What's really in PGP 5.5?

-----BEGIN PGP SIGNED MESSAGE-----

Ryan Anderson wrote:
>>What you call "fragility" is properly called "security".  Would you
>>describe 128-bit keys as more "fragile" than 40-bit keys?  Why is PGP,
>>Inc. inventing propaganda terms for the authorities?
>
>Okay, call it security but the point is, if you're protecting documents vital 
>to your company using encryption (say the design of a new product) and the 
>person who knows the passphrase dies, you've just lost a great deal of money.
>
>PGP for Business Security gives the business a way to have a backup key that 
>can read that person's information.  Frankly, it might not be able to read 
>anybody elses, from the description given.
>
>This is a feature that any business (that understands encryption) will want.
>
>They can easily store the secret keys of the other id someplace secure, and 
>never retrieve them until someone dies.  Disavow knowledge of them to 
>government, etc.  heck, the keys can even be in the primary person's 
>posession.   (Perhaps stored in a safe-deposit box, or without a passphrase, 
>etc.)

Decrypting files and decrypting messages are not the same problem.
The PGP product has SMTP support - it is explicitly designed to weaken
transmitted messages.  Just like Clipper.

[Jon Callas wrote:]
>>It is fascinating to me that every example you use does not involve
>>decrypting transmitted messages.  Yet, that is the feature which is
>>under discussion.
>
>Amazingly, he gave an example where, had encryption been used, the
>project would have stopped, and restarted because the person with the
>keys was incapacitated.  Are you just being combative here?

No.  I am making the distinction between files and messages.

It is quite easy to use a key for encrypting files which is different
from encrypting your messages.  The key for encrypting files can then
given to the appropriate company officer.  It has been possible to use
PGP in this mode forever.

What is new is the feature set which will make it very easy to impose
GAK on the customers of PGP.  According to Mr. Schneier, the code is
just sitting there in all their products waiting to be compiled in.

>>The demand for the ability to decrypt encrypted messages in the
>>corporate environment can easily be measured with this test: how
>>many companies have a policy that requires employees to record all
>>outgoing mail?
>
>Well, any company giving stock advice (and governed by SEC rules on
>stock tips, etc.) is already require to have all outgoing mail
>approved (e-mail and snail), so does it matter if they record it or
>not?

The answer to the question I posed is "No".  The example you cite
requires the company to review message before transmission, not
afterward.

After transmission, the recipient of the message can submit it as
evidence to the SEC should there be an impropriety.  This is no
different from the way it works now.

There is probably a context issue here.  It sounds like you do not
share the context within which many people on this list are operating.

The Big Problem for PGP, Inc. right now is that in 18 months or so
their products may be illegal to use within the United States.  While
many of the rank and file in the company may be willing to accept
corporate dissolution under those circumstances, the board, the
investors, and the senior management are unlikely to welcome that
option.

So, they are preparing for that situation.  They are releasing
products with GAK support right now.  By the time it becomes mandated
by law, it will be easy to compile the well tested features into all
of the products, or just discontinue their line of privacy products
and sell only their Big Brother Inside products.

The corporate key becomes the government key with the stroke of a pen.

PGP, Inc. is betting that they'll win either way.  If GAK goes down in
flames again, they can still trade on their reputation for being The
Good Guys.  If GAK becomes mandated by law: "We didn't want to do it,
but we had no choice."

Some of us may be in the habit of thinking of PGP, Inc. as the hero.
However, the problem with organizations is that they are easily
corrupted.

Does this mean everybody in the organization is now working for the
government?  No, it does not.  Most likely there are many PGP, Inc.
employees who are feeling uncomfortable with this new development but
they are being given the same lines we are getting.  Many of these
people are probably getting an additional "inside scoop" from their
leaders that explains the "real" strategy.  Hopefully, the rank and
file at PGP knows that people you like and see in the hallways every
day will still lie to you if they think their career depends on it.

Monty Cantsin
Editor in Chief
Smile Magazine
http://www.neoism.org/squares/smile_index.html
http://www.neoism.org/squares/cantsin_10.html

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQEVAwUBNDuw65aWtjSmRH/5AQGwLgf7BBYy2HO30eDXO2YLqkgtuOxAcnQU19AO
DOKIa3jwU723ICgWD00dPW1GegKL5LaRfp3a+Xtd+qjHaFs0IlXr4kJGv1wiCJhu
sRKxmSi3g1aGIfiF48uMef5bbzPL27b3u78rJds+E44LP9zEnieBcACBp+3977W0
0N0k7ADY6ME+zmc4f5D5aLdJa8gprOCPj1nF6esspkMuDDu5iH/YY/GT6VprRCub
NojEXPS/uB6tgTS4vDfenL8UWUUqjXQof8MrRoMAG6dsjvNEyU5WlhbnDgyh9nF5
DbuP9+oA8DHYSmEGzTxD4oG7mGwb+NMMuYOyg3xC1H8ou2Ajv0kfOg==
=UtlQ
-----END PGP SIGNATURE-----

-----BEGIN PGP SIGNED MESSAGE-----

Anonymous, claiming to be Monty Cantsin, writes about PGP 5.5:
>
>Decrypting files and decrypting messages are not the same problem.
>The PGP product has SMTP support - it is explicitly designed to weaken
>transmitted messages.  Just like Clipper.
>
No, it isn't "just like Clipper." Messages encrypted with PGP 5.5
can be decrypted or verified by PGP 5.0 or other implementations that
can decode the PGP message and encryption format. Also, PGP 5.0 or other
complient implementations can send encrypted and signed messages to
PGP 5.5 users.

The transmitted message is as "weak" as the quality of the encryption,
the number of people who have access to the secret key, and the
quality of their passwords. It is also as "weak" as the physical
security that prevents passwords from leaking (i.e., by Tempest
or black bag wiretaps inside the secure user's workstation).

Given all of the other risks, encrypting to a corporate private
key doesn't seem to me to significantly increase the risk.

The real risk, to my view, is that some future implementation will
require the secondary key, and will require that that secondary
key be stored in a "government accessable" database. This, however,
seems a bit remote and, given that PGP publishes their source code,
reasonably easy to detect.

Martin Minow
minow@apple.com


-----BEGIN PGP SIGNATURE-----
Version: PGP for Personal Privacy 5.0
Charset: noconv

iQCVAwUBND0KUG23+ciinrc5AQGvMgP9HKE4s/3FbjfFpvfC230wvHkjDd2HWyC1
lhVa/i9ayO8pZC/k+V06pGTZLcZ89a4X2r3fQGjj0QNghuotkV9xVVD9AzLTLX88
YPp2DQWDCjkWq4PzBB8IR6c+rH6AbuqtDAhfas/Rto/9DI7EJWqi3dZbi7tFm7jj
g2nTAFk9VVo=
=DkbK
-----END PGP SIGNATURE-----