[ppml] Incentive to legacy address holders

[John Santos]

On Mon, 9 Jul 2007, Ted Mittelstaedt wrote:

> 
> 
> >-----Original Message-----
> >From: John Santos [mailto:JOHN at egh.com]
> >Sent: Monday, July 09, 2007 7:25 PM
> >To: Ted Mittelstaedt
> >Cc: ppml at arin.net
> >Subject: RE: [ppml] Incentive to legacy address holders
> >
> >
> >On Mon, 9 Jul 2007, Ted Mittelstaedt wrote:
> >
> >>
> >>
> >> >-----Original Message-----
> >> >From: John Santos [mailto:JOHN at egh.com]
> >> >Sent: Monday, July 09, 2007 3:59 PM
> >> >To: Ted Mittelstaedt
> >> >Cc: Leo Bicknell; ppml at arin.net
> >> >Subject: RE: [ppml] Incentive to legacy address holders
> >> >
> >> >
> >> >On Mon, 9 Jul 2007, Ted Mittelstaedt wrote:
> >> >
> >> >>
> >> >>
> >> >> >-----Original Message-----
> >> >> >From: ppml-bounces at arin.net [mailto:ppml-bounces at arin.net]On
> >Behalf Of
> >> >> >John Santos
> >> >> >Sent: Sunday, July 08, 2007 6:47 PM
> >> >> >To: Leo Bicknell
> >> >> >Cc: ppml at arin.net
> >> >> >Subject: Re: [ppml] Incentive to legacy address holders
> >> >> >
> >> >> >
> >> >> >Under these circumstances, I can't see any sense in doing anything
> >> >> >else but what we are doing now, continuing as a legacy, non-RSA-
> >> >> >signing holder.
> >> >> >
> >> >>
> >> >> I guess you think your pretty smart in that you have outlined a
> >> >> situation you think isn't solvable in IPv4.
> >> >>
> >> >> So, when all your customers have switched over to IPv6 and are
> >> >> demanding that you do the same, it appears to me you will be still
> >> >> in exactly the same circumstances.  You customers will still
> >be natting
> >> >> under IPv6 - if you don't think so, go ask them now.
> >
> >You don't know my customers.  They strongly believe in "if it ain't
> >broke, don't fix it."
> >
> 
> Yeah, sounds exactly like mine too.  Believe it or not I've been told
> "the reason we don't apply Microsoft security patches is because if
> it ain't broke, don't fix it"  (this was when talking to a customer
> that their line was clogged with outbound spam because their exchange
> server had been cracked into)
> 
> >(In case you totally have the wrong end of the stick, my customers
> >are *NOT* buying any sort of internet service from us.  We use the
> >internet as a tool for supporting our customers.  They typically
> >have enormous internal networks, and may eventually implement v6
> >on them, but there is no prospect they'll be turning off v4 for
> >decades.  Switching to v6 for this function would be a pointless
> >waste of time for both us and them.)
> >
> 
> No, I understood this. We have dealt with similar private-to-private
> interconnects ourselves
> and I'm aware that it is very seductive to use legal numbers for such
> interconnects to avoid clashes with private number space.
> 
>  One of the main drivers for going to IPv6
> is, of course, it gives so much numbering that it should make no difference
> if
> a bit of the public numbering goes away into these kinds of connections
> forever.
> 
> I will also point out that staying with IPv4 for your
> interconnect is also a solution, if the other parties don't want to
> update.  Once the Internet switches over to
> IPv6 the IPv4 you have in the interconnect will be worthless anyway,
> so there's a great argument to leaving it alone, and nobody will
> care if it's legacy or not.  Obviously you will have problems sourcing
> traffic from it into the rest of the world but generally most
> interconnects of these types aren't sourcing anyway.
> 
> However, the mistake you made is trying to apply your situation to
> the global problem with legacy numbering.  You set up a fairly narrow

No, I did not say anything about any global problem.  I just described
my situation and said it did not seem to fit in with what was being
discussed.  I don't know how many others are in the same situation,
it might be extremely rare or it might be very common.

> situation, and in this post you have added even more conditions to
> narrow it even further.  Doubtless if we were to discuss it further
> and discuss the usual solutions used for this situation, you would bring
> reasons why you can't do them which would even further narrow the
> scope of the example.  Eventually so many solutions would have been
> brought up and shot down that it would be obvious to anyone that
> your situation is so unique it's completely inapplicable to the
> larger discussion of legacy number holders, and you would have succeeded
> in invalidating the original analogy you tried to make in the first
> place.

What analogy?  I didn't make any analogy.  I just described reality.

> 
> >
> >Totally bogus analogies.  Why don't you propose the police go back
> >to their records and charge with drunk driving anyone they stopped
> >with a breathalyzer reading below what was then the threshold but
> >is now above the threshold?
> >
> 

Oh, this analogy?  You mean my mocking response to *your* ridiculous
analogies (which you conveniently snipped?)

> Nobody is arguing that in 1993 your now-legacy assignment was assigned
> incorrectly or that you shouldn't have had it in 1993, or 1994 or
> so on.  But the point that has been repeatedly made over and over on
> this list is that the IP numbering SCHEME is a SHARED scheme.
> 
> You cannot deny that the Internet would not function if nobody agreed to
> respect numbering allocations - you yourself respected them when you
> got yours originally.
> 
> What I think your blind spot is, is that your implying that conditions
> on the Internet haven't changed from 1993.  I think a few of the old timers
> on this list (and keep in mind I was running UUCP back in 1982) seem
> to have a problem with the idea that their baby grew up into the 800 pound
> gorilla.
> 
> You have to treat the 800 pound gorilla differently, you don't let him
> sit on your lap like he could when he was a baby gorilla.  Life changes
> and we all have to change with it.  Me, I absolutely deplore a lot of
> changes that have happened on the Internet, for example I think it's a
> terrible thing that child predators are able to use it nowadays to get
> victims, that wasn't going on a decade ago that I remember.
> 
> The numbering rules that were in effect in 1993 cannot stand.  As proof of
> this the entire IPv4 numbering scheme itself has been tossed in the
> garbage can, and replaced by IPv6.  Yet, there's still people out there
> that if they got a chance would turn the clock back to 1993 and
> bring the old 1993 rules into 2007 and beyond.

I'm not one of them.  But I think some babies (whether they grew up
to be 800 pound gorillas or bonobos) got thrown out with the bath
water...  

And if IPv4 has been replaced by IPv6, why do you care about legacy
v4 assignments anymore?

(Down below, it sounds like you don't... :-)

> 
> >> Face the facts.  Your getting something for nothing.  Your getting
> >> tracking and visibility in a system you aren't paying for - in fact,
> >> in a system that -I'm- paying for.  (or more accurately, my employer,
> >> who due to paying for this system has less money he can pay me, and
> >> so forth)  You certainly don't seem appreciative of this.
> >
> >I never said I wasn't willing to pay my fair share for *something*
> >(like v6 addresses.)  I'm not willing to pay, agree to terms I did
> >not originally agree to, and risk losing my /24 for no discernable
> >benefit to me.
> >
> 
> As others have claimed if you sign an RSA for IPv6 it doesen't affect
> your IPv4 holdings.  I would ask, have you even e-mailed hostmaster at arin.net
> and asked any of these questions?
> 

Not yet.  I don't need v6 *yet*.  I probably will someday.  I've got
an O'Reilly book on my desk that I'll read someday when I have time. :-)

So if I sign up for a v6 allocation and sign the RSA and pay my $100
per year, will I still be in danger of losing my v4 allocation as
various policy proposals being discussed here seem to indicate?  Will
I in fact increase the danger of that happening?  (Leo seemed to
indicate in another subthread that I'm using enough of my /24 to be
safe, currently 126 hosts in my DNS, but I know some of them are
defunct.)

> >
> >>
> >> And, as I asked before, how are you going to move your setup to
> >> IPv6?
> >>
> >
> >I didn't answer this before because I don't spout nonsense off the
> >top of my head, unlike some people I could name, and I need to do
> >a bit of research before answering, but at least three possibilities
> >come to mind:  1) I believe there is a class of addresses that can
> >be generated from IPv4 addresses, and I can just use those.
> >2) Apply for v6 addresses through the normal process.  3) If I
> >don't qualify for 2 because my network is too small, then form a
> >cooperative with some of the 20,000 other legacy class C holders,
> >pointlessly duplicating the work of ARIN, etc. but aquiring enough
> >v6 addresses for all of us.
> >
> >If any of this is wrong, or unworkable, *you* are the one who insisted
> >on an answer...  If you're so damn smart, what would you do?
> >
> 
> I don't see anything wrong with #2.  But keep in mind that I also
> feel the requirements in the following:
> 
> http://www.arin.net/registration/guidelines/micro_alloc.html)
> 
> are unworkable and favor large companies.  But you see there's a lot
> of politics going on.  One of the biggest problems I think is the
> insistence on aggregation.  This is why the
> requirements for getting a micro allocation are unworkable for most
> organizations, the people that wrote them want to force every potential
> small holder to request from upstream.  (except, of course, then the
> small holder is them - why if your a holder that runs a public exchange
> you can get a micro allocation)
> 

I think it would fall under the Exchange Point Operator category.
But I'm not sure.  The other two categories clearly don't apply
(Critical Network Infrastructure, it's critical to us, and to our
customers, but not to the world at large, and Non-Routed Core
Addressing, which requires that you already have a v6 allocation.

> I mean - think of this!  We have people fighting with me on this list
> that IPv4 is so important post-transition that ARIN must keep track of
> IPv4 allocations forever!  They are happy to spill some 100,000+
> IPv4 route entries into the public BGP table post-IPv6 for the next 50
> years - yet the same folks buy off on the policy that there are too many
> route entries so we must restrict the micro allocations!

As I pointed out previously, my /24 (and future v6 allocation) doesn't
need to go out to the public BGP table (though it would be nice if it
did, I can live with it remaining on our private networks.)

> 
> Clearly there's contradictions in the policies, and far more in how some
> people view things.  As I said last week nobody wants to take any steps
> to push IPv6 implementation, they are all expecting the other guy to
> just do it without trouble.  And nobody had any response to that.  They
> just bitched up a storm with the idea that it might be a good idea to
> one of these days just stop paying attention to IPv4.  But I ask you, how
> do we even start uncovering these problems if the legacy holders don't
> want to get engaged?  (and I'm not talking about you, I'm talking about
> the legacy holders who are out there and who aren't even reading, much
> less participating, in the discussion)  You at least are partipating in
> the discussion.
> 
> Ted
> 
> 
> 
> 

-- 
John Santos
Evans Griffiths & Hart, Inc.
781-861-0670 ext 539