New ARIN ipv6 allocation policies

Iljitsch van Beijnum iljitsch at
Mon Sep 4 08:31:39 CEST 2006

On 4-sep-2006, at 7:55, Andrew Alston wrote:

>> Yes, those academic networks that think they can get by without paid
>> transit in IPv6 and then dump everything they can't move over peering
>> in those triple-continent tunnels are extremely annoying.

> I just have one question here, what about those places that wish to
> develop, wish to move forward, and have no OPTION but to tunnel.  I'll
> openly admit that I run trans-continental tunnels to two continents to
> announce my v6 space.  Why do I do it that way?  Because there is *no
> one* on the continent doing native v6, infact, my tunnelled v6
> connectivity is to my knowledge the largest active v6 deployment in
> sub-saharan Africa.

> It's all well and good to sit and complain about the tunnelled
> connectivity when you have the ability to do it some other way, but  
> not
> all of us have that luxury.

If you look above you'll see that my annoyance is unleashed at a very  
specific class of tunnel users.

I always say "if a tunnel is the fastest way to move IPv6 packets,  
tunnel away". There are even cases where tunnels can be superior to  
native connectivity, due to ASIC support and other implementation  

But sending packets from North America to Japan over a tunnel and  
then from Japan to Europe over another tunnel does NOT qualify as  
reasonable tunnel use. I've seen this many times, most recently at  
the IETF meeting in Montreal when I tried to reach my server at home  
in Holland. Since I have an IPv6-enabled server and my Mac  
automatically gets IPv6 connectivity from the IETF meeting network  
over which it will access email and do SSH, this type of sub-standard  
IPv6 connectivity really gets in my way, sometimes to the degree that  
I have to manually fall back to IPv4.

But if you directly tunnel from your location to the place where the  
IPv6 packets have to go this isn't an issue as IPv6 performance  
should be nearly identical to IPv4 performance.

More information about the ipv6-ops mailing list