1

If I host a service over IPv6 only, what percent of clients will be able access it?

I know about these measures of roll-out among ISPs: http://www.worldipv6launch.org/measurements/

And Google's measure of IPv6 access among users: https://www.google.com/intl/en/ipv6/statistics.html

What are the estimates for when IPv6 access becomes (close to) universal among commercial and residential clients?

  • 2
    serverfault.com/q/421445/126632 Nothing has changed, you should still dual stack. – Michael Hampton Sep 21 '14 at 18:11
  • @MichaelHampton Uptake appears to have doubled since. However, the more pertinant part of my question is how long before IPv6 access becomes universal among clients? – Oliver Moran Sep 21 '14 at 18:13
  • 1
    You already know the answer to the first question. The answer to the second question requires a psychic. – Michael Hampton Sep 21 '14 at 18:15
  • 2
    My estimate is between years and decades YMMV. – user9517 Sep 21 '14 at 18:20
  • 1
    I think as soon we definitely ran out of IPv4's (with IANA not having any from their "give-back" pool, with the regional registries running out of them, and with the first ISP's running out), we will see an uptake in global deployment. – MichelZ Sep 21 '14 at 19:31
3

This is impossible to predict. A few big players enabling IPv6 can make a huge difference. Take a look at the historical graph of IPv6 user availability in Belgium. Things can progress quickly, but such jumps are hard to predict.

It also depends on where your audience is. If you only target Belgian users then ±30% can access an IPv6-only website today. The chance of Belgium reaching 90% is greater than some other country doing the same, unless there is only a very limited number of ISPs in that country and they all enable IPv6. It all depends on ISPs providing IPv6 and CPEs getting replaced by IPv6 capable devices.

If your audience is the whole world then it is going to take many years before IPv6-only becomes an option...

| improve this answer | |
  • 4
    PS: I am very unhappy having to give this answer. I have been working with many companies to implement IPv6 but many ISPs just aren't doing their job and avoid taking responsibility :( – Sander Steffann Sep 21 '14 at 23:32
3

The best model for this sort of growth is logistic growth. Last year I did some calculations on how growth would have had to look in order to complete the transition in time. I.e. completing the transition between the time the IPv6 standard was finalized and the time IPv4 addresses ran out.

Comparing that to actual deployments told me, that we were 13 years behind schedule. A year later I checked again, and we still appeared to be 13 years behind schedule. If we keep being 13 years behind that graph, dual stack will overtake IPv4-only by 2018, and we'll reach 99% in 2023. Those are currently my best estimates.

| improve this answer | |

Not the answer you're looking for? Browse other questions tagged or ask your own question.