RE: [dhcwg] DUID on a Virtual Host

> On Mar 1, 2007, at 5:45 PM, David W. Hankins wrote:
> > It would be perfectly acceptable to use a derivative of a key as a
> > DUID.  If you actually wanted to make use of the key of course,
> > in authentication or encryption or what have you, there would need
> > to be some additional mechanism to transfer its content.
> 
> No, we didn't agree on that.   You assert that the key is 
> unique, and  
> that therefore the fingerprint (derivative) is unique, but in fact  
> nothing of the sort is true - the only reason the fingerprint works  
> as an identifier is that there are additional disambiguation  
> mechanisms that (a) make the likelihood of an undetected collision  
> acceptably small and (b) provide a path for both detecting and  
> dealing with a collision.   Unfortunately, these mechanisms require  
> the intervention of an intelligent agent (a person) and can't really  
> be automated in the way you're suggesting.

Forgive me if I am missing the point completely, but is't
handing out unique pieces of information to clients exactly
the reason why we have DHCP servers in the first place? For
example, why can't it be such that a client could walk up
to the server and say:

 "My name is James Horatio Blankenship Esquire the Third."

and the server replies:

 "That's great; we'll just call you "Jim" from now on."

So, as long as the server accepts the client's (long) public
key as an identifier and gives it back a (short) nickname to
be used in subsequent transactions it should be OK - right?

Fred
fred.l.templin@boeing.com 

_______________________________________________
dhcwg mailing list
dhcwg@ietf.org
https://www1.ietf.org/mailman/listinfo/dhcwg