🚀 go-pugleaf

In article <6880@transfer.stratus.com>, cme@ellisun.sw.stratus.com (Carl Ellison) writes:
|> All this talk about total isolation and how the enemy can translate your
|> public keys given that all communication is trapped is a fine parlor game,
|> perhaps, but hardly relevant.
|> 
|> The criteria for trusted systems pay quite a lot of attention (perhaps more
|> than is deserved) to detecting and preventing covert channels.  It's a
|> *very* hard problem -- and that assumes that you have total ownership of
|> the computer system and all you're trying to prevent is communication from
|> within the one system -- and you're allowed to announce that you've forbidden
|> access -- and the user isn't allowed to complain.
|> 
|> To achieve complete spoofing of the kind being discussed here, it's not a
|> computer system for which covert channels are being prevented but a social
|> system and the enemy has to achieve it without letting anyone know that the
|> censor/intruder is active -- in a society in which free speech is a guaranteed
|> right, no less.
|> 
|> How about taking this discussion to alt.tv.mission.impossible?


While in many cases, it is possible to detect the fact that
all communications are being trapped, that is not the case if
the computer in question is being used in a dedicated applications
environment with non-computer-literate users or worse still in an
isolated location with no on-site operators, but only a security guard 
at the door.

For example, a commercial system at an oil-drilling site or a defense radar
system at an isolated radar site could easily be vulnerable to complete 
trapping and spoofing of communications.  Even a bank in downtown Manhattan
could be vulnerable, if no one in the branch was familiar with the signs
that spoofing is happening!  The common attitude is that if the computer
says so, then it must be right.  

This is a people problem, not a mathematical
problem, but encryption must be used in the real world, not in an ivory
tower.  That is why keys need to get changed and procedures are very critical.
The British would likely not have broken Enigma during WWII, if the Germans
had used the system in its theoretically best ways.  Most of the breaks
came from bad procedures and lazy crypto operators.