Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
×
The Internet

India has Highest Rate for IPv6 Adoption, While US Ranks Sixth (aelius.com) 147

Nearly half of the U.S. (47.5%) has migrated to IPv6, ranking the U.S. sixth in percentage of users who have migrated to the latest version of the Internet Protocol. That estimate is based on an analysis of data about Google users — which also finds that India leads the world, with a 61.67% adoption rate. Inside.com's developer newsletter reports: The figures come from the latest Google IPv6 Statistics data organized and ranked by BBC Radio and Music's Lead Technical Architect for his blog. Malaysia and French Guiana came in second and third, respectively...

IPv6 is considered far more secure, faster, and powerful than its predecessor, IPv4, while offering more IP addresses. Test your IPv6 connectivity here.

After third place there's a close three-way race, where the U.S. has dropped from the #4 position earlier this week. As of today the U.S. is now in 6th place...

Fourth is France (48.38%) and Fifth is Taiwan (48.0%).

Canada ranks #20 (36.59%) and the UK ranks #24 (33.27%).
This discussion has been archived. No new comments can be posted.

India has Highest Rate for IPv6 Adoption, While US Ranks Sixth

Comments Filter:
  • by Dorianny ( 1847922 ) on Sunday August 22, 2021 @07:44PM (#61718839) Journal
    It would be more interesting to see a breakdown into mobile-handset, other networking. Even in the U.S Wireless carrier deployment is pretty good, It is enterprise and home networking that's significantly lagging. The IPV6 adoption lead in India could simply be that wireless-phone only homes are far more typical there than in the U.S
    • by jfdavis668 ( 1414919 ) on Sunday August 22, 2021 @07:49PM (#61718849)
      Was thinking the same thing. We have made sure everything at work is IPv6 ready, for many years now, and yet we haven't made the switch. Every mobile device I ever checked is running IPv6.
      • by Dorianny ( 1847922 ) on Sunday August 22, 2021 @08:06PM (#61718893) Journal
        "Was thinking the same thing. We have made sure everything at work is IPv6 ready" Are you sure about that? What we discovered in our networks was that there are so many hardcoded ip's in the configs and applications that most of everything would break, The networking is all good to go for IPV6 only but so far only our DMZ systems are truly ready.
        • by AmiMoJo ( 196126 ) on Monday August 23, 2021 @05:25AM (#61719761) Homepage Journal

          For a lot of people it's just not worth the hassle. IPv4 works, NAT is okay for them, so why change? Larger orgs seem some benefit but even there it's not like things are actually broken at the moment.

          • For a lot of people it's just not worth the hassle. IPv4 works, NAT is okay for them, so why change? Larger orgs seem some benefit but even there it's not like things are actually broken at the moment.

            The point is maintaining a network that provides the same opportunities for everyone not just lucky rich people who managed to gobble up most of the number resources.

            With exhaustion of IP resources assignment is now entirely market based with highest bidder winning and everyone else eating dust.

      • IPv6 is considered far more secure, faster, and powerful than its predecessor, IPv4, while offering more IP addresses.

        This statement is not just false it's insane. IPv6 has almost exactly the same security architecture as IPv4 and it objectively requires more bytes of overhead than IPv4 leaving less for user data. The only thing in that sentence which is true is that it offers more IP addresses.

    • by jrumney ( 197329 )

      Here (New Zealand) it's the opposite. I am with the same company for my Home and Mobile internet, Home is dual stack, IPv6 and CG-NAT IPv4, mobile is just CG-NAT IPv4.

      • the "NAT" is the key to the startling ongoing use of IPv4. Some environments use the limited IPv4 space to discourage users from having exposed, routable IP addresses, and to compel people's personal devices and system servers to remain behind a locally managed NAT.

    • by Bert64 ( 520050 )

      The reason for that is that mobile networks came later, have more users and the users are inherently mobile, so allocating ipv4 to each subscriber was simply impractical. You ended up with cgnat instead which is costly, problematic and reduces performance (although the lack of performance was less of an issue with slower mobile networks).

      Fixed line networks in developed countries generally don't use cgnat, are not seeing significant new growth and the incumbent providers already have enough ipv4 to service

    • by tlhIngan ( 30335 )

      Exactly.

      The IPv6 transition for phones happened during 3G. 3G supported both IPv4 and IPv6, with IPv6 preferred.

      LTE is entirely IPv6 based. Since it's data only, they mandated that LTE shall use IPv6 exclusively.

      5G remains IPv6 only for many reasons, including its intended use for IoT applications and thus needing even more IP addresses.

      Carriers use CGNAT to help connect the IPv6 mobile network to IPv4 only internet hosts.

      It was easier because carriers used 3G to familiarize themselves with IPv6 equipment,

  • Faster? (Score:5, Insightful)

    by fahrbot-bot ( 874524 ) on Sunday August 22, 2021 @07:55PM (#61718865)

    IPv6 is considered far more secure, faster, and powerful than its predecessor, IPv4, while offering more IP addresses.

    [citation needed]

    • IPv6 is considered far more secure, faster, and powerful than its predecessor, IPv4, while offering more IP addresses.

      While I disagree with the characterization "far more secure" IPv6 does have significant advantages.

      Bypass CGN

      Naive brute force scanning of global address space is infeasible

      Improves peer to peer communications (Games, voice and video chat, remote access..etc) Enabling capability and enhancing privacy vs routing data channels thru someone else's servers.

      SPI inherently more secure than NAT: no packet mangling or exploitable ambiguous ALGs.

      IPv6 has potential for better route aggregation (limited by whims of TE

      • by Bert64 ( 520050 )

        Massive address space allows for remote networks to cause local broadcast traffic.

        IPv6 does not use broadcast traffic.
        It uses multicast, and more specifically solicited node multicast for neighbor discovery:
        https://en.wikipedia.org/wiki/... [wikipedia.org]

        So while theoretically scanning a large sequential address range will generate neighbor discovery traffic, in reality most of it will go nowhere as most of the solicited node multicast addresses won't exist on the switch and will be discarded. So it just becomes a flood of junk traffic like any other DoS attempt.

        Less efficiency per packet due to overhead of 128-bit addressing. While this is mostly irrelevant and a rounding error for data transfers the majority of packets transmitted over the Internet are quite a bit less than 100 bytes.

        Theoretically, although there are variou

      • by tlhIngan ( 30335 )

        The RIAA/MPAA/etc love IPv6, because it destroys the legal argument that you cannot identify a person with an IP address. Which is true for IPv4 as one IPv4 can hide a family or a more people behind it.

        But you can be sure with IPv6, they'll be getting prefixes from your ISP to contact you to identify a specific address or to preserve the IP address logs so they can identify the PC or device they think is pirating.

        And since that sort of notification means you have to preserve the logs, joe home user will get

    • [citation needed]

      https://datatracker.ietf.org/ [ietf.org] There you go. Unfortunately there's no single document that will answer your question, but if you go through the RFCs you can find many standards that detail precisely what additional IPv6 security features are, how it improves routing efficiency, and what all the additional capabilities are over IPv4. I hope I don't need to provide a citation about how it offers more IP addresses, but just in case the Kahn Academy has some introduction to mathematics videos where you can lear

      • the Kahn Academy has some introduction to mathematics videos where you can learn about the powers of 2 as well.

        No need to be snarky. :-)

        if you go through the RFCs you can find many standards that detail precisely what additional IPv6 security features are,

        My comment, as the subject suggests, was about "faster" not "security" -- I'm familiar with this stuff, but not an expert...

        • Re:Faster? (Score:5, Informative)

          by thegarbz ( 1787294 ) on Monday August 23, 2021 @06:03AM (#61719799)

          Sorry am in a bad mood this morning :-)

          The faster part comes from massively increased routing efficiency. IPv4 was never designed to be segmented in the way that it currently is. The idea was that large blocks should represent large networks, and each sub division a smaller network within the larger network meaning that when a packet is routed looking up where it needs to go is supposed to be trivial. The current routing tables have 900k entries already meaning every packet needs to be looked up in a massive table in order to be passed on. IPv6 is supposed to slim that back down as there's ample ways to subdivide the far longer address.

          It's like postcodes in a way. If you receive a letter with the postcode starting 75XXX you know it needs to be sent to the Texas post office. That only works if there isn't explicably a 75342 in New York, otherwise you need to look up both numbers to be sure which state post sorting centre to send to.

          Sidenote: In 2023 we're going to hit a sort of Y2K style bug. Routing tables are expected to exceed 1,048,576 entries which was incidentally the limit of the table size on a lot of core routers. Network providers are scrambling to upgrade these old ones so they don't hit a cap which could potentially lead to advertised routes not being accepted. Incidentally IPv6 will suffer from this as well as the routing table sizes are only 128kB not 1MB for similar model routers (reads: old ones) and while 1MB should be more than enough for everyone, it turns out 128kB is not likely to be.

          That's my admittedly limited understanding of BGP anyway :-)

    • We’ve observed that accessing Facebook can be 10-15 percent faster over IPv6. [fb.com]
      His contention was that, on average, IPv6 network connections are established 1.4 times faster [www.sidn.nl]
      Google's stats [google.com] have a per-country measurement of the latency hit of v6, and for almost all countries it's either 0ms or negative.

      Here's a few.

  • by JcMorin ( 930466 ) on Sunday August 22, 2021 @07:55PM (#61718867)
    The ratio of IPv4 "owned" by the USA is ridiculously high compared to other countries. The ratio directly affect the need to switch to IPv6.
  • India doesn't have SpaceX.

  • I have a data center in Toronto where IPv6 isn't available. The world hasn't ended so far.

    • I have a data center in Toronto where IPv6 isn't available. The world hasn't ended so far.

      Perhaps not for you but a /16 is now priced at a million or more at auction.

  • I have a couple of servers living in a web co-host facility that has very good IPv6 support and is close to high traffic hubs. And for some cosmic error of advancement that will never be explained, Comcast has really decent support for IPv6 for its residential cable subscribers.

    End result is I use IPv6 a lot and it turns to be a lot more convenient than IPv4 ever was. No NAT, no problem with address space allocation (I have 14 IPv4 addresses that the co-host gave me), no having to re-use addresses all

    • Actually, what keeps IPv4 alive, are ISPs that are gutless. We need 1 large ISP to switch to IPv6, and within 2-4 years, all others would follow.
      • by Bert64 ( 520050 )

        There are many ISPs which already provide IPv6, but they also always provide some form of backwards compatibility mechanism too - either dual stack, NAT64, DS-Lite etc.
        IPv6 works better, and achieves better performance (sometimes significantly so) on these providers, but the users aren't aware why so they don't demand IPv6 from other ISPs. LinkedIn did a study a few years ago and determined that IPv6 was usually quicker, sometimes up to 40% (likely caused by cgnat):
        https://www.linkedin.com/pulse... [linkedin.com]

        I've had

        • Any site that doesn't enable IPv6 these days is giving a second class experience to many users, and likely drives users away to other "faster" sites.

          What would it take to convince users to actively avoid websites that don't support IPv6? The website Why No IPv6? [whynoipv6.com] lists Reddit, Twitter, Amazon, Twitch, Pornhub, XVideos, Yahoo! Japan, eBay, GitHub, PayPal, and Stack Exchange as among the most popular laggards outside China. A lot of these are communication platforms that don't have significant "faster" competition because network effect. I'm guessing many have hesitated because adding IPv6 would amount to amnesty on IP range bans issued for serious violati

  • Is the large number of call centers across India.

    And of course the lack of available V4 addresses in useful chunks.

    • There are more than 4 million [npr.org] IT workers in India. That number includes call center workers and others. There are something like 750 million [statista.com] smartphones in India, so the sheer number of smartphones is most likely to be driving the need for IP address space.
  • Seriously, if a large ISP would switch to IPv6, dropping IPv4, or perhaps charging more for IPv4, it would force all to ISP to switch quickly.
    • A large ISP would probably then try to charge you for each IP address you use, because they're shitty shits who do shitty things.

      • I've never heard of any ISP charging customers for individual v6 addresses. I am aware of lots of ISPs not doing that. I'm sure you could find one somewhere if you looked (probably a server host or something similar...), but it's certainly very uncommon.

        ISPs that provide less than a /56 are more common, but even then it seems to be less a matter of charging people and more a matter of just failing to provide a minimum service. "ISPs will charge you for each individual IP" is scaremongering rather than somet

    • I only get a /64 from my ISP, which means only one vlan can be ipv6. I have six primary vlans at home. Spectrum won’t provide /56’s for residential.

  • In my experience, wide configuration of "proper" rDNS for IPv6 addresses in use seems lacking.This makes some e-mail checks (forward/reverse/IP/HELO) not work, and that's my primary area of interest/concern, as such checks are, from my perspective, a must when it comes to combatting spam these days.

    i see this issue as an impedement to full, IPv6-only transition for many e-mail server operators.

    • > In my experience, wide configuration of "proper" DNS seems lacking

      I hope you don't mind that I fixed that statement for you?

    • by Bert64 ( 520050 )

      Reverse DNS is only used for spam checking of SMTP connections, there is virtually nothing else that uses it.
      If you are hosting a mail server, you setup RDNS, if you're doing anything else then you probably won't bother. Where there are stray PTR records on legacy IP that is often caused by address reuse from the days when more people used IRC, or when people actually bothered to set RDNS. Nowadays virtually noone does.

      There's nothing inherent in RDNS either, it's just assumed that anyone configuring a mail

  • The nation with the most IPv4 space isn't ready to trade in their valuable addresses?

    Maybe enterprise routers will pony up and pay for licenses for the firmware updates that include IPv6. I keep running into colos that still don't have good native IPv6 support, which trickles down into VPS services that force you to have an IPv4 address.

    I'm sure in 10 years we'll have completed the transition ... mostly.

    • by Bert64 ( 520050 )

      All equipment made for many years is IPv6 capable by default. If a provider is running equipment which is not IPv6 capable then this is a VERY BAD SIGN. This equipment is likely to be hugely out of date, probably entirely end of life and unsupported by the vendor at this point which means it's likely to have very serious security holes.

      A lack of IPv6 support is an indicator of SERIOUS problems at a provider, the security problems mentioned above plus the general lack of support for modern protocols. I would

      • If a provider is running equipment which is not IPv6 capable then this is a VERY BAD SIGN.

        I can get you a list of colos that are used for enterprise that have this problem. I wish it was just a distant memory, but I setup a system in June this year and was shocked when I had IPv6 problems. In the end I had to setup a tunnel broker. A very bad sign in deed, but typical.

        This equipment is likely to be hugely out of date

        What happens is vendors provide a support branch where you get security updates for an old version of their OS. I know Juniper and Cisco did this, I suspect other vendors as well. If the new OS supports old hardware, you usually ha

        • by Bert64 ( 520050 )

          Cisco support old versions for years with security updates, but at some point those updates stop and the device becomes end of life so you can't even get paid support for it anymore. Devices which lack IPv6 have long since fallen into the EOL category, multiple generations of hardware with IPv6 support are also EOL too. If you have something which doesn't support IPv6 from a major vendor like Cisco then it is way out of support and has long since passed the point of having any security updates.

          • Sorry, but that's not really true. It depends on what is meant by IPv6 support. Yes, the equipment can switch IPv6 and with a little effort can tunnel them. But without configuration in RIP a proper route can't be established, so your IPv6 networks end up being isolated.

            I will admit, it's been a while since I worked for Cisco, but I'd be surprised if that stodgy company's policies have changed at all.

  • What has happened is that IPv6 has been enabled on lots of systems without their owners realizing it, and those systems use IPv6 to talk to Google and a very short list of other web sites that support IPv6 - not Amazon, not eBay, not GitHub, not PayPal, not Reddit, not Slashdot. If you turned off IPv6 right now, everything would still work. Most people wouldn't even notice a difference. There is nothing you can do with IPv6 that you can't do with IPv4, except things where IPv4 is deliberately excluded to pr
    • That's kinda poorly informed. There's a significant number of sites out there with v6 (not as many as there should be, mind), and if you have v6 a significant amount of your traffic (more than half by byte count, on average) will flow over v6.

      The point of v6 isn't to let you do new things, but to allow you to continue doing things you used to be able to do on v4, but can't due to address space exhaustion. Sometimes (but not always) there are workarounds for v4, but they're costly and a pain in the neck. Sim

      • We can have the same argument in ten years. More people will have CGNAT then and fewer people will care. The only way an IPv6 "only" network is usable right now is if it has a transition mechanism tacked on which allows it to access the IPv4 internet. You can't even upgrade Linux Mint between releases without IPv4.
    • by tepples ( 727027 )

      There is nothing you can do with IPv6 that you can't do with IPv4

      Except accept incoming connections in countries where all ISPs lack enough IPv4 space for all subscribers. This means you can't run a personal website, media server, or game server from home without paying a significant amount of money per month for a VPN to relay requests between the Internet and your home server behind carrier-grade NAT.

      • You can forward incoming connections with a VPS and it doesn't cost "a significant amount of money". Also, of all the people who use the internet, statistically speaking nobody needs that. It's a niche application. Everybody uses the cloud nowadays, so it's all outbound connections, which work fine with NAT. If you really need P2P connections without relays, you can use NAT hole punching. IMHO we will never reach a point where IPv4 will be optional for server operators. It is worse than letting Internet Exp
      • Except accept incoming connections in countries where all ISPs lack enough IPv4 space for all subscribers. This means you can't run a personal website, media server, or game server from home without paying a significant amount of money per month for a VPN to relay requests between the Internet and your home server behind carrier-grade NAT.

        You explained yourself why IPv6 adoption rate of ISPs are so slow. Why take time and effort such that customers no longer need to pay premium when they want to host public server at home / office? This is anti-profit.

  • What happened to the all-time leader Belgium ?
    Looks like there's something wrong with the Google IPv6 stats : https://www.aelius.com/njh/goo... [aelius.com]

    • My belgian major ISP (proximus) doesn't provide native ipv6 on my (business grade) connection.

      • by wimg ( 300673 )

        Telenet and several others have been for many years. Belgium was number 1 for a long time.

      • by tepples ( 727027 )

        Have you reported this lack of IPv6 service to your ISP? If so, what was the reply? And what other business-class ISPs are available at your address?

        • Yes, answer is: "We do not support ipv6 on your line". I guess they need to migrate most "older" business customers to newer backbones and that would probably involve losing/changing my fixed ipv4.
          I can't really change my ISP if I want to keep my fixed IPv4s.

    • What happened to the all-time leader Belgium ? Looks like there's something wrong with the Google IPv6 stats : https://www.aelius.com/njh/goo... [aelius.com]

      Something odd happened between 2021-08-11 (49.2%) and 2021-08-18 (17.85%) so it does look like there is a measuring problem this week. They should still be ahead of the US.

  • All this IPv6 adoption, and Google still refuses to implement proper DHCPv6 support in Android
    https://issuetracker.google.co... [google.com]

FORTRAN is not a flower but a weed -- it is hardy, occasionally blooms, and grows in every computer. -- A.J. Perlis

Working...