Re: [pkng] More thoughts: Privacy

I agree that whatever PKNG does should support privacy to
the extent possible. (Hence the "selective field
confidentiality" item in my looong wish-list earlier.)

Its not clear that that'll be tremendously useful in that
many cases, but I can at least think of some. For example,
if a cert-like blah-thing were to allow me to encrypt any
field, and if I had tls-extractor stuff working, then a
TLS client could send the blah-thing with some fields
encrypted in a TLS handshake (assuming that we can squish
the blah-thing into a TLS handshake) and later on send
the decryption key to the TLS server encrypted under
a key derived from the TLS extractor stuff. (Or I guess
if we want to imagine messing more with TLS, then
something derived from the pre-master secret could have
been used by the client to encrypt parts of the blah-thing,
and a TLS ClientHello extension could signal that to
the TLS server.)

A side-effect there would be that it would then be safe
to send the blah-thing in the initial handshake, and so
there'd be no need for TLS renegotiation, and so we'd
avoid the currently topical TLS renegotiation bug.

S.

Massimiliano Pala wrote:
> Hi Peter,
> 
> I understand that is a matter of Policy. Unfortunately the user has no
> control over what the provider of the OCSP service does with the data
> that might track one's activities (and potentially location) throughout
> the day.
> 
> Some organizations like government agencies might overlook this problem
> and leak data that is supposed to be kept secure. The OCSP is an example,
> but my point was that if we can manage to introduce a
> user-machine-enforceable
> (not relying only on policies decided by someone else once and for all)
> we could have better coverage of the USER's need. And, also, we would
> simplify policy writing and agreements for 3rd parties provided services.
> The same simplification would occur in managing federated identities and
> access to external resources - whenever possible, of course!
> 
> Just my 2-cents.
> 
> Later,
> Max
> 
> 
> On 11/17/2009 04:10 PM, Peter Saint-Andre wrote:
>> I think this is a policy issue for providers of PK* services (just as it
>> is for providers of email services, IM services, microblogging services,
>> and all the rest). Now, I agree that better disclosure of policies will
>> help humans make more informed decisions (if they care to be informed).
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> pkng mailing list
> pkng@irtf.org
> https://www.irtf.org/mailman/listinfo/pkng