> From: William Herrin <[email protected]>
>> That would make an MPLS tag an "address". Somehow I don't think so.
> Why not? What characteristics of what you think of as an address are
> missing from the typical use of an MPLS tag?
Um, because to me, an address, like its real-world prototype, says _where_
you are, in some way that anyone can grok, independent of either who they,
where they are, or how they get to that destination. (Yes, NAT complicates
the issue for IPv4 addresses.)
An MPLS tag identifies (at best) a particular path. It identifies a
particular destination only as a _side-effect_ (the same way that my
land-line phone number identifies my street address as a side-effect).
> From: Tony Li <[email protected]>
>>> If we really need a name for some particular crisp definition ...
>>> ... pick a new word/phrase, don't try and morph an existing word.
>> If the chair is amenable to that idea, so am I.
> Normally, I would love that idea. However, our specific request is that
> we define the terms locator and identifier.
I don't see a conflict here. We can define locator and identifier, as
requested - and on our own, for our own use, we can define a term to mean
'field(s) in the packet used to make next-hop forwarding decisions' - along
with terms for any other meanings that seem useful to have terms for, meanings
_other_ than the defined meanings for address, locator and identifier.
Noel
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
