Re: PI Policy Task Force: dead?

[Gert Doering <]

Hi,

coming back to this:

On Tue, Apr 15, 2003 at 06:28:17PM +0200, Kurt Erik Lindqvist wrote:
> > I still fail to see an example of something that is:
> >
> >  - smaller than a /24
> >  - not a root name server
> >  - but needs PI space
[..]
> 
> Ok, TLD servers are a bad example, not because I think they do not need 
> portable space, 

*Why* do they need portable space?

> but because the problem with changing glue-records are 
> ore administrivia than anything else.  

Yep.  If the process of inserting glue records is broken, please don't
break the PI policy to get around this.

> As I am currently working for a 
> IXP, getting address space is a problem. And no, IXPs don't always have 
> a clear upstream provider. And I would not like to have to get one just 
> to sort this out.

I would agree that an IXP peering mesh makes sense (which is why they
[plus the DNS root] are about the only thing that is admittedly "special"
in the IPv6 policy).

But then, an IXP peering mesh is usually big enough to warrant a /24 or
more.  

If they are *so* small that a /28 would suffice, it's usually not overly
"impartial" or "neutral" anyway (so could use upstream space).

As for the announcement of the IXP mesh - that's an open issue whether 
this is desireable or not.  200-odd IXPs round the world should not do
much harm anyway.

[..]
> The TLD server issue a side, my main problem is that we have a policy 
> that says we should hand out useless addresses.

Define "useless"?  People might want to use PI space for VPN connections
and actually do not *want* them to be visible globally.

Gert Doering
        -- NetMaster
-- 
Total number of prefixes smaller than registry allocations:  58512  (58485)

SpaceNet AG                 Mail: netmaster@localhost
Joseph-Dollinger-Bogen 14   Tel : +49-89-32356-0
80807 Muenchen              Fax : +49-89-32356-299