Are you sure?
https://www.google.com/intl/en/ipv6/statistics.html
https://www.facebook.com/ipv6/?tab=ipv6_total_adoption
https://stats.labs.apnic.net/ipv6/IN
https://radar.cloudflare.com/embed/IpVersionBubble?location=in&chartState={}
Are you sure?
https://www.google.com/intl/en/ipv6/statistics.html
https://www.facebook.com/ipv6/?tab=ipv6_total_adoption
https://stats.labs.apnic.net/ipv6/IN
https://radar.cloudflare.com/embed/IpVersionBubble?location=in&chartState={}
Holland drops…
https://stats.labs.apnic.net/ipv6/NL
But they state capable, not if it’s used.
Minor difference.
Facebook is mostly used by people with phones and other small devices, I do not see how them using IPv6 is relevant.
Also, when you change the map to per country the numbers sink to about 50%.
Not very good for a protocol that is over 30 years old.
Perhaps that is because the outcome wouldn’t be convenient for you? Just like the statistics from Belgium probably don’t suit you.
The overall statistics show, without question, that there is a considerable potential usage of IPv6.
That justifies some changes to the NTP pool (namely AAAA records everywhere instead of just on 2.pool.ntp.org), to match the actual situation better.
And that’s all I’m advocating for.
So, what I’m asking you (and the community as a whole) in this thread is not for your personal opinions about IPv6—because IPv6 is here, and that’s a fact—but rather to objectively discuss the technical merits or disadvantages of my proposal, based on the understanding that we all prioritize the importance of the NTP pool.
Or not for you
As the figures show 50% acceptance…one would expect it to be the other way round.
Most people do not know they run dual-stack and the dual-stack websites like Google, Facebook, etc just count their own hits.
It doesn’t say anything about IPv6 except their requests.
Then 50% is a pretty low number.
I’m kindly asking you to spare me your opinions, and to instead help the NTP pool community further with technical knowledge and facts.
What will break if we add more AAAA records besides the one on 2.pool.ntp.org
?
(Something that is necessary both to address the potential CGNAT issues discussed elsewhere in this thread and to fully unleash the potential of the vast number of IPv6-capable NTP servers in the pool.)
Do we know why this is still not done yet?
I guess only @ask does.
One reason in my understanding is that some issues that were seen with IPv4 should be addressed before IPv6 is enabled more widely to avoid the issues being replicated there. Don’t have the thread at hand, but I once made a summary of my understanding what Ask’s task list is before he would get to enabling IPv6, and others chimed in with additional items (and views as to how to maybe strip down/streamline/simplify the list a bit).
E.g., it seems some of the vendor zones are particular about whether they are ok to get IPv6 or not, so improvements on the self-mangement side for vendor zones were on the list to avoid the currently mostly manual admin-side process of managing vendor zones.
Edit: It was actually earlier in this very thread that I posted my summary.
If you did program, you’d know that one can write code exclusively for IPv6 and the kernel would encapsulate IPv4 addresses as IPv6 ones, as explained in RFC 3493. Then again, you obviously never read an RFC or anything on IPv6, as your rants plainly demonstrate.
Modern cellular protocols from 4G LTE to 5G NR use only IPv6.
The US Government has removed IPv4 from 50% of their networks, and will have 80% moved to IPv6-only by next year. That’s a population of 2.95 million people.
My home ISP only assigns public IPv6 addresses, with IPv4 reachable only via CGNAT. We’re IPv6-only at work, too. It wasn’t worth the complexity of running dual-stack.
With IPv4 addresses now costing USD$7,000 to USD$11,000 for a /24 (the smallest block that can be routed via the Internet), new ISP’s are essentially priced out of the IPv4 market. IPv6 addresses are essentially free.
I understand you, but this is not the case in Europe.
I have Static IPv4 and 6 here at home.
All my VPS servers also have static at no extra costs.
A static IPv4 costs near to nothing extra.
Even at home, like I have, no extra charge.
It is.
One cannot simply obtain IPv4 addresses anymore. There is a waiting list. If someone needs them, like when they want to start a new ISP, they either have to buy them on the second hand market or lease them (or wait for almost two years and still be out of luck, most likely).
Is that still a thing? I can understand 10+ years ago when the fear of the new was still running high.
See the part about “views as to how to maybe strip down/streamline/simplify the list a bit”…
Please explain why companies like Microsoft still can have 993 IPv4 networks, this is just 1 company.
Why isn’t anybody taking action against this misuse of addresses while others suffer.
If they stop this abuse of IPv4, there are plenty addresses.
Different discussion. But it’s wrong.
Advantage of the first-comer.
When IPv6 wins that IPv4 is switched off then having IPv4 address will have no value anymore. All this inequality is going to vanish.
I guess you forgot that they are also a service provider, with data centers in many countries and provide cloud / Azure, mail hosting, etc. to thousands of companies / organisations.
That will never happen. After 30 years IPv4 is still the dominant protocol.
So they have to change/reinvent IPv6 to make it replace IPv4.
The way it is now, I do not see it ever replace IPv4.
No I didn’t, but they have 16 mil adresses.
Not only them, MIT or IBM too,
I do not believe they use 50% of the IP’s in the USA. As they have 50% reserved.
Maybe it’s expensive there, but over here, they do not bother to activate IPv6 by default.
Not do they promote for you to use it.
And if you order a VPS, you get a IPv4 static for free with it.
I do know the prices went up in Holland, but not in Germany or Belgium.