Dino Farinacci allegedly wrote on 04 09 2009 12:47 AM:
>> Noel Chiappa wrote:
>>
>>> There is simply no way to produce a definition which both i)
>>> completely and
>>> accurately describes all the uses to which an IPvN address is
>>> currently put,
>>> and ii) is short and crisp.
>>
>>
>> So true.  And it would seem that i) is a clear indication of the
>> amount of overloading that we've applied to the IPv4 namespace.
> 
> Is a group address a locator or ID?
> 
> For LISP, it is considered an EID since it doesn't map to topology and
> when a host moves around and stays joined to the group, the group
> address doesn't change. So it is close to your definition of "Identifier".
> 
> Dino

Are you talking about multicast group addresses?  They are a perfect
example of a forwarding selector.  At each hop, forwarding uses them to
decide on a next-hop, but they do not name a particular point of attachment.

Just because they are not 'locators' doesn't mean they are identifiers.

What was it Noel said about trying to fit everything into too few names?
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg

Reply via email to