Re: [v6ops] PI heresy [draft-ietf-v6ops-ula-usage-recommendations - work or abandon?]

[George Michaelson <[e-mail]>]

I'm sorry Brian, but as I said in another reply, this is just specious.
It's true in as much as IF all SME's wound up in BGP we'd have a heat-death
problem.

But a) they won't arrive tomorrow, or next year: it takes time,  and b)
other solutions to routing will emerge. Like, giving them scoped /48 out of
a block which has export limits under some covering prefix, and limits the
local visibility of their BGP independence to within some constrained
context.

It never ceases to amaze me that I tell a yak-herder my BGP failures,
forcing local path selection, with no subsequent alteration in routing of
concern to them, *and we're not even shipping packets*

Tony Hain amongst others proposed models of geographic and scoped
assignment which were shot down in flames by the usual suspects. Because
its "not how we do it now" people don't like talking about this sensibly,
but there are rational models of address management which preserve BGP like
behaviour, but don't expose all routing decisions globally, that only have
local meaning.

G



On Fri, Nov 13, 2015 at 5:53 AM, Brian E Carpenter <
[e-mail]> wrote:

> Please let's not pretend we're discussing the ULA draft any more.
>
> On 13/11/2015 07:32, Owen DeLong wrote:
>
> ...
> > How is a /48 PI hard to acquire? Where?
>
> As others have implied, that isn't the point. The point (and I expect
> we had this argument on the RRG list some years ago) is that we can't
> put 10,000,000 /48 entries into BGP. So if we go around encouraging
> all small enterprises to get a PI prefix we are all doomed.
>
>    Brian
>
> _______________________________________________
> v6ops mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/v6ops
>