"The secret code is 00000000"

[Steven M. Bellovin]

Although I'm not a professional in the nuclear weapons field, I have 
done a fair amount of research on the subject of PALs (Permissive 
Action Links); you can find a summary of my results at
http://www.research.att.com/~smb/nsam-160/pal.html.  I'll be updating 
the page soon, to add more information I've received via FOIA and from 
published papers I've seen more recently, but none of the updates will 
change any of the discussion below.  (The parent directory of my Web 
page discusses the possible relationship of PALs to the history of 
public key cryptography.)

The first thing to realize is that Blair's short note, on the 00000000 
launch code, confuses PALs with "use control systems".  A PAL is 
integral to the weapon itself; a use control system regulates the 
launch vehicle.  This is important because PALs are within the security 
perimeter -- the tamper-resistant barrier -- of the bomb; if a bomb 
were detached from a missile, the missile couldn't be launched 
without bypassing the use control system, but the bomb could be 
detonated if other safety mechanisms were bypassed.  These safety 
mechanisms are designed to reflect "human intent" and proper 
environmental conditions -- a missile-launched bomb, for example, 
should experience a period of high acceleration, then free fall, then 
decelaration and heat.  But the inputs from these sources are outside 
the barrier, and hence could presumably be spoofed.  For various 
reasons, I do not think these signals are cryptographically protected, 
though I've seen some indications (not yet on the Web page) that the 
human intent signal might be.  (I should note that Blair is very well 
respected in this field; I cite one of his books in my bibliography.  
I'm frankly a bit puzzled by his note.)

The second critical point about PALs is that they were *not* intended 
to guard against what I'll call the "Dr. Strangelove scenario".  While 
there certainly was tension between parts of the military and the civilian 
authorities -- Curtis LeMay did his best to provoke World War III, and 
his successor as the head of the Strategic Air Command, Thomas Power, 
was described as worrisomely unstable by his own colleagues -- 
preventing such misbehavior was not the primary goal of PALs.  If 
nothing else, President Kennedy never could have sold the idea to 
Congress under those circumstances.  Instead, the problem was to 
retain U.S. control of nuclear weapons that were physically in the 
hands of our allies; furthermore, there was a desire to permit forward 
deployment of tactical nukes in West Germany, in positions that were at 
high risk of being overrun in the early stages of a Warsaw Pact 
invasion.  The former was politically vital: not only was Congress 
concerned about our allies (France was seen as politically unstable; 
one of their own nominal nuclear tests was, in fact, scuttling a bomb 
before the rogue generals in the Algerian campaign could get hold of 
it), but German access to nuclear weapons was *extremely* threatening 
to the Soviets.  (Think of Tom Lehrer's line in "MLF Lullaby":

	Heil--hail--the Wehrmacht, I mean the Bundeswehr

and recall that this was less than 20 years after the end of World War II.)

The Pentagon, on the other hand, was attracted by the forward 
deployment feature.  It was geographically obvious that the West German 
frontier was indefensible against a massive armored invasion from the 
east, but it was politically impossible to state that or to act as if the 
real plan was to fall back to the Fulda Gap.  The only solution seen 
was tactical nuclear weapons (which gave us such charming things as 
nuclear artillery shells and backpack-carried nuclear land mines).  But 
these had to be deployed with the forward units, which might easily be 
overrun.  Worse yet, a junior officer might use a nuke without 
authorization, out of desperation.  PALs solved that problem, too -- 
the devices couldn't be used without the unlock codes, either by our 
forces or by the Soviets.

To sum up (this note is already far too long; see my web page for 
details and bibliographic citations), the threat that PALs were 
intended to deal with was physical capture of the devices; it had 
nothing to do with our own launch officers.  The Pentagon was very 
worried about PAL or procedural malfunctions preventing use of nuclear 
weapons (command and control of the military during a nuclear war --
including ending the war! -- is a subject that has received a great
deal of study; there's a vast literature on it); given that, I'm not 
particularly surprised by the 00000000 code.  Blair's 1977 article on 
the physical risk to our missile silos illustrated that there was a 
capture risk to them, too; official reaction was apparently swift and 
(at least partially) appropriate.  Blair got it right; the Pentagon had 
been wrong.


		--Steve Bellovin, http://www.research.att.com/~smb


---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at metzdowd.com