[ppml] 2001-2 Revisited (and 2002-3 too ;)

[Dave Diller]

> How does this better than 2002-3, which was approved (albeit in amended 
> form) ? The intent for 2002-3 was that you would get a /24
> _from ARIN_, which IMHO is better  than from your upstream.
> 
> Now, it was changed to a /22, and I haven't submitted my proposal yet 
> (so I don't know if it has ever been exercised), but given 2002-3, why 
> do we need 2001-2 ?
> 

By my read, 2002-3 won't work for someone who can only justify a handful of IPs 
at each of several sites.  It requires the same justification as any other 
request for IPs from ARIN, just with different bit boundaries.  No multihoming 
special-case caveats like 2001-2.

Speaking of 2002-3, how will it be put into practice?  There seem to be two 
ways it could play out:

* Current MH policy is "Used a /21?  Here's a /20".  2002-3 could simply be 
shifting that two bits to "Used a /23?  Here's a /22".

* Single-homed policy is "Used a /20?  Here's a /20".  2002-3 could both move 
the bits down as well as unify the policies.

It dsoesn't say which of the two wil be put into play - whether the /21-->/20 
is an integral part of the MH policy concept or simply a now-outmoded 
exception.  At risk of steering the topic astray, clarification on the 
mechanics would be useful.

-dd