IPv6 multihoming

[Brian E Carpenter]

On 2011-02-08 22:55, Siraj 'Sid' Rakhada wrote:
...
> To be honest, in my scenario, the idea was actually to split (at most)
> upto /40s, and hopefully aggregate them up to /39s or bigger, depending
> on growth in certain areas. It's just thought processes at the moment.
> I wouldn't want to put out a solution that deaggregated into /48s - if I
> ever got forced to do that, I definitely wouldn't want my name anywhere
> near it. :)

It seems to me that there are two cases of deaggregation of a PA prefix:

1. One or more customers of the ISP in question decide to multihome,
using the RFC 4116 method in IPv6. So the customer persuades their
second ISP to advertise their /48.

In this case, if the operators in general or in parts of the Internet
choose to filter out the route, the customer partially loses their multihoming,
but that's all. They don't get what they think they are paying for, but
they still have connectivity most of the time.

2. A customer leaves the ISP and insists on taking their /48 with them.
Well, in IPv6, they're not supposed to do that. And if they do, and the
prefix gets filtered, they lose connectivity. Boo hoo.

So from the 10,000 metre view - it doesn't matter if some or even most
operators filter PA /48s. The IPv6 Internet doesn't break, and for the
time being, until one of the better methods becomes viable, big sites
can get PI prefixes that will not be filtered.

   Brian