Re: [GLOBAL-V6] Global IPv6 PI policy proposal

["Per Heldal" <heldal@eml.cc>]

On Fri, 14 Apr 2006 18:40:01 +0200, "JORDI PALET MARTINEZ"
<jordi.palet@consulintel.es> said:
> 
> Following that approach, if we agree on a /48, some folks could decide to
> deaggregate it to /56 or whatever, and announce it. I think we need clear
> rules to avoid people deaggregating, and I will say nothing smaller than
> /32
> should be announced. 

There are no enforceable rules. RIRs can only publish a set of
prefix-filtering-recommendations to accompany their assignment-policies. 
(which should exactly match the size of blocks assigned unless you want
to allow room for de-aggregation).

[snip]

> > 
> > My recommendation is that a /16 be set aside out of each RIR's
> > address reserve for PI assignments to smiplify filtering.
> 
> Yes, whatever figure is needed, I will agree on that.

Size doesn't matter, just make sure assignment-windows and
filtering-guidelines are clear and consistent.


> 
> > 
> >> Expiry for those assignments: In the case of assignments done under this
> >> proposal in order to address the multihoming issue, they will need to return
> >> the block in a maximum period of 3 years after a technically correct
> >> alternative valid and deployable solution becomes accepted by the community.
> >> Alternatively, to avoid renumbering, some of the organizations affected by
> >> this, could become an LIR, if they qualify for it.
> > 
> > Any attempts to reclaim address space by force could be difficult,
> > expensive and disruptive.  "Temporary" assignments also discourageuse by
> > large organizations. Even with the LIR conversion option it
> > sounds like there is too much uncertainty for a large org to make
> > a commitment to IPv6.
> 
> I don't agree that this is the case here:
> 1) Is better to be subjected to one renumbering in several years than if
> those organizations need to renumber every time they decide to change
> their
> upstream.

What organizations want is predictability, hence ability to have
reliable planning. Periodic changes in tech-areas are generally
accepted, even expected. Given a notice-period of X-years that shouldn't
be much of a problem.

New routing-technology for MH must be supplemented by tools to
support/simplify re-numbering and other administrative issues before it
is accepted as a replacement for PI-blocks. That is both a relevant
criteria when the decision is to be made, as well as an important
argument in favor of making it a viable policy/procedure.


//per
-- 
  Per Heldal
  http://heldal.eml.cc/