[arin-ppml] [arin-announce] Policy Proposal: Open Access To IPv6

[Garry Dolley]

On Fri, May 29, 2009 at 09:02:57PM -0700, Matthew Kaufman wrote:
> Ted Mittelstaedt wrote:
>> Matthew Kaufman wrote:
>>> Ted Mittelstaedt wrote:
>>>>
>>>> By contrast, the problem of table bloat in routers is very real.  
>>> Absolutely, but should not be ARIN's problem. ARIN should make it 
>>> possible for organizations (be they ISPs or not) to obtain unique 
>>> allocations from ARIN's portion of the IPv6 address space. It should be 
>>> easy to get allocations which are aggregated in such a way that they are 
>>> less expensive to route, but that shouldn't drive policy.
>>
>> OK then taking that argument to it's logical conclusion then let's just
>> assign every last man and woman on the face of the earth an IPv6 number
>> drawn out of a random lottery pool and let them plug in and have at it!
> Well, you could do exactly that. And they could have their own chunk of 
> IPv6 space *and* a separate, routeable PA address as well. And we *still* 
> wouldn't run out.

Sure, you could make the assignment; but it would just be a dumb
number.  Current infrastructure wouldn't route it.

>>>> You
>>>> are essentially asking thousands of orgs out there to put money into
>>>> tens of thousands of routers out there to replace them with new gear
>>>> that will hold and manage a fantastically gigantic IPv6 table,
>>> No. We are asking ARIN to fairly allocate reasonably-sized parts of IPv6 
>>> to entities that ask, in return for no more than reasonable compensation.
>>
>> A single snowflake by itself harms no one.
>>
>> A trillion snowflakes creates an avalanche that kills people.
> Lots of people having non-routeable or non-routed addresses isn't going to 
> kill anyone.

The people who don't care about the routeability of their addresses
can just use ULAs [1].  I really don't see what the practical issue is.

> Heck, even a big explosion in routing table size won't kill anyone. Maybe 
> someone will take the time to solve the problem more elegantly than it has 
> been so far, and it won't even matter.

Then send me the cash to get SUP720-3BXL so I can hold up to a
million IPv6 routes in hardware.

Sure it won't kill anyone, but I'm not going to pay for your desire
for a /32; and why should anyone else?


1. RFC 4139, "Unique Local IPv6 Unicast Addresses"
   http://tools.ietf.org/html/rfc4193

-- 
Garry Dolley
ARP Networks, Inc. | http://www.arpnetworks.com | (818) 206-0181
Data center, VPS, and IP Transit solutions
Member Los Angeles County REACT, Unit 336 | WQGK336
Blog http://scie.nti.st