[arin-ppml] Use of "reserved" address space.

[Leo Bicknell]

In a message written on Fri, Jun 25, 2010 at 11:01:02PM -0700, Owen DeLong wrote:
> Respectfully, I think ComicCon is a cheap and undeserved pot-shot
> as you yourself admit in your footnote. I challenge you to list the outreach
> locations which you think are ineffective.

Ineffective implies a binary choice which is not the case in this
situation.  Each outreach location comes at some cost (actual travel,
staff time, and the opportunity cost of other things they could do)
and some benefit via the folks you reach and the action that they
take.

ARIN has made this value judgement in a competent manor, they started
with the most effective venues, and as they have added more have
moved down the list in terms of value to add more locations.  This
is absolutely the right approch.

The value though cannot be objectively measured.  It is my opinion
they have moved far enough down the list that there is no added
benefit in going further down the list, and indeed I take a skeptical
eye to one or two of the places they are going.  That said, I fully
expect someone else to make different value judgements, rating some
conferences lower than I do, and some higher.

In short, in my opinion this avenue is fully tapped.  There is no more
sigificant good to be done via this method, and indeed I think taking a
hard look at who showed up at each location and re-evaluating priorities
for next year would be an excellent thing for staff to do, and I'm sure
they already are doing just that.

My point, which you seem to have totally glossed over is that there
are other avenues like putitng forth these issues as informational
presentations at places like the IETF, or directly to vendors that
I think might have a much greater chance of reaching the folks who
could actually do something.

In a message written on Fri, Jun 25, 2010 at 09:53:11PM -0500, Stephen Sprunk wrote:
> I'm not as worried about networking gear (which is, for the most part,
> run by reasonably competent professionals) as I am end systems.  It took
> over ten years for Microsoft to ship an OS that fully supported IPv6,
> and there are still lots of folks out there running WinXP, not to
> mention Win95/98/Me/2k.  Plus, there are all the assorted embedded
> devices (e.g. printers, toasters, phones, etc.) for which the
> manufacturers have no future software updates planned--if they're still
> in business.  If it were even possible to get everything upgraded, it
> certainly wouldn't be done soon enough to matter.  Maybe if we had
> started on this ten years ago it would make sense, but now it's far too
> late.  IMHO, it's better to push manufacturers and network providers to
> IPv6 rather than distract them with yet another hack that will only keep
> IPv4 rolling along for another year or so.

Fortunately, you think in an end-to-end model, as we need more
people to do that.  However, in this case that is a detraction.

For instance, my home connection today gets one public IP, and then
sits behind NAT on RFC 1918 space.  I believe this is amazingly
typical.  If I got a 240/4 space public address my home gateway
needs to know how to use it, but none of my workstations, printers,
phones, toasters or other devices will ever know, happily using
10/8 internally.

Much like deploying IPv6 does not require a flag day or 100% device
upgrade neither does using presently reserved space.

In a message written on Sun, Jun 27, 2010 at 01:10:52PM +0200, Eliot Lear wrote:
>    You've covered a lot of the ground of the discussion.  I presented this
>    draft on behalf of our author group to the int-area, and got pushback
>    from the room in the following forms:
>      * It would take forever to fix printers, fridges, and other
>        appliances, along with routers, firewalls, and other middle boxes.
>      * /4 is a honking lot of private address space that would benefit
>        few.
>      * /4 really doesn't buy us much time in terms of staving off or
>        easing a transition
>      * There are several v4/v6 transition protocols that base assumptions
>        about private address space.
>      * Better to focus on v6 transition.

Point #1 is addressed by my answer to the previous poster.

Point #2 and #3 are valid concerns, particularly if the proposal is to
make it more RFC1918 like space.  If it's public space, then the concern
is quite silly, as a /4 could give us 2-3 more years, which I think is
non-trival.

Point #4 is the reason I am responding.  I would like to know which
transition protocols base assumptions on what is private address space.
I must admit I am no deep expert on all the transition protocols, but I
don't quickly see anything in the most popular ones that treat private
address space as special in any way.

Point #5 is a trueism, and says nothing about the actual landscape of
what is happening.  It's one thing to have an idealistic position, it's
quite another to hold on to it when it no longer matches reality.

If IP space "trading" becomes popular, and prices become expensive
there will be real economic pressure to free up this space.  It's
the IP equivilant of $120 a barrel oil making people want to drill
in ANWR.  Sure, it's not the ideal space to get the resource, but at
some economic point a lot of people see it as worth while.

The worst case scenario here is pretty bad.  IPv6 adoption is slower
than we would like, IPv4 trading is wildly popular, the price spikes,
and things like using reserved space are rushed through.  That will 
be a real mess.

-- 
       Leo Bicknell - bicknell at ufp.org - CCIE 3440
        PGP keys at http://www.ufp.org/~bicknell/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 826 bytes
Desc: not available
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20100627/5c7a5571/attachment.sig>