IPv6 Deployment: Religion and Reality
Someone left the following comment on one of my blog posts a few days ago:
IPv6 to a network engineer is like Communism to a Marxist. It would come in such a distant future that it would be in a form we can barely picture accurately. […] So my money is on NAT444, at least in the US.
Meanwhile on planet Earth (in 2014):
- More than 2.5% of Internet users use IPv6 to reach Google and the growth seems to be exponential;
- 6.15% of US customers (and 11% of Swiss customers) use IPv6 to reach Google. APNIC reports a bit higher number (more than 7%)
- 60% of US transit autonomous systems run IPv6, and 40% of the US content (weighted based on Alexa index) is reachable over IPv6 (source: Cisco’s IPv6 statistics);
- Almost 20% of autonomous systems in RIPE region, and 13% of AS in ARIN region advertise IPv6 prefixes;
- 464XLAT is included in Android 4.4 and T-Mobile uses IPv6-only connectivity for mobile users with the new phones (464XLAT makes Skype work over IPv6-only networks).
We’re getting close to the point where content providers ignoring IPv6 will get penalized, and some engineers are bound to face the five stages of IPv6 grief. Do keep in mind that a proper IPv6 deployment takes anywhere from 6 months to 3 years.
How Do I Get Started?
I created numerous IPv6 webinars to help you design and deploy IPv6 in your network, and if you’re looking for free resources, you’ll find plenty of them on the Internet:
- I created several IPv6-focused presentations for Slovenian IPv6 summits;
- RIPE and ARIN have created numerous IPv6-specific documents;
- Internet Society (ISOC) has IPv6-specific Deploy360 program that includes tutorials, case studies and training.
We need a truly great IPv6 only app or service to really drive adoption,
Doesn't exist at all though, despite how flamingly obvious it is.
In every client I've looked at, when support is advertised at all it turns out to be either fictitious, or IPv4-only. This despite the ugly latency problems that going through a registrar introduces.
I find it more than a little suspicious, really. There seems to be absolutely phenomenal resistance on the part of client developers to doing SIP without a registrar, even though the not-inconsiderable bandwidth the registrars consume generally gets paid for either out of their own pockets or those of their sponsors. Just who the hell goes that far out of their way to insist that strangers take their money, whether they want to or not?
Nobody, that's who.
So I'm forced to consider that maybe there's something more going on here than meets the eye. It's not too hard to guess at it, either. Who would want to get in the middle of all those conversations? Who is known to have compromised SIP's biggest proprietary rival, Skype? Yeah.
Notice my communism analogy. Marxists (and Ivan's native Yugoslavia has a few prominent ones) are quite certain of its eventual arrival, but are hesitant to predict it in its details.
To be honest, I was merely parroting my teacher who said that a few months ago. He said that IPv6 will probably not become dominant in the US.