IPv4? That Will Cost You


After my recent articles on Network Computing, I got an email from Fred Baker.  To say I was caught off guard was an understatement.  We proceeded to have a bit of back and forth about IPv6 deployment by enterprises.  Well, it was mostly me listening to Fred tell me what he sees in the real world.  I wrote about some of it over on Network Computing.

One thing that Fred mentioned in a paragraph got me thinking.  When I heard John Curran of ARIN speak at the Texas IPv6 Task Force meeting last December, he mentioned that the original plan for IPv6 (then IPng) deployment involved rolling it out in parallel with IPv4 slowly to ensure that we had all the kinks worked out before we ran out of IPv4 prefixes.  This was around the time the World Wide Web was starting to take off but before RFC 1918 and NAT extended the lifetime of IPv4.  Network engineers took a long hard look at the plans for IPv6 and rightfully concluded that it was more expensive to run IPv6 in conjunction with IPv4 and instead it was more time and cost effective to just keep running IPv4 until the day came that IPv6 transition was necessary.

You’ve probably heard me quote my old Intro to Database professor, Dr. Traci Carte.  One of my favorite lessons from her was “The only way to motivate people is by fear or by greed.”  Fred mentioned that an engineer at an ISP mentioned to him that he wanted to find a way to charge IPv4 costs back to the vendors.  This engineer wants to move to a pure IPv6 offering unless there is a protocol or service that requires IPv4.  In that case, he will be more than willing to enable it – for a cost.  That’s where the greed motivator comes into play.  Today, IPv6 is quickly becoming equivalent in cost to IPv4.  The increased complexity is balanced out by the lack of IPv4 prefixes.

What if we could unbalance the scales by increasing the cost of IPv4?  It doesn’t have to cost $1,000,000 per prefix.  But it does have to be a cost big enough to make people seriously question their use of IPv4.  Some protocols are never going to be ported to have IPv6 versions.  By making the cost of using them higher, ISPs and providers can force enterprises and small-to-medium enterprises (SMEs) to take a long hard look at why they are using a particular protocol and whether or not a new v6-enabled version would be a better use of resources.  In the end, cheaper complexity will win out over expensive ease.  The people in charge of the decisions don’t typically look at man-hours or support time.  They merely check the bottom line.  If that bottom line looks better with IPv6, then we all win in the end.

I know that some of you will say that this is a hair-brained idea.  I would counter with things like Carrier-Grade NAT (CGN).  CGN is an expensive, complicated solution that is guaranteed to break things, at least according to Verizon.  Why would you knowingly implement a hotfix to IPv4 knowing what will break simply to keep the status quo around for another year or two?  I would much rather invest the time and effort in a scaling solution that will be with us for another 10 years or more.  Yes, things my break by moving to IPv6.  But we can work those out through troubleshooting.  We know how things are supposed to work when everything is operating correctly.  Even in the best case CGN scenario we know a lot of things are going to break.  And end-to-end communications between nodes becomes one step further removed from the ideal.  If IPv4 continuance solutions are going to drain my time and effort they become as costly (or moreso) that implementing IPv6.  Again, those aren’t costs that are typically tracked by bean counters unless they are attached to a billable rate or to an opportunity cost of having good engineering talent unavailable for key projects.

Tom’s Take

Dr. Carte’s saying also included a final line about motivating people via a “well reasoned argument”.  As much as I love those, I think the time for reason is just about done.  We’ve cajoled and threatened all we can to convince people that the IPv4 sky has fallen.  I think maybe it’s time to start aiming for the pocketbook to get IPv6 moving.  While the numbers for IPv6 adoption are increasing, I’m afraid that if we rest on our laurels that there will be a plateau and eventually the momentum will be lost.  I would much rather spend my time scheming and planning to eradicate IPv4 through increased costs than I would trying to figure out how to make IPv4 coexist with IPv6 any longer.

2 thoughts on “IPv4? That Will Cost You

  1. Yikes- who sets the costs? The government? God help us. I think its an interesting angle and agree with the need for creative motivators, but too much is taxed and fee’d already in a society that seems to funnel money back to the same pockets over and over. And those pockets tend to be worn by sleazebag politicians that want to dictate how to do things they have no clue about. Burn this blog post, and we must never speak of it again.

  2. This assumes that what you get with IPv4 is equal to what you get with IPv6 today, but it isn’t. If I deploy IPv6 I only connect to a small fragment of the total internet. If I deploy IPv4 I connect to 99.99% of the internet.

    I agree that thinking economically is the right way, but it should have been started sooner.

    The REAL problem with IPv4 address was how it was distributed, each according there need pseudo-communist style. That causes a finite resource to be used inefficiently. Look how much black is on this map http://internetcensus2012.bitbucket.org/hilbert.html

    And that map matches my enterprise experiences. Lots of IPv4 addressing isn’t being used, but hoarded. Because of acquisitions many of the largest enterprises have huge blocks of unused IPv4 addresses. I know of one major enterprise that just in the past 18 months took the effort to remove public IPv4 addresses from it’s printers. (and this is one that should know better)

    This situation favors existing players and limits new ones. Which is the real problem we should be addressing, which has many parts to the solution which will likely include IPv6.

    But lets make sure we are keeping the real need (how do new people / institutions connect to the internet with finite addressing) in focus and not confuse part of the solution (IPv6) with the need.

    I suggest there be an open marketplace for IPv4 addressing. This will give incentives to IPv4 address hoarders to release their addresses, and eventually, as real IPv4 usage increases (not just allocations) this will in turn raise the price (in a market fashion) of IPv4 address making IPv6 more attractive.

    The cart (IPv6) has been put before the horse (Internet access to newcomers) for too long. IPv6 is the ‘perfect’ virtually infinite solution that is the enemy of the good.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s