RE: PI addressing in IPv6 advances in ARIN

["Bound, Jim" <Jim.Bound@hp.com>]

ALso if those prefixes are for a nuclear power plant I want them to have
proper time to renumber for sure.  It is relevant to what business
impact it has specifically in mission critical environments.  I am
removing all but IETF lists in my responses.
/jim 

> -----Original Message-----
> From: owner-v6ops@ops.ietf.org 
> [mailto:owner-v6ops@ops.ietf.org] On Behalf Of Iljitsch van Beijnum
> Sent: Friday, April 14, 2006 10:41 AM
> To: Durand, Alain
> Cc: Craig Huegen (chuegen); v6ops@ops.ietf.org; 
> ppml@arin.net; shim6@psg.com; global-v6@lists.apnic.net
> Subject: Re: PI addressing in IPv6 advances in ARIN
> 
> On 14-apr-2006, at 16:02, Durand, Alain wrote:
> 
> >> Maybe I'm being overly simplistic, but the policy can set 
> a recovery 
> >> timeframe in its allocation of PI space to end users and 
> the market 
> >> forces can drive the recovery based on the impact to the 
> >> infrastructure.
> >> If only a few hundred prefixes are handed out, it might 
> not be enough 
> >> of a problem to force recovery.
> 
> > One could argue that if we are only talking about a few hundred 
> > prefixes, Why do we care reclaiming them?
> 
> If the number of prefixes becomes large enough to be 
> problematic, reclaiming those to solve these problems isn't 
> going to work. For one thing, it's likely that such problems 
> won't be experienced to the same degree by different people: 
> people with a few large routers (and deep pockets) will be in 
> a much better position than people with a larger number of 
> smaller routers (and less money). Also, policy development is 
> done regionally while the results are suffered globally. But 
> even discounting all of that, if ARIN were to decide that the 
> prefixes must be revoked, it will take a significant amount 
> of time before that actually happens. It gets worse when 
> people start to sue.
> 
> So in practice the only thing that can happen is that if the 
> problem is severe (i.e., that 51% of all people in the (rich) 
> ARIN region feel the pain) the policy is changed so that no 
> _new_ prefixes of this type are given out, and then we have 
> to wait for the increase in router performance over time to 
> make the problem disappear.
> 
> If it gets really bad a quick "ipv6 prefix-list no-v6-pi deny 
> ::/0 ge 48" will clean up the routing table and a timeout or 
> two later you're back on IPv4 when trying to reach the 
> holders of these PI blocks.
> 
>