Re: [ekmi] Reminder of tomorrow's EKMI TC conference all

[Arshad Noor <arshad.noor@strongauth.com>]

Hi Tomas,

The protocol, as it stands today, expects the GKID to be a
simple ServerID-KeyID tuple.  The next DRAFT that I will
provide the TC will change this to DomainID-ServerID-Keyid
to make the GKID unique across the internet.

In theory, having the client application requesting a key
using a URI such as ek://<DomainID-ServerID-KeyID> can be
passed to the Symmetric Key Client Library (SKCL).  The SKCL
in turn can strip out the "ek://" and just pass on the GKID
in the conventional format to the SKS server.

However, my feeling is if we make an accomodation for the
IEEE, then there may be requests for other accomodations
for other forms.  Then it might start getting unwieldy.

I think it would be more efficient for the infrastructure, to
state that the GKID is expected to be in the "DID-SID-KID"
format, and let implementers decide what source forms they want
to accept from applications.  However, when the client library
(or the application itself) finally sends the request to the
SKS server, it must be in the "DID-SID-KID" format as we will
specify in the XSD.

Hope that answers the question.

Arshad

Tomas Gustavsson wrote:
> 
> I have read the draft though. Would it also be possible in theory to use 
> an SO_GUID in the URI format in EKMI?
>