IPv4 address exhaustion

[Myrph]

So we all know the internet is growing at a ridiculous rate and that the current IPv4 address standard was eventually going to run out. When referring to the internet 4 billion addresses doesn't seem quite as many any more, especially as that roughly equates to 1 address for every 2 people on the planet.

On 1st February 2011, however, we hit the big milestone, the point of no return on the IPv4 allocation. With the recent allocation of two IPv4 blocks to the Asia-Pacific Network Identification Centre, We are officially down to our last 5 blocks of IPv4 addresses, a situation that has initiated action dictated by a rule that stated when there are only 5 blocks left, one will be given to each of the world's 5 regional Internet registries. This is expected to happen in the next couple of days, and when it happens there will officially be no IPv4 addresses left in the hands of the highest power, we will have officially run out of IPv4 addresses.

At least, on the grandest scale we will have run out. The regional registries will still have unallocated addresses that they are about to receive, so we aren't looking at an imminent collapse of the internet, but it does mean that we are going to have to start moving towards an alternative to IPv4 and quickly.

Fortunately this will be fairly easy, because the next step already exists! IPv6 is the planned successor to IPv4 and uses an 128 bit address, compared to the 32 bit address used by IPv4, which will support 34 million billion billion billion (american standard, 3.4 times 10 to the power 34) addresses. I suspect, however, the transition will be less than smooth and almost certainly late.

That being said, we may also be able to reclaim addresses with are no longer in use, but the process will likely be long and expensive and probably much less efficient than switching to the new system.

Regardless, I think its an interesting situation to find ourselves in and at the moment I'm surprised there hasn't been more media coverage of, what I believe to be, quite a significant technological event, and certainly one which most likely won't ever happen again, at least not for a very long time.

Sites of interest:
Wikipedia - IPv4 Address Exhaustion
Official current state of IPv4 /8 allocations, as maintained by IANA
IPv6.com - Knowledge Center for Next Generation Internet IPv6
potaroo.net: IPv4 Address Report with countdown
Computerworld News - Address allocation kicks off IPv4 endgame

[Metcarfre]

Can we not mine more tubes?

[gcninja]

I had thought that they had already prepared for this with the inclusion of IPv6 in network cards and drivers? So whats the hold up? why not when the IPv4 addresses hit, say 6, start on IPv6 and handing them out to the Big 5 (YGO reference lol) and say start using these when you hit 5 instead of waiting til the last minute. they KNEW this was going to happen... geez I procrastinate, but I don't run the INTERNET <_<

[Myrph]

Actually, we were on 7, 2 were sent to APNIC and we are now on 5. Most things now are compatible with IPv6, however as of the last time it was surveyed, there was a less than 1% uptake. Maybe now people will start switching across.

[Ottoman]

I did some math a little while ago to explain this to my friends.

Code: Select all

Size difference between the IPv4 and IPv6 address spaces:

Take an ordinary sheet of notebook or printer paper. Fold over or trim off about 4 cm from the bottom. Place the sheet on the floor. This represents all possible IPv4 addresses.

At this scale, the set of all IPv6 addresses is the entire surface area of Earth.

So, yeah, it's a bit of an improvement. I can't wait until people actually start using it.

[Metcarfre]

Can we not mine more tubes?

NB - this is the actual extent of my knowledge of how the Internet works. I believe I am in the vast majority.

[Slack Mesa]

Can we not mine more tubes?

The problem is that the Council on Tube Numbering Standards decreed long ago that each tube must have its own unique number written on the end. And then they only left a very small space for the number-barely enough to write 10 digits. So we're not running out of tubes, we're running out of numbers.

More precisely, IPv4 gives each device (computer, router, cellphone, Internet-attached vending machine, etc.) a 32-digit binary number as its unique ID. There are approximately 4.2 billion 32-digit binary numbers. So if every person on the planet had an Internet-connected phone, there wouldn't be enough numbers to uniquely address them all.

One thing that's helped delay the arrival of doomsday is NAT, a system whereby an organization (a company or university, for example) can assign private numbers to computers within its walls and make them all look like the same IP address to the outside world. The private addresses don't have to be unique worldwide, just unique within that organization.

One thing that's helped hasten the arrival of doomsday is the way the numbers are managed. If you call one of the organizations that manage the IP address space (ARIN in North America, for example, or RIPE in Europe) and convince them that your company deserves fifty thousand IP addresses, they won't give you fifty thousand random numbers. Instead, they'll give you a contiguous block of approximately sixty five thousand numbers (2 to the 16th power). This system of giving out big blocks of contiguous numbers is actually important to how routing in the Internet scales. But it's inherently wasteful: a company might get a big block of IP addresses and then not use all of them.

With IPv6, the Council on Tube Numbering Standards decided to leave more space at the end of each tube to write the number. An IPv6 address is a 128-digit binary number. That's more than enough room to give every person on earth several billion unique numbers of their own.

The challenge with IPv6 is that if your website is reachable only by IPv6, people whose computers only support IPv4 won't able to visit you any more. Thus the companies that make operating systems and networking devices began supporting IPv6 many years ago. But even some devices introduced in the last 5 years don't support IPv6. For example, iPhones that are still running iOS versions prior to 4.0 don't have IPv6. Even when the technology is available, getting hundreds of millions of people and organizations to upgrade is a time-consuming process.

[Gollom]

Can we not mine more tubes?

NB - this is the actual extent of my knowledge of how the Internet works. I believe I am in the vast majority.

You're probably right. And thats okay, because the details are really only interesting to networking nerds like me. I think thats pretty much true of everything, though.

To your average consumer, there won't be any real changes. Your ISP may need to replace/update your modem, and there may be an added step if you're still using Windows XP (or god help you, 2000), but for the most part, you won't even notice the change.

It'll be pretty cool when we finally do make the switch over. Although I'm glad I don't have to worry about any of the higher level stuff, because that seems like it might be unfun.

[Edit]Beaten. Oh well.

[Metcarfre]

Can we not mine more tubes?

*brilliance*

OOOooooooohhhhhh...

Somewhat more seriously - again, coming from a complete nub - is this an issue that can be resolved via a mere browser update? Or is it more fundamental, like certain older hardware will simply be obsolete? Will these changes to The Internet mean all existing addresses need to conform to the new standard, or will IPv4 addresses still be in IPv4?

WILL MY IBOOK G4 STILL BE ABLE TO LOOK AT FUNNY PICTURES OF KITTEHS?!!?

(It lost the ability to watch videos of kittehs some time ago)

[gcninja]

its more of a hardware/software(driver) stand point, ie: the iphone can do it IF you have OS 4+. You can go to your network setting and right click your ethernet/wireless adapter and click properties. you see in the box how it lists like
Client for MS networks
QoS Packet
etc
Internet protocol Version 6
Internet protocol Version 4

Thats all you need and you're set, if you don't see that you have an old computer and probably need to update your drivers.

Most personal computers running recent operating system versions are IPv6-ready. Most applications with network capabilities are not ready but could be upgraded with support from the developers. Java applications adhering to Java 1.4 (February 2002) standards have support for IPv6
Most equipment would be IPv6 capable with a software or firmware update if the device has sufficient storage and memory space for the new IPv6 stack. However, manufacturers may be reluctant to spend on software development costs for hardware they have already sold when they are poised for new sales from IPv6-ready equipment.[citation needed]

In some cases, non-compliant equipment needs to be replaced because the manufacturer no longer exists or software updates are not possible, for example, because the network stack is implemented in permanent read-only memory.

so its mostly the big companies who have to worry

Also, on the website stuff, it wont be the site taht has to change, if anythign the DNS provider (the people who work in the background) will have to update their systems if I'm reading this right

[Gollom]

Can we not mine more tubes?

*brilliance*

OOOooooooohhhhhh...

Somewhat more seriously - again, coming from a complete nub - is this an issue that can be resolved via a mere browser update? Or is it more fundamental, like certain older hardware will simply be obsolete? Will these changes to The Internet mean all existing addresses need to conform to the new standard, or will IPv4 addresses still be in IPv4?

WILL MY IBOOK G4 STILL BE ABLE TO LOOK AT FUNNY PICTURES OF KITTEHS?!!?

(It lost the ability to watch videos of kittehs some time ago)

Does that fossil still run? Impressive. Assuming you're running the latest version of Mac OS X it supports(Leopard, unless it's REALLY old), yes. It could look at teh funny pictures of teh kittehs! If it still had that ability, anyway.

The change will be fairly gradual, because IPv6 won't really work on its own until everyone, or at least a vast majority, are using it. Also note that just because the last five blocks have been handed out, doesn't mean we'll immediately use them. It could be another couple years before we literally run out of IPv4 addresses. I think the ETA was 2015, but that's basically an educated guess. So we have some time to adopt IPv6 and most of it will occur "behind the scenes" so to speak. It'll be a couple decades yet before IPv4 disappears entirely; this simply marks the point where all new IP address will be exlusively IPv6. So it's pretty cool, and marks a really neat point in the internet's life, but it's nothing to worry about. Funny kitteh pictures will still be posted, it's just that the numbers in the background will change a bit.

[gcninja]

HEY I have a wallstreet that still runs so

[Metcarfre]

Yee-uh, ahm rollin' up PowerPC style, WHAT!?

Actually, it probably only can't play some vids because every time I update Firefox it says, "you should update flash!" and I say, "do it for me, DICK. If I wanted to update stuff on my own I wouldn't be a Mac user, would I?"

[Elomin Sha]

We need fatter pipes.

[empath]

"We require additional addresses; spawn more IPs" what?

But as gcninja has said, we end-users and our stuff are all pretty much ready to go with this, They (the 'backend' companies and providers and whatnot) just simply have to get through their burro-cratic red tape and flip the switch so to speak...

...so I expect this to get successfully completed about three months after the last IPv4 address gets used up.

[Dutch guy]

Only a handful of providers are actually ready for a switchover top IPv6. So expect the end of the world as we know it

[empath]

Oh! So I can hope for something better?!

[Sable]

Hi.

Network engineer, here. That means it's my job to deal with things like IP addresses.

This is not the end of the world, or the end of the Internet. It will make being an ISP slightly complicated in the short run, and very complicated in the long run. The change from IPv4 to IPv6 will probably be fairly smooth on the end-user's perspectives, but will entail a long, painful process from people like me.

Running an ISP isn't easy. Infrastructure costs money - a LOT of money. But even beyond that, the network itself is mind-bogglingly complex. It is a mess of technologies in various states of refresh, different software revisions, manned by not enough people with not enough time in most cases.

The transition to IPv6 is going to be unbelievably, almost overwhelmingly complicated. There will have to be phased rollouts, step-by-step and segment-by-segment upgrades, sidegrades, and outright replacements. There has to be a consistent end-user experience and that is by far going to be the hardest part.

You're all network-connected people, with reasonably new machines (even the iBook G4). What about Grandma, with her 486 still running Windows 3.11? That doesn't understand IPV6. What about all the third-world countries that might not have infrastructure that understands IPv6? The Internet is not homogenous, and anyone who's spent a nanosecond in tech support knows that end-users never have an ideal setup.

The slow transition from IPv4 to IPv6 isn't only because of "burro-cratic" corporate overlords who don't understand what's going on. There are, and there will be, people caught with their pants down when there's no more workarounds, when the choice HAS to be made; there always is (On a related note: Ask Me About Narrowbanding And The FCC). It is a genuinely gigantic, complex problem that involves players around the world who have to move or adapt in something like tandem, and that level of cooperation is frankly not easy.

There's a pithy saying about how to eat an elephant: You do it one bite at a time.

The IPv6 transition isn't an elephant.

It's more like an entire continent of elephants and all you have is a pair of chopsticks and a Swiss Army knife.

[Keab42]

It's been fascinating to see us reach this point. They've been talking about this for years, they've known it was happening but they've stood by and not bothered to take up IP v6.

I think it's interesting because it's exactly the same approach we're taking to similar resource shortages in the world.

[Danielle Pepin]

We need fatter pipes.

Yup we do. Especially in Canada to help prevent that UBB thing in the other thread.

[Danielle Pepin]

Hi.

Network engineer, here. That means it's my job to deal with things like IP addresses.

This is not the end of the world, or the end of the Internet. It will make being an ISP slightly complicated in the short run, and very complicated in the long run. The change from IPv4 to IPv6 will probably be fairly smooth on the end-user's perspectives, but will entail a long, painful process from people like me.

Running an ISP isn't easy. Infrastructure costs money - a LOT of money. But even beyond that, the network itself is mind-bogglingly complex. It is a mess of technologies in various states of refresh, different software revisions, manned by not enough people with not enough time in most cases.

The transition to IPv6 is going to be unbelievably, almost overwhelmingly complicated. There will have to be phased rollouts, step-by-step and segment-by-segment upgrades, sidegrades, and outright replacements. There has to be a consistent end-user experience and that is by far going to be the hardest part.

You're all network-connected people, with reasonably new machines (even the iBook G4). What about Grandma, with her 486 still running Windows 3.11? That doesn't understand IPV6. What about all the third-world countries that might not have infrastructure that understands IPv6? The Internet is not homogenous, and anyone who's spent a nanosecond in tech support knows that end-users never have an ideal setup.

The slow transition from IPv4 to IPv6 isn't only because of "burro-cratic" corporate overlords who don't understand what's going on. There are, and there will be, people caught with their pants down when there's no more workarounds, when the choice HAS to be made; there always is (On a related note: Ask Me About Narrowbanding And The FCC). It is a genuinely gigantic, complex problem that involves players around the world who have to move or adapt in something like tandem, and that level of cooperation is frankly not easy.

There's a pithy saying about how to eat an elephant: You do it one bite at a time.

The IPv6 transition isn't an elephant.

It's more like an entire continent of elephants and all you have is a pair of chopsticks and a Swiss Army knife.

Indeed I'm wondering with my knowledge of how dynamic IP update regularly if one day my laptop won't recognize the new one if my ISP serves it an IPv6. My laptop has trouble recognizing when it receives a new one as it is.

[Lisra]

Slack Mesa:
That was very well explained. Are you in any way related to a teaching profession?

[empath]

{valid points about real-world issues with implementation}

The slow transition from IPv4 to IPv6 isn't only because of "burro-cratic" corporate overlords who don't understand what's going on. There are, and there will be, people caught with their pants down when there's no more workarounds, when the choice HAS to be made; there always is (On a related note: Ask Me About Narrowbanding And The FCC). It is a genuinely gigantic, complex problem that involves players around the world who have to move or adapt in something like tandem, and that level of cooperation is frankly not easy.

Fathead executives (you know, the ones that say "do it." or "don't do it."?) resisting co-operation that would help them? You DON'T SAY; imagine that! (I'd have said "don't get me started on the evils of corporate...well, ANYTHING involving corporations", but...yeah... )

Still sounds like mules being recalcitrant.

There's a pithy saying about how to eat an elephant: You do it one bite at a time.

The IPv6 transition isn't an elephant.

It's more like an entire continent of elephants and all you have is a pair of chopsticks and a Swiss Army knife.

Yeah, and to carry the analogy, your bosses have been spending quite a while sitting at the table and yet NOT EATING.

I'd say "The sooner you start tucking into that continent-sized appetizer platter, the sooner you can finish."

[Master Gunner]

Given that some aspects of the internet's infrastructure are over 40 years old, and the innate complexities of it that would put the Total Perspective Vortex to shame, they could have being gearing up for the switch to IPv6 back in '98 and still be scrambling today with all the things that need to be changed, checked, and tested to make sure it works, because the engineers and programmers of 20 or more years ago didn't anticipate this (or didn't care because it wasn't their problem). When dealing with such projects, they'll always take longer than you have, even if you take into account that it will take longer than the time available.

[empath]

Regardless, when sitting in front of a meal you HAVE to eat, it STILL looks like childish pique to CONTINUE TO REFUSE TO EAT.