[AusNOG] IPv6: "Objections to sale"

Mark Newton newton at atdot.dotat.org
Wed Mar 6 19:51:54 EST 2013

Sales people go through a series of stages to convince a customer to buy.

The first stage is to establish a rapport, so that the salesperson is seen as a 
trustworthy and credible voice.  Pretty much impossible to sell anything high-value
if you can't do that.  So that's why salespeople have big expense accounts, spend
a lot of time at expensive restaurants, invite customers to golf courses, and hand
out promotional toys:  It's all about setting themselves up as your friend.

In the process of doing that, they gain enough info to move on to the second stage,
which is to identify specific problems you have and products in their portfolio
which could solve them.  Thus begins the pitch...

Even if the pitch is successful, and they've convinced you that they have a product
that you ought to buy, you're not going to sign on the dotted line until you've been
through the third stage, which is overcoming your "objections to sale."  

Perhaps you don't like the price; the salesperson will negotiate on that. Maybe there's
a feature $competitor has which you like; the salesperson will either show you their
equivalent features, or try to convince you that $competitor's feature is buggy or
unnecessary.  Maybe you don't have budget the salesperson either comes back when
you're working out next FY, or gives you an offer you can't refuse to induce you to
buy immediately.It's an iterative process, where one by one the reasons you'd have for
not spending money are worn down.

At the end you have an identified need, a product to fill it, no reason not to buy
the product, and a natural human inclination to want to make your friend happy.
All the psychological boxes ticked, you make the sale.  The magic of capitalism.

The good salespeople execute the stages in parallel, but even then you can pick
the progression if you're looking for it.  They're basically coin-operated humans,
all running variants of the same software :-)

We're all going through something similar with IPv6.  The first stage is pretty 
straightforward, it was accommodated years ago by people like Geoff Huston banging
the drum (it was his RIPE talk about IPv4 allocations increasing in geometric time
instead of exponential time that got Internode into the game.  I think I still
have the email message I sent to Simon about it sitting in my sent folder!)

The second stage is pretty straightforward too:  IPv4 disappearing.  Some of
you don't believe it is disappearing half as quickly as everyone has made it out
to be, but even those people would be dishonest if they didn't at least acknowledge
the legitimacy of the case.  No IPv4 + availability of IPv6 = problem + solution.

So now we're up to stage 3:  Objections to sale.

Back in 2003, people said things like, "Router support is terrible and buggy." Then
the router vendors basically fixed that one, and it's now next to impossible to buy
business-oriented routing hardware that doesn't do reliable IPv6, and even some
of the cheap consumer gear is doing it now too.  Tick the box, move on.

Back in 2005, there were policy objections: "I can't get PI space from my RIR,"
or, "This is stupid, am I seriously supposed to give 4 billion times more addresses
than the entire IPv4 Internet to every Ethernet segment?"  The RIRs came to the
party with that one, policies were changed, education was delivered.  Perhaps the
IPv6 Forum played a part here, but so did "buzz."

In 2007 it was, "I can't get IPv6 from my upstream."  Solved.  In almost any
geography on the planet you can now buy native dual-stack transit from lots of

In 2008 we had objections like, "My load balancers don't support it."  "My
firewalls don't support it."  None of them were good reasons to do nothing at all,
they were merely good reasons to leave the bits behind the firewalls and load
balancers on IPv4.  Solved problem now, lots of options, if your vendor doesn't do what
you need or tries to charge extra for v6 you can just tell them that you'll never
buy another thing from them again for as long as your company lives unless they
deliver, and make purchasing decisions appropriately thereafter.

In 2009 the big message was, "Can't do it, it's expensive."  Well, no.  It isn't.
It might be under a pretty small number of specific deployment modes, but experience
shows there are plenty of other deployment modes where it doesn't need to be
expensive at all, even for networks larger than yours. Totally optional problem,
you get to choose whether or not it applies to you, if you think this is important
just choose differently and get happy.  We've all known this for years, entire 
conferences have been held to emphasise it.

In 2010 it was, "I can't connect my eyeballs."  I like to think Internode played
a none-too-small part in solving that one.  Vendor support is there, the features
are available at both the BRAS end and the CPE end for anyone who wants them.  Done
and dusted.

So let's look at where we are today:

One by one, reasons for not doing IPv6 have been knocked down.  The original reasons
for doing it still exist.  They're actually worse now because APNIC has run out
and new addresses are expensive on the open market.  All the predictions are coming

But there are still objections to sale.

The difference between the previous objections and the current ones, though is 
the triviality of the latter.

We're actually down to, "I'm not doing IPv6 because I can't make the phones work
without IPv4."

Seriously?  Is that it?

We geeks like to see technical solutions to problems, and that's the way we've
approached this so far.  Each technical objection to sale has been beaten to death
by technical hammers, the state of play is so much better than it was ten years ago
that it's not funny.  Anyone who patiently explained why they weren't proceeding in
2003 has certainly had every single last one of their issues dealt with in the 
intervening decade.

But we geeks are also sometimes blind to non-technical issues, and I think I'm
starting to see what's going on here.  I think there is actually a major
non-technical reason holding back progress in this area:

You don't want to do it.

That's it.  It's that simple.  It explains why you aren't messing around with it
at home like you did with Linux boxes in the late 1990's.  It explains why there's
so much butthurt about the price of IPv4, given that you aren't visualizing a future
when IPv4 is considered obsolete.  It explains why you so desperately want to
believe that it's hard and expensive and mysterious and impossible, despite the
availability of so many people who've actually done it and tell you the complete
opposite.  It explains why someone like Don can say, "[complexity] is exactly
the stuff that's driving SMB geeks away," even though a defining attribute of being
a geek ever since Michelangelo dreamed of helicopters has been that they'll be 
energised and turned on my complexity.  It's why when the phones work on IPv6
you'll say you still can't deploy it because the MIT CoffeeCam is still on IPv4.
It's why it's 2013 and you haven't even deployed in in the lab. It's why Bevan
will donate $50k on training, and at the end of it half of you will use your 
newfound expertise and insight to figure out new reasons to not do it[1].

You just don't want to do it.

Now, I'm not being judgmental about that choice.  It's as valid as any other.  We
all place limitations in our lives around the things we want to get involved in and
the things we don't, it's normal human behavior.

But I think we need to stop crapping on about technology and complexity, and 
perhaps be a bit honest about what's going on here.  The technical objections
that are being raised are proxies for something more fundamental and psychological.
We can play whack-a-mole with them for as long as we like and it won't change
the reason underlying all of it:  There are some people on this list who, to 
varying extents, have adopted v6 already, and there are others here who,
consciously or otherwise, will try every trick in the book to avoid doing it.

We eat, sleep, live and breath complexity.  We use all manner of scripting 
languages, databases, network devices, operating systems, routing protocols,
arcane stuff that the typical person in the street (or Whirlpool nick) can't
ever hope to understand.  A lot of what we do doesn't work very well, and 
we apply effort towards making it work better.  Very few of us have ever used
"The implementation sucks!" as a reason for not doing something that needed to
be done.

We all earn our livelihoods from managing this complexity.

But not with IPv6.

Have any of you applied enough introspection to figure out why this particular
piece of internet technology, among all others, is the one that you don't seen
to want to touch?

  - mark

[1]  Thank you, Bevan -- the gesture is a worthy one regardless, even if only
     for the remaining half who will benefit from it.

More information about the AusNOG mailing list