AppNeta | The Path

Posts Tagged ‘IPv6

The last time I wrote about IPv6 was almost a year ago.  And soooo much has changed since then!  Actually, no. Nothing has changed.   IPv4 address space exhaustion is still imminent (late 2011).  There still is very little happening to convince most network operators to make the requisite changes soon.  And there was that financial crisis thing…

IPv6 is still a train wreck in slow motion.

First off – the projected exhaustion date – Geoff Huston maintains a great page with the best projection model that I know of:

He makes regular adjustments as circumstances change.  Currently his models are pointing at roughly Oct 2012 when IANA runs out of IPv4 address space, and Nov 2012 when the RIRs start running out.  Those dates have moved out slightly in response to IANA’s change in allocation policy (go read his IPv4 report).

Under two years though – definitely under two years.  Not enough time to start and complete an IPv6 implementation plan, if you think you’ll be affected.  Who does not think they will be affected?  It is just a question of how much you will be affected.
Most of the IPv6 pundits have declared that words like “transition” no longer apply.  Partly because transition was never really an option – that is, we never really intended to replace IPv4 with IPv6.  Instead they will be coexisting from now forward.

These same pundits are unhappily declaring that the original plan to use dual stack as a transitional mechanism has officially failed.  Dual stack was certainly available early enough – but no one took the goal of IPv6 adoption seriously enough to accomplish it in time.  So that door has closed.

Instead, we are faced with a future filled with NAT (Network Address Translation) devices.  They will allow local networks to reuse IPv4 addresses behind a wall maintained by NATs which can translate them to IPv6 or other IPv4 address.  But worse than NATs at the edge is the prospect of NATs in the core – realizing that they have little choice, providers are now girding their loins with carrier-grade NATs that can partition their networks into sub-domains composed of re-used IPv4 address space.

For anyone that knows how the Internet was built and the principles of its founding fathers, this is pure heresy – for example, it shatters end-to-end and points at an increasingly fragile future for Internet expansion, one that is literally held together with paper clips and string.

For those of you who have not yet lost interest in this thread, there remains the question of why this ever happened in the first place.

Once again, Geoff Huston has written a particularly insightful piece in his blog that is definitely worth reading:

Geoff explores the prospect that IPv6 represents an instance of market failure.  That is, there never was any incentive to adopt IPv6 and it should never have been left up to the markets to promote it.  Instead, he suggests that it may instead have been better to treat IPv6 as a public good in the common interest.  In which case, there are a number of well-established means for ensuring that it is propagated.

It may all be just 20/20 hindsight.  However it makes for some very interesting reading (if you like economic analysis).

So what are you doing about IPv6?  At least go read Geoff’s blog.  He publishes once a month on various network issues.  They are invariably well written and researched articles.

Yawn.  Back to sleep now.

The Internet Engineering Task Force, the body responsible for defining most IP-related standards, has been making a concerted effort of late to promote the transition to IPv6.  They have developed Best Practices, staged IPv6-only events for media profile, and developed subsequent RFCs (standards) that help make IPv6 easier to adopt.

The latest RFC to be released takes a bold step forward by leveraging the pervasiveness of social networks.  RFC 5514, entitled “IPv6 over Social Networks”, has recently been released (April 1, 2009) and promises to catapult IPv6 into common use.  Although an experimental RFC (as opposed to a full-fledged standard), this new work offers a breakthrough implementation.

The sole author of the RFC is Eric Vyncke of Cisco Systems (Belgium).  Not always the source of network innovation, Cisco clearly has taken a lead in driving IPv6 into the public consciousness and the Internet.  While as a standard it is available for all vendors to implement, it is hopeful that Cisco itself will continue to lead with some interesting new products that employ this RFC.

Like most standards, this one is a bit of a challenge to read.  However, it is well worth perusing to get a sense of what is in store:

If the arrival of IPv6 can be likened to “climate change for networks”, then Google has become the corporate incarnation of “Al Gore for the Internet”.  Not only does the search engine giant have an apparent interest in furthering the adoption of IPv6, but it has taken steps to make sure everyone else does too.

Last week, Google held its IPv6 Implementers conference in San Jose.  It had a limited attendance, was paid for by Google, and was aimed at technical folks who had specific interest in how, less if or when, to implement IPv6.  Speakers offered views from companies such as Google, Hurricane Electric, BechTel, Comcast, Cisco, Microsoft, Beijing Internet Institute, and NTT America.  Our own IPv6 researcher Dr. Vilcu attended and was impressed by everything – from the people, to the content, to Google’s cafeteria food.

While the some of the usual “terrifying” stories of IPv4 depletion were being passed around (including updated predictions IANA will run out of IPv4 addresses as early as September 2010), there were more reassuring stories suggesting that the transition IPv6 is not as hard as some feared.  For example, Google’s Lorenzo Colitti described their transition as relatively care-free and low cost.

On the other hand, there were disappointingly few details provided for critical how-tos such as addressing plans and network management.  It was clear that IPv6 still has not reached the mainstream consciousness.  On a third hand, Colitti’s demonstration of IPv6 on Android hinted at Google’s interest in mobile networking, the only real contender for “killer app” for IPv6.

The following week, Colitti appeared again at the Internet Society sponsored “Seven Stages of IPv6 Denial” panel featured at the Internet Engineering Task Force conference in San Francisco.   The IETF meetings develop and advance a broad range of protocols like IPv6 – however, IPv6 has been increasingly in the spotlight 0ver the last few years.  Even the attractive conference T-shirt, bearing a faux concert tour motif,  was used to promote IPv6 over IPv4 (referred to as “sold out”).

Colitti shared the stage with well-known faces from Comcast, IETF, ARIN, and Ericsson, amongst others.  And he repeated many of the same messages – and others such as a claim that there was a business case for IPv6 however indirect the revenue might be.  The others relied on familiar themes that did not add anything new for those already in the choir.

With little or no persuasive economic impetus to transition to IPv6 (it is primarily being described as a strategic move), at least companies like Google are stepping up to give people reasons to move earlier.  Spreading the knowledge seems to be their intent with the implementer’s conference.  And by offering an IPv6-specific search portal, they have raised its visibility.   Some companies have made similar contributions to the cause – for example, Microsoft is limiting its slick new DirectAccess capability in Windows 7 to being IPv6-only (DA offers users secure intranet access with0ut using a VPN).

[Note:  On the other hand, is not yet IPv6 accessible, and all indications are that home systems like XBox will not be running IPv6 any time soon.]

The general noise these days sounds like a pack of sheep milling about discussing how good the grass is over on the other side of the hill.  Even as the last of the grass is being finished off where they are, few seem inclined to  make the hike.  What is notable th0ugh is that Google clearly has some vested interest in the pack making the move sooner than later and in an orderly fashion.  It remains to be seen if it is a visionary sheep or a dapper wolf.

Or perhaps it simply wants to be there when the first early adopters arrive – as part of its goal of world domination without doing evil.

Certainly every little bit helps – but what would be really interesting is a  IPv6-only Day where access to Google’s indispensable search engine is briefly cut-off for IPv4 users.  Sort of a warning shot over the world’s bow…..  now that would get some attention.

Links of Interest:

So IPv6 – now there’s a train wreck in slow motion that has been simultaneously fascinating and utterly boring.  Every six months the media checks back to whoop and holler about how now there are only 32 /8s remaining in the IPv4 address space!  We are running out of Internet!  Transition to IPv6 before it’s too late!  Run for your lives!

And then everyone goes back to sleep for another six months.

I’m sure every CIO and network operations manager has half an eye on the IPv6 situation.  Wondering when it is supposed to become important to them.  If ever.  Unfortunately, it is important and it is simply hard to tell why.  Or what to do about it.

In that respect, I tend to think of IPv6 as “climate change for the Internet” – it affects everyone, it is happening whether you believe in it or not, what is actually happening is really hard to tell, and wild rumors are flying thick and heavy about what the impact will be.  At least we don’t have to deal with “IPv6 deniers”.  But otherwise, the implications are about the same.

One way or the other, everyone needs to make personal and organizational decisions about how to deal with it.  For most, the question is what and when.  For some, it is also why.  In the end, it is nearly impossible to predict the outcomes exactly – but none of the forecasts are pleasant.

The general wisdom arising from front-line groups like the Internet2 and NANOG communities is that the typical organization will need about 24 months to make an effective cut-over from IPv4 to whatever blend of IPv4 and IPv6 makes sense.  Two years.  That’s a pretty long time to execute a cut-over strategy.  Maybe too long insofar as it might give sluggards the impression that they still have time for a nap.  But if you work backwards from the projected date of the IPv4 Exhaustion Event (I4EE), there is a little over 2 years left before there simply aren’t ANY IPv4 addresses left to give out.

Waiting until then will be a bit like running out of gas in the middle of the Nevada desert.  In this day and age, you are unlikely to die but you will be plenty uncomfortable and it will likely cost you a lot more in money and time than you might otherwise choose.  But maybe I’d be better off scaring you with the prospect of imminent death and destruction.

Run for hills!  The worms are stampeding!  The ice cream is melting!  My coffee is cooling!

No – it is not really compelling.  Somehow, you have to convince yourself that there are series of inevitable steps to be taken.  And you had better plan to take them while you can do it effectively.  So what are they?

  1. Evaluate your situation – take the time to figure out what IPv6 means to your business, look at where the impacts will be, and determine the requirements, in order of priority.
  2. Develop a transition plan – in order of priority, set milestones for the transition of the key parts of your network, budget for them, and put a team together to execute it.
  3. Execute the plan.
  4. Undertake regular reviews – at each major milestone, reflect on progress to date in the transition, accommodate new information or unexpected complications, and update your go-forward plan.
  5. Do not hesitate.  Do not procrastinate.  Do not wait around for someone else to fix the problem for you.

And if somehow you still don’t know what the problem is, or why it matters to you, do the research.  Now.  Read our whitepaper for starters.  Check out the various information sites.  Talk with early adopters like DREN or Boeing to understand the issues.

According to the quasi-official IPv4 Exhaustion Counter, we have just over 2 years left and counting.  Consider this your last warning before it becomes increasingly costly to get off the train before it hits the wall…. albeit in slow motion.

Check it out:

Follow us on Twitter!

Error: Twitter did not respond. Please wait a few minutes and refresh this page.

%d bloggers like this: