Hey there! Sign in to join this conversationNew here? Join for free

[COMPETITION] When will the IPcalypse happen? Watch

    Offline

    1
    ReputationRep:
    3rd February 2011. :moon:

    :ninja: What does it mean by IP adresses running out?

    Well done to P_T and wiz.
    Offline

    2
    ReputationRep:
    Goddamn, just 2 more days!

    (Original post by Wednesday Bass)
    Next time we run out of IP addresses? Not likely to happen for a long time with IPv6. :p:
    You say that, but you know what caused exhaustion in the first place? /8's being handed out left right and centre. E.g. Apple got 17.x.x.x (http://www.iana.org/assignments/ipv4...ress-space.xml)
    Let's just think about that. The Apple company itself has 16,581,375 public internet accessible addresses it can use. That's 335 for each employee. This completely ignores the existence of company private networks (10.x.x.x etc) are undoubtedly in use and assumes every single Apple company machine is publicly accessible on the internet.
    Why does it need this much address space? It doesn't and it undoubtedly is not even close to using the allocation. The same thing is true of many other /8 allocations.
    In fact, it's unlikely that ipv4 addresses will truly run out for a long time. This is merely possible IANA allocations running out.

    We can but hope that stupidity is resolved this time around, and no more massive address blocks are handed out to companies/organisations who don't need them. Have to wait and see what happens on http://www.iana.org/assignments/ipv6...ss-space.xhtml
    Offline

    2
    ReputationRep:
    (Original post by Chrosson)
    Goddamn, just 2 more days!


    You say that, but you know what caused exhaustion in the first place? /8's being handed out left right and centre. E.g. Apple got 17.x.x.x (http://www.iana.org/assignments/ipv4...ress-space.xml)
    Let's just think about that. The Apple company itself has 16,581,375 public internet accessible addresses it can use. That's 335 for each employee. This completely ignores the existence of company private networks (10.x.x.x etc) are undoubtedly in use and assumes every single Apple company machine is publicly accessible on the internet.
    Why does it need this much address space? It doesn't and it undoubtedly is not even close to using the allocation. The same thing is true of many other /8 allocations.
    In fact, it's unlikely that ipv4 addresses will truly run out for a long time. This is merely possible IANA allocations running out.

    We can but hope that stupidity is resolved this time around, and no more massive address blocks are handed out to companies/organisations who don't need them. Have to wait and see what happens on http://www.iana.org/assignments/ipv6...ss-space.xhtml
    True, but by sheer numbers alone, even going at the current rate IPv6 will last a good few years. I do hope they learn from the mistakes of IPv4 as you've pointed out. I had no idea Apple owned that many IP addresses, seems ludicrous.
    Offline

    0
    ReputationRep:
    oh no im too latee
    • Offline

      14
      Sheesh, talk about panic buying. I think we'll see IPv4 bouncing and scraping along the bottom, as opposed to simply stopping dead. That said, IPv6 migration isn't for the faint hearted as you've got to consider:

      • Overhaul of your entire network infrastructure
      • IPv6 compatibility of your current equipment (do your routers, switches, SAN system, hardware IP phones, corporate Blackberry fleet, hardware firewalls or even your upstream telco support IPv6?). Upgrades can be expensive and time consuming, and any translation hardware that does exist is non-standard and a pointless expense for a one-off transition period
      • Routing protocols - currently, BGP only supports IPv6 through messy extensions IIRC
      • Software upgrades - does your IDS/IPS/host based firewall/IP telephony/whatever support IPv6 connectivity? (Yes, dual-stack operating system environments might help, but under Windows and FreeBSD you have to explicitly define an IPv6 socket in your code - only Linux as far as I know supports hybrid sockets. Even then, you need to deal with IPv4 mapped IPv6 addresses - messier than Chernobyl)
      • The removal of NAT and making all your machines publicly routable, it can be argued, makes you more susceptible to traditional attacks such as port scans and DDoS


      I'm sure most network managers would love to click their fingers and a seamless migration takes place, but I regret to inform you that this is the real world where logic, common sense and pragmatism is deadly important.

      Spoiler:
      Show
      Classful IPv4 walks into a bar. The landlord says "What can I get you?", and classful IPv4 replies "A strong CIDR please, I'm exhausted and seem to be all over the place these days."

      :facepalm2:


      (Original post by Chrosson)
      Goddamn, just 2 more days!

      You say that, but you know what caused exhaustion in the first place? /8's being handed out left right and centre. E.g. Apple got 17.x.x.x (http://www.iana.org/assignments/ipv4...ress-space.xml)
      Let's just think about that. The Apple company itself has 16,581,375 public internet accessible addresses it can use. That's 335 for each employee. This completely ignores the existence of company private networks (10.x.x.x etc) are undoubtedly in use and assumes every single Apple company machine is publicly accessible on the internet.
      Why does it need this much address space? It doesn't and it undoubtedly is not even close to using the allocation. The same thing is true of many other /8 allocations.
      In fact, it's unlikely that ipv4 addresses will truly run out for a long time. This is merely possible IANA allocations running out.

      We can but hope that stupidity is resolved this time around, and no more massive address blocks are handed out to companies/organisations who don't need them. Have to wait and see what happens on http://www.iana.org/assignments/ipv6...ss-space.xhtml
      /8s were handed out like complementary sweets in the early days, primarily because people got drunk on the idea and nobody anticipated or expected the massive explosion in growth that we've seen. The one hundred thousandth network was connected in 1996 whereas, at that time, it wasn't expected until 2020 or something (IIRC).

      Apple have built that new data centre in Maiden, NC so some of the pool will likely be reallocated there. ~16m publicly addressable hosts is nuts, though; no way the data centre will be housing that much anytime soon let alone the rest of Apple.

      If you want Apple to return their unused addresses, then you might consider e-mailing Steve Jobs like all the other nerds and ask him for a sysadmin job. You can have lots of fun maintaining an Apple-wide NAT setup

      (Original post by Wednesday Bass)
      True, but by sheer numbers alone, even going at the current rate IPv6 will last a good few years. I do hope they learn from the mistakes of IPv4 as you've pointed out. I had no idea Apple owned that many IP addresses, seems ludicrous.
      I should hope so, considering 2^{128} = 3.40\times10^{38}. To put this into perspective, there are 1.33\times10^{50} atoms in the whole world Even in the realm of science fiction, that number of operational, publicly routable computers on earth is ludicrous. From an essay I wrote last semester (which got 93% :party:):

      "In 2010, the population of the world was expected to reach 6.8 billion people[18], meaning that the IPv6 address space can provide 295 unique addresses for each individual on the planet, thus likely solving the address shortage presented by IPv4 for many years to come."
      [18] is http://www.census.gov/main/www/popclock.html and according to my reference list it was accurate as of November 2010. Even if the world's population explodes to ridiculous levels, you're still not going to run desperately short.

      Any limitation comes not with the number of addresses available, but the allocation policy. We've already seen what happens when you go back a few years and hand out /8s like there's no tomorrow (oh, classful networking, how hideous you were).

      Currently, the default allocation seems to be that end users will get a /64 (where you only have one subnet) or up to /48 (http://www.apnic.net/policy/ipv6-address-policy.html). Any more than this and you need to have a pretty good reason, to be honest; the days of just grabbing half of the addressable space for yourself are over and people have learned from it :bhangra:

      Anyhow, NAT was originally designed as an evil kludge which has essentially become permanent for some people. NAT must die in a fire .
      Offline

      2
      ReputationRep:
      (Original post by ch0llima)
      Sheesh, talk about panic buying. I think we'll see IPv4 bouncing and scraping along the bottom, as opposed to simply stopping dead.
      I agree with this.
      As I understand it, IANA has methods by which it can recall unused ip address blocks and reassign them.
      One of the first to go will probably be the DoD. With 4 /8 blocks with the acronym DoD in them (> 66 million publicly addressable ips, or approx 1.5% of all possible IPv4 ips , most of which I believe aren't even publicly accessible) I think they're going to be given a slap.

      Wait....http://royal.pingdom.com/2008/02/13/...ense-has-them/
      I assume this includes things such as "Defense Information Systems Agency"
      To put this into perspective, >20% of all possible IPv4 addresses, and a greater % of possible public IPv4 addresses.
      :lolwut:
      They must be getting pressure to give up some.


      (Original post by ch0llima)
      Currently, the default allocation seems to be that end users will get a /64 (where you only have one subnet) or up to /48 (http://www.apnic.net/policy/ipv6-address-policy.html). Any more than this and you need to have a pretty good reason, to be honest; the days of just grabbing half of the addressable space for yourself are over and people have learned from it :bhangra:
      About that....
      http://royal.pingdom.com/2009/03/26/...pv6-addresses/
      A drop in the ipv6 ocean admittedly, but still not a great start.
      Offline

      15
      ReputationRep:
      (Original post by Chrosson)
      Goddamn, just 2 more days!


      You say that, but you know what caused exhaustion in the first place? /8's being handed out left right and centre. E.g. Apple got 17.x.x.x (http://www.iana.org/assignments/ipv4...ress-space.xml)
      Let's just think about that. The Apple company itself has 16,581,375 public internet accessible addresses it can use. That's 335 for each employee. This completely ignores the existence of company private networks (10.x.x.x etc) are undoubtedly in use and assumes every single Apple company machine is publicly accessible on the internet.
      Why does it need this much address space? It doesn't and it undoubtedly is not even close to using the allocation. The same thing is true of many other /8 allocations.
      In fact, it's unlikely that ipv4 addresses will truly run out for a long time. This is merely possible IANA allocations running out.

      We can but hope that stupidity is resolved this time around, and no more massive address blocks are handed out to companies/organisations who don't need them. Have to wait and see what happens on http://www.iana.org/assignments/ipv6...ss-space.xhtml
      Makes more sense than the Ford Motor Company getting a /8 block. At least Apple make devices which at the time had the potential to utilise a large number of its allocated addresses - the day Ford make cars with built-in Internet access will see my views change a little. It was understandable that initially IANA wouldn't know how quickly addresses would be required, and when it comes to IPv6, we still won't be able to predict it.

      Lessons to be learnt, sure, but they won't be as simple as handing out smaller blocks.
      Offline

      2
      ReputationRep:
      (Original post by DarkWhite)
      Makes more sense than the Ford Motor Company getting a /8 block. At least Apple make devices which at the time had the potential to utilise a large number of its allocated addresses - the day Ford make cars with built-in Internet access will see my views change a little. It was understandable that initially IANA wouldn't know how quickly addresses would be required, and when it comes to IPv6, we still won't be able to predict it.

      Lessons to be learnt, sure, but they won't be as simple as handing out smaller blocks.
      Fair enough, Ford is a better example. Apple was just one among many. I disagree about them making devices which could potentially use their allocation though - unless every customer with an Apple product used a VPN through Apple's servers there was no way they were all going to get used up by the company itself.
      Point taken about IANA not knowing back then...

      And personally I do think handing out smaller block sizes is probably the biggest lesson to be learnt here. There is still no way of predicting, true, but by being a bit more frugal with handing ranges out and aggressive with recalling them, IPv4 could probably have lasted a bit longer. Obviously being more careful doesn't solve everything and things are never quite that simple, but I think it's pretty close. What lessons do you believe are more important?
      Offline

      1
      ReputationRep:
      I think that twitter feed is just a copy of someone else's extrapolation (HE) and I can't see any real basis of how they came up with those numbers.

      IANA allocate addresses in /8's (16 million addresses) so how can the number of addresses available drop by a couple of million a day. If it's trying to count at the RIR level then there's around seven /8's still left, They've only just been allocated by IANA to the RIR's (I'm not even sure if the last five have officially been allocated yet! - but they'll be on each for the RIRs)

      For some of the other points. IPv6 is going to be relativly secure at the moment due to it's sheer size. Try port scanning 2^64 hosts that's what you've got to do for one subnet. Chances are that subnet's going to have no more than 1000 hosts on it.

      In some respects IPv6 isn't being handed out in what would be seen as a good way. Anyone who wants a network block gets a /48 to use and generally subnets are /64 in size. That gives 65536 subnets to use per allocation (I've got that size allocation on a VM because I needed more than 1 IPv6 IP!). However this does mean IPv6 is done in a uniform way.

      In terms of the transition. I'd hope most people have been aware of the change over for a while and at least in businesses equipment is IPv6 capable. Home equipment is likely to be what's not IPv6 capable at the moment. I expect it's going to be several years before everyone has an IPv6 deployment and even longer before IPv4 is totally dead.

      In terms of how the addresses were handed out originally that comes down to how the addresses were done initially. Originally you could only get allocations in /8, /16, /24. If you needed more than 65k addresses you got a /8. Some of the people with /8's might be able to hand some addresses back but that could require them to do a lot of work renumber systems. Making changes to networks like that can be a time consuming task.

      In terms of competitions surely the next ones should be along the lines of which RIR will run out first? When will RIPE run out ? etc.
     
     
     
    Reply
    Submit reply
    TSR Support Team

    We have a brilliant team of more than 60 Support Team members looking after discussions on The Student Room, helping to make it a fun, safe and useful place to hang out.

    Updated: February 4, 2011
  1. See more of what you like on The Student Room

    You can personalise what you see on TSR. Tell us a little about yourself to get started.

  2. Poll
    Did TEF Bronze Award affect your UCAS choices?
    Useful resources

    Articles:

    The Student Room tech wikiTech forum guidelines

    Quick link:

    Unanswered technology and computers threads

    Sponsored features:

    Web Legend

    Win a Macbook Air!

    Blog about setting up a website for a chance to win in our Web Legend competition.

    Groups associated with this forum:

    View associated groups
  3. See more of what you like on The Student Room

    You can personalise what you see on TSR. Tell us a little about yourself to get started.

  4. The Student Room, Get Revising and Marked by Teachers are trading names of The Student Room Group Ltd.

    Register Number: 04666380 (England and Wales), VAT No. 806 8067 22 Registered Office: International House, Queens Road, Brighton, BN1 3XE

    Quick reply
    Reputation gems: You get these gems as you gain rep from other members for making good contributions and giving helpful advice.