New ARIN ipv6 allocation policies

Iljitsch van Beijnum iljitsch at
Sun Sep 3 18:57:53 CEST 2006

On 1-sep-2006, at 20:21, Doug Barton wrote:

>> Yes, the fact that IPv4 is a big mess and now IPv6 gets the chance to
>> become one as well is very sensible. Never mind all the work in  
>> the IETF
>> to come up with something that works better in the long run.

> It doesn't matter how elegant a design is if not one actually wants to
> deploy it. For the most part, the very people whose interests you  
> claim to
> be protecting have spoken clearly, and said that without PI space  
> IPv6 is a
> non-starter. It's well past time that we acknowledged this, and  
> moved on.

The latter is obviously untrue as the changelog for the policy is  
dated less than a week ago. It's important that people walked into  
this with their eyes open, and chose to reject shim6 before any  
meaningful evaluation of its performance was possible, and before  
there was much reason to deploy IPv6 in the short term. So even if  
this turns out to have been the right decision, that eventual fact  
can't be known at this point so the decision was at the very least  

I lend no credence to claims that IPv6 PI is required for IPv6  
deployment: there won't be much IPv6 deployment even with IPv6 PI.  
Look at ISPs that can get PA (and people like Google who were able to  
get it as well although I don't see how they qualify): the vast  
majority of those isn't deploying IPv6 either.

>> But why am I complaining, I don't have to pay for those bigger  
>> routers.

> You do actually, even if it's indirectly via higher fees, etc.

Is this an attempt to cheer me up...?

> But, that's the cost of progress.

What progress? IPv6 inter-domain routing is now officially the same  
as IPv4 routing only with longer addresses.

The unfortunate thing is that people haven't been willing to look  
beyond their own shadow here, and accept temporary inconvenience to  
gain long-term benefits. This is true to some degree in the IETF, but  
much more so in the RIR policy groups. This makes it clear (at least  
to me) that the way the decision making process for these issues is  
organized today is fundamentally flawed.

The good news about the operational community is saying routing table  
size isn't an issue is that researchers can now focus on other  
problem areas in inter-domain routing.

 From another message:

>> No.  It is called "repeating history".  "Progress" would have been  
>> IPv6
>> being designed and implemented with a scalable routing architecture.

> If your definition of "scalable routing architecture" includes not  
> tying
> an enterprise (be it medium sized ISP, big company, whatever) to a  
> single
> provider for their IPv6 bandwidth, then once again, I agree with  
> you, but
> it still doesn't matter. The equation is simple. We cannot solve the
> routing problems till we get experience.

No; we can't solve anything until we start trying. So far, all  
initiatives to that effect have been shot down. Experience is what  
you gain with something that's already there. With idr, the stakes  
are so high that deployment must be so predictable that it's almost a  
formality at the end. Trying various things and see what works isn't  
acceptable here.

I'm very interested to see what will happen in the upcoming IAB  
routing workshop.

More information about the ipv6-ops mailing list