Why "carnivore" type systems can't be (entirely) open source

[Brian Gladman]

From: "Owen Lewis" <oml@eloka.demon.co.uk>
To: <ukcrypto@chiark.greenend.org.uk>
Sent: Thursday, February 08, 2001 12:30 PM
Subject: Re: Why "carnivore" type systems can't be (entirely) open source


> > You cannot know how fast or how slow others have been in finding the
> > weaknesses in the GSM cryptography.
>
> As you will know, expert prognoses of that type are continually made by
> those competent to do so. Sometimes, when they have prognosticated
> incorrectly, that failure becomes public knowledge some considerable time
> after the event. The public have never (yet) got to hear of their
successes.
> In the case of GSM, the cipher + STO held up for very nearly the required
> time; close enough not to matter and to be counted a success, IMO, given
the
> limited purpose for which it was only intended.

My contention is that you do not know the minimum time that it 'held up'
for.  All that you know is the maximum time - that is the time at which the
weakness became public knowledge.

> > It was the communications functionality that that made GSM a success. It
> > seems to me extremely unlikely that its security (or lack of it) had any
> > significant impact on its success one way or the other.
>
> Which aspects of functionality had you in mind?

Cost effective and convenient mobile communications available within a wide
geographic scope (from a single handset?).

   Brian