multiple prefixes

Doug Barton dougb at
Mon Feb 11 19:31:47 CET 2013

On 02/11/2013 10:22 AM, Brian E Carpenter wrote:
> Doug,
> On 11/02/2013 17:26, Doug Barton wrote:
>> On 02/11/2013 06:09 AM, Tim Chown wrote:
> ...
>>> Any organisation that cares enough about renumbering
>>> implications of changing provider should be able to obtain/afford
>>> PI.
>> The fact that this is demonstrably untrue has been demonstrated many
>> times.
> Can you cite the evidence for that? I've seen exactly the opposite
> assertion recently too, and I am curious to know the facts of the
> matter.

You mean other than the dozens of times operators have made the 
statement that they care about renumbering? For most enterprises, 
_especially_ the medium sized ones that fall into the category between 
"small enough that renumbering isn't a problem" and "large enough to 
have dedicated IT staff to handle it" this is a huge concern, and one of 
the chief benefits of NATv4.

If all the times it has come up on IETF lists, NANOG, etc. don't do it 
for you, my current day job involves working with enterprise customers 
dealing with DNS/DHCP/IPAM. And that's enterprises of sizes small'ish 
all the way up to multinationals. They all care about renumbering, and 
dealing with the pain of it (even from one 1918 block to a different 
one) is one of the reasons they hire us.

NAT is the common case, even on the production side nowadays. In 
addition to the perceived security benefits, ease of renumbering 
external address space is right up there in the top 2. For us not to 
acknowledge this reality only serves to slow the pace of IPv6 adoption.


More information about the ipv6-ops mailing list