Dick Hardt wrote:
> On 19-Jan-06, at 5:27 AM, Ben Laurie wrote:
> 
>> John Merrells wrote:
>>> An identity information exchange should involve just three parties: the
>>> user, their agent, and a relying party. The user’s agent is where they
>>> authenticate themselves and a repository where they store their identity
>>> information, and the relying party is an entity requesting identity
>>> information.
>>
>> This seems overly prescriptive. In particular, it would appear to
>> exclude any kind of temporary certificate. It also excludes proxies. Oh,
>> and the case where authentication occurs elsewhere.
> 
> Hey Ben, would you take the time to write up simple use cases for your
> three points so that we (or at least I) can understand them?

Temporary certificate

This is to satisfy the minimality requirement. User has cert including
date of birth, say, and wants to prove he's over 21. So, he (or his
agent) shows the cert to some CA that produces a temporary cert for him
saying he's over 21, which he or his agent then shows to the relying party.

Note that this only half gets you unlinkability if the certs are
anything conventional because the CA can link the permanent and
temporary certs.

The CA is, of course, a fourth party in the transaction.

Proxy

Not sure exactly what to say about this, except that a proxy could sit
between any of these parties, and the language above assumes that it can
do so both transparently and securely. Which may not be so (that is, it
may have to be non-transparent to remain secure) if it adds
functionality, like caching, or anonymising.

Authentication Elsewhere

It may turn out that for whatever reason I have to use multiple agents,
so I'd like to authenticate them via my meta-agent. Unlinkably.

Cheers,

Ben.

-- 
http://www.apache-ssl.org/ben.html       http://www.thebunker.net/

"There is no limit to what a man can do or how far he can go if he
doesn't mind who gets the credit." - Robert Woodruff

_______________________________________________
dix mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/dix

Reply via email to