[ppml] 240/4

[Michel Py]

Hi Paul,

I'm not trying to bash at you, but I have a few things to comment on:


> Paul Wilson wrote:
> The suggestion was to use the space for private use,
> not for global unicast.

This does not make much of a difference. Anyone who needs more than
10.0.0.0/8 for a private network has routing issues the same as every
large public operator. Maybe the difference is they would like the IS-IS
and OSPF parts handling 240/4 faster than the eBGP.


> The cost of redesignating the class E address space seems very low,

This is exaggerated; moderately low is the best you can hope; there are
gazillions of lines of code that needs to be patched. Does not happen
overnight.


> and without any downside

Wishful thinking, but no game. Bottom line is, a network that you can't
plug in any "legacy" (meaning, earlier than 2009) M$ or Ci$co hardware
on (because you have a class E address) is not a network; there are way
too many vested egoes in both these companies to expect they would
implement 240/4 gracefully. If you have to wait IOS 23.x and Windows
2014 to have an IPv4 stack that handles 240/4 you are SOL.

<selfish bastard>
Let me put it in other words: I have no problem with 240/4 becoming
either an extension of private addresses or global unicast, as long as
it's not on _my_ network nor any I have to talk to :-D
It's like software from a well-known vendor: as long as other people
"beta-test" it before it reaches SP2, works fine with me :-D 
</selfish bastard>


Bottom line is: I am among the large number of people who supported
and/or co-authored something about reclaiming class E for a better use
than experimental. That being said, this late in the game, it does more
harm than good. Everyone is better off waiting for v4 allocations to
expire, in order for the market to figure out if:
a) Shortage of v4 addresses will be enough to launch v6 or
b) Market will move to living on v4 stockpiles, more NAT
   and so on so we can declare v6 dead and move on.

The way I see it: the sooner the better. We have been arguing for the
last 10 years about IPv6, time to make the call if it's a valid solution
or not. Time to let the market make the call. In the big scheme of
things, a 240/4 address is not something I would want to use on my
network today, too many bugs to sort out for the very little extra time
it would provide. The world will move either to IPv6 or to more/double
NAT, I have waited long enough. Let's exhaust IPv4 and see what happens.


Michel.