On 17  Jun 2010, at 15:10 , Paul Jakma wrote:
> On Thu, 17 Jun 2010, RJ Atkinson wrote:
> As per the docs though, ILNP is an end-host technology.

I would not agree with that statement.

Instead, I'd say that ILNP does not *require* 
router changes.  Avoiding a requirement to
change routers enables cooperating nodes
to incrementally upgrade, which seems very important.

However, it is definitely also true that ILNP enables 
implementation/deployment of new router features that 
aren't practical today with IP.

For example, Locator Rewriting by site border routers 
is clearly not an end-host technology.  ILNP's approach
to Locator Rewriting is different from IP NAT in several
ways, for example in that one need not be concerned with 
recalculating TCP/UDP checksums or deploying complex IPsec 
NAT-tunnelling schemes.

The locator rewriting capability is optional to implement, 
and optional to deploy, so it doesn't gate ILNP,
but the feature itself is clearly documented.  Among other places, 
it is described in some detail in the MILCOM 2009 paper
on 'Site-controlled Multi-Homing & Traffic Engineering'.

> Intermediate routers continue to treat the packets as IP, ...

Not necessarily true, please see one counter-example above.

Yours,

Ran


_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg

Reply via email to