Slashdot Mirror
NRO Warns They Are On Final IPv4 Address Blocks
eldavojohn writes "According to the Number Resources Organization, they will have issued their final twelve IPv4 blocks in a few months. Each block is 16 million addresses and represents 1/256th of the total addresses issued. We are now down to 12 blocks left in the global pool for issuing to Regional Internet Registries, who will then assign the last addresses that will run out sometime later in 2011. The pool of free addresses works out to be less than half of where we were in January. The new numbers from the NRO indicate estimated global pool IP address exhaustion in a few months, a year earlier than they estimated at the beginning of 2010."
When it gets expensive to continue using IPv4, which may not be until well after we "run out."
You're not seeing some magic IP address fairy making them last longer, you're seeing armies of senior IT pros working until after dark trying to sort this all out and deal with things because the pointy-haired bosses on top have been seeing that IPv4 is 'good enough.' As long as IPv4 looks easier and cheaper on paper than IPv6, that's what we'll be using.
Well, NAT saved us from a certain doom, and also provides extra security (might act as a firewall).
I don't see IPv6 deployed 100% any time soon. Increasing the number NATed Internet users might be the only feasible solution, at least in short term.
Álvaro
You are misstaken, notable predictions have predicted the following:
May 21, 2007: ARIN predicts sometime in 2010
June 20, 2007: LACNIC sets final date to januari 1, 2011
June 26, 2007: APNIC sets the date to sometime in 2010
April 15, 2009: ARIN says sometime before 2011
So for the last 3-4 years there has been a fairly good estimate on when they are supposed to run out.
I'd like to see you try.
jhw
Less than one year (12 months)
For sure before the end of next year, but probably not by the end of this year.
My bet is in Feb or March of 2011.
Keep in mind, despite having 12 /8 blocks left, that really means 6.
Once there are only 6 blocks left, whoever purchases #6 has ended the game, because the remaining 5 left are automatically to be given to the other world registries at that same moment.
So in reality those last 6 blocks will all go at the same time.
So 6 more /8 purchases and we will be out of space.
They just sold off 12 /8's in the past few months, so it will take half of 'a few months' at the same rate, even though I suspect it will go faster now that there is a crunch for it.
I would not count on it. ISPs are increasingly consumption-oriented services; I would guess that instead of deploying IPv6, we will start to see ISPs offer lower prices for customers who agree to be NATed (or perhaps, demanding higher prices from those customers who refuse to be NATed).
Maybe there is some hope at the universities, though...
Palm trees and 8
Will be siites and services with only ipv6 addresses, that won't be able to be accessed from ipv4
Pretty much. The largest german consumer ISP recently announced its plan to enable an IPv4/IPv6 dual-stack on all DSL connections by the end of 2011. Source in German.
Several server hosters already implemented IPv6 during the last few months.
It's really overdue. All mobile ISPs that I've seen so far only offer NAT'ed Internet access. Horrible.
You don't want that question answered. Just like when a car's headed for a sheer cliff, you don't want to know exactly when it'll go over it. You want to avoid ever having to have that question answered.
The reason the day of recekoning's been being pushed back is because the IT techies, even as they've been warning of the inevitable cliff, have also been doing everything they can to push the deadline back. They know there's going to inevitably be problems making the switchover to IPv6, and they're trying to buy as much time as possible so we'll have time to fix any glitches, but sooner or later they're going to run out of ideas and tricks and the deadline's not going to move anymore. Ideally by that point it shouldn't matter because we've taken the warning and done what's needed to avoid the cliff entirely. But if everyone keeps assuming that, just because the deadline's been pushed back once, it'll keep being pushed back indefinitely, well, suddenly going into free-fall as the car's wheels pass over the cliff-edge is not a good feeling.
You want really impressive examples? Look back to the big fireball over Cape Canaveral that a few seconds before was STS-51-L (Challenger), or the big fireball over Texas that a few minutes before was STS-107 (Columbia). Challenger blew up because the managers at NASA knew the O-rings were eroding and would sooner or later be breached, and they brushed this off with "Well, it hasn't happened yet so it won't happen ever.". Columbia disintegrated during re-entry because managers at NASA knew pieces of heavy foam insulation were striking the leading edges of the wings during launch and sooner or later one of those strikes would fatally damage the heat-resistant panels, and they brushed this off with "Well, it hasn't happened yet so it won't happen ever.". When we run out of IPv4 addresses the results won't be quite so pyrotechnic, but if we keep saying "Well, it hasn't happened yet so it won't happen ever." we will end up regretting it.
And yet none of those would make more than a dent.
They're allocating /8s, even the addition of several /8s would only extend the time frame by a few to several months, compared to the siginifigant effort required to reclaim them.
a handful of selfish greedy people are no match for millions of selfish, greedy people -u4ya
It has been discussed already - if the class As that were allocated to corporates back when anybody with the money could buy a class A regardless of need were reclaimed, it wouldn't provide more than a few months of extra capacity.
Yeah, this gets posted EVERY TIME there's an article about IPv4 address exhaustion, and every time the answer is the same - increasing assignment efficiency will at most buy us a few months, perhaps a year or two, of time. It doesn't solve the problem, only postpones it a little longer.
In truth, when the addresses are exhausted, I expect all the holders of /8's to start auctioning off their unused allotments to the highest bidder. There's a reason none (or most) of them have not given addresses back voluntarily - they are about to become a very scarce, very valuable commodity for trade. Those companies who got in early and got a Class A will make maybe hundreds of millions or even billions of dollars auctioning off the addresses. When companies who have IPv4 address blocks are going into bankruptcy or up for sale, the value of their allotments will start to be accounted for as assets.
Which, I think, is one reason that some tech companies are not pushing harder for IPv6 adoption - they stand to make a lot of money off of artificial scarcity.
> Well, NAT saved us from a certain doom, and also provides extra security
NAT is a horrible hack. It might be a good solution for some things, but to fix the addressable space option, it is a disaster.
Talk about an almost entirely useless "broadcast" only Internet. Is that what you want?
> (might act as a firewall).
Even worse. I don't even want to begin to explain to you why you are wrong about this. The broad adoption of UPNP makes the idea that NAT provides you with a useful firewall complete idiocy....
If you want a firewall, make a firewall. Do not rely on NAT. Ever.
Why do we have to have this conversation every single time the issue comes up? gods...
We have allocated 14 /8 networks since January of 2010 (source: http://www.iana.org/assignments/ipv4-address-space/ipv4-address-space.txt )....meaning we go through about 1.5 /8s every month. Reclaiming a /8 will take more than a couple weeks, so the simple fact is that reclamation isn't worth the effort: we would burn through several /8s in the time it would take us to reclaim one of them.
I hope I'm wrong, but I've come to the conclusion there will be no quick transition to IPv6. When the last blocks get allocated, I think we'll enter a period of several years at least where IPv6 is *starting* to get rolled out, but is not rolled out yet, and companies who desperately need public IP addresses for their servers will pay thousands of dollars to buy IPv4 addresses from the hoarders. It's not like the Internet will suddenly end when IP address exhaustion is reached, it will just become much harder to get a public IP for servers or for making your home computer accessible to the outside world.
Carrier Grade NAT will probably start to be used by large ISPs, further extending the life of IPv4 by making it so that instead of getting 1 public IP address for your home/small business network, you now get zero public IP addresses for your home/SB network. Through stuff like that, millions of IP addresses will be 'reclaimed' and made available. . . at a price.
The increased price *will* give an incentive, finally, to companies and people to start adopting IPv6, but we're going to go through an expensive transitional period for some period of time while that happens.
The sad thing is, I'm ready to use IPv6 today (and am using it a little through a tunnel broker), but there's no indication from my ISP that they ever have any plan to turn on IPv6 in their routers. The only U.S. ISPs I've heard of who are planning to test IPv6 are Comcast and Earthlink.
To build on this post, we've gone through 14 /8s just since January of 2010. Reclaiming a /8 would buy not even a month, and it would take more than a month to reclaim it.
Reclamation is wasted effort. Implement IPv6.
What exactly is supposed to happen? Does it mean that new devices can't hook to the internet - or does something happen to everything that's currently running?
We will just NAT the NATed NATed NeTed NAT and run the entire internet on a single IP address TRA-LA!
Then there's the free market cool-aid crowd who can't see why bidding wars driving the price of a single IP into the thousands a year is a big deal.
Next up, the "It's so HAAAAAAAAaaaaaRRRRRRRRrrrd!" crowd who don't understand why they should burn their geek card for saying that. That and their close relatives who still haven't realized that very simple firewall rules grant 100% of the security NAT does.
I wish my ISP was as enlightened as this German one and they are one of the biggest in the UK as well and as recently as last March they had no plans to migrate to IPv6.
One thing holding it back from the Consumers is the lack of Comsumer ADSL Modem/routers that support it. AFAIK, the Draytek Vigor series is about the only ones that do it. Sigh
I'd rather be riding my '63 Triumph T120.
Really? I kept hearing "We'll run out of IPv4 addresses in five years" about four years ago, and so forth.
I've been hearing about the end of IPv4 space since my very first ISP handed me XXIV.XII.CXXIV.VI via DHCP
Trolling is a art,
I currently run a business-class DSL connection with a block of 5 static IPs. I only use two. So, one may ask if there's any way to reclaim the other three.
The answer is quite simply no. There are technical reasons why you can't assign IPv4 addresses in blocks less than 5 but more than 1. Nor is there any clear way I could share the extra addresses with someone else. The other three addresses are simply lost. Multiply that by hundreds of thousands, perhaps millions, of similar installations at ISPs and colos, and then you see why this is a problem that goes way beyond those misappropriated /8's.
Not a typewriter
This has been debunked so many times, in this thread and others, that I'm fully in favor of banning anyone who mentions it ever again.
Not a typewriter
So let me get this straight.. In the beginning we had a very simple very open design. Any host can talk to any other host on any port. Then, over the years bouts of paranoia, fear, and idiocy have created default drop firewalls and nat devices that fundamentally break the open nature of the internet, protocols that rely on that nature break when presented with that stupidity, and somehow it's the fault of the protocol designer?
How would you suggest we operate? Instead of using my internet connection to accept connections from my peers should I proxy through a 3rd party? Should I use a ridiculous hack like upnp to beg the nat device for a forward? What happens when we're all behind default drop inbound firewalls w/ a nat'd address generously provided by our ISP? Suddenly and even though you have an internet connection and I have an internet connection we can no longer communicate directly with each other? Do you not see this as a problem? Is this still a protocol issue?
No. IPv4 specifies that the host portion of the address with all-0's is the network address, and the all-1's address is the broadcast address for that subnet. If you assign these to an actual host, you will break things very badly. Since a /31 would contain only address 0 and 1, it has no addresses that can be assigned to a host. The /30 subnet is the smallest block that can be given out.
Not a typewriter
NAT-ing will only get us so far. In order to route to one private network to another the address must either be bridged and on the same subnet, or not bridged on a separate subnet. 10.0.0.0/8 networks will live a little longer than 192.168.0.0/16, but not by much. ISP A Natting all their customers to 10.10.1.0/24 and ISP B Natting all their customers to 10.10.1.0/24, nobody from ISP A will be able to talk to ISB B unless they create an explicit bridge between themselves. The potential for a abuse and misunderstanding of this is going to be rife.
Too much NAT and it's going to be much more than a PITA.
- Dan.
~ People that think they are better than anyone else for any reason are the cause of all the strife in the world.
Get a patent on allowing IP addresses from 256.x.x.x -> 999.x.x.x
or perhaps, demanding higher prices from those customers who refuse to be NATed ...or perhaps just refusing to assign public addresses to anybody. "Don't like it? Tough. Call your congressman."
jhw
From what I've read, the internet back-bone is ready for IPv6, it's just the ISPs that need to start using it.
I know my ISP is handing out IPv6 addresses. Charter Comm. They do get routed to a broker, but if I do a tracert ipv6.google.com I get several hop responses from my ISP with valid public IPv6 addresses before going to a broker, then google.
I even get a DNS name a long with my IPv6 IP. Bit Torrent even starts using IPv6 when I hook up this way. Yay for no port forwarding!
The best part is I don't even need to setup anything. A 100% fresh Win7 install, plug into my cable modem, and IPv6.Google.com works instantly.
Alas, I can't use it though. My ISP still limits my cable modem to only respond to the first MAC address it picks up on the network. Once I get DD-WRT on my Netgear 3700(it's still very buggy), I will let my router talk to the modem and my router can handle IPv6 on my network.
Here's a decent list of SOHO routers with IPv6 support.
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
OSX 10.5 supports ipv6 just fine, so did 10.4, not sure what version introduced V6 support...
XP also supports ipv6, although it's not installed by default and you can't use v6 exclusively.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
They spent HOW long advertising those free-or-highly-subsidized digital converter boxes and people still threw away perfectly functional TVs?
Regardless, no. Both WinXP (unless you're seriously out of date on your software updates) and OS X 10.5 support IPv6 just fine. Of course that's separate from hundreds of badly-coded apps that somehow shoehorned themselves into the IPv4 stack, but that's hardly OS-dependant.
How are sites slashdotted when nobody reads TFAs?
I want IPv4 to run out. The sooner the better. When Y2K was about to come around, all the businesses who had old code some of it from the '60s, started hiring programmers like crazy. They needed to convert all the dates from two digit year to 4 digits. A massive effort but still only a very small amount of the total codebase that was out there needed to be modified.
Fast forward to 2010, 4-byte IPv4 address running out. A new protocol exists but much of the old software and networks cannot use them. The only solution is to hire a massive number of programmers and rewrite the software..
Think of this, every piece of software on every computer that accesses the internet, has to be rewritten. How big is that codebase? A lot larger than Y2K. I can see this pulling in programmer after programmer like some huge vortex, in a race to be done before last address is given out..
You see why I welcome the new of IPv4. The end of the recession in the tech industry and plethora of new job.
I think the number of IPv6 addresses is supposed to allocate something like 2^34 IPs per atom in the universe, or some equally absurdly large number. I think we'll be OK for a while if that's actually the case.
From Wikipedia:
The very large IPv6 address space supports a total of 2^128 (about 3.4×103^8) addresses—or approximately 5×10^28 (roughly 2^95) addresses for each of the roughly 6.8 billion (6.8×10^9) people alive in 2010.[13] In another perspective, this is the same number of IP addresses per person as the number of atoms in a metric ton of carbon.
How are sites slashdotted when nobody reads TFAs?
Yeah right, they'll offer lower prices. :-P
I think they've known all along that they will raise prices and as soon as the media starts bombarding everyone with a crisis, they will roll out new more expensive services.
Meanwhile PR departments from a number of major ISP's will go on and on as to why IPv6 is not feasible.
Actually, it's almost the reverse problem... New devices (mostly) universally support IPv6, which has plenty of unallocated IP Space (we can allocate 200 quadrillion IPv6 addresses per square inch of land on the planet) popular and actively maintained services either have already, or will soon move over to providing services on an IPv6 address. ICANN has already switched over their root DNS Servers to resolve IPv6, and most larger ISPs are following suit. So, if you've got a new device on an ISP who has updated their DNS servers to work with IPv6, and you're accessing a popular website that has been updated to IPv6, you might already be using IPv6 and never notice the difference.
There's a lot of ifs in that statement though. Plus there's a pile of legacy OSes and TCP/IP stacks that won't work with IPv6, so while you might be able to access Amazon, Google, and Facebook, it may be that your corporate payroll system is run off an old Windows NT4 system, which isn't IPv6 capable, so your whole corporate network is held up on the IPv6 migration because that NT4 system isn't IPv6 capable, and the payroll system isn't compatible with Windows Server 2008.
Plus, even some modern equipment/software from low-price vendors is lacking IPv6 support, because it hasn't been cost-effective to add it. Current versions of Windows, Linux, MacOS, Android, and iOS all support IPv6, but the custom software stack in the Avaya IP-based phone on my desk probably doesn't. Nor does the $20 ZyXEL WiFi gateway that I picked up 2 years ago off the cheap shelf at Frys
How long do you give it until ipv6 address space exhaustion?
It only seemed inconceivable that IPv4 would run out because it was never expected to be used by pretty much everybody - it was originally thought the Internet would only be used in some government labs, military faclities, academia, and a few defense contractors. For the scope of that particular problem, 32 bits was conceivably "enough". However, the Internet grew out of that "scope" and become a globally available common communications system. Suddenly the old scope was no longer sufficient, so the IETF *foresaw* the problem 15 years ago that the 32-bit space was no longer sufficient.
So, we now have to decide what the new theoretical scope might be, right? But since the Internet already grew rapidly out of the first scope, what's to say any new scope is 'sufficient'. Well, look at it this way: 2^32 is approximately 4 Billion. Number of people on earth is approx 6 or 7 Billion - expected to grow to about 10 Billion in a couple decades. Now, not everyone necessarily needs their own IP address, but if you have *enough*, it's desirable to give an IP address to everyone (and, here come the people saying that using NAT increases security, so it's bad to give people public IP addresses, even though in reality a firewall with a default-deny inbound policy provides the exact same level of protection as NAT, while still allowing you the *option* of allowing traffic in if you choose - which for different types of applications, like games, VoIP/Video calling, direct file transfer between users, etc. can be very useful).
Of course, not all 6 or 7 Billion people can afford computer/electronics which would need an IP address. But, on the other hand, many people 'consume' more than one IP address - I have a computer at work, a computer at home, and a cell phone - that's potentially 3 addresses for just me; I might want to setup a home media player/DVR, which might need an additional IP address, and maybe I want to setup my own file server for sharing photos, videos, etc with friends and family. In addition to the addresses needed for users, you also need additional addresses for servers. Some servers will need a bunch of IP addresses because they are serving multiple domain names (that is, they are pretending to be a lot more servers than they are).
Well, we can't say for sure, but having somewhere around 20-30 Billion addresses available seems like it would be 'enough' *for now*, but as you ask, if you increased to 20 or 30 Billion, how do you know you wouldn't "run out in (say) 50-100 years".
So, the IPv6 people didn't increase the number space to 20 or 30 Billion. Every time you add one more bit, you *double* the address space. So, they started with approx 4 Billion, and they *doubled* it 96 times.
2^128 is approximately 340,000,000,000,000,000,000,000,000,000,000,000,000 (that is, 3.4 * 10^38). addresses. To give you some perspective, let's say that we invented a Faster-Than-Light ship, and began a great age of Galactic Expansion. Further, we somehow developed Faster-Than-Light communications, and wanted to connect all the planets in the Milky Way to the Intra-Galactic-Net. Wikipedia gives an estimated range of 100 - 400 Billion stars in the Milky Way. Let's use the *worst case* scenario, and say that we have 400 Billion. Further, let's say that there is an average of 1 habitable planet per star (which is likely grossly over-estimating the number of planets), so we say there are 400 Billion habitable planets in the Milky Way. That would give us about 850,705,917,302,346,158,658,436,518 (that is, about 8.5 * 10^26) IPv6 addresses *per Planet*. Which means that, really, we have enough IP addresses, probably to colonize hundreds or thousands of galaxies.
Are you *really* still worried about us running out of IPv6 address space? In what application could you possibly propose we do run out? Individually addressing atoms? Addressing photons?
If you wanted to be nit-picky, you could say that 64-bits of each 128-bit address is reserved for individual hosts - that the IPv6 specification isn't r
42 years
O.o
Not quite. I have a router that does NAT. I leave UPnP turned on, and I trust my security.
A NAT makes quite a good firewall against outside attacks (port scans and the like). Leaving UPnP tuned on means that you trust what is inside your own network -- you do not currently have any worms/rootkits/malware, and you are not going to visit sites that host that sort of thing. It works great for me! No having to manually open up ports to use a torrent client to get the latest Ubuntu.
Yes, some "trusted" sites may get compromised and I could get a "drive-by" malware install. But that has not happened yet.
"-1 Troll" is the apparently the same as "-1 I disagree with you."
Certainly all those who use bittorrent, and/or run their own webserver and/or mailserver.
Privacy begins with
We use up almost 2 /8's every month.
Don't you mean 1/4 every month? Remember! Always simplify your fractions.
Mod me up/Mod me down: I wont frown as I've no crown
Those who use xbox live or playstation network as well (yay for p2p game serving)
So untrue.
If we would reclaim the ips of all porn sites and store the media centralized, say like 127.0.0.1, then we would have nothing to worry about ip shortage.
Or the internet whatsoever.
So let me get this straight.. In the beginning we had a very simple very open design. Any host can talk to any other host on any port. Then, over the years bouts of paranoia, fear, and idiocy have created default drop firewalls and nat devices that fundamentally break the open nature of the internet, protocols that rely on that nature break when presented with that stupidity, and somehow it's the fault of the protocol designer?
Well, no. You got the beginning part right. However, the reasons for NAT are off.
It's more like years of:
Besides all that, assuming that you can abuse ports/connections willy-nilly is overly optimistic, even on an open network. Growth will mean more users on a box (so your 20 ports being used can multiple by 200 users, and starts to add up). As well as more inter-network connections (more users from your uni dealing with users at another uni, across the state, country, continent, or even world), and thus bottlenecks. Reducing your traffic can be very important here.
It turns out that NAT devices share many problems with other aspects of our modern internet, even exacerbating some otherwise-existing issues. It's not idiocy. There is much malice here, but not on the part of NAT engineers/devs.
Personally, I'll likely continue using a NAT device even after the entire world is IPv6, though I'll obviously have to find an IPv6 NAT router, if only to provide a relatively trivial-to-set-up firewall between my TCP/IP printer and WDTV device, neither of which I can otherwise control too easily from prying hackers, and said hackers. I don't want some nimwit in Nigeria to start printing their 419 scams directly to my printer. And I don't know what vulnerabilities are in the WDTV device, I don't trust it much, so, again, keeping it away from inbound connections is probably a good thing.
I could set this all up with a real firewall instead. But NAT provides it simply enough, and UPnP is, well, universal enough to make it easy to configure.
We hear plenty of people acting as if we can duct tape IPv4 for ever and plug their ears at the shear mention of IPv6. The truth is instead of spending energy trying to hold afloat a sinking ship, it may be time to start putting the gang-plank out to that shiny new boat that can take us the rest of the way. It doesn't make sense to wait for the boat to be sunk before jumping ship, since you will find yourself having deal with bigger issues. Then again overpopulation and lack of natural resources may have started world war three in a few years, so none of this is worth worrying about ;)
For those of you that have already decided that its time to make the move, what steps have you put in place to ensure you get to IPv6 in one piece.
BTW Akamai is already working on upgrading its network to support IPv6 and have a target date of 2011. The admit that its going to be a tough challenge, but at least they have recognised it makes sense to start moving now, rather than later.
Jumpstart the tartan drive.
Eh, not really. IPv4 will be gone. If you are an ISP, and you pursue Carrier Grade NAT (CGN) as your solution, you growth limit yourself. It's equivalent to fixing your available bandwidth permanently - you can't add more customers past a certain point without significantly degrading performance for all customers. In a few years, you'll need to deploy IPv6 anyway; your customers will pay a price for the capital cost of your CGN gear, then your customers will pay a further price for the capital cost of your v6 gear.
If you're only concerned about web+mail, deploy dual stack lite. Browsers and mail clients do IPv6 transparently already. CPE devices support v6 out of the box at the sub-$100 price range (Netcomm, Billion, and, uh, the one used in the big v6 trial by xs4all in the Netherlands). Going DS-Lite means that as more software supports v6, and more services appear on v6, the pressure on your public v4 addresses drops over time. You can sustain DS-Lite throughout transition. The capital cost is similar to CGN, and the ongoing expenses of v6 are generally covered by your existing v4 expenses (ie, bits you pay going over a v6 session are bits you no longer pay for over your v4, and if your upstream is charging you more for v6 it's time to go provider independent!)
Some of the services that don't work over CGN include, by the way, XBox Live, BitTorrent, many network games, and most VOIP solutions. Some services do work over CGN, but rely on a reasonable proportion of Internet users having a public address to do so, and thus aren't long term viable: Skype, some of the smarter BitTorrent clients that do hole punching. Some services rely on emerging protocols for dealing with CGNs, like FaceTime: ICE, STUN, and TURN.
You can get a taste for life under a CGN by configuring your home NAT device to ignore uPnP requests, and disabling any manual forwarding settings.
Also, the summary is full of shit regarding the changing estimation. The linked articles are pretty clear that it's still early 2011. Available metrics (http://www.potaroo.net/tools/ipv4/ is one of the best) show a pretty unchanging date; that link, in fact, includes a few graphs down the bottom showing the change in predicted date over time. If you're an ISP, you've got a reasonably reliable date to plan around, and it should see you unrestricted on your IPv4 clear through to 2012, plenty of time to get ipv6 upstream (typically free or very cheap, when taken alongside your v4) and implement dual stack in your core.
So what happens to all the companies that have already spent thousands of dollars to get an IPv4 block get their addresses taken away from them?
How would you deal with all the internet sites that are now completely unroutable?
If both sides are NATd how would you communicate?
How would you get around the port restriction of NAT? You're assuming 1 ip == 1 computer.
Why would to waste time coming up with some contrived solution that takes much longer and is less supported then simply switching to IPv6 without problems?
If every person in the world had a personal network the size of the Internet, and every machine on it was routable, then IPv6 would still be doing sparse addressing - we'd have used approximately the square root of the possible IPv6 addresses.
I am TheRaven on Soylent News
It probably won't be not being able to read your Facebook page. It'll more likely be that one day your Internet connection stops working because your ISP doesn't have enough IPv4 addresses to give one to every subscriber, they can't get any more netblocks, and you happened to be the guy whose computer was turned off when someone else wanted the last free address. Or your company suddenly can't submit it's payroll because the company that processes their payroll started providing IPv6 address resolution and, while your company's machines understand it quite well, the corporate firewall and filtering appliance doesn't and isn't capable of passing the traffic through. And it may very well be comparable to driving your family's car over a cliff with them in it when payday arrives, you need to write that rent check and the paycheck deposit isn't in your checking account and Payroll can't tell you when they'll be able to fix the problem. Note that this isn't theoretical, there've already been problems with Web sites who started providing AAAA records becoming intermittently or permanently inaccessible to people whose machines understand IPv6 but whose ISPs don't yet support it. The software's fully capable of falling back to IPv4 when IPv6 isn't available, but treated the case where IPv6 was available but didn't work as a network failure.