Re: key recovery options for PKIX-3 CAs?

[Stephen Kent <kent@xxxxxxx>]

Tony,

	I do not question the plausibility of the scenario you described;
it'salways useful to examine possibe worst case scenarios in evaluating
systems.  Your point, and that of some others who object to making CA key
gen a mandatory option, is that this would significantly facilitate
government imposition of the controls you describe.  However, since the CA
key gen feature does NOT call for the CA to retain the private key, the
example is flawed.  That is, a compliant CA would not have to retain the
private key it generated (nor an ability to regenerate it).  So, licensing
PKIX part 3 compliant CAs, the year 2000 event in your example, does not
achieve the effect you cited.  To achieve the predicted effect, vendors
would have to choose to retain this data, a design "feature" outside the
scope of the standard.  In that case, your wrath should be directed against
any vendor who chooses to take this approach.  However, if we predict
vendors adopting this tact, wny not assume that they will do so
irrespective of the standard?  This seems to be a typical slippery slope.

	I'll assume that the smoke-filled room allusion was in reference to
something other than the deliberations taking place on this list. Besides,
I don't smoke and always request non-smoking seating in restaurants,
airplanes, etc. :-).

Steve