Re: comments on draft-ietf-shim6-proto-05.txt

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

Hi Harold,

El 02/11/2006, a las 11:43, Harold Grovesteen escribió:

> > > > the assumption here is that there is no situation of merging hosts,
> > > > meaning that if two contexts are initiated with a disjoint set of
> > > > locators, they will remain disjoint during the lifetiem of the 
> > > > context,
> > > > and the situation where two contexts have disjoint sets and then 
> > > > later
> > > > on one of them adds new locators that interject with another 
> > > > previously
> > > > disjoint context is not supported. This is what the above 
> > > > assumption is
> > > > expressing... makes sense?
> How does this assumption relate to various host-based load-balancing 
> and fail-over designs that currently depend upon a single IPv4 address 
> for a set of services?

i am not sure i fully understand your comment here...

In the shim6 scenario, a multihomed site obtains multiple PA prefixes, 
one per provider, so hosts within the multihomed site obtain multiple 
addresses, again one per provider. So, this is implies that selecting 
one address for using in a communication determines the ISP used for 
routing these packets. So, load sharing mechanisms would rely in using 
different addresses so that packets flow through different providers.

would this clarify your question?

regards, marcelo


>   Is there a need for server-side host-to-host coordination of 
> locators ala VRRP or similar mechanisms?
>
> Since the multi-homing direction was determined in the multi6 WG, for 
> various reasons, I have not followed the minute details of protocol 
> development.  So, if this has been discussed elsewhere, a simple 
> pointer will suffice to bring me upto speed.
>
> Thanks,
> Harold Grovesteen