever and uses it.

In this case, would code written in the USA be exportable?  

Wei's library would have been perfect for something like this... too bad. :-(

=================================================================93=======
 + ^ + |  Ray Arachelian | Amerika: The land of the Freeh. | \-_    _-/  |
  \|/  |sunder@escape.com| Where day by day, yet another   |  \  --  /   |
<--+-->|                 | Constitutional right vanishes.  |6 _\-  -/_  6|
  /|\  |    Just Say     |                                 |----\  /---- | 
 + v + | "No" to the NSA!| Jail the censor, not the author!|     \/      |
=======/---------------------------------------------------------VI------/
      /  I watched and weeped as the Exon bill passed, knowing that yet /
     / another freedom vanished before my eyes.  How soon before we see/
    /a full scale dictatorship in the name of decency? While the rest /
   /of_the_world_fights_FOR_freedom,_our_gov'ment_fights_our_freedom_/

On Fri, 28 Jul 1995, Greg ROSE wrote:
> A few yuears ago I asked Matt Blaze if he would
> publish CFS with the sryptography removed, and he
> told me that AT&T's lawyers also believed this to
> be true. (So, of course, his answer was "No".)
> The hooks are as important as the crypto code.
> 
> Interestingly though, Kerberos made it to
> Australia (Bond University I think) legally.

I was the person who put the encryption back into that version of 
kerberos (which is now called eBones).  They removed all encryption calls.

They had actually pulled out all calls to the des routines, so we had a 
'working' authentication system that encrypted nothing.
This version was called Bones (they ran a program called parania over 
Kerberos, and that left Bones :-).  When I left, we had Kerberos working 
but I had not tested against 'true' kerberos.  I belive it has been fixed 
by 'those that have followed' and now fully interoperates with MIT 
kerberos v4.  So the 'international' version of kerberos is fully legal.

BTW I wrote libdes (my DES library) as part of this work.  Luckily I have 
    escaped from Kerberos/eBones when I left Bond Uni but my nights are
    still haunted with memories of trying to follow the code :-).

eric (who is having far more fun putting an SSL package together :-)
--
Eric Young                  | Signature removed since it was generating
AARNet: eay@mincom.oz.au    | more followups that the message contents :-)