[ppml] Policy Proposal 2007-21: PIv6 for legacy holders with

Stephen Sprunk stephen at sprunk.org
Mon Oct 29 15:00:09 EDT 2007

Thus spake <mcr at xdsinc.net>
>>>>>> "David" == David S Madole <david at omd3.com> writes:
>    David> As Ted pointed out recently, there is currently a
>    David> disincentive for legacy holders to implement IPv6 at all
>    David> versus dragging out IPv4 as long as they can.
>  There are no obvious carrots for IPv6, only sticks against IPv4.
>    David> One simple way to remove this disincentive would be to offer
>    David> IPv6 addresses to legacy holders on the same terms as their
>    David> IPv4 addresses, or maybe not the same but something that
>    David> removes or lessens the disincentive.
>  From what I understand, IPv4 RSA holders already have access to
> IPv6 PI space.

There is no "IPv4 RSA".  Some IPv4 registrations are under an RSA, while 
others (most legacy ones) are not.  Anyone who signs a standard RSA (i.e. 
not the LRSA) "has access" to IPv6, provided they meet the bar given in the 
NRPM.  According to Leslie's preso at ARIN XX, out of 209 IPv6 requests in 
the last 12 months, only six were denied (not counting the five that were of 
the wrong type and corrected), so the bar doesn't appear to be particularly 

> I am curious to know what the distribution of legacy IPv4
> holder are: how many /24s, how many /16s, etc?

You can send a request to hostmaster if staff doesn't respond here directly.

>  I think that the many multitude of /24s, probably don't need more than
> a /48, and there are many ways to get that.

If you want a PI /48, there's only one way to get it: qualify for a v4 
assignment under current policy (which requires 25% utilization of a /22 if 
multihomed, /20 if not).

> Of the /16s and /8 legacy holders, I would think that getting IPv6
> PI space directly would be easy.

It's hard to imagine that the bar for PIv6 wouldn't be met by those folks. 
The only "problem" is the folks that have a legacy /24, which they most 
likely wouldn't qualify for under current policy.

>    David> Taking this point a little further, it's largely the legacies
>    David> that got IPv4 to take off and got the Internet built, they
>    David> could be the ones to do the same for IPv6 too perhaps if
>    David> given an incentive rather than a disincentive.
>  Doubtful.
>  Just make IPv6 space as easy to get as IPv4 was back in 1988.

We're a bit short on routing slots these days; we can't afford to give a 
block to everyone who asks, though we're not far off that with a rejection 
rate <3%.

>    David> After all, even ARIN, who should be leading the way I would
>    David> think, isn't even offering whois on IPv6 yet as far as I can
>    David> tell, even though RIPE has been doing so for almost five
>    David> years.
>  I didn't know there was no whois yet.

# host -t any whois.arin.net
whois.arin.net has address
whois.arin.net has address

WHOIS does, of course, contain v6 records; it's just not accessible via v6. 
That should be fixed.  (The same problem doesn't apply to www.arin.net)


Stephen Sprunk         "God does not play dice."  --Albert Einstein
CCIE #3723         "God is an inveterate gambler, and He throws the
K5SSS        dice at every possible opportunity." --Stephen Hawking 

More information about the ARIN-PPML mailing list