IPv6 PI allocation

[Roger Jorgensen]

On Fri, 18 May 2007, Nick Hilliard wrote:
> What worries me is not the nat, but the nat-upon-nat and 
> nat-upon-nat-upon-nat, and the ever increasing layers of bad networking 
> layered upon worse.
>
> The issue that interests me is whether this will stabilise into something 
> which mostly kind-of works or whether people will realise that this will be a 
> stereotype good-money-after-bad situation and eventually cut their losses and 
> migrate to v6, when they realise the brokenness and therefore direct 
> financial loss that nat introduces at a larger scale.

NAT will work until the network reach a certain size, after that there are 
no ways around with NAT that don´t create more problem tahn it solve. At 
work we have to connect several larger organization that also use rfc1918 
space... we´re using all of the blocks but still we get conflict. 
Our current workaround are to place the really critical services everyone 
need to reach on public IP space we´ve been assigned by RIPE and just 
proxy everything through services there...

Of course this don´t scale very well and don´t work with all services we 
have. Problem is, there are services that require traffic both way, 
or the other way than our NAT are set up. Right now we have to do NAT the 
other way to, no real choice at all. A really ugly design but the only 
way we can do it with IPv4. Of course we could get more IP space from 
RIPE but that mean renumber a way too high amount of huge and complex 
organization....not to mention we´re talking about a huge amount of IP 
space to.

in short, the only real option we have is IPv6.



-- 

------------------------------
Roger Jorgensen              | - ROJO9-RIPE  - RJ85P-NORID
roger at jorgensen.no           | - IPv6 is The Key!
-------------------------------------------------------