I think we need to keep all this terminology stuff simple. And not
have too many terms. What comes out of this group will pro-market or
negative-market locator/ID separation. And you don't want people to
react "this locator/ID separation stuff is more trouble then it is
worth".
Dino
On Mar 30, 2009, at 9:27 AM, Noel Chiappa wrote:
From: Dino Farinacci <[email protected]>
From a network perspective, if a switching node makes a forwarding
decision based on an [field] in a packet header, then that [field]
is a
locator.
Dino, I hear what you're saying: I have made a very similar point
before,
but.... I called those fields 'forwarding tags'. (Not that I have any
particular deep attachment to that particular name, I'd be happy
with pretty
much any term.)
The reason I picked a new term there was simple: I think we need
terms for
_both_ of these concepts (what others call 'locators', and 'forwarding
tags'), and overloading both meanings onto one term is just confusing.
Like I said earlier, it's true that one can use a MAC address, or
any similar
unique identifier, for flat routing. (In which case it's what Tony
defined as
an 'identifier', for the interface - i.e. just a name which denotes
one
particular instance of a set of objects.)
However, for _non-flat_ routing, one needs a set of names with more
_structure_ in them, and it's names of that type (ones with built-in
structure
which is related to the _location_ of the thing they are naming),
which most
people mean when they say 'locator'.
And clearly, one can't use the same term for both of these (quite
different)
concepts, without great confusion...
Noel
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
