IPv6 Usage Reaches Historic 50% Across Google Services 73
IPv6 usage briefly reached 50% across Google services for the first time, marking a major milestone for a protocol created in 1998 to solve IPv4's address shortage. Tom's Hardware reports: [...] IPv6 was dismissed early on as a headache-inducing, hard-to-implement complication that would hardly ever gain any traction -- despite offering 2^128 possible numbers, solving all network number assignments in one fell swoop. That changed over time by force of necessity, and Google's tracking graph shows that for a brief moment in time on March 28, 50% of worldwide users accessed the service over an IPv6 connection, marking a historic first. APNIC's stats show that the protocol is in use by 43% of the world, with Asia and the Americas inching ever close to those 50%. Cloudflare, meanwhile, shows that 40% of traffic is done in IPv6, an actually impressive figure if you consider it's measuring actual transferred packets rather than just counting addresses.
The tried-and-true IPv4 and its well-known 123.456.789.123 format from 1980 offers ~4.3 billion addresses in theory, and around 3.7 billion in practice. That always sounded like a lot, but nobody could have predicted just how rapid the explosion of the Internet would be. IANA, the entity controlling the North-American IPv4 space, ran out of IPv4 addresses around 2011, while its European equivalent RIPE NCC could spare no more four-octet addresses nearly seven years ago in 2019. Asian, African, and Latin-American IP registries equally ran out during that timeframe.
The tried-and-true IPv4 and its well-known 123.456.789.123 format from 1980 offers ~4.3 billion addresses in theory, and around 3.7 billion in practice. That always sounded like a lot, but nobody could have predicted just how rapid the explosion of the Internet would be. IANA, the entity controlling the North-American IPv4 space, ran out of IPv4 addresses around 2011, while its European equivalent RIPE NCC could spare no more four-octet addresses nearly seven years ago in 2019. Asian, African, and Latin-American IP registries equally ran out during that timeframe.
Always felt they could just add one more set (Score:2)
Re: Always felt they could just add one more set (Score:5, Funny)
Yeah sweet, fits great on all those 40 bit architectures or chunks nicely into four 10 bit chunks for all those 10 bit ones.......
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Re: (Score:2, Interesting)
Or maybe just 6 sets of 256. Why not.
I've always considered IPv6 to be one of the biggest engineering failures in history, except for maybe Therac-25.
But I'm sure someone here will graciously explain why I'm wrong and how IPv6 will save the world.
Re: (Score:2)
But I'm sure someone here will graciously explain why I'm wrong and how IPv6 will save the world.
You're wrong because it's not supposed to save the world, only to save us from having to half-ass update to IPv8 in a few years.
Re: (Score:2, Insightful)
Guess it failed at that, then, too - because IPv8 has been proposed, and it's actually something approachable compared to the management and comprehensional shitstack that IPv6 is.
https://www.ietf.org/archive/id/draft-thain-ipv8-00.html
They went back and addressed the issue from first principles instead of relying on a presumption which has not proven to be fully true, in turn resulting in a mismatch of capabilities and implementations across platforms which don't play nicely with each other (and subsequent
Re: (Score:3)
That IPv8 proposal was confirmed to be AI written, submitted by an anonymous entity in the Bahamas.
It calls for OAuth for authorization of devices but IP operates at Layer 7, OAuth is Layer 3. It's AI slop - complete garbage.
Re: (Score:2)
" IP operates at Layer 7, OAuth is Layer 3"
what?
Re: (Score:2)
Re: (Score:3)
Guess it failed at that, then, too - because IPv8 has been proposed, and it's actually something approachable compared to the management and comprehensional shitstack that IPv6 is.
They went back and addressed the issue from first principles instead of relying on a presumption which has not proven to be fully true, in turn resulting in a mismatch of capabilities and implementations across platforms which don't play nicely with each other (and subsequently, unfortunately, make it difficult to move forward with either v6 or transition to anything else).
Oh give me a break, literally anyone can submit an ID, this has not been adopted by any WG and most of them go absolutely nowhere. Saw this come through the announce list the other day and can attest to the fact this is a pointless waste of time.
There is a never ending stream of people angry they didn't just somehow magically extend the address space instead of inventing a new protocol clinging to all kinds of tunneling address extension schemes. This stubbornness is never ending and remains today even af
Re: (Score:2)
Yeah, but -which- ipv6 is implemented everywhere?
That's the problem. You deploy IPv4 and it works. You deploy IPv6 and... half the internet is black, and almost nothing works. And it's comparably difficult to use - and that makes it meaningfully worse, in both regards.
To get any of the good features of IPv6, the things that make it worth the time, you've got bolt on a half dozen different higher level (OSI) services.
Re: (Score:2)
Yeah, but -which- ipv6 is implemented everywhere?
What does this even mean? Are you saying there is some sort of interop problem between IPv6 implementations? Are there multiple IPv6 Internets?
That's the problem. You deploy IPv4 and it works. You deploy IPv6 and... half the internet is black, and almost nothing works.
There is only one Internet. Not all of it is presently available via IPv4 or IPv6 alone which is the reason dualstack deployments are common.
And it's comparably difficult to use - and that makes it meaningfully worse, in both regards.
It is all the same shit. Most people have no idea what version(s) of IP protocol they are even using neither should/do they care.
To get any of the good features of IPv6, the things that make it worth the time, you've got bolt on a half dozen different higher level (OSI) services.
What makes IPv6 worth deploying is maintaining a global network of peers. Nothing more or less
Re: (Score:2)
"Otherwise IPv6 is not in any meaningful way different, better or worse than IPv4."
Tell that to the tech who can't read hex or the guy trying to find a network range in logs using sed/grep on an 80x30 crash cart terminal in a DC, because something broke at 1am.
It is wholly unsuitable to straight up replace IPv4 for these reasons: it isn't a human-accessible protocol.
That's what I meant by "which IPv6?" SLAAC, RAs, DHCPv6, authoritative DNS AAAA with reverse are all basic table stakes to make it useful, whic
Re: (Score:2)
Tell that to the tech who can't read hex or the guy trying to find a network range in logs using sed/grep on an 80x30 crash cart terminal in a DC, because something broke at 1am.
It is wholly unsuitable to straight up replace IPv4 for these reasons: it isn't a human-accessible protocol.
Can't read hex? What? And who is speaking of "straight up replace IPv4"? Who is saying that? What are you responding to? If you want to use IPv4 on your private network knock yourself out. Nobody cares. IPv6 is only needed for public Internet.
That's what I meant by "which IPv6?" SLAAC, RAs, DHCPv6, authoritative DNS AAAA with reverse are all basic table stakes to make it useful, which already grossly exceed what small IPv4 business networks have for v4, and there's still another dozen services required to get full interoperability with v4.
"SLAAC, RAs" is the same shit. DHCPv6 is optional and AAAA records are no different than A records. All of the elements are the same and the protocols work the same way with only minor structural differences. Any router that is going to provide routing and DH
Re: (Score:2)
I'm a big fan of IPv6 and enable it everywhere I can, HOWEVER, there is a problem when it comes to firewalling. SLAAC is basically a requirement, since so many things don't support DHCPv6 (such as Android). So now how do you block or allow access to/from a device using SLAAC, especially if that device uses privacy extensions? Device addresses are unpredictable and change frequently.
Then you get issues with things like Matter devices, which only work on link local IPv6 addresses. You can't segment them into
Re: (Score:2)
Or how about, 4 sets of 16-bit numbers.
Re: (Score:2)
What is IPv6 and why would I ever need or want that? (I'm being facetious, of course I've had it pushed in my face at least for the past two decades...)
You have to be over 16 to use IPv6 (Score:3)
Re: (Score:2)
Coming soon: the year of IPv6! (Score:2)
Before you know it, IPv6 will be everywhere, there will be nowhere to hide.
Re: (Score:2)
Before you know it, IPv6 will be everywhere, there will be nowhere to hide.
The last place on earth safe from IPv6 will be DNS glue records.
Re: (Score:2)
IPv6 was dismissed early on as a headache-inducing, hard-to-implement complication that would hardly ever gain any traction
And whaddaya know, IPv6 turned out to be a headache-inducing, hard-to-implement complication that hardly gained any traction.
What stops IPv6 from being universal (Score:4, Interesting)
Comcast is my ISP and my issues with them aside they implemented IPv6 perfectly. Back when I was running a virtual lab I could bring up any number of endpoints in the cloud and at other sites and could get 100% connectivity anywhere I wanted without dealing with any NAT complications and everything easy to account for and manageable with firewall rules.
And no bot harvester ever found a single system of mine to initiate ssh attacks on. How could they and why should they when there still are so many vulnerable IPv4 endpoints around?
I could understand back when several popular OSes didn't support IPv6 very well but that stopped being true a decade ago. Yet new deployments every day with IPv4 address only provisions.
What's it going to take to kill IPv4?
Re: (Score:3)
Re: (Score:2)
Re: (Score:2)
You want to kill it? Make the mainstream routers enable it by default out of the box. It's literally that easy.
They literally do. I've never come across a router that doesn't have IPv6 enabled by default if it is presented with an IPv6 address from it's WAN connection. The problem is the latter part. I finally got IPv6 last year (my ISP sent me a new modem). In 2025 some. 30 years after the standard was written.
Re: (Score:2)
Re: (Score:3)
I am OK with IPv4. All my devices support it, the addresses are easy to remember. Everything works.
Though the local v6 addresses can be easy to remember as well, like fd00::0:1, it would be more annoying to remember a public IP as it would be longer.
The fact that my numpad does not have letters and : would make it more annoying to type, but whatever.
Still, IPv4 for me is good enough. There are not enough IPv6-only services that would make me consider adding it to my network and duplicating all the firewall
Re: (Score:2)
Still, IPv4 for me is good enough. There are not enough IPv6-only services that would make me consider adding it to my network and duplicating all the firewall rules. That's for my home network. It goes many times that for the work network, because that would be even more difficult and time-consuming for very little benefit right now.
Firewalls allow multiple addresses to be aliased to a name allowing one set of rules to be applied across address families without duplication. Family specific bits are IPv4 NAT forwarding which generally would not be necessary for IPv6.
Re: (Score:2)
I looked into IPv6 for home use a while back, and basic stuff like adding new devices to the network and finding them is a bit of a nightmare.
On IPv4 with DHCP it's easy enough to just scan the entire subnet looking for the new device, if you don't know its IP address. With IPv6... There are three of four different ways devices can make themselves discoverable, and Windows support for them is a bit limited.
IPv4 works, it's familiar, it's easy. That's why there hasn't been a big migration. There just aren't
Re: (Score:1)
What's it going to take to kill IPv4?
Outlaw or get rid of NAT. Then IPv4 will fail.
Re: (Score:2)
> What's it going to take to kill IPv4?
Something that needs bi-directional out of the box streams I guess.
Re: (Score:2)
Probably the fact that it was designed to be a duct tape solution which was incomprehensible by anyone with an IQ under 120, additionally requiring a myriad of services to make it worth the time or effort to fully implement, further exacerbating its adoption and interoperability.
Re: (Score:2)
Some colos don't even provide v6 yet.
Some ISP's still don't suport it.
Some classes of network gear only recently got hardware v6 support. Older gear still in service pushes v6 onto the CPU. Probably why those colos and ISP's don't support it.
And I still see important v6 fixes coming in hardware changelogs, mostly on the LAN management stacks (neighbor discovery, mDNS, etc.)
That said, the pieces are finally coming together in the past two years, roughly.
I would bet 2030* will see v6 at around 75% of traffi
Re: What stops IPv6 from being universal (Score:2)
Verizon's 2 Gbps Fios home service doesn't support IPv6, due apparently to them picking hardware for this particular service that doesn't handle it. Their other service does handle it, including slower home Fios, but not that in particular.
Re: (Score:2)
What's it going to take to kill IPv4?
Everyone being as good as Comcast. Vodaphone user here, I don't get an IPv6 address on my phone despite it being supported by Android.
So bizarre (Score:3)
All of my services (at home and for business, including self-hosted and collocated servers) have been dual-homed with IPV4 and IPV6 for decades. I just can't grasp services that don't include IPV6 connectivity in this modern world.
Is this traffic mostly phones? (Score:2)
Re: (Score:2)
It's interesting to look at Google's graph (https://www.google.com/intl/en/ipv6/statistics.html). You can see that there's a greater proportion of IPv6 use at weekends, when the ratio of phone to work use is higher. And if you look around March 2020, when covid lockdowns started, the difference between work days and weekends reduces substantially.
Re: (Score:2)
Re: (Score:2)
There is likely a much higher density with phones. But anyone with a modern ISP and modern router has the majority of their home devices just using IPv6 natively, seamlessly.
An unintended side effect.. (Score:3)
..of the shortage of IPV4 addresses and NAT is that IOT devices need to connect to servers, often with subscriptions, for remote access.
I should be able to connect directly with my IOT devices using IPV6 and the devices should be secure enough to exist on the public internet.
Re: (Score:2)
That's probably the first legitimate reason I've seen for IPv6.
Re: (Score:1)
And if we had some pixie dust, we could all fly. Coulda woulda shoulda, in real life, sadly, technical decisions are often made by MBAs who make their assistant print out emails for them because the don't know how to use Outlook.
Re: (Score:2)
Horseshit. The technical reasons were pushed by *us*, the nerds on Slashdot, the ones who for some reason conflated NAT with firewalls, the ones who bitch and moan that IPv6 addresses can't be remembered. The world was full of greybeards, tech-heads, and IT geniuses who rallied *against* the adoption of IPv6.
It was the law of unintended consequences, MBAs had nothing to do with the sad state we're in now.
Re: (Score:2)
..of the shortage of IPV4 addresses and NAT is that IOT devices need to connect to servers, often with subscriptions, for remote access. I should be able to connect directly with my IOT devices using IPV6 and the devices should be secure enough to exist on the public internet.
Or not. You can still have a stateful firewall with IPv6, and it will provide exactly as much security as a NAT device. There's no reason to require that all of your devices be able to exist on the public Internet, which is actually a pretty tall order -- especially for IoT devices that tend not go get updated as much as they should.
Re: (Score:2)
The difference is that if the customer is on IPv6, the customer is more likely to have a globally unique address. This means the customer is at least technically capable of forwarding an inbound port across a stateful firewall, provided the ISP doesn't deliberately interfere with port forwarding the way T-Mobile US (a wireless ISP using 5G NR) does with its home Internet service. The TV commercials don't mention that T-Mobile home Internet is an outbound-only service.
Re: (Score:2)
Out of interest, how do you discover the IPv6 addresses of your IoT devices, so you can connect to them?
Re: (Score:2)
the devices should be secure enough to exist on the public internet.
ROFLCOPTER
The people making those devices want money. There is not enough of a feedback loop to force any manufacturer to consider security. Combined with proprietary software, you are fucked if you want your devices to be secure against external attacks.
ipv6 by default (Score:1)
ipv6 is just the default for most mobile and home networks nowadays. And it just works. Too bad it just start to be on for some services like AWS that are supposed to be modern but you still define ipv4 network and if you really want, you can now add ipv6 on top.
Re: ipv6 by default (Score:2)
TBH, AWS is pushing for IPv6. You do have to pay a premium for public IPv4 adresses. That said when they did start charging, not all AWS services were IPv6 capable yet, IIRC.
At least learn your industry (Score:2)
"The tried-and-true IPv4 and its well-known 123.456.789.123 format from 1980 offers ~4.3 billion addresses in theory, and around 3.7 billion in practice." Oh yes, the definitely well known format that includes 2 octets of invalid information. Also wtf does format matter? Why are we reporting on this crap? Math is math, numbers are numbers, regardless of how you express them.
Re: (Score:2)
you can usually ping 2130706433
it can helpful for people that are too mentally tied to dotted quads to understand how ipv4 really works
Fictional address (Score:2)
The octets of invalid information mark the address as fictional, as opposed to being the live address of a real machine. Telephone subscribers in several area codes started receiving prank phone calls after the 1982 release of "867-5309/Jenny", a song by the band Tommy Tutone containing a live numeric address on the US phone network. This led US TV show producers to start using the 555 (or KLondike 5) exchange, which was largely set aside for fictional use.
Um .. (Score:1)
IPv6 was dismissed early on as a headache-inducing, hard-to-implement complication that would hardly ever gain any traction
Ha! What fools those scoffers were!
Instead, it took the world by storm with lightning speed ... er, to briefly over 50% in almost 30 years!
It's REALLY not that hard prople. (Score:3)
So the addresses are bigger, so what?
In most cases, all you care about is the prefix, which is just 64 bits, expressed as 4 groups of 4 hex digits. This is IT, is hex really beyond people's grasp?
Most of the admin is done one terminals that support cut and paste anyway.
Re: (Score:2)
This is IT, is hex really beyond people's grasp?
Any sufficiently advanced technology is indistinguishable from magic.
hex1 /hks/
North American English
noun
a magic spell; a curse.
You really needed to phrase that in a different way. Between Clarke and the Oxford dictionary, yes.. it is beyond people's grasp.
I'm a huge proponent, but I can't answer this.. (Score:2)
I've got a network on a boat. I've got a router that uses My ISP for internet connectivity when I'm at shore and uses cellular for backup.
If I'm using IPv6 on my devices and the router switches the connection to the internet, all of the local devices would need to get new addresses via SLAC?
The nice thing about using NAT on IPv4 is that the local devices are mostly oblivious to how they are talking to the internet.
My home network is nearly pure IPv6 (Score:2)
Re: (Score:2)
Re: (Score:2)
I have something like ~1.2 trillion times the number of routable addresses that the entire IPv4 space has. Not all are reachable, of course, just the services that need incoming access and they're each on their own isolated DMZ.
Azure & dual-stack is the problem (Score:2)
I have a need for thousands of VMs for a load test in Azure. But Azure requires dual-stack for IPv6 to work -- which completely defeats the purpose! The DevOps team tells me we are out of IPv4 addresses. If everything IPv6 also requires an IPv4 address, then IPv6 is useless. I am told that AWS VMs do not have this problem.
IPv6 hype (Score:1)
Re: (Score:2)
The sales pitch for IPv6 has always been fraudulent. The address space problem for 32 bit IPv4 was solved 30 years ago with Network Address Translation ( NAT). Each 32 bit address in the IPv4 space can be a gateway for up to 16 million addresses in the 10.x.x.x or 64k addresses in the 192.168.x.x space. The REAL address space for IPv4 is 56 bits. Pre-existing users of IPv4 have never run out of address space. The problem is that IANA squandered the IPv4 address allocations, and many organizations have WAY more addresses than they need.
This is gibberish, there are far more people and devices than IPv4 addresses and it is impossible to uniquely address them with a 32-bit address space no matter how carefully number resources are allocated. The 56 bit nonsense reminds me of Eugene Terrell's unhinged rants about ternary logic on the IETF lists decades ago.
Another advantage of NAT is that it hides your internal IP addresses like a cloaking device, but IPv6 breaks that by making internal.addresses visible to the internet.
It does no such thing. All hosts have privacy extensions enabled by default where the host portion is constantly changing.
Virtually no one who has pre existing IPv4 space is tearing it up and replacing it with IPv6.
Of course not neither is anyone suggesting doing this. Most depl
Why don't the just use name instead of numbers /s (Score:2)
sql databases and 128-bit int column type (Score:1)
IPv8 was published yesterday (Score:1)
Original inventor of IPv6 has admitted its bad (Score:1)
From memory the size of IPv6 header is 40 bytes. That was stupid decision, it should have been naturally aligned 32 bytes.
I believe NAT does not exist in IPv6. Again, another stupid decision. NAT is good for ISP customers because it gives some privacy, as they all use the same public IP.
They should have simply extended IPv4 to 48bits. 2^48 PUBLIC IP addresses is enough f