[ppml] those pesky users...

Stephen Sprunk stephen at sprunk.org
Mon Mar 26 15:30:33 EDT 2007


Thus spake "Johnson, Ron" <RJohnson at newedgenetworks.com>
> Ok, I am going to put on my smarty pants...
> Since the release of RFC-1918 and CIDR, the pressure on
> address space has decreased significantly.
> For the most part we are able to live within our means.
> We make our downstreams justify address requests, we press
> for 1918 use with NAT when ever possible, and we recycle
> disconnected customer blocks regularly.

I'd say "good for you", but you don't have a choice ... all of the above is 
required by the v4 policies :)

> As a mid-sized ISP, we just don't feel the drive to change our entire
> address scheme for no good reason. My management certainly
> does not want large scale disruptions in service to perform what
> is seen as unnecessary renumbering.
>
> If ARIN were in fact offering an exchange program where we
> would receive vastly increased sized IPv6 allocations for a greatly
> reduced fee over the IPv4 blocks. A case could be made on
> the economics of changing to v6.

Adding v6 to your network isn't just "renumbering".  You add it on top of 
v4.  Current ARIN policies/fees basically boil down to "if you have v4, you 
can get v6 for free", plus the v6 blocks are positively huge compared to the 
v4 blocks you're spending endless man-hours shoehorning yourself into.

Unfortunately, until there's a good way for v6-only hosts to talk to v4-only 
hosts, or everyone goes dual-stack (ha!), you can't turn off IPv4.  And 
that's a major part of the problem: going to v6 today means you have to 
maintain two protocols, which means instead of the reduced cost and 
complexity that v6 should bring, you see your costs double.

In the long term, getting up to speed on v6 now will save you money and 
headaches, but for managers who are oriented towards this quarter's profits 
and ignore anything after that (sadly, most of them), it makes no sense to 
do it until the death of IPv4 is imminent.

S

Stephen Sprunk      "Those people who think they know everything
CCIE #3723         are a great annoyance to those of us who do."
K5SSS                                             --Isaac Asimov 





More information about the ARIN-PPML mailing list