Slashdot Mirror


IPv6 Deployment Picking Up Speed

An anonymous reader writes "The Internet's addressing authority (IANA) ran out of IPv4 Internet addresses in early 2011. The IPv6 protocol (now 15 years old) was designed exactly for this scenario, as it provides many more addresses than our foreseeable addressing needs. However, IPv6 deployment has so far been dismal, accounting for 1% of total traffic (the high-end of estimates). A recent paper by researchers at the Cooperative Association for Internet Data analysis (CAIDA) indicates that IPv6 deployment may be picking up at last. The paper, published at the Internet Measurement Conference (IMC) shows that the IPv6 network shows signs of maturing, with its properties starting to resemble the deployed IPv4 network. Deployment appears to be non-uniform, however; while the 'core' of the network appears to be ready, networks at the 'edges' are lacking. There are geographical differences too — Europe and the Asia Pacific region are ahead of North America."

158 comments

  1. Stop the Presses! by Anonymous Coward · · Score: 3, Informative

    North America fails to take up an International Standard.

    That's NEVER happen. Except with everything.

    1. Re:Stop the Presses! by canadiannomad · · Score: 1

      I fail to see how this is offtopic...

      --
      Hmm, the humour and sarcasm seem to have been be lost on you.
    2. Re:Stop the Presses! by Ravaldy · · Score: 1

      North America doesn't have a shortage yet so nobody is acting. It's just human nature. I know most major ISPs in Canada still have plenty of blocks to keep going for a while.

    3. Re:Stop the Presses! by Anonymous Coward · · Score: 1

      Except of course, North America HAS taken up the international standard.
      Not taking up the standard would be the US deciding on an alternate standard. This is one of those instances where the United States IS taking up the same standard.

      IPv6 has always been driven by need. For historical reasons the United States has owned a disproportionate portion of the IPv4 address space. This means that in the United States, the NEED for IPv6 has not been quite as desperate as in other portions of the globe. The US is not shirking a standard, it's just not migrating as fast. There's a huge difference.

      PS - How dare you toss our Canadian and Mexican neighbors into this. If you're going to rip on the US, make sure you leave them out of it.

    4. Re:Stop the Presses! by Creepy · · Score: 3, Insightful

      To be fair, America has adopted standards, but hasn't always standardized on them, and sometimes invents a standard that is outdated by the time the rest of the world adopts it.

      For instance, metric is used in hospitals, at NASA, in many sciences, etc. It was even taught in school until Ronald Reagan in his infinite wisdom and reverence decided America was too f**king stupid to learn it (sorry about the sarcasm injection - it was a REALLY bad time for me to switch, as I was half way into learning metric when it happened and we all of a sudden had to learn these nonsensical English units - I'm still all for switching to metric).

      CDMA predates GSM, and some providers bet big on it early in America. Nothing America can really do about it except wait for it to age and be replaced, hopefully with an international standard. Data already has been merged with LTE.

      Almost all cable providers use DOCSYS international standard.

      IPv6 is supported by some ISPs and CLECs, but many that supported PPPoE like mine bought IPv4 only hardware. The former owner of this hardware, Qwest, said they would never implement IPv6. Their current owner, CenturyLink, is rolling out IPv6 support, but only currently in areas that were not formerly Qwest. Meanwhile, my IPv4 and IPv6 addresses are registered and just waiting for IPv6 to be supported to go live (I hacked the router to get its IPv6 address just in case this is a server only issue - the underlying hardware supports it, just not the PPPoE connection).

    5. Re:Stop the Presses! by davester666 · · Score: 1

      Here in Canada, the main ISP's all seem so proud of their progress which is....

      Their home page is accessible via IPv6.

      Hell, the router/modem combo we got early this year isn't IPv6 capable.

      Obviously, they have a long-term plan for transitioning to IPv6 and they are still on the first step "When Should We Start Switching To IPv6".

      --
      Sleep your way to a whiter smile...date a dentist!
    6. Re:Stop the Presses! by jrumney · · Score: 1

      CDMA predates GSM, and some providers bet big on it early in America.

      CDMA deployment in US might predate GSM deployment in US, but GSM development in Europe dates back to 1982, while CDMA development for mobile phone use started in 1995. Perhaps you are thinking of DAMPS, the digital cellular system used in the US before CDMA came into the picture.

    7. Re:Stop the Presses! by adolf · · Score: 1

      I like your rant, and would like to add to it P25 ('merkin, barely works) vs. TETRA (everyone else, seems to have worked for years).

      CDMA predates GSM, and some providers bet big on it early in America. Nothing America can really do about it except wait for it to age and be replaced, hopefully with an international standard. Data already has been merged with LTE.

      But back in context, it amuses me that my DSL provider still doesn't have the ability to give me real IPV6 connectivity, but my CDMA/LTE cell phone scores 10 out of 10 on IPV6 at http://test-ipv6.com/ on Verizon with no trickery.

      (Not that I give two shits about my IPV4 vs IPV6 on my cell phone......)

    8. Re:Stop the Presses! by MikeBabcock · · Score: 1

      I've asked all three major ISPs local to me if I can get IPv6 addressing and they don't even know why I'd want it.

      --
      - Michael T. Babcock (Yes, I blog)
    9. Re:Stop the Presses! by CAIMLAS · · Score: 1

      IPv6 is yet another example of a niche technology favored by a small elitist segment of a diverse but elitist professional trade categorization, pushing change for the sake of change, against the wishes of everyone else.

      Sorry, but when pretty much everyone except the people who gobble up marketing and training material doesn't think IPv6 deployment is a good idea, it's probably not a good idea. We have decades of equipment considerations to phase out, and that's not even counting applications which won't work with IPv6.

      --
      ~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
  2. Where the heck is IPv7 by Anonymous Coward · · Score: 1

    Thats what I want to know, IPv6 is old hat! Any respectable IP _must_ have functionality equal to TOR built right into the specs!

    1. Re:Where the heck is IPv7 by Anonymous Coward · · Score: 1

      closest thing I know of: https://github.com/cjdelisle/cjdns/

    2. Re:Where the heck is IPv7 by Cyphase · · Score: 1

      Can you post a source for that quote?

      --
      by Cyphase ( 907627 )
    3. Re:Where the heck is IPv7 by Cyberia · · Score: 1

      Wow, seriously? An anonymous first post, and you blow it with this? *Facepalm*

    4. Re:Where the heck is IPv7 by michelcolman · · Score: 1

      IPv6 ought to be enough for anybody.

    5. Re:Where the heck is IPv7 by Anonymous Coward · · Score: 1

      Definitely. Here you are

    6. Re:Where the heck is IPv7 by Ravaldy · · Score: 1

      It's not a good comparasion. Changing measure systems doesn't stop people from measuring. On the other hand, more IP blocks can be a show stopper.

    7. Re:Where the heck is IPv7 by Anonymous Coward · · Score: 0

      How can we model every atom in the universe if we can't assign an IPv6 address to it, huh?
      And don't forget the quarks!

    8. Re:Where the heck is IPv7 by FilmedInNoir · · Score: 2

      Excellent! I'll add that as citation for the Wikipedia article and then you can cite that article as proof.

      --
      Sig. Sig. Sputnik
    9. Re:Where the heck is IPv7 by Anonymous Coward · · Score: 0

      Not everyone gets massive erections because they get a first post. Who gives a shit?

    10. Re:Where the heck is IPv7 by Anonymous Coward · · Score: 0

      I know what you mean. Last time I got a first post I lost my boner right away. Then the moles escaped and it was all downhill from there. (I was up a hill).

    11. Re:Where the heck is IPv7 by Anonymous Coward · · Score: 0

      You register an account and all you post is THIS! *Facepalm*

    12. Re:Where the heck is IPv7 by unixisc · · Score: 1

      Contrary to others, I think that ultimately, there will be an IPv7. It will be compatible w/ IPv6 in that it will still use 128-bit addressing. What I think may change are the boundaries - they might make the global prefix the top half, and split the lower half b/w subnets and hosts. If they do it on IPv6, problem is that not all IPv6 equipment will recognize that, since the standard recognizes the entire lower half as being the hosts address. Changing the rev of the header but leaving everything else the same, except the boundaries of the network vs subnet vs host would make it easier for future equipment to support both IPv6 and IPv7

  3. 2013 could be... by Bradmont · · Score: 5, Funny

    The year of IPv6 on the desktop!

    1. Re:2013 could be... by Anonymous Coward · · Score: 3, Funny

      If we make it past December 21st...

    2. Re:2013 could be... by Anonymous Coward · · Score: 1

      Why do you have a networking protocol on your desktop?

    3. Re:2013 could be... by NatasRevol · · Score: 4, Funny

      It looks cool.

      --
      There are two types of people in the world: Those who crave closure
    4. Re:2013 could be... by Kjella · · Score: 2

      Nope. People have IPv6 enabled browsers so they can connect to IPv6 enabled websites, but how many have some sort of legacy software that doesn't in any way understand or support IPv6 - perhaps there's not even an input field for an IPv6 address. Of course people will now chime "dual stack" but it has practically all of the annoyances while not solving the problem since it means pairing every IPv6 address with an IPv4 address. And by annoyance I mean like some stupid software, I don't remember what would prefer the IPv6 address over the IPv4 address then leading to a delay before it would connect via IPv4. I couldn't be arsed to find some other solution, so IPv6 is completely disabled on my machine. And so far I've had zero reason to change that. The only people feeling the hurt are those not getting an IPv4 address.

      --
      Live today, because you never know what tomorrow brings
    5. Re:2013 could be... by rvw · · Score: 1

      The year of IPv6 on the desktop!

      It should be branded iPadV6 and then sold in clean stores all over the world. I bet it would be commonplace in no time, especially when the Chinese try to copy it.

    6. Re:2013 could be... by DarwinSurvivor · · Score: 1

      According to the summary, the Chinese are *ahead* of us!

    7. Re:2013 could be... by tlhIngan · · Score: 1

      And by annoyance I mean like some stupid software, I don't remember what would prefer the IPv6 address over the IPv4 address then leading to a delay before it would connect via IPv4. I couldn't be arsed to find some other solution, so IPv6 is completely disabled on my machine. And so far I've had zero reason to change that. The only people feeling the hurt are those not getting an IPv4 address.

      That would probably be your resolver, since most programs use getaddrinfo() to do a DNS lookup (which returns a linked list of addrinfo structures containing the addresses (in very convenient sockaddr format, so no more pesky casting of sockaddr_in/sockaddr_in6/sockaddr_whatever to sockaddr anymore). It's protocol independent and gets you IPv6 support "for free" (getaddrinfo()/getnameinfo() were created to extend gethostbyname and the like for IPv6). Of course, it doesn't help as DNS records are often returned by IPv4 only so IPv6 connectivity cannot be verified even though that DNS query returned IPv6 records.

      I suppose the biggest issue I have with IPv6 is having to have the network renumbered whenever my ISP decides it's time to renumber the network (roughly annually), and since the ISP gives you your prefix, it means every host has to get a new address. Sure IPv6 gives you the ability to add more addresses (like the private address space) per host, but you lose out in the convenience of being able to ping the router and know you can get traffic through it (because you'll ping the router's private address (or worse yet, link-local) without realizing some software somewhere screwed up and either handed out bad internet-accessible DHCPv6 addresses, or a rogue device is sending bad router solicitations. So no connectivity with connectivity.

      It's one of the niceties of NAT - that your inner network is isolated from the outer network so the whims of the outer network provider don't impact your inner network (or minimally - reboot the damn router).

      Yes I know IPv6 is better, but I'm sure a lot of people don't really care to have end-to-end connectivity (and more will APPEAR to have it, but it doesn't work because of firewalls and such), and just want a simple box that they can plug in and replace their existing Linksys router with and not worry about a thing - router hands out DHCPv6/router solicitations, hosts configure themselves, and router NATs the inner network from the outer. If you can ping the router, you can get connectivity and not have to figure out if it's because you're using the link-local or private or the actual internet-routable address. (Try explaining that concept to your parents).

    8. Re:2013 could be... by Creepy · · Score: 1

      Already pretty much debunked because an older, longer calendar exists, but you can always give me all your earthly belongings, just in case.

    9. Re:2013 could be... by BlueBlade · · Score: 1

      The idea with IPv6 is that, even though your network prefix will be assigned to you by your ISP and is subject to change (for example, if you move to a new ISP), you typically won't configure any device with a fixed prefix. You'll assign them a host address (through DHCP, router advertisements or static) and the the prefix will be assigned to your router only. For example, on a cisco router, you would use :
      ipv6 general-prefix ISP-prefix XXXX:XXXX:XXXX::/48

      Everything else will be using that general prefix, gotten from the core router. If, for some reason, you later move to a new ISP with a new prefix, then you only have to change your general prefix. Your internal network adresses won't change, they'll remain the same, except with the auto-appended new general prefix. Pretty much just like what you'd get with NAT right now.

      It's very elegantly designed, but it takes a while to wrap your head around all its intricacies, especially if you're very used to the IPv4 way of doing things.

      --
      Religion is the best example of mass psychosis
    10. Re:2013 could be... by Anrego · · Score: 1

      Especially if you're very used to the IPv4 way of doing things

      I suspect that's a big part of the reason why adoption has been so slow. IPv6 is annoyingly different. You pretty much just have to force yourself to accept that you have to do things differently, and a lot of people don't like that.

      NAT is ugly, but people are very comfortable with the way it works.

    11. Re:2013 could be... by unixisc · · Score: 1

      Nope. People have IPv6 enabled browsers so they can connect to IPv6 enabled websites, but how many have some sort of legacy software that doesn't in any way understand or support IPv6 - perhaps there's not even an input field for an IPv6 address. Of course people will now chime "dual stack" but it has practically all of the annoyances while not solving the problem since it means pairing every IPv6 address with an IPv4 address. And by annoyance I mean like some stupid software, I don't remember what would prefer the IPv6 address over the IPv4 address then leading to a delay before it would connect via IPv4. I couldn't be arsed to find some other solution, so IPv6 is completely disabled on my machine. And so far I've had zero reason to change that. The only people feeling the hurt are those not getting an IPv4 address.

      This! Fact remains that since most people are now either on Windows 7 or OS-X or Android or iOS, most devices support IPv6. If an ISP supports IPv6 now, then a customer could easily request a link, and hook all the devices he wants to that. As it is, the default local networking protocol on Windows 7 is IPv6.

      Issue is that most websites are not IPv6 enabled, which would make accessing them difficult from IPv6-only computers, which is why people still need to be dual stacked, which somewhat defeats the purpose of moving to IPv6 from the POV of the ISPs. The pressure needs to be on websites to be available on IPv6 as well, but right now, too few of them are.

      Other problem is equipment. In my office, we recently got a new network and a bunch of IP phones (8x8 polycom). There is nothing on the phones or the site about IPv6 support, but the phones were preconfigured and worked just fine, albeit under NAT. But I've heard that when IP phones are set up, they need to have public addresses to them. This is the biggest place where IPv6 would be handy - once you have a switch i.e. an IPv6 link in an office network, each port would be a free IPv6 address, as opposed to the handful of IPv4 addresses one gets from the ISP. I would have gone w/ IPv6 had the phones supported it, but they don't.

    12. Re:2013 could be... by squiggleslash · · Score: 1

      I wouldn't say it's annoyingly different. One of the best parts of IPv6 is that once you have the core infrastructure set-up (ie a NDP-broadcasting gateway with a valid prefix - which is actually much, much, simpler than the IPv4 "NAT router with DHCP server" that everyone is used to) everything "just works".

      - You don't have to assign static IP addresses. They're already static.
      - You don't have to forward ports. NAT isn't getting in the way any more. Everything sees everything else.
      - Discovery takes care of things like default domains.

      It kinda reminds me of a story about OS/2's early development. IBM wanted to test the developer-friendliness of the API, so they brought in a group and asked them to do similar tasks under OS/2 and other OSes such as Mac OS. Supposedly the OS/2 development took slightly longer, because the OS/2 developers kept asking questions along the lines of "How do I do this {hack I do under Windows, GEM, and Mac OS}" and wouldn't understand the answer "You don't, it's not necessary, we've eliminated the need to do that hack."

      To be fair, while it "just works" it comes at the price of needing to understand that you now, thanks to the fact your network now works properly, need to take care of security. Still, IPv6 includes half the tools (mandatory IPSec support, for instance), and your operating system the others (software firewalls.) All of these are things that you should be doing anyway, but you don't because NAT lulls you into a false sense of security, a sense broken the first time you let someone with a virus infested laptop use your network.

      --
      You are not alone. This is not normal. None of this is normal.
  4. Come on slashdot ... by Anonymous Coward · · Score: 5, Insightful

    This is not what I expected from you when facebook and google enabled it long ago ...

    1. Re:Come on slashdot ... by canadiannomad · · Score: 5, Informative

      The problem I have faced is that none of my server ISPs will even let me get an IPv6 address even if I know they have it and I beg. That goes for major service providers too. I'm looking at you Amazon Cloud and RackSpace. Amazon kinda has it, but only if you use one of their load balancers.

      --
      Hmm, the humour and sarcasm seem to have been be lost on you.
    2. Re:Come on slashdot ... by mellon · · Score: 2

      I use linode.com. They have IPv6 (and have for quite some time). Now if only voip.ms (and oh, say, /.) would support it...

    3. Re:Come on slashdot ... by dkf · · Score: 2

      The problem I have faced is that none of my server ISPs will even let me get an IPv6 address even if I know they have it and I beg.

      They'll come round once they start having problems getting more IPv4 addresses from their upstream providers, at which point it will start to hit their bottom line (as they need to have all their cloud instances individually direct-routable for configuration and management purposes). We're getting close to that, but aren't there yet.

      I wouldn't base any long term plans on them staying IPv4 only...

      --
      "Little does he know, but there is no 'I' in 'Idiot'!"
    4. Re:Come on slashdot ... by marcosdumay · · Score: 2

      They'll come round once they start having problems getting more IPv4 addresses from their upstream providers

      They won't, they'll just put everybody behind a NAT, with the added bonus of breaking bittorrent, VoIP, or any other protocol that actualy uses bandwidth.

      IPv6 will only come later, and just for the places that have any competition between ISPs.

    5. Re:Come on slashdot ... by Kergan · · Score: 1

      Note that Slashdot is in no way exceptional here:

      http://ipv6-test.com/validate.php?url=apple.com
      http://ipv6-test.com/validate.php?url=microsoft.com
      http://ipv6-test.com/validate.php?url=yahoo.com
      http://ipv6-test.com/validate.php?url=oracle.com
      http://ipv6-test.com/validate.php?url=twitter.com

      In light of the above bad examples, I was actually surprised that Internic, ICANN and the White House were IPv6-ready...

    6. Re:Come on slashdot ... by Anonymous Coward · · Score: 0

      Step 1 - Pony theme (check)
      Step 2 - Unicode support
      Step 3 - IPv6 support
      Step 4 - ?
      Step 5 - Profit?

    7. Re:Come on slashdot ... by jandrese · · Score: 1

      The US government has a mandate to support IPv6, which is why most .gov sites can be reached over IPv6. Otherwise they would be in the same boat as everybody else.

      What we really need is for Comcast, Verizon, Time Warner, etc... to enable IPv6 on their networks. Almost nobody is going to set up a tunnel broker for their home connection, it's way to esoteric and most home routers are crap anyway.

      --

      I read the internet for the articles.
    8. Re:Come on slashdot ... by Anonymous Coward · · Score: 0

      Comcast will have their entire network enabled by the end of 2013. If you haven't done it yet, upgrade your home router to something that supports DHCPv6. (So any new router).

      "IPv6 has been launched on all Arris DOCSIS 3.0 C4 CMTSes, covering over 50% our network. We are targeting completion of the rest of the network by mid-2013. Our progress has led to nearly 2.5% of our Xfinity Internet customers actively using native dual stack. Additionally, IPv6 traffic has increased 375% since World IPv6 Day in June 2011. Following World IPv6 Launch in June 2012 Comcast also observed that approximately 6% of the 2012 Olympics served over YouTube to Comcast customers was over IPv6. " - http://comcast6.net/

    9. Re:Come on slashdot ... by Anonymous Coward · · Score: 0

      I've found that US web companies are dragging their heels with ipv6 more than the rest of the world. I've been on to our host (Verio/NTT) for years about getting ipv6 connectivity to our solution but they still haven't rolled anything out. Even our CDN (Akamai) who is meant to be promoting ipv6 wants to charge an extra fee (quite a large one) to enable ipv6 on our account.

    10. Re:Come on slashdot ... by Anonymous Coward · · Score: 0

      They'll come round once they start having problems getting more IPv4 addresses from their upstream providers, at which point it will start to hit their bottom line (as they need to have all their cloud instances individually direct-routable for configuration and management purposes).

      What? All they need is a single IP for each physical datacenter. All the consumer needs to do is be able to reach the edge of their network. The consumer should not know or care if the request is being sent to cluster A or B within that network.
      Applications shouldn't be concerned with IP addresses to start with. If we quit mixing multiple layers of the OSI model in places we shouldn't be, not only would ipv6 not be as immediate of a need but things like NAT would become a complete non-issue. Hell, we already basically ignore the port number in a packet and go directly to the URL request for web requests, that's how it ought to work for basically anything.

  5. New Rule: by ledow · · Score: 4, Interesting

    New Rule:

    Websites are only allowed to try to garner page-views on IPv6 when all the websites that article is posted on are available over IPv6.

    1. Re:New Rule: by mellon · · Score: 1

      Good luck with that. :)

    2. Re:New Rule: by NatasRevol · · Score: 1

      Maybe get a Congressman from California to pass that as a new law.

      --
      There are two types of people in the world: Those who crave closure
    3. Re:New Rule: by Anonymous Coward · · Score: 0

      sorry we are all busy driving up to washington

  6. Slow news day by AltF4ToWin · · Score: 1

    This is one.

  7. IPv6 was no big deal by Cimexus · · Score: 3, Informative

    I've been on native IPv6 for a couple of years on my home DSL connection. It works very well - only thing I had to do was check the 'enable IPv6' option in my modem/router and everything 'just worked'. It is rather nice not having to deal with NAT and port forwarding etc.

    I'm in Australia (so within the Asia-Pacific/APNIC region, which as the summary mentions, is a bit ahead of the curve when it comes to IPv6 adoption. Most of the major sites are fully IPv6 now too (e.g. all the Google sites, Facebook, etc. etc.) But the point is, done properly, it should be a completely seamless transition to enable dual-stack (and eventually to turn off IPv4, though I'm sure that won't happen for decades!). Hell I usually forget I'm even on IPv6, unless I happen to do a ping/tracert to an IPv6 host and see all those long-ass IPs :)

    C:\>tracert www.google.com

    Tracing route to www.google.com [2404:6800:4006:800::1014] over a maximum of 30 hops:

        1 1 ms 1 ms 1 ms gateway [2001:44b8:(snip!)]
        2 7 ms 7 ms 7 ms loop0.lns20.cbr1.internode.on.net [2001:44b8:9010::5]
        3 7 ms * 7 ms gi0-0-2.cor3.cbr1.internode.on.net [2001:44b8:9010:14::1]
        4 11 ms 11 ms 11 ms te6-0-0.bdr1.syd4.internode.on.net [2001:44b8:9010:e::2]
        5 11 ms * 11 ms te0-0-0.bdr1.syd7.internode.on.net [2001:44b8:b070:1::11]
        6 11 ms 11 ms 11 ms gi1-2-121.cor2.syd7.internode.on.net [2001:44b8:b060:121::2]
        7 11 ms * 12 ms gi6-0-0-101.bdr1.syd7.internode.on.net [2001:44b8:b070:104::1]
        8 12 ms 11 ms 12 ms 2001:4860:1:1:0:1283:0:4
        9 13 ms 13 ms 12 ms 2001:4860:0:1::1fb
      10 13 ms 12 ms 11 ms 2404:6800:4006:800::1014

    1. Re:IPv6 was no big deal by Anonymous Coward · · Score: 1, Insightful

      Yes your home DSL network is completely and thoroughly analogous to an enterprise implementation.

      Just check some stuff off in the UI and run a dual stack...

    2. Re:IPv6 was no big deal by jones_supa · · Score: 0

      I've been on native IPv6 for a couple of years on my home DSL connection. It works very well - only thing I had to do was check the 'enable IPv6' option in my modem/router and everything 'just worked'. It is rather nice not having to deal with NAT and port forwarding etc.

      I've been on native IPv4 for ages on my home DSL connection. Only thing I had to do was tick the "NAT" and "UPnP" checkboxes and it just worked. Provided me also with a simple firewall as a nice side effect. ;)

    3. Re:IPv6 was no big deal by Anonymous Coward · · Score: 0

      So are you suggesting that, because you don't mind having NAT on your home network, the rest of us should give up on trying to move the Internet forward to a more sustainable scheme?

    4. Re:IPv6 was no big deal by Anonymous Coward · · Score: 2, Informative

      > Provided me also with a simple firewall as a nice side effect. ;)

      I rather suspect you're a troll given how often and exhaustively this has been refuted previously on this site, but oh well.

      NAT is not a firewall. The stateful firewall in your home router is a firewall. NAT isn't . There are plenty of technologies to punch holes into NAT, usually developed because NAT is such a fucking pain to deal with for many protocols.

    5. Re:IPv6 was no big deal by Cimexus · · Score: 1

      Well any decent router that offers NAT will also have an actual firewall as well. Which will continue to work on IPv6 ... so it's not like you're unprotected or anything. I'm not one of those rabid "NAT is an awful hack" people, but it's just another tool in the box that ceases to really have a purpose in the IPv6 world.

      Anyway, if your point is that you don't NEED IPv6 (yet), then you're absolutely right. I was merely offering my experience of IPv6 adoption as an end user. And that it was pretty painless (both for me personally, but also for the ISP - they ran quite a few blog articles about the migration and since it was done in a steady, staged manner in the course of the normal business of replacing hardware and reconfiguring networks, it didn't really cost them too much extra time or money).

      The reason I did this is that it seems to me that some organisations are very stuck in the mud about IPv6 - they dig their heels in and take the attitude that it's preferable to use workarounds to make the existing IPv4 space last longer, than just bite the bullet and get IPv6 up and running. Note that I don't mean people that simply have a "what I have now works fine, so why bother" opinion - that's a perfectly reasonable position to take. Or those that have a real reason they can't do it (legacy hardware/software, in enterprises particularly). But there are also those that actively try to reject IPv6 for no reason other than that they seemingly don't like the idea of it. They aren't just ambivalent about it - they go out of their way to find reasons not to use it. The reluctance they have for implementing IPv6 is not commensurate with the actual difficulty of doing so. Hell I was bit like that myself, but my ISP did things well and it was so brain-dead easy that I kinda feel silly for having been like that.

    6. Re:IPv6 was no big deal by Anonymous Coward · · Score: 0

      I just tried my luck on hacking your gateway at 2001:44b8:(snip!). It worked as well as self-castration.

    7. Re:IPv6 was no big deal by Cimexus · · Score: 1

      Haha. Am reasonably confident my router's firewall is up to the task, but because the IPv6 /64 assigned to my router (and the /56 prefix assigned to my network) is static, figured I'd better remove it, just in case :)

    8. Re:IPv6 was no big deal by jones_supa · · Score: 1

      Naah! I just like to flip things around. I have nothing against IPv6 adoption in general.

    9. Re:IPv6 was no big deal by Rockoon · · Score: 1

      only thing I had to do was check the 'enable IPv6' option in my modem/router and everything 'just worked'.

      I would have enabled it long ago 'cept my router doesnt support it, not does it support any of the alternative (open) firmwares available.. this in spite of both previous and later models of this same line of routers being supported by those alternative firmwares.

      I would buy a new router, 'cept this one works and has a decent enough range/power/sensitivity that the bedroom machine on the other side of the wet-wall and about 40 feet away doesnt have issues. It needs a reboot every month or so due to some resource leak, but thats not a good enough reason to risk playing whack-a-mole with a new router that may not work as well.

      --
      "His name was James Damore."
    10. Re:IPv6 was no big deal by wallbase · · Score: 1

      Well duh, it's Internode. Of course you're gonna have an easy time with IPv6. OK the modem/router has to support it of course, but with Internode at least it's easy enough to enable it on your account. I'd expect nothing less from those guys.

      --
      Dude...
  8. A new apocalypse. by concealment · · Score: 2

    Dear Media,

    Every week, there's a new apocalypse in the news.

    AIDS. Global Warming. Copyright violations. Vodka enemas. Terrorism. ???. Prophet. (I mean... profit.)

    The IPv6 lolocaust is not going to impress us unless there are concrete figures about exactly when and how it's going to devastate us.

    Then, we can plan for it.

    Until then, it reeks of hype.

    Love,
    The consumers

    1. Re:A new apocalypse. by Sique · · Score: 1

      This is an information for the experts who deal with ip address assignments everyday. It's news for nerds, not news for consumers. Ideally, a consumer should never even encounter an IP address, be it IPv4 or IPv6.

      --
      .sig: Sique *sigh*
  9. What about Slashdot? by quarkoid · · Score: 1

    It would be nice if Slashdot itself was available on IPv6... After all, you would have thought that a site reporting on the latest and greatest in tech would have managed to adopt a technology fifteen years old by now!

    1. Re:What about Slashdot? by KingMotley · · Score: 2

      You mean like an edit button?

    2. Re:What about Slashdot? by somersault · · Score: 1

      An edit button would drive trolling levels off the scale here.. though it would be nice to correct those moments where you click submit and notice an incredibly stupid typo. I use preview when I remember, but sometimes things just slip by..

      --
      which is totally what she said
    3. Re:What about Slashdot? by DarwinSurvivor · · Score: 1

      An edit button that only works for 10 seconds would solve SO many problems and give trolls very little leverage to play with.

    4. Re:What about Slashdot? by squiggleslash · · Score: 1

      They're going to implement it just as soon as Slashcode's Unicode support is ready...

      --
      You are not alone. This is not normal. None of this is normal.
    5. Re:What about Slashdot? by mcgrew · · Score: 1

      There is an edit button. It's labeled "preview". If you could edit your comments after posting, you could post a funny comment, have it modded to +5, and then change it to a GNAA troll.

      Just use the preview button and pay attention.

  10. Provider slowness. by PlusFiveTroll · · Score: 5, Insightful

    IPv6 Capable operating systems: check.
    IPv6 Capable router: check.
    IPv6 Capable cable modem: check.
    IPv6 Capable internet service: .........

    Maybe one of these years the cable company will get this figured out, sigh.

    1. Re:Provider slowness. by Anonymous Coward · · Score: 0

      Some people still have legacy hardward and/or legacy software, hence one of the problems. Although, while my machines aren't necessarily IPv6 friendly, such as my Linksys CIT400 iPhone (it's a Skype phone, sharing the same trademark as Apple's iPhone), I would suggest as many devices as possible be IPv6 as default to help move things along. We're going to have to use both IPv4 and IPv6 during a long transition period, in my opinion, but eventually it will happen.

    2. Re:Provider slowness. by ledow · · Score: 1

      Unfortunately, the next step:

      IPv6 Capable end-points:

      is also missing. The website you are reading is IPv4-only and hasn't bothered to publish an AAAA record in all the time it's been posting IPv6 articles.

      And how long does it take to IPv6 enable a website nowadays, even if only in a basic "testing" mode before you try to redo all your blacklist scripts, etc.? About ten minutes.

    3. Re:Provider slowness. by mrvan · · Score: 1

      xs4all (Netherlands) does provide ipv6 and I find it quite useful to have direct links between computer that would otherwise be difficult to reach (e.g. between my computer at home (ipv4 NAT by the ADSL router) and a virtual server at work for which I didn't get an ipv4 address and hence only has pulic ipv6.

      Also, it is useful to be able to connect directly to my home box from outside, and there are multiple ssh enabled machines on my LAN. Of course, I could give them all different ports and forward them using the router, but that is just so much more hassle...

    4. Re:Provider slowness. by canadiannomad · · Score: 1

      I don't know who hosts slashdot, but I bet they are the ones that are slagging.

      --
      Hmm, the humour and sarcasm seem to have been be lost on you.
    5. Re:Provider slowness. by NatasRevol · · Score: 1

      Just curious, but do you see lots of hack/hit attempts on those now exposed boxes like you do if you put an IPv4 box on the internet with ssh open?

      --
      There are two types of people in the world: Those who crave closure
    6. Re:Provider slowness. by compro01 · · Score: 1

      Cable companies have it somewhat easy. DOCSIS 3 requires hardware to support IPv6.

      --
      upon the advice of my lawyer, i have no sig at this time
    7. Re:Provider slowness. by Cimexus · · Score: 1

      Yeah it'll be dual stack for a long, long time. Though must say, I currently have 11 devices connected to my internal home network, and 9 of them have a globally addressable IPv6 address. The Nintendo Wii and the WDTV Live are the only non-IPv6-capable devices in the house, apparently. The rest are all fine and grabbed a v6 address with no additional config needed.

      (2 Windows machines, 1 Mac OS X, 2 Linux, 2 iPhones, 1 iPad and a D-link NAS, presumably running some embedded Linux).

    8. Re:Provider slowness. by characterZer0 · · Score: 1

      Does DOCSIS 3 require that the ISP actually route IPv6 packets? Because if it does not, they will not.

      --
      Go green: turn off your refrigerator.
    9. Re:Provider slowness. by Alioth · · Score: 2

      Network scanning is much, MUCH harder to do with IPv6. Assuming a reasonably random assignment of v6 addresses, your local subnet has 2^64 possible addresses, in other words twice as many bits as the entirety of the world's IPv4 address space. But remember twice as many bits doesn't mean merely twice the effort, it actually means it would take 4 billion times as long to scan a *single* IPv6 subnet as it would to scan the *entire* IPv4 internet.

    10. Re:Provider slowness. by RazzleDazzle · · Score: 1

      Even for a couple of servers that do not have an external firewall filtering packets for my IPv6, there is basically zero packets besides those going to applications hosted on my servers, and they have published DNS records for web and DNS. Some basic PCs I have online see zero packets from random internet hosts on IPv6.

      The IPv6 address space is literally too large to crawl within any useful amount of time. If you figure an average LAN will have 18,446,744,073,709,551,616 addresses (a /64 block). Let's say you scanned 1000 IP addresses per second (very optimistic for a single PC) it would take you 584,942,417 years to complete scanning just a single LAN. Then are so many /64 LAN blocks that it is very likely you're scanning an network block that does not have any hosts to begin with.

      There will have to be other means to gather active/in-use IP address such as looking at server logs that clients connect to, email headers, DNS records, soliciting traffic from the client machines via some application/trojan/virus, network traffic sniffing, etc. All of these means already exist for IPv4 so there is nothing new there.

      --
      ZERO ZERO ONE ZERO ONE ZERO ONE ONE! Just brushing up for my next big invention: Ethernet over Voice (EoV)
    11. Re:Provider slowness. by rsmith-mac · · Score: 1

      Blame Cisco.

      The major ISPs are generally ready, but they can't do anything until Cisco actually rolls out IPv6 capable head-end gear. Among other things, Cisco has already all but missed their 2012 deadline for having IPv6 working on their CMTSes.

    12. Re:Provider slowness. by NatasRevol · · Score: 1

      Thanks. Makes sense given the size.

      I see the next round of software like netcat being much faster/parallel to deal with these huge sizes.

      --
      There are two types of people in the world: Those who crave closure
    13. Re:Provider slowness. by sp332 · · Score: 1

      Setting up an IPv6 tunnel is not hard to do. A couple minutes and you'll have IPv6 internet access. tunnelbroker.net (just for example) walks you through it, then you can install 6orNot in your browser to show off :)

    14. Re:Provider slowness. by imlepid · · Score: 0

      You might be surprised to find out how many people fail in one, if not multiple of the points you mentioned. Take, for example, me:

      IPv6 Capable operating systems: Not really. I run Mac OS X 10.6, which, wile "IPv6 capable" does not have support for a critical IPv6 component DHCPv6.
      IPv6 Capable router: Not really. My router does not support IPv6 without some serious hacks. Plus it doesn't support DHCP-PD at all.
      IPv6 Capable cable modem: Yes, but only because I just (two months ago) bought a new modem.
      IPv6 Capable internet service: Yes, and it's been available from my ISP for a long time.

      The major problem with the majority of devices is not the "first level" IPv6 support (e.g. ability to get an IPv6 address via SLAAC) but second level and beyond (DHCPv6, etc). IPv6 is a protocol which is still very young and not "fully" supported by most software/hardware, mostly because it is still changing. It will be a long while before IPv6 has the maturity of IPv4. I just laugh when I read marketing drivel with statements like "IPv6 supported!" because until they provide more details, I just assume that it means it can self-assign a link local address and that's all.

    15. Re:Provider slowness. by PlusFiveTroll · · Score: 1

      Oh, I'd add more then a few IPv6 end points to the net if the cable company would provide IPv6 to the business fibre service here. For now we stack services on ip addresses via NAT. Fun to scan an IP and see Windows and Linux services living at the same address.

    16. Re:Provider slowness. by SuricouRaven · · Score: 1

      Not at all practical even then. You'd need some way other than scanning. Maybe a set of broadcasts designed to provoke a response in some way from common servers, tricking them into revealing their presence. Or passive monitoring.

    17. Re:Provider slowness. by jandrese · · Score: 1

      You can always set up a tunnel broker if you want to play around with IPv6. It's ugly and messy, but it gets you on the network. I've been using this very service for over two years now and have never had a problem with it. As far as supported devices, I have an original 2G iPhone, a Nintendo Wii, an ancient Sharp Zarius, and a pair of TiVos that don't support IPv6, but everything else was able to autoconfigure an address and use it right away.

      Also, it is entirely possible to run a pure IPv6 network today though a combination of DNS and packet translation. Basically, your IPv6 only hosts do a name lookup for a host using the local DNS server. The server queries but only finds a A record, which it then translates into a AAAA record by appending a well known prefix and returns that to the host. That host then opens a socket to the IPv6 address specified and sends the packet through a static NAT setup on your gateway to translate the packets back to IPv4, and also translate return packets to IPv6. It sounds esoteric, but this setup works (I use it in an IPv6 only test environment) and isn't hard to configure at all if you're using a Linux gateway. I've been using the Trick or Treat Daemon for DNS conversions and Tayga for the IP packet translation.

      --

      I read the internet for the articles.
    18. Re:Provider slowness. by compro01 · · Score: 1

      No, but it requires the hardware to support it, making it considerably easier to convince then to start routing.

      --
      upon the advice of my lawyer, i have no sig at this time
    19. Re:Provider slowness. by CAPSLOCK2000 · · Score: 1

      I've been using IPv6 for over 10 years and scans are very rare. I've just checked this months logs and I cannot find a single attempt to connect to ssh from an unknown host (it has happened in the past). I think that if that machine didn't have an 'easy' IPv6 adress but a randomly generated one it would never have happened.

    20. Re:Provider slowness. by Anonymous Coward · · Score: 0

      Maybe because adoption of ipv6 is so low that no one bothers yet.

      Heard the same mentality about mac and viruses, even apple had to change the statement on their site after 500K machines kept getting infected.

    21. Re:Provider slowness. by squiggleslash · · Score: 1

      6to4 works on the majority of ISPs too and doesn't require any arrangements with a tunnel broker. I use it myself - originally on Earthlink, now on Comcast, and not had a single problem.

      Modern routers (as in the cheap D-Link crap that combines a Wi-fi hub, PPPoE/DHCP client, and local DHCPD server in a $30-50 box) actually support 6to4 out of the box too. Unless you're one of the unlucky few who uses an ISP that actually blocks 6to4 there's not a lot of reason to avoid it.

      --
      You are not alone. This is not normal. None of this is normal.
    22. Re:Provider slowness. by squiggleslash · · Score: 1

      I think the majority of vendors are avoiding DHCPv6 because it's a hack and doesn't really fit the "Just works" aspects of IPv6. NDP generally does exactly what's needed. Virtually all operating systems support it out of the box.

      I'm finding the majority of cheap routers (D-Link $50 Wi-fi + DHCP etc types) I see these days have some level of IPv6 support, usually allowing the setting up of 6to4 etc.

      --
      You are not alone. This is not normal. None of this is normal.
    23. Re:Provider slowness. by kqs · · Score: 1

      Big numbers are big.

      Let's say you have a 10gig connection between you and the target network, for your use only. 10 gig means you can send 10 billion bits per second. And let's say each one of those bits could test one IP address on that one small 64 bit subnet (which is crazy, but why not).

      In that case, it would only take you about 6 years. To scan 10% of the subnet. And most providers are giving out hundreds or thousands of subnets to each house.

      Parallelism and speed increases will not help here!

    24. Re:Provider slowness. by WaffleMonster · · Score: 1

      Cable companies have it somewhat easy. DOCSIS 3 requires hardware to support IPv6.

      From what I understand cable companies are sitting on their thumbs waiting for multiple vendors including Cisco to fix broken code at the CMTS so operators can actually deploy IPv6 to their customers.

    25. Re:Provider slowness. by Anonymous Coward · · Score: 0

      IPv6 Capable operating systems: check.
      IPv6 Capable router: check.
      IPv6 Capable cable modem: check.
      IPv6 Capable internet service: .........

      Maybe one of these years the cable company will get this figured out, sigh.

      You forgot the Billing, Inventory, and Provisioning systems, not to mention all the internal tools used to support, log, and monitor those systems.
      I work for an ISP, and while yes our switches, routers, CMTS's, cable modems, and OS's are all capable of ipv6, those are just the basics needed to establish a network. Those are the trivial things to get v6 working on, it's all of the other "back end" systems needed to run the operations which are the tricky part. And that doesn't even start to address enterprise issues, such as all the proprietary in-house or 3rd-party systems used by the offices, phone support people, etc.

    26. Re:Provider slowness. by Anonymous Coward · · Score: 0

      Oh, I'd add more then a few IPv6 end points to the net if the cable company would provide IPv6 to the business fibre service here. For now we stack services on ip addresses via NAT. Fun to scan an IP and see Windows and Linux services living at the same address.

      Why the hell would you do that? The only thing the IP scan should reveal is your external-facing firewall giving the middle finger to the intrusion attempt. I also don't know why you'd need to rely on something as weak as NAT to stack services in the first place. Get a decent firewall and set it up and you don't need multiple IP's in the first place.

      The port number should represent a type of protocol, not a particular instance of a protocol, but many applications use it as an absolute indicator of an instance. So your firewall/loadbalancer has to dig into the packet's payload and determine which internal resource to send it to, which is what we typically do these days for things like hosting services where you have multiple "sites" all running in VM's on a single physical machine. We should be focusing on resolving this by adding some type of additional header field (or sub-header) representing a particular instance of an application, not just exposing every system directly to the outside world.
      This is why people view NAT as being "bad" or hard to work with- it's not NAT's fault at all, the problem is because we're still using port numbers and trying to treat the internet as an endpoint to endpoint network. We've advanced our networks and systems to the point where an "endpoint" may be a single device, a cluster of devices, a virtual cluster, or even an entire network, but we have not addressed this fundamental shift in design philosophy in the TCP/IP layer. The solution is to fix that approach, not to simply throw more IP addresses at the problem, regardless of whether we're using v4 or v6.

      Don't take this the wrong way- I'm looking forward to ipv6. v4 really is getting outdated, and it's just far too limited in terms of the available ranges and how we've carved blocks of IP's out in the past. But most people are looking at it as a solution for the wrong problem.

    27. Re:Provider slowness. by Anonymous Coward · · Score: 0

      Even for a couple of servers that do not have an external firewall filtering packets for my IPv6, there is basically zero packets besides those going to applications hosted on my servers, and they have published DNS records for web and DNS. Some basic PCs I have online see zero packets from random internet hosts on IPv6.

      The IPv6 address space is literally too large to crawl within any useful amount of time. If you figure an average LAN will have 18,446,744,073,709,551,616 addresses (a /64 block). Let's say you scanned 1000 IP addresses per second (very optimistic for a single PC) it would take you 584,942,417 years to complete scanning just a single LAN. Then are so many /64 LAN blocks that it is very likely you're scanning an network block that does not have any hosts to begin with.

      There will have to be other means to gather active/in-use IP address such as looking at server logs that clients connect to, email headers, DNS records, soliciting traffic from the client machines via some application/trojan/virus, network traffic sniffing, etc. All of these means already exist for IPv4 so there is nothing new there.

      That's a lot of text to say "security by obscurity".
      Look, back in the "old days" of slow modems and CPU's and limited bandwidth, the idea of scanning all 65k ports on the entire ipv4 network was equally silly. But look what happened- people started using the same port # hard coded into applications, and it's only in relatively recent years, and largely spurred by people trying to deal with NAT, that applications are flexible on their port assignments. I foresee application developers and network operators getting lazy again in the future and basically doing the same damn thing. It's nice to have a shitload of IP's to pick from, but when everybody defaults their devices to address number 420 or 69 or 42 (depending on if you're pothead, sex fiend, or Hitchhiker fan) it renders the entire thing moot. And once someone does sniff a source IP or just accidentally stumble on it or make a good guess, poof! there goes your "protection".

    28. Re:Provider slowness. by Anonymous Coward · · Score: 0

      Modern port scanners don't hit every port even on an ipv4 network. They pick the commonly used ones. People will get lazy and you'll end up with your Dell printer demanding that it only be assigned to one specific address (or small range of addresses) on the subnet. People will pick addresses on the subnet which are "clever" or otherwise more memorable, etc. It won't be all that different because people tend to create patterns of behavior which can be anticipated. Best Practices will say to use randomized address schemes, but best practices are often ignored.
      Just as an example, I was playing around with v6 on an internal test network and what I found is that the ipv6 DHCP server in the common brands of consumer routers still starts at the first available address and simply increments one address for each new device request, as opposed to random assignment out of the entire available scope. So as long as the scanner starts at the beginning, if obscurity is all you're relying on they'll have all your active IP's in less than a second.

       

    29. Re:Provider slowness. by Anonymous Coward · · Score: 0

      If you want to start testing ipv6 before your ISP support native ipv6 in your area, you can use tunnel. Try http://www.tunnelbroker.net from HE.

      They also have a very cool certification program. It's some kind of tests that prove to the other you have a fully working ipv6 environment. You need to setup an http web server, dns records and route emails over ipv6. It's very interesting and not that easy to achieve.

      I completed all the levels and just got my ipv6 free t-shirt. It's really nice.

      Have you tried to ping6 www.facebook.com ? Their ipv6 address is cool. ;)

      If you want, you can also check my ipv6 monitoring page that compare the loading speed in ipv6 and ipv6 of google and facebook. The page is only available to ipv6 users. http://graph.jeandebogue.com

      Check it out and if you would like to add your own network to the monitoring page, please let me know at jean@wedebugyou.com

      Also check my bad ipv6 address.

      ping6 mybadassipv6.jeandebogue.com

  11. rest of that headline by slashmydots · · Score: 1

    ...and I'm already sick of typing that many extra digits to ping something.

    1. Re:rest of that headline by Ultra64 · · Score: 1

      What is DNS?

  12. Re:Slashdot slowness. by Anonymous Coward · · Score: 0, Insightful

    If you have an IPv6-only host, you can't reach slashdot.org

    You can get google.com and facebook.com.

    Also missing: amazon.com, microsoft.com, hotmail.com, nytimes.com, kernel.org, github.org, lwn.net etc etc.

    Your IPv6 experience is going to be pretty poor.

  13. Just using it by Anonymous Coward · · Score: 0

    Using aaisp.net.uk and it all just works,

    Facebook, google, youtube, world of warcraft... all work perfectly on ipv6. More than half my data now goes over it without me doing *anything* to make it work :)

  14. the real game changer: 4G by anarcat · · Score: 3, Interesting

    The game changer here is that US cell phone companies have finally figured out that 4 layers of NAT isn't exactly a great way to manage a growing network, and are switching to IPv6 for their 4G networks. That is millions of customers right there, using IPv6 without even knowing about it.

    Pieces are falling into place, it's just a matter of time now. And if you lobby your ISP instead of complaining about it, you may get it native too soon enough.

    BTW: for those worried about the switch, let me just mention that both ipv6.google.com and www.kame.net (common test IPv6 addresses) are reachable in *less* latency and *less* hops than their ipv4 counterparts. IPv6 rocks.

    --
    Semantics is the gravity of abstraction
    1. Re:the real game changer: 4G by Anonymous Coward · · Score: 0

      T-Mobile USA has IPv6 available on all accounts on all Samsung phones, turn it on https://sites.google.com/site/tmoipv6/lg-mytouch

    2. Re:the real game changer: 4G by Anonymous Coward · · Score: 0

      The reason for this is as much that LTE mandates IPv6 support as the carriers realizing the flaws in NAT.

  15. CPE equipment from ZyXEL by fredan · · Score: 1

    ZyXEL's hardware sucks for CPE since they cannot do IPv6.

    1. Re:CPE equipment from ZyXEL by RazzleDazzle · · Score: 1

      Many ZyXEL CPEs can do IPv6. Check out their NBG4615, which is your typical home wifi/router appliance that supports IPv6. I think all or most of their current ADSL/VDSL CPEs all support IPv6 out of the box too.

      --
      ZERO ZERO ONE ZERO ONE ZERO ONE ONE! Just brushing up for my next big invention: Ethernet over Voice (EoV)
    2. Re:CPE equipment from ZyXEL by Anonymous Coward · · Score: 0

      Depends on the model, all newer ones should have it now. Look at the 26xxHNU series.

    3. Re:CPE equipment from ZyXEL by fredan · · Score: 1

      Model P-660HN-F1Z does not support IPv6.
      Model P-2612HNU-F1 does not support IPv6.

    4. Re:CPE equipment from ZyXEL by Anonymous Coward · · Score: 0

      Both of these were discontinued before this decade started.

      Funnily enough though my Beige G3 Macintosh running Mac OS X 10.2 ran IPv6 fine. Go figure.

  16. Why don't US companies implement IPv6? by Nutria · · Score: 1

    Inertia.

    I work for a large company that's had a domain since the Elder Days of domain registration, and there's just no way that it'll migrate over to IPv6. Too many computers and routers (including many legacy) and there's no actual need to do it: 10.*.*.* and 196.168.*.* networks abound and work just fine.

    --
    "I don't know, therefore Aliens" Wafflebox1
    1. Re:Why don't US companies implement IPv6? by Anomalyst · · Score: 1

      Must be nice that to have a company sandbox next to your keyboard in which to hide your head.

      --
      There is no right to feel safe thru security vaudeville at the expense of everyone's freedom, privacy and tax money.
  17. Insightful, but missing the point of the article. by concealment · · Score: 1

    Ideally, a consumer should never even encounter an IP address, be it IPv4 or IPv6.

    This is insightful and I agree for the most part. However, a consumer facing a "black box" is helpless. It's better to make the box simpler and more accessible so they can fix minor problems when they arise, since they arise with every technology we have on a regular basis.

    This is an information for the experts who deal with ip address assignments everyday. It's news for nerds, not news for consumers.

    Here I disagree.

    This a species of fear-inducing news articles designed to panic the consumers and induce them to put pressure on vendors, legislators and service providers.

    Hence my original comment.

  18. Ran out of IPv4? by Anonymous Coward · · Score: 1

    I've provisioned several thousand servers in the past month or so spread widely over the world and had no problems with availability of IPv4 addresses. Where's the problem? Is seems there's plenty of addresses for servers and clients can use NAT if there's a squeeze. No need for IPv6.

    1. Re:Ran out of IPv4? by Anonymous Coward · · Score: 0

      When you tell me that you provisioned several million desktops in the past month with IPv4 addresses, you can tell me there are plenty of addresses.

      NAT is a hack to get by, it's not very good, certain applications - especially anything point to point like VoIP - are broken by it. We haven't really had enough IP addresses since the late nineties and have had to resort to hacks like NAT to get anything done at all.

    2. Re:Ran out of IPv4? by Anonymous Coward · · Score: 0

      Try getting hold of some PI (Provider Independant) space. You can't. There's none left.

      Want your servers to have resilient links across multiple ISPs. You can't. You need PI space for that.

      Want to start a new ISP that doesn't do IPv6, and does it's own routing? You can't.

      You must have been fortunate enough to have an ISP that could provide you with some of their remaining space.

  19. Linux wasn't ready for IPv6 when I tried it by Anonymous Coward · · Score: 0

    I have IPv6 at home, but had to disable it: a bug in NetworkManager caused it to misbehave when it encountered a IPv6 DHCP on wifi. Long story short, I had a kernel panic every 30 minutes (+/- 10 min). Windows 7 via wifi? No problem. Linux via ethernet cable? No problem. Will have to check if it was patched since I last tried it.

    1. Re:Linux wasn't ready for IPv6 when I tried it by PlusFiveTroll · · Score: 1

      I have IPv6 at home, but had to disable it: a bug in NetworkManager caused it to misbehave when it encountered a IPv6 DHCP on wifi. Long story short, I had a kernel panic every 30 minutes (+/- 10 min). Windows 7 via wifi? No problem. Linux via ethernet cable? No problem. Will have to check if it was patched since I last tried it.

      That sounds like neither a bug in networkmanager nor with DHCPv6. Neither of those would cause a kernel panic. That's a bad wireless network card driver. Network manager did have issues with DHCPv6, but it's more on the lines of not setting the routing correctly

    2. Re:Linux wasn't ready for IPv6 when I tried it by Anonymous Coward · · Score: 0

      Network manager did have issues with DHCPv6,

      Network manager had issues with pretty much everything for more than a decade. It was one of the most buggy pieces of essential software I've ever come across. It's improved In the last year or so though it still has lots of bugs e.g. changing settings usually requires an undocumented restart/reboot and just to add to the confusion the main menu has sub-menu titles that are indistinguishable from sub-menu entries.

  20. I got the very last IPv4 address! Score! by Anonymous Coward · · Score: 0

    Suck on it, all you losers!

  21. This must be some new definition of "resemble"... by jonadab · · Score: 1

    > its properties starting to resemble the deployed IPv4 network

    Oh, so you mean there are a number of popular services, which are so well known as to be basically household names, that are only available via this protocol and no other, right? And the number of people who use the network both at work and at home rivals that of any other network, including the phone network? Right?

    No?

    Because that is what it would mean for the deployed IPv6 network to resemble the deployed IPv4 network.

    Holding my breath I am not.

    --
    Cut that out, or I will ship you to Norilsk in a box.
  22. North American not behind in IPv6 by YesIAmAScript · · Score: 1

    http://www.networkworld.com/news/2012/092412-ipv6-side-262674.html

    US has the most IPv6 users, North America has the most IPv6 traffic.

    Sure, it's still small in absolute magnitude, but it's a start.

    --
    http://lkml.org/lkml/2005/8/20/95
    1. Re:North American not behind in IPv6 by dave420 · · Score: 1

      Of course, as the US has over 300m people. Compare it to the EU, for example, which is a more comparable size, and you realise that the US is not ahead.

    2. Re:North American not behind in IPv6 by Anonymous Coward · · Score: 0

      Well, China & India have even more people. But less IPv6 traffic.

  23. IPV6 and the **AA folks by Anonymous Coward · · Score: 0

    Doesn't switching to IPV6 virutually guarantee the **AA spies would be able to tell exactly which computer was used to download supposed infringing files?

    1. Re:IPV6 and the **AA folks by Ultra64 · · Score: 1

      No

    2. Re:IPV6 and the **AA folks by l_bratch · · Score: 1

      IPv6 privacy extensions go some way to solving that problem: http://en.wikipedia.org/wiki/IPv6#Privacy

    3. Re:IPV6 and the **AA folks by WaffleMonster · · Score: 1

      Doesn't switching to IPV6 virutually guarantee the **AA spies would be able to tell exactly which computer was used to download supposed infringing files?

      This is what privacy addresses are for. They are enabled by default on windows and can be enabled manually on the MAC. Basically computers get a random address that keeps changing over time.

      I must say however using torrents to download illegal crap is pretty stupid nowadays you should assume 100% coverage from both LEA and "*AA spies".

      Transfer files directly between you and your friends. This is not only safer for you but by denying feedback channel to "the man" it helps protect the Internet in your country against excuses for opressive legislation by officials who are in bed with big media interests.

    4. Re:IPV6 and the **AA folks by Anonymous Coward · · Score: 0

      his is what privacy addresses are for. They are enabled by default on windows and can be enabled manually on the MAC. Basically computers get a random address that keeps changing over time.

      Basically what this means is that in order to avoid using (many to one) NAT, you're.... using (one to one) NAT. Or dropping your connection every time your IP changes. (yes, there are multiple kinds of NAT)

      Transfer files directly between you and your friends. This is not only safer for you but by denying feedback channel to "the man" it helps protect the Internet in your country against excuses for opressive legislation by officials who are in bed with big media interests.

      What? No. You get a subnet, it's linked to you. Rotating IP's on your windows box isn't going to do anything any differently than using NAT on v4 and rotating local IP addresses on your windows box.

  24. Hah! by Shaman · · Score: 3, Informative

    Companies are still shipping network gear that is IPv4 only. Find me a fixed-wireless device that supports IPv6! Sure they're layer 2 devices, but the units themselves don't have IPv6 addressability.

    IPv6 will take a long, long time. Maybe 10 years for major crossover. The fanbois and the advocates get shriller every day, but moving to IPv6 - even dual-stack - from an existing network is currently *hard*.

    --
    ...Steve
    1. Re:Hah! by Anonymous Coward · · Score: 0

      OpenWRT supports IPv6, so there is lots of gear that works : http://wiki.openwrt.org/toh/start . I use a (refurbished) WRT160NL (and the Hurricane Electric tunnel). When my ISP offers native IPv6, then I'll do that.

    2. Re:Hah! by iggymanz · · Score: 1

      hard? sudo apt-get install miredo

      or get a free account with tunneling broker, maybe six steps to set up your whole house with one of your machines supplying tunnel and addresses for the others.

      my AT&T DSL supports IPv6 too, but with the broker I get static addresses

    3. Re:Hah! by Anonymous Coward · · Score: 0

      or get a free account with tunneling broker,

      Tunneling brokers are a security fail. Do you really want all your traffic going through a single, unaccountable third party that makes no guarantees? Particularly one that is probably not in your country and thus is not a legally accountable? And is probably deep packet inspecting to cover their costs?

      No thanks. I want IPv6 but until my local ISP and my modem offers it natively it's not going to happen. I've experimented with using IPv6 locally but it's pretty useless without ISP support.

    4. Re:Hah! by Anonymous Coward · · Score: 0

      OpenWRT does IPv6 (http://wiki.openwrt.org/doc/howto/ipv6), and there is tons of gear that works with it (http://wiki.openwrt.org/toh/start). For example, I use a (refurbished) WRT160NL. I use a free Hurricane Electric tunnel to get IPv6. It works great (just install the "6in4" package and configure it).

    5. Re:Hah! by WaffleMonster · · Score: 1

      Tunneling brokers are a security fail. Do you really want all your traffic going through a single, unaccountable third party that makes no guarantees? Particularly one that is probably not

      I agree with your general message people are better off waiting till they have native connectivity but for those of us who want to dink around or have a reason to use IPv6 now I don't see any problem with a tunnel broker.

      I trust hurricane electric as much as I trust any other anonymous router/tier x ISP along the network path. The Internet is inherently insecure regardless of who your ISP is... the way I see it your better off assuming every router on the Internet hates you and therefore use end to end precautions to protect integrity and privacy of your data.

      in your country and thus is not a legally accountable? And is probably deep packet inspecting to cover their costs?

      In my case they are not only in my country they actually have a tunnel router at an IXP most data thru my ISP transits anyway... latency between IPv4 and IPv6 is virtually the same. Been working flawlessly so far.

      Hopefully comcast gives me my native prefix soon and the tunnels can go away.

    6. Re:Hah! by CAIMLAS · · Score: 1

      How does this even begin to approach the GP's concerns? It doesn't.

      Sorry, I do not want to have to maintain two address pools because a handful/quarter/third/half of the devices on my network do not fully or properly support IPv6. It's anathema to "internet protocol".

      It's the same reason why we hated on IE for so many years, and why technology like OpenVPN and OpenSSL became not only commonplace but have become preferred over the likes of isolationist technologies like IPSEC. (That's what IPv6 is, an isolationist technology.) Hell, that's why Linux has become so popular - it works, with no fuss, in most situations.

      --
      ~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
    7. Re:Hah! by iggymanz · · Score: 1

      what kind of obsolete junk do you have on your home network. the concerns are the same as with an ISP. use encrypted traffic for things you care about, firewall ports you care about at the tunnel endpoint. it's easy, everything in my house runs IPV6 and IPV4 with no issues and no increased security risk over IPV4 via ISP

  25. Not the end of the world by Holliday · · Score: 1

    What if the Mayans predicted the day we'd run out of IPv4 addresses?

  26. Most servers still only ipv4 by kvnslash · · Score: 1

    I've been on ipv6 for 6 months or so using hurricane electric to tap into the ipv6 world. When I browse around the net, the only ipv6 sites I really come across are google and facebook..

  27. Non-sensical customary units of fail by Artemis3 · · Score: 3, Interesting

    I learned under metric, for me those "customary" units of height are very hard to grasp.

    In metric, everything is in tens, you add or subtract zeros, thats it.

    A meter contains 10 decimeters (rarely used), a decimeter contains 10 centimeters, a centimeter contains 10 milliliters, etc.
    http://physics.nist.gov/cuu/Units/prefixes.html

    Customary/Imperial units are a mess, and to make matters worse, you don't use a single unit but TWO different ones for measuring things (feet AND inches?). What the hell is an inch? half a feet? quarter? decimal? no... its freaking 1/12. OF COURSE you don't fit 12 feet in a yard, that would be too easy, its 3... AND you also don't fit 12 pica in an inch, but 6...

    To make sense of your nonsense, we have to convert to a single unit first (eg. inches), and THEN move to metric, that is not a trivial mental operation for many.

    Another American annoyance is paper sheet sizes. But there are many more areas for frustration in those outdated customs.

    Let them sink in their isolation, is what we say here.

    --
    Artix
    Your Linux, your init.
    1. Re:Non-sensical customary units of fail by Zan+Lynx · · Score: 2

      Everyone seems to understand meters, centimeters, millimeters and kilometers; bytes, kilobytes and megabytes.

      Yet when I try to describe the distance from Denver to Chicago in megameters my friends look at me funny.

      I find it especially lame when astronomers describe distances in millions of kilometers. Are they too stupid to understand metric and use gigameters?

    2. Re:Non-sensical customary units of fail by JohnFen · · Score: 2

      I wonder how much of this is just visualization problems.

      I'm an old fart who was taught from a young age to measure things in inches, pounds, etc. I understand metric units just fine and use them almost daily, but still when I'm given a measurement in metric, I don't have an instinctive feel for how much the measurement is. i have to convert to the old units to be able to picture it.

      Same for things like gigameters. Yes, I know what that means, but it's not a unit that is commonly used and so there's no immediate visualization of it. If I hear that a start is 2 million kilometers away, I have (once I roughly convert to miles, anyway) a kind of "feel" for that. If I hear that it's 2 gigameters away, then I have to do an extra conversion -- to kilometers -- before I can get that sense.

    3. Re:Non-sensical customary units of fail by Anonymous Coward · · Score: 0

      Yet when I try to describe the distance from Denver to Chicago in megameters my friends look at me funny.

      So, they understood better when you used megainches instead?

    4. Re:Non-sensical customary units of fail by Anonymous Coward · · Score: 0

      In metric, everything is in tens, you add or subtract zeros, thats it.

      And that's exactly why metric sucks. The core concept of dealing with units in the imperial system is dividing by two. 1/2", 1/4", 1/8", 3/8", 15/16" - it makes good sense and it's great to work with when you're actually building stuff. I'll be the first to admit that the imperial system is far from the best implementation of a binary measurement system, but it's the best one we've got, so we use it. Nobody uses 1/4 or 1/8 centimeters, and that's what kills the metric system for people like me. Dividing by two makes a lot more sense than dividing by ten. It's as true for units of measurement as it is for computing.

    5. Re:Non-sensical customary units of fail by Anonymous Coward · · Score: 0

      1/2", 1/4", 1/8", 3/8", 15/16" - it makes good sense and it's great to work with when you're actually building stuff

      Not if your mentally retarded like myself.

      With metric tools the bigger things have bigger numbers and smaller things have smaller numbers. With garbage like 3/8 and 15/16s you have to do mental aerobics first setting both items in common terms just to understand which is bigger and which is smaller.

    6. Re:Non-sensical customary units of fail by Anonymous Coward · · Score: 0

      Customary/Imperial units are a mess, and to make matters worse, you don't use a single unit but TWO different ones for measuring things (feet AND inches?). What the hell is an inch? half a feet? quarter? decimal? no... its freaking 1/12. OF COURSE you don't fit 12 feet in a yard, that would be too easy, its 3... AND you also don't fit 12 pica in an inch, but 6...

      Of all the things to complain about with Imperial, base-12 should not be one of them. Base-12 is superior to base-10; the advantages are fairly succinctly described on the wikipedia article for duodecimal. You only think base-10 is easier because that's what you were taught growing up; you were handicapped by your culture. There's a good reason why there are 12 inches in a foot, 2x12 hours in a day, and 12 months in a year.

  28. Vegas area MIA by Anomalyst · · Score: 1

    COX claims they have "plenty addrsses" and IPv6 is not scheduled for deployment "any time soon"

    --
    There is no right to feel safe thru security vaudeville at the expense of everyone's freedom, privacy and tax money.
  29. 1% by Anonymous Coward · · Score: 0

    1%

  30. Re:This must be some new definition of "resemble". by kqs · · Score: 1

    Interesting definition. So which is it: either Google, Facebook, Youtube and Netflix are not basically household names, or the deployed IPv4 network does not resemble the deployed IPv4 network.

  31. I know how you feel here. by chronokitsune3233 · · Score: 1

    Can you use a 24-hour clock system? Perhaps you can make use of the Celsius scale at least? It's not necessarily easy to learn to feel something reflexively, especially when everybody around you is used to something else.

    I've been using a 24-hour clock system for the past two years, yet I still think of the time as "3 P.M." rather than 15:00 (or 15h or 15h00 or whatever notation you wish to use). That's just the way things work.

    Likewise, I've preferred Celsius for little more than a year, but since everybody here in the US uses the Fahrenheit scale I am forced to convert to the Fahrenheit scale to make someone understand the temperature. After all, I know the mnemonic "30 is hot, 20 is nice, 10 is cold, 0 is ice" but when I personally think of temperature, I can't look at my phone and see 20 and know how it FEELS. I know how 70F feels. But 20C? Not sure exactly. I'm guessing that without completely thinking in terms of only Celsius and 24-hour time, I won't be able to make the switch. I think the same is true of other units of measure as well. After all, there have been 2-liter bottles of soda pop since I was a child, and I've been seeing 1-liter bottles of Coca-Cola in convenience stores as well as 20-ounce bottles. I didn't notice the difference until I held it in my hand because they appeared to be relatively the same pretty much. In actuality, the difference is nearly 12 fl. oz. more liquid in the 1-liter!

    (warning: rant follows)

    Oh, and I love how there are different "ounces" as well. Whereas you can compare a liter to a cubic centimeter (they're both units of volume), you can't necessarily compare an ounce (1/16 of a pound) and a fluid ounce (~29.6 milliliters) because one is used for weight while the other is used for volume. It would be like comparing kilograms to cubic centimeters. Two identical boxes, one filled with stuff while the other is completely empty, possess the same volume but different mass, which means the weight is different. Simply put, you can't trust that volume and weight are identical. As a result, there are two very different units that some can get confused. After all, an ounce is supposed to be an ounce, right?

    I wish someone would remind me why the US doesn't even seem to attempt the adoption the much more convenient metric system because cost can't be the only factor. Several attempts have been made already, and it's taught in schools now. It should be put to practical use beyond the government sector.

    --
    I have been a captive in America my entire life. Everybody and everything uses customary units instead of metric.
    1. Re:I know how you feel here. by JohnFen · · Score: 1

      I'm glad I'm not the only one. :)

      Can you use a 24-hour clock system?

      Yes, but I do have to do the math to convert it. Fortunately that's a really easy conversion.

      Perhaps you can make use of the Celsius scale at least?

      This is the most difficult of all for me, actually. Without doing math, I do have a vague sense of what a kilometer "means" (a bit more than half a mile), and what a meter means (a bit more than a yard), but I have exactly no sense of what the various Celsius numbers feel like. They all sound "cold" to me. I have to convert to Fahrenheit to understand them on a physical level.

      None of this affects being able to use the metric system, of course. I've been using it for decades. It's just not instinctual.

      After all, I know the mnemonic "30 is hot, 20 is nice, 10 is cold, 0 is ice"

      Oooh, I learned a new mnemonic! Thank you.

  32. IPv6 Deployment Picking Up Speed by Anonymous Coward · · Score: 0

    The Internet's addressing authority (IANA) ran out of IPv4 Internet addresses in early 2011. The IPv6 protocol (now 15 years old) was designed exactly for this scenario, as it provides many more addresses than our foreseeable addressing needs. However, IPv6 deployment has so far been dismal, accounting for 1% of total traffic (the high-end of estimates). A recent paper by researchers at the Cooperative Association for Internet Data analysis (CAIDA) indicates that IPv6 deployment may be picking up at last. The paper, published at the Internet Measurement Conference (IMC) shows that the IPv6 network shows signs of maturing, with its properties starting to resemble the deployed IPv4 network. Deployment appears to be non-uniform, however; while the 'core' of the network appears to be ready, networks at the 'edges' are lacking.
    regards.
    bizworldusa