Re: Architectural approaches to multi6

[marcelo bagnulo <marcelo@it.uc3m.es>]

Hi Ran,

On Mon, 2003-03-24 at 22:19, RJ Atkinson wrote:
> On Monday, Mar 24, 2003, at 16:09 America/Montreal, Kurt Erik Lindqvist 
> wrote:
> > But in principal the division of solution space could be
> >
> > - multiaddress
> > - Mobility based
> 
> Does "Mobility-based" mean MIPv6 or something very similar ?
> 
> I would say that MIPv6 is an instance of Locator/Identifier
> separation.

I am not so sure about that. I would say that in MIP identifier and
locators are not separated. I mean, a given address can work both as a
locator and as an identifier, depending on the situation. Perhaps, I
would say that the identifier role and the locator role are separated.
Please, note that this is not related to the fact that locator space and
address space are the same (IP addresses) but that sometimes a given
address is used as a locator and other times as an identifier.
This is true also when considering transport layer solution such as the
modifications to TCP to support multiple addresses per connection.
This is different from the LIN6 case, where there is a specific type of
addresses reserved to serve as identifiers, in which case identifiers
and locators are clearly separated.

Thanks, marcelo


>   I'd guess there are lots of potential instances
> of any given approach.  I was trying to take a much higher
> altitude perspective on the potential approaches that are
> fundamentally different from each other, then work top down.
> 
> > - Location/Identifier separation
> > - Based on today (basically improved routing, but I haven't seen any 
> > real proposals on this).
> 
> I'm not sure what "based on today" means.  So if someone
> does have a proposal for this, a high-level outline would
> be pleasant.
> 
> Ran
-- 
marcelo bagnulo <marcelo@it.uc3m.es>
uc3m