Re: Thinking differently about names and addresses

[Rob Austein <sra at hactrn.net>]

At Tue, 1 Apr 2003 16:58:33 -0500, Keith Moore wrote:
> 
> On Tue, 1 Apr 2003 16:26:47 -0500
> "J. Noel Chiappa" <jnc at ginger.lcs.mit.edu> wrote:
> 
> >     > From: Keith Moore <moore at cs.utk.edu>
> > 
> >     > HIP only solves part of the problem ... it doesn't provide any
> >     > way of mapping between that identity and an address where you
> >     > can reach the host.
> > 
> > A system doesn't have to provide mechanisms to look up mappings from
> > <any-name> to <any-other-name> to be useful; 
> 
> agreed.  but it does need to provide such mechanisms in order to provide
> useful endpoint identifiers.  without a mechanism to map the endpoint
> identifier to an IP address, such identifiers are useless in referrals
> between application components.

Correct, if you assume that there will be situations in which you only
have the endpoint identifier and you need the IP address.  That's not
the only conceivable way of passing around a referal, it's just the
one that's closest to being a drop-in replacement for a referal
mechanism based on raw IP addresess.

In other words, whether you need this particular mapping depends on
precisely how you intend to use the endpoint identifiers, which is
what I think Noel was saying.