Yeah that was my thinking, like now my parents as the example I keep using sorry, probably hit google on ipv6, not because they "Adopted it", because their ISP enabled it.
I suppose if you zoom out, and forget individuals, that counts as adoption because their ISP adopted it, but passive inclusion vs active adoption are different.
As a regular user you cannot "adopt" IPv6. You can go out of your way to "disable" it, but when it comes to adoption, I specifically mean high level adoption, specificly, on the level of ISPs and datacenters who need to invest in hardware costs to adopt IPv6.
So ya, i think your 30 to 50 years is a more reasonable estimate compared to my 10~20 years...
20% within the past 5 years on that graph, though.
adoption rate is not linear, so in my opinion, it would take 30+ years, unless some big giant steps in to say we are switching fully to IPv6 and you can only access our slower servers over IPv4.
You're correct, it's not linear, but that actually goes against your argument - the last 5 (or rather some 10) years show an acceleration in adoption, not a slowdown.
In the first 5-10 years, we saw about 5% growth, in the last 10 years, around 45% growth.
In my knowledge, first 10 years, since 1999 were the innovators/early-adopters.
Since 2017 to now, we are in the early-adopters/early-majority portion, where you see increased growth. It is expected to stall and fall rapidly in near future.
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
@somik said: As a regular user you cannot "adopt" IPv6. You can go out of your way to "disable" it, but when it comes to adoption, I specifically mean high level adoption, specificly, on the level of ISPs and datacenters who need to invest in hardware costs to adopt IPv6.
yeah I guess, its probably silly of any of us to argue the future of IPv6 based on the meaning of a probably badly chosen word or a at least one without a qualifying operator.
The UK is adopting digital ID, I personally am not
@somik said: As a regular user you cannot "adopt" IPv6. You can go out of your way to "disable" it, but when it comes to adoption, I specifically mean high level adoption, specificly, on the level of ISPs and datacenters who need to invest in hardware costs to adopt IPv6.
yeah I guess, its probably silly of any of us to argue the future of IPv6 based on the meaning of a probably badly chosen word or a at least one without a qualifying operator.
Yea, it'll end up with some powerful politician pushing for IPv6 and everyone following to adopt. Until thæn, IPv4 is king, IPv6 is the jester
@backtogeek said:
The UK is adopting digital ID, I personally am not
Singapore has fully adopted digital ID and in most cases, your phone is all you need to verify your identity on most physical places (hospitals, police, office, etc.), including on many local websites.
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
Yeah that was my thinking, like now my parents as the example I keep using sorry, probably hit google on ipv6, not because they "Adopted it", because their ISP enabled it.
But with that logic, they didn't "adopt" ip4 either.
I do not know your parents, but I'm guessing they have no idea if they run ip4 or ipv6, and they do not care either. And this is true for the vast majority of users, they do not know and they do not care. If their ISP enables ipv6, they will happily adopt it as long as Netflix or Youtube shows up on their screen. If someone told them they need to run ipv11 to get to Instagram they would happily "adopt" that too.
We, as computer literate people, tend to forget that we are not the norm. 99% of users do not know what an ip address is, and they do not care. Ipv6 adoption happens when the big players, carriers and ISP's, decides that it will happen - it is not up to the users. Most users will never demand ipv6 in the same way that they never demanded ip4, they just want a functioning service.
The big players on the other hand know that ipv6 will happen so adoption started years ago. In many countries over half of the traffic is already ipv6, and in some it's as high as 70 or even 80%.
Small home LAN's will probably run ip4 for a long time to come because it works and it is what people know, but internationally ipv6 has already surpassed ip4 in many places, so it's kind of hard to talk about "adoption rate" without specifying exactly what that means.
@rcy026 said: But with that logic, they didn't "adopt" ip4 either.
well, yes, that's a fair point.
@rcy026 said: We, as computer literate people, tend to forget that we are not the norm. 99% of users do not know what an ip address is, and they do not care. Ipv6 adoption happens when the big players, carriers and ISP's, decides that it will happen - it is not up to the users.
Also good point well made, i also am guilty of mixing up "general people" with "hosting customer" people of all shapes, because again, even companies that order services will select ipv6 = yes when ordering even if they dont know what it is because its free, add $1 and I am sure the "adoption" at that specific level, would be far less, in the same way I am SURE if all ISP's said "Add IPv6 for only $1 per year" the individual adoption... or lets say "use" would be FAR less and that would directly hit the google chart square between the eyes... not that I think they are trying to prove anything with it I suppose. it just is what it is.
@rcy026 said: But with that logic, they didn't "adopt" ip4 either.
well, yes, that's a fair point.
@rcy026 said: We, as computer literate people, tend to forget that we are not the norm. 99% of users do not know what an ip address is, and they do not care. Ipv6 adoption happens when the big players, carriers and ISP's, decides that it will happen - it is not up to the users.
Also good point well made, i also am guilty of mixing up "general people" with "hosting customer" people of all shapes, because again, even companies that order services will select ipv6 = yes when ordering even if they dont know what it is because its free, add $1 and I am sure the "adoption" at that specific level, would be far less, in the same way I am SURE if all ISP's said "Add IPv6 for only $1 per year" the individual adoption... or lets say "use" would be FAR less and that would directly hit the google chart square between the eyes... not that I think they are trying to prove anything with it I suppose. it just is what it is.
All true. But the opposite is also true, if all ISP's said "Add ip4 for only $1 per year" we would have close to 100% ipv6 adoption within a year. It really is not up to the users, they do not know and they do not care. Trying to put the decision on the users is like asking your 3-year old what kind of oil to put in your car - they may ride in the car but they have no idea what makes it work.
The people that argue that "customers do not ask for ipv6" is not making sense, because customers do not ask for ip4 either. They do not ask for ip at all, they ask for Tiktok or Insta or streaming or whatever it is average people use the internet for. Once all "the big ones" get to ipv6 and the carriers and ISP's realize they do not need to hand out expensive ip4 to every customer to get Tiktok or Insta to work, I think we will see a rapid increase in ipv6 only networks. To some extent, we are already there.
So ya, i think your 30 to 50 years is a more reasonable estimate compared to my 10~20 years...
20% within the past 5 years on that graph, though.
adoption rate is not linear, so in my opinion, it would take 30+ years, unless some big giant steps in to say we are switching fully to IPv6 and you can only access our slower servers over IPv4.
You're correct, it's not linear, but that actually goes against your argument - the last 5 (or rather some 10) years show an acceleration in adoption, not a slowdown.
In the first 5-10 years, we saw about 5% growth, in the last 10 years, around 45% growth.
In my knowledge, first 10 years, since 1999 were the innovators/early-adopters.
Since 2017 to now, we are in the early-adopters/early-majority portion, where you see increased growth. It is expected to stall and fall rapidly in near future.
Expected by whom?
And what's this graph even about? It makes absolutely no sense - IPv6 usage doesn't just fall. It may continue to grow more slowly and steady at some point, but fall? I think you've got it all wrong.
Based on the link we discussed, the IPv6 adoption trend is clearly on the rise. If this trend (an increase of 20% over the last 5 years to reach 50% adoption) continues, and there's no reason to believe it won't, then in the next 5 years we'll likely see another 20% increase, reaching around 70%. After that, the trend might slow down as we approach the remaining 10% of users.
Google graph show you next
2010 - 2016 10%
2017 15%
2018 20%
2019 25%
2020 30%
2021 33%
2022 35%
2023 40%
2024 45%
2025 50%
So based on that, something like this is more realistic projection:
2026 55%
2027 60%
2028 65%
2029 70%
2030 75%
2031 80%
2032 85%
2033 90%
... and then slowing a bit
(Speed above is based on the trend from the last 5 - 10 years. Growth may not be exactly the same in the next 5 - 10 years, but there is no fall in adoption. It may rise slower or faster, but it can’t fall, as your graph wrongly shows)
Expected by whom?
And what's this graph even about? It makes absolutely no sense - IPv6 usage doesn't just fall. It may continue to grow more slowly and steady at some point, but fall? I think you've got it all wrong.
Sorry, my bad. The graph I used requires some background knowledge of engineering concepts and isn't clear for general public to understand.
The graph shows adoption rate.
So it shows that initially the number of adopters per month start going up, thæn stalls and thæn there are less adopters per month once it reaches around 50% and starts to fall. So the numbers you provided, yes, this graph shows exactly that.
This graph should be easier to understand since it shows both the adoption rate (same as the top graph) & the total users (what most people understand):
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
@rcy026 said: All true. But the opposite is also true, if all ISP's said "Add ip4 for only $1 per year" we would have close to 100% ipv6 adoption within a year
That’s not how protocol adoption works. IPv6 uptake isn’t blocked by IPv4 pricing, it’s blocked by legacy hardware, software compatibility, and operational inertia. Slapping a $1 fee on IPv4 wouldn’t magically rewrite decades of infrastructure.
Lets say I run a small–mid-size regional ISP serving between 5,000 and 50,000 customers in a non first world country.
Like most ISPs, I own my own hardware and I have invested $ 500k into my hardware infrastructure. I am fully IPv4 compatible, but my hardware is old and does not support IPv6.
Looking at the market, I need to invest somewhere between $ 200k and $ 400k to upgrade my hardware to be IPv6 compatible. Here is a approximate breakdown of my costs:
Item
Cost
Core routers (2–3 units)
$50,000–$200,000 each
Aggregation switches (10–50 units)
$2,000–$10,000 each
Customer-edge / CPE replacement
Optional, but $50–$150 per customer
Engineering, testing, migration
$20,000–$100,000
From my point of view,
1. My customers are happy that they can access the internet. They can play games and do everything they want.
2. Upgrading to IPv6 is a huge cost and headache, not to mention potential downtime
3. My team is not trained to manage IPv6 network, so I need to send them in for further training, meaning more money spent.
4. I do not see any way to profit or earn more money from upgrading to IPv6.
So tell my why I would invest hundreds of thousands of dollar into something like IPv6?
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
@rcy026 said: All true. But the opposite is also true, if all ISP's said "Add ip4 for only $1 per year" we would have close to 100% ipv6 adoption within a year
That’s not how protocol adoption works. IPv6 uptake isn’t blocked by IPv4 pricing, it’s blocked by legacy hardware, software compatibility, and operational inertia. Slapping a $1 fee on IPv4 wouldn’t magically rewrite decades of infrastructure.
I do not think hardware or software is a big problem, ipv6 has been supported by most vendors for decades.
I do agree that operational inertia is a big problem, I would simple call it laziness.
@somik said:
Lets say I run a small–mid-size regional ISP serving between 5,000 and 50,000 customers in a non first world country.
Like most ISPs, I own my own hardware and I have invested $ 500k into my hardware infrastructure. I am fully IPv4 compatible, but my hardware is old and does not support IPv6.
Looking at the market, I need to invest somewhere between $ 200k and $ 400k to upgrade my hardware to be IPv6 compatible. Here is a approximate breakdown of my costs:
Item
Cost
Core routers (2–3 units)
$50,000–$200,000 each
Aggregation switches (10–50 units)
$2,000–$10,000 each
Customer-edge / CPE replacement
Optional, but $50–$150 per customer
Engineering, testing, migration
$20,000–$100,000
From my point of view,
1. My customers are happy that they can access the internet. They can play games and do everything they want.
2. Upgrading to IPv6 is a huge cost and headache, not to mention potential downtime
3. My team is not trained to manage IPv6 network, so I need to send them in for further training, meaning more money spent.
4. I do not see any way to profit or earn more money from upgrading to IPv6.
So tell my why I would invest hundreds of thousands of dollar into something like IPv6?
Just out of curiosity, what kind of hardware you running that does not support ipv6?
@AuroraZero said:
That’s not how protocol adoption works. IPv6 uptake isn’t blocked by IPv4 pricing, it’s blocked by legacy hardware, software compatibility, and operational inertia. Slapping a $1 fee on IPv4 wouldn’t magically rewrite decades of infrastructure.
I do not think hardware or software is a big problem, ipv6 has been supported by most vendors for decades.
I do agree that operational inertia is a big problem, I would simple call it laziness.
As far as I know, most hardware vendors started IPv6 support only between 2011 to 2015. So anything made in the last "decade" should support IPv6. Anything made in the last 15 years, might support it, might not.
However in third world countries, ISPs buy older hardware that other companies are discarding, they end up buying obsolete hardware that are definitely not made in the last decade. You'll be lucky if they buy 20 year old hardware. It could be even older. So IPv6 support is not expected to be there.
Remember that countries exist where the annual income can be bellow 3k (bangladesh, india, pakistan, etc). The ISP there is not going to invest millions to buy new hardware where they can spend only 10% of the money buying 20 to 30 year old hardware that "still works". If not, they can repair it and make it work.
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
@rcy026 said: All true. But the opposite is also true, if all ISP's said "Add ip4 for only $1 per year" we would have close to 100% ipv6 adoption within a year
That’s not how protocol adoption works. IPv6 uptake isn’t blocked by IPv4 pricing, it’s blocked by legacy hardware, software compatibility, and operational inertia. Slapping a $1 fee on IPv4 wouldn’t magically rewrite decades of infrastructure.
I do not think hardware or software is a big problem, ipv6 has been supported by most vendors for decades.
I do agree that operational inertia is a big problem, I would simple call it laziness.
You're overlooking a major constraint. A lot of industrial and legacy systems simply can't be upgraded to IPv6 without serious risk. It's not laziness, it's about stability. Replacing hardware or rewriting firmware in environments like manufacturing or healthcare isn't trivial. The cost and potential downtime make IPv4 the safer choice for many. Adoption isn't just about availability, it's about operational risk.
@rcy026 said: All true. But the opposite is also true, if all ISP's said "Add ip4 for only $1 per year" we would have close to 100% ipv6 adoption within a year
That’s not how protocol adoption works. IPv6 uptake isn’t blocked by IPv4 pricing, it’s blocked by legacy hardware, software compatibility, and operational inertia. Slapping a $1 fee on IPv4 wouldn’t magically rewrite decades of infrastructure.
I do not think hardware or software is a big problem, ipv6 has been supported by most vendors for decades.
I do agree that operational inertia is a big problem, I would simple call it laziness.
You're overlooking a major constraint. A lot of industrial and legacy systems simply can't be upgraded to IPv6 without serious risk. It's not laziness, it's about stability. Replacing hardware or rewriting firmware in environments like manufacturing or healthcare isn't trivial. The cost and potential downtime make IPv4 the safer choice for many. Adoption isn't just about availability, it's about operational risk.
I can assure you I'm not overlooking anything. Helping manufacturing/heavy industries, healthcare and governments modernize their networks have been part of my job for 10-15 years now.
My experience is that running 15-20 year old firmware and hardware is a much bigger risk than upgrading it, and since my customers pays millions to get it done they seem to think so to. Actually, industries and healthcare are pretty good when it comes to keep up to date, which is understandable since a stopped steel plant can cost millions to restart and a stopped hospital could literally cost human lives. They do not run 20 old year firmware or software, and in the rare cases that they do, it's in extremely controlled environments securely isolated.
Of course, ipv6 compatibility is not what drives them, it's more of a bi-product. Security is one of the major reasons they do not run old equipment, as in the case of healthcare and especially hospitals they are often required by law to stay up to date. But since they keep up to date with just about everything, ipv6 just comes natural.
We should keep in mind that things might look a bit different in different parts of the world. I work in EU (mostly Germany, Scandinavia and UK) so I can only speak about what it looks like here.
@somik said: Sorry, my bad. The graph I used requires some background knowledge of engineering concepts and isn't clear for general public to understand.
Sorry, mate, but that's clearly not how stats work, not even basic elementary-school statistics. I'm getting the impression that you used some random tech graph that has nothing to do with IPv6 adoption and are trying, through some strange maneuvers, to generalize it to IPv6 adoption.
You still haven't answered: Expected by whom? Can you at least answer this?
I wanted to know because I'm curious whether you're just talking out of your ass or if there's actually some indicator for this.
Not to mention that the point at which adoption slows down makes a difference - again, it's on the rise, increasing by 20% in just the last five years solely.
What makes you think that within the next 5 years IPv6 won't grow by another 20%, or within the next 10 years by, say, 35%, making it the de facto dominant protocol with 85% adoption?
@AuroraZero said:
That’s not how protocol adoption works. IPv6 uptake isn’t blocked by IPv4 pricing, it’s blocked by legacy hardware, software compatibility, and operational inertia. Slapping a $1 fee on IPv4 wouldn’t magically rewrite decades of infrastructure.
I do not think hardware or software is a big problem, ipv6 has been supported by most vendors for decades.
I do agree that operational inertia is a big problem, I would simple call it laziness.
As far as I know, most hardware vendors started IPv6 support only between 2011 to 2015. So anything made in the last "decade" should support IPv6. Anything made in the last 15 years, might support it, might not.
However in third world countries, ISPs buy older hardware that other companies are discarding, they end up buying obsolete hardware that are definitely not made in the last decade. You'll be lucky if they buy 20 year old hardware. It could be even older. So IPv6 support is not expected to be there.
Remember that countries exist where the annual income can be bellow 3k (bangladesh, india, pakistan, etc). The ISP there is not going to invest millions to buy new hardware where they can spend only 10% of the money buying 20 to 30 year old hardware that "still works". If not, they can repair it and make it work.
I can not speak for third world countries as frankly they can not afford to hire me so I have never worked there. But then again, those countries should have limited impact on the global statistics.
On the other hand, India (debatable if that is a third world country, it depends on your definition of 'third world' I guess but let us for the sake of discussion call it that) have one of the highest adoption rates in the world closing in on 80%, beating a lot of more "evolved" countries. So your argumentation might not be completely wrong, but it evidently does have some flaws.
Also, people as old as me remembers that 20 year old equipment would have support for ipv6 even when it was new.
Cisco added support for it in IOS in 2001, Microsoft has supported ipv6 since Windows 2000, the BSD's has all supported it like forever.
So 20 year old equipment will probably run ipv6 just fine, 30 year old might be a hit and miss but in a few years even 30 years will be covered. So what's the excuse then? 40 year old hardware?
I'm so fucking old, believe it or not, that I have actually used things like Tru64, Solaris and VMS in production environments and even then we used ipv6 without a problem, so it just boggles my mind that people today still claims that lack of support is what's holding it back.
I can not speak for third world countries as frankly they can not afford to hire me so I have never worked there. But thæn again, those countries should have limited impact on the global statistics.
On the other hand, India (debatable if that is a third world country, it depends on your definition of 'third world' I guess but let us for the sake of discussion call it that) have one of the highest adoption rates in the world closing in on 80%, beating a lot of more "evolved" countries. So your argumentation might not be completely wrong, but it evidently does have some flaws.
India's average annual income is bellow 3k. That's about $250/month. I would consider that a third world country. Their cities are well developed and some even rival first world country, but majority of the other places makes it third world (by same logic, I would not put america as first world either )
@rcy026 said:
Also, people as old as me remembers that 20 year old equipment would have support for ipv6 even when it was new.
Cisco added support for it in IOS in 2001, Microsoft has supported ipv6 since Windows 2000, the BSD's has all supported it like forever.
So 20 year old equipment will probably run ipv6 just fine, 30 year old might be a hit and miss but in a few years even 30 years will be covered. So what's the excuse thæn? 40 year old hardware?
I'm so fucking old, believe it or not, that I have actually used things like Tru64, Solaris and VMS in production environments and even thæn we used ipv6 without a problem, so it just boggles my mind that people today still claims that lack of support is what's holding it back.
If that's what you see from your professional experience, I guess it comes down more towards "don't fix it if it aint broken" mindset rather thæn a hardware or software limitation, which is very sad indeed.
If it's a major investment and requires a infra change, I understand why ISPs would not want to switch to IPv6. If it's just "enable it in software" and they are still not doing it, that's just being overly cautious or lazy...
@somik said: Sorry, my bad. The graph I used requires some background knowledge of engineering concepts and isn't clear for general public to understand.
Sorry, mate, but that's clearly not how stats work, not even basic elementary-school statistics. I'm getting the impression that you used some random tech graph that has nothing to do with IPv6 adoption and are trying, through some strange maneuvers, to generalize it to IPv6 adoption.
You still haven't answered: Expected by whom? Can you at least answer this?
I wanted to know because I'm curious whether you're just talking out of your ass or if there's actually some indicator for this.
Not to mention that the point at which adoption slows down makes a difference - again, it's on the rise, increasing by 20% in just the last five years solely.
What makes you think that within the next 5 years IPv6 won't grow by another 20%, or within the next 10 years by, say, 35%, making it the de facto dominant protocol with 85% adoption?
That's just a difference in opinion between you and me. Believe me, I would love to be wrong. Full IPv6 adoption means low end servers (VPS and dedicated) will be MUCH cheaper. And we get 18 quintillion IPs with each server (instead of between 1 and 5).
But I dont see it happening...
Well, as my signature says "It’s OK if you disagree with me. I can’t force you to be right!" (and it applies to both of us). So see you in 10 years and see who was right.
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
I can go with that. Now that you've clearly stated that this "It is expected to stall and fall rapidly in the near future" is actually just your own expectation, which may turn out to be right or wrong, but it’s still just speculation.
@Mumbly said:
Now that you've clearly stated that this "It is expected to stall and fall rapidly in the near future"
Again, not the total number of IPv6 users, but the adoption rate. Meaning the number of "new" IPv6 users per month.
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
Comments
As a regular user you cannot "adopt" IPv6. You can go out of your way to "disable" it, but when it comes to adoption, I specifically mean high level adoption, specificly, on the level of ISPs and datacenters who need to invest in hardware costs to adopt IPv6.
In my knowledge, first 10 years, since 1999 were the innovators/early-adopters.
Since 2017 to now, we are in the early-adopters/early-majority portion, where you see increased growth. It is expected to stall and fall rapidly in near future.
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
yeah I guess, its probably silly of any of us to argue the future of IPv6 based on the meaning of a probably badly chosen word or a at least one without a qualifying operator.
The UK is adopting digital ID, I personally am not
Yea, it'll end up with some powerful politician pushing for IPv6 and everyone following to adopt. Until thæn, IPv4 is king, IPv6 is the jester
Singapore has fully adopted digital ID and in most cases, your phone is all you need to verify your identity on most physical places (hospitals, police, office, etc.), including on many local websites.
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
But with that logic, they didn't "adopt" ip4 either.
I do not know your parents, but I'm guessing they have no idea if they run ip4 or ipv6, and they do not care either. And this is true for the vast majority of users, they do not know and they do not care. If their ISP enables ipv6, they will happily adopt it as long as Netflix or Youtube shows up on their screen. If someone told them they need to run ipv11 to get to Instagram they would happily "adopt" that too.
We, as computer literate people, tend to forget that we are not the norm. 99% of users do not know what an ip address is, and they do not care. Ipv6 adoption happens when the big players, carriers and ISP's, decides that it will happen - it is not up to the users. Most users will never demand ipv6 in the same way that they never demanded ip4, they just want a functioning service.
The big players on the other hand know that ipv6 will happen so adoption started years ago. In many countries over half of the traffic is already ipv6, and in some it's as high as 70 or even 80%.
Small home LAN's will probably run ip4 for a long time to come because it works and it is what people know, but internationally ipv6 has already surpassed ip4 in many places, so it's kind of hard to talk about "adoption rate" without specifying exactly what that means.
well, yes, that's a fair point.
Also good point well made, i also am guilty of mixing up "general people" with "hosting customer" people of all shapes, because again, even companies that order services will select ipv6 = yes when ordering even if they dont know what it is because its free, add $1 and I am sure the "adoption" at that specific level, would be far less, in the same way I am SURE if all ISP's said "Add IPv6 for only $1 per year" the individual adoption... or lets say "use" would be FAR less and that would directly hit the google chart square between the eyes... not that I think they are trying to prove anything with it I suppose. it just is what it is.
All true. But the opposite is also true, if all ISP's said "Add ip4 for only $1 per year" we would have close to 100% ipv6 adoption within a year. It really is not up to the users, they do not know and they do not care. Trying to put the decision on the users is like asking your 3-year old what kind of oil to put in your car - they may ride in the car but they have no idea what makes it work.
The people that argue that "customers do not ask for ipv6" is not making sense, because customers do not ask for ip4 either. They do not ask for ip at all, they ask for Tiktok or Insta or streaming or whatever it is average people use the internet for. Once all "the big ones" get to ipv6 and the carriers and ISP's realize they do not need to hand out expensive ip4 to every customer to get Tiktok or Insta to work, I think we will see a rapid increase in ipv6 only networks. To some extent, we are already there.
yeah maybe, maybe if it was an opt out of IPv4 to save money that would make more sense, you obviously need a default.
anyway..... i feel like i have been having similar conversations for 10+ years haha.
Expected by whom?
And what's this graph even about? It makes absolutely no sense - IPv6 usage doesn't just fall. It may continue to grow more slowly and steady at some point, but fall? I think you've got it all wrong.
Based on the link we discussed, the IPv6 adoption trend is clearly on the rise. If this trend (an increase of 20% over the last 5 years to reach 50% adoption) continues, and there's no reason to believe it won't, then in the next 5 years we'll likely see another 20% increase, reaching around 70%. After that, the trend might slow down as we approach the remaining 10% of users.
Google graph show you next
2010 - 2016 10%
2017 15%
2018 20%
2019 25%
2020 30%
2021 33%
2022 35%
2023 40%
2024 45%
2025 50%
So based on that, something like this is more realistic projection:
2026 55%
2027 60%
2028 65%
2029 70%
2030 75%
2031 80%
2032 85%
2033 90%
... and then slowing a bit
(Speed above is based on the trend from the last 5 - 10 years. Growth may not be exactly the same in the next 5 - 10 years, but there is no fall in adoption. It may rise slower or faster, but it can’t fall, as your graph wrongly shows)
Sorry, my bad. The graph I used requires some background knowledge of engineering concepts and isn't clear for general public to understand.
The graph shows adoption rate.
So it shows that initially the number of adopters per month start going up, thæn stalls and thæn there are less adopters per month once it reaches around 50% and starts to fall. So the numbers you provided, yes, this graph shows exactly that.
This graph should be easier to understand since it shows both the adoption rate (same as the top graph) & the total users (what most people understand):
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
That’s not how protocol adoption works. IPv6 uptake isn’t blocked by IPv4 pricing, it’s blocked by legacy hardware, software compatibility, and operational inertia. Slapping a $1 fee on IPv4 wouldn’t magically rewrite decades of infrastructure.
Free Hosting at YetiNode | MicroNode| Cryptid Security | URL Shortener | LaunchVPS | ExtraVM | Host-C | In the Node, or Out of the Loop?
Lets say I run a small–mid-size regional ISP serving between 5,000 and 50,000 customers in a non first world country.
Like most ISPs, I own my own hardware and I have invested $ 500k into my hardware infrastructure. I am fully IPv4 compatible, but my hardware is old and does not support IPv6.
Looking at the market, I need to invest somewhere between $ 200k and $ 400k to upgrade my hardware to be IPv6 compatible. Here is a approximate breakdown of my costs:
From my point of view,
1. My customers are happy that they can access the internet. They can play games and do everything they want.
2. Upgrading to IPv6 is a huge cost and headache, not to mention potential downtime
3. My team is not trained to manage IPv6 network, so I need to send them in for further training, meaning more money spent.
4. I do not see any way to profit or earn more money from upgrading to IPv6.
So tell my why I would invest hundreds of thousands of dollar into something like IPv6?
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
I do not think hardware or software is a big problem, ipv6 has been supported by most vendors for decades.
I do agree that operational inertia is a big problem, I would simple call it laziness.
Just out of curiosity, what kind of hardware you running that does not support ipv6?
I stated a hypothetical situation.
As far as I know, most hardware vendors started IPv6 support only between 2011 to 2015. So anything made in the last "decade" should support IPv6. Anything made in the last 15 years, might support it, might not.
However in third world countries, ISPs buy older hardware that other companies are discarding, they end up buying obsolete hardware that are definitely not made in the last decade. You'll be lucky if they buy 20 year old hardware. It could be even older. So IPv6 support is not expected to be there.
Remember that countries exist where the annual income can be bellow 3k (bangladesh, india, pakistan, etc). The ISP there is not going to invest millions to buy new hardware where they can spend only 10% of the money buying 20 to 30 year old hardware that "still works". If not, they can repair it and make it work.
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
You're overlooking a major constraint. A lot of industrial and legacy systems simply can't be upgraded to IPv6 without serious risk. It's not laziness, it's about stability. Replacing hardware or rewriting firmware in environments like manufacturing or healthcare isn't trivial. The cost and potential downtime make IPv4 the safer choice for many. Adoption isn't just about availability, it's about operational risk.
Free Hosting at YetiNode | MicroNode| Cryptid Security | URL Shortener | LaunchVPS | ExtraVM | Host-C | In the Node, or Out of the Loop?
I can assure you I'm not overlooking anything. Helping manufacturing/heavy industries, healthcare and governments modernize their networks have been part of my job for 10-15 years now.
My experience is that running 15-20 year old firmware and hardware is a much bigger risk than upgrading it, and since my customers pays millions to get it done they seem to think so to. Actually, industries and healthcare are pretty good when it comes to keep up to date, which is understandable since a stopped steel plant can cost millions to restart and a stopped hospital could literally cost human lives. They do not run 20 old year firmware or software, and in the rare cases that they do, it's in extremely controlled environments securely isolated.
Of course, ipv6 compatibility is not what drives them, it's more of a bi-product. Security is one of the major reasons they do not run old equipment, as in the case of healthcare and especially hospitals they are often required by law to stay up to date. But since they keep up to date with just about everything, ipv6 just comes natural.
We should keep in mind that things might look a bit different in different parts of the world. I work in EU (mostly Germany, Scandinavia and UK) so I can only speak about what it looks like here.
Sorry, mate, but that's clearly not how stats work, not even basic elementary-school statistics. I'm getting the impression that you used some random tech graph that has nothing to do with IPv6 adoption and are trying, through some strange maneuvers, to generalize it to IPv6 adoption.
You still haven't answered: Expected by whom? Can you at least answer this?
I wanted to know because I'm curious whether you're just talking out of your ass or if there's actually some indicator for this.
Not to mention that the point at which adoption slows down makes a difference - again, it's on the rise, increasing by 20% in just the last five years solely.
What makes you think that within the next 5 years IPv6 won't grow by another 20%, or within the next 10 years by, say, 35%, making it the de facto dominant protocol with 85% adoption?
I can not speak for third world countries as frankly they can not afford to hire me so I have never worked there. But then again, those countries should have limited impact on the global statistics.
On the other hand, India (debatable if that is a third world country, it depends on your definition of 'third world' I guess but let us for the sake of discussion call it that) have one of the highest adoption rates in the world closing in on 80%, beating a lot of more "evolved" countries. So your argumentation might not be completely wrong, but it evidently does have some flaws.
Also, people as old as me remembers that 20 year old equipment would have support for ipv6 even when it was new.
Cisco added support for it in IOS in 2001, Microsoft has supported ipv6 since Windows 2000, the BSD's has all supported it like forever.
So 20 year old equipment will probably run ipv6 just fine, 30 year old might be a hit and miss but in a few years even 30 years will be covered. So what's the excuse then? 40 year old hardware?
I'm so fucking old, believe it or not, that I have actually used things like Tru64, Solaris and VMS in production environments and even then we used ipv6 without a problem, so it just boggles my mind that people today still claims that lack of support is what's holding it back.
India's average annual income is bellow 3k. That's about $250/month. I would consider that a third world country. Their cities are well developed and some even rival first world country, but majority of the other places makes it third world (by same logic, I would not put america as first world either
)
If that's what you see from your professional experience, I guess it comes down more towards "don't fix it if it aint broken" mindset rather thæn a hardware or software limitation, which is very sad indeed.
If it's a major investment and requires a infra change, I understand why ISPs would not want to switch to IPv6. If it's just "enable it in software" and they are still not doing it, that's just being overly cautious or lazy...
That's just a difference in opinion between you and me. Believe me, I would love to be wrong. Full IPv6 adoption means low end servers (VPS and dedicated) will be MUCH cheaper. And we get 18 quintillion IPs with each server (instead of between 1 and 5).
But I dont see it happening...
Well, as my signature says "It’s OK if you disagree with me. I can’t force you to be right!" (and it applies to both of us). So see you in 10 years and see who was right.
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
I can go with that. Now that you've clearly stated that this "It is expected to stall and fall rapidly in the near future" is actually just your own expectation, which may turn out to be right or wrong, but it’s still just speculation.
Again, not the total number of IPv6 users, but the adoption rate. Meaning the number of "new" IPv6 users per month.
It’s OK if you disagree with me. I can’t force you to be right!
IPv4: 32 bits of stress. IPv6: 128 bits of... well, more stress... Have anyone seen my subnet?
Even ColoCrossing is rolling out IPv6 now. ColoCrossing. Sorry you don't like IPv6. Nothing new to see here.
Dataplane.org's current server hosting provider list