[ppml] The WIANA registry

[william at elan.net]

> > Given the above, I expect WIANA to say at some point in the 
> > future "Look, we are already using this space and no-one else 
> > is; so just let us use it." All without going through the 
> > mutually agreed upon processes, procedures and justifications 
> > that everyone else must go through.
> > 
> > This, I believe, is the real problem with this 'Registry'..
> 
> This seems like a perfectly sensible strategy to me and that is why there
> needs to be a REAL mechanism to stop such squatting and tactic.  This takes
> the implicit (though explicit would be more satisfying) cooperation of the
> Registries and the ISP community.
And how do you see that? What can you possibly do?

They have made it clear they are not using ips for internet connectivity.  
They would not do as RFC suggested and use one of other ranges (10/8), so 
whatever problems their users run into (like not being able to access real
1/8 ips if they were assigned) is what they get for doing something that 
is not standard.

As far as anything else - ARIN tries to maintain good balance between 
assigning ips and actually getting involved in routing and network connectivity
- in my opinion it would be completely inappropriate for ARIN to tell 
their members you MUST not route to particular ip block, etc. ARIN can 
tell them this block should not be routable through whois and best is to 
maintain that database better and to provide better means of accessing it.

As far as WIANA - I think best we can do is NOT get involved in this situation
(dont not help by having everybody talk about it and giving them such free 
publicity - and do note that if ARIN is against them it may very well work 
in their favor as far as public opinion goes). If WIANA is successfull 
in getting users to use their system, they may realize best they can do 
is to apply for actual ip block range from APNIC, ARIN, RIPE, etc or use 
procedures for experimental allocation (but not expect they would get 1/8 
likely something much smaller), but I would expect they would not be 
successfull - their system of having to route individual ips across the 
net would lead to very large routing table and this is just not technically
workable when it gets large enough. Better way would be to use static ips 
within ipv6 for devices on user end (out of /64) and have good way 
of doing Dynamic DNS with ipv6 A6 and DNAME records that works immediatly 
as soon as they connect, and this is something for IETF to work on.

-- 
William Leibzon
Elan Communications Inc. 
william at elan.net