NetNews Usenet Archive 1992 #26/NN_1992_26.iso/spool/sci/crypt/4735

Internet Message Format | 1992-11-14 | 5.8 KB

Xref: sparky sci.crypt:4735 alt.privacy:2240 comp.org.eff.talk:7007
Path: sparky!uunet!know!mips2!news.bbn.com!noc.near.net!news.Brown.EDU!qt.cs.utexas.edu!yale.edu!jvnc.net!darwin.sura.net!zaphod.mps.ohio-state.edu!cs.utexas.edu!sun-barr!ames!agate!linus!linus.mitre.org!bistromath.mitre.org!ptrei
From: [e-mail] (Peter Trei)
Newsgroups: sci.crypt,alt.privacy,comp.org.eff.talk
Subject: Re: A Silver Bullet to Limit Crypto?
Message-ID: <[e-mail]>
Date: 13 Nov 92 18:27:30 GMT
References: <[e-mail]> <[e-mail]> <[e-mail]>
Sender: [e-mail] (News Service)
Organization: The MITRE Corporation
Lines: 119
Nntp-Posting-Host: bistromath.mitre.org


     We've really got two separate problems:

     1. The availability of technology which makes communications
surveilance (CS) increasingly cheap, easy, and undetectable. Can we
ensure that it will not be misused, by governments and others?

     2. The wide spread use of strong cryptography (SC) and it's
enabling technology (personal computers). Can we ensure that things
which should not be hidden are not?

     These should be viewed against the background of the increasing
shift towards digital forms of communications and data storage.



     The Communications Surveilance (CS) problem:

     In the past, CS could be and was performed, both legally and
illegally. However, it was sufficiently difficult and resource
intensive that it was restricted to valuable targets.  The average
citizen had no reason to worry about having his or her phone calls
watched for evidence of adultery (still a crime here in
Massachusetts).

     The new technologies change all that - they are an enabler for
casual intrusion and snooping. Digital text can be trivially searched
on the fly, and it is often rumored that overseas phone lines can be
searched by voice recognition (VR) systems (I have no idea if this is
true - I think Bamford talks about it). Even if VR systems are not
practical today, we must assume that tomorrow they will be.

     Many people have little or no faith in procedural restraints on
governmental CS; there have been too many cases of illegal wiretaps.
In this group, Phil Karn has stated that he was told that the
government is resisting digitally encrypted cellphones, despite the
possibility of tapping the landline part of a call, because "A
warrant would be required to gain the telephone company's assistance,
and warrants were considered to be "too impractical" in many cases."

     If this is true, it appears that those who lack faith in the
government have a case.  Even if todays government were composed of
saints and angels, they worry that a future government may not be so
scrupulous.

     The upshot of this is that there is a defensible position that:

"Not only should the government not tap casually without a warrant; it
should not *be* *able* to tap casually, period. Tapping should remain
difficult and expensive enough that it will only be used on targets of
high perceived value. Procedural barriers against improper tapping are
not, and will never be, adequate - they can be eliminated at the
stroke of a pen, or simply ignored."



    The Strong Cryptography problem:

    Strong cryptography (SC) and its relatives (digital signatures,
zero knowledge proofs, et al) are a two edged sword. On one hand,
they can be used as a tool for message security:

* That it comes from whom it says it does.
* That it has not been tampered with.
* That no one but the but the sender and intended recipient can read it.

    Most of the readers of this group are aware of the potential
services requiring these attributes. SC *WILL* be widely used in the
near future for a myriad of purposes.

    Many of the services proposed or already in use (eg, Kerberos)
make use of "session keys", which are generated on the fly and used
for brief periods.

    On the other hand, if a document is encrypted with SC, and you
can't obtain the key, you are SOL, regardless of how legitimate is
your need to read it. In an earlier post, you made an analogy to
fencing off a part of your land, and declaring that no one, not even
the government, could enter it without your permission. You felt that
this was absurd - no government could or would tolerate such a
situation.

    They may not have a choice. SC is something new under the sun: as
far as we know, it IS strong. It's discovery within the mathematics
may have fundamentally changed what we can hope to know about a
non-cooperating system or person.  To complicate matters, SC is
*easy*. The software can be developed and run on any personal
computer.

     A blanket ban on the use of SC is utterly impractical - it would
turn the US into an information backwater, as the rest of the world
avoids using our markets, networks, and systems due to their lack of
security, and of services predicated on SC. The US is already losing
in the world marketplace due to the pathetic ban on DES software
exports.

     Requiring key registration is almost as bad. Much of the world
information traffic would again avoid the US due to the hassle it
would present, doubts about the trustworthyness of the key registry,
and the lack of services requiring briefly held session keys.

     My feeling is that the cat is out of the bag. Strong cryptography
is out there, it can and will be used, and it's use cannot be
prevented.  We might be able to check it's use in the US, but doing so
will cripple the US as an economic power.  In the future, law
enforcment organizations will have to rely on other ways to obtain
information, whether they like it or not.

     Ms. Denning: you are designing stable doors, long after the horse
has bolted.

                            Peter Trei
                            [e-mail]


Disclaimers:

1. I do not have access to non-public information concerning communications
   surveiliance.