Well, maybe not the end of the Internet, but the end of the IP Address space, as we know it, is near. You know, IP addresses are the 4 blocks of number separated by periods, i.e. 10.0.0.1; well, there are only 4 billion of them. Sometime next year, 2012, Internet Assigned Numbers Authority (IANA) will have no more IP addresses to give to the Regional Internet Registries (RIRs). About a year after that, the RIRs will have no more to give out to Internet Service Providers (ISPs), such as Cox Communication and Go Daddy.
The scientists knew IPv4 wouldn’t last forever. I doubt they realized it would only last just over 2 decades. But, they knew, and started working on a new protocol, and by the mid 90s, the first draft of Internet Protocol Version 6 emerged.
The 2003 dotcom bust made it look like we had decades left, but then consumption rate raised to 160M per year by 2005. It became apparent we only had a few years left pretty quickly after that.
This article by CNN indicates the reason we’re running out so fast is cell phones and RFID tags, and all of a sudden we realized it. Hah. It’s been apparent for years and it’s not RFID tags and cell phone, that’s only a small part of the story. There are a number of IPv6 evangelists, such as Tony Hain, that have helped predict the depletion date in A Pragmatic Report on IPv4 Address Space Consumption, published in 2005. The pataroo.net tool has been updating it’s prediction daily based on a scientific estimation that considers many factors. Hurricane Electric has gadgets you can add to your web page, iGoogle, iPhone and more. These tools are geared towards the IANA depletion, so add about 1 year to what you see here, and that’s when ISPs will no longer be able to obtain new IP address allocations.
One more neat tool is the IPv4 Depletion tool found here. This tool predicts IANA to run out on 2011-03-26 (as of today, 2010-07-27). Watch the video on the page to understand how to use the tool. This provides a good summary of variables involved in making a prediction. Makes it clear why it’s impossible to make an accurate prediction, and why the date keeps sliding. Keep in mind, though, the more time that goes by, the more accurate the prediction gets. The day we run out, we’ll be 100% sure! Also note, I’ve been tracking this for some time, and between 2005 and 2009, the date kept getting further and further out, and between 2009 and now, the date keeps getting sooner and sooner. Last time I checked, just a couple months ago, the prediction was 2011-11. It’s apparent the consumption rate is increasing right now, and there is no reason it’ll slow down. In fact, there will likely be a landrush at the end that’ll eat up the last few months in days. I wouldn’t want to be a RIR in the last few months dealing with ISPs trying to get their last allocations early.
What does this mean to you? Depends on who you are:
Casual Internet User – might need to buy a new router to enable IPv6 on your home network. But, don’t start looking yet, they’re not out yet, and the home ISPs are not handing out IPv6 allocations, yet, anyway. If you can’t wait, you could get an allocation from a tunnel broken, like Hurricane Electric, though. To enable this, you need an IPv6 capable router, and the only way to get that now is to buy a router that you can flash the firmware with something like DD-WRT (get ready to study for days on end). An easier alternative is to install Freenet6 and enable just your desktop or laptop through a tunnel. Or just wait for the industry to catch up.
Have a web site – might want to start looking for a web hosting provider that offers dual stack hosting. This is currently only provided by a few niche providers. Hasn’t hit mainstream yet, so hang in there, it’s coming soon.
SMB with some colo equipment – Operating systems have supported IPv6 for some time now. Network gear, on the other hand, is a whole other story. Investigate what you have and whether it supports IPv6. If it’s supported in software only, it may as well be unsupported. Software only features on network gear is only good enough for light testing. It needs to be supported by the hardware. You may have an investment to make in upgrading your gear. Also need to look into getting an IPv6 allocation and transit. You should be able to get that now. Investigate the softwares running on your servers to see what will support IPv6 and what won’t. Apache, for example, is IPv6 ready. MySQL, is not.
Enterprise with datacenter(s) Identify what services need to run on dual stack, or IPv6 only, and trace those through the network gear and determine what upgrades/replacements are required. Investigate the softwares running on your servers to see what will support IPv6 and what won’t. Create an IPv6 rollout plan, and a series of projects to accommodate the plan. Start early. If you haven’t started already, you’re in trouble.
It’s surprising to me how under prepared the industry is for this inevitable event. Part of the problem that large ISPs and Enterprise companies face is large infrastructures running on legacy equipment that doesn’t support IPv6. They will need to make large investment in upgrades. There are also security concerns. And even though Operating Systems have natively supported IPv6 for some time, much of the software that runs on these OSes are not IPv6 compatible.
For web hosting, another reason you don’t see dual stack (that’s IPv4 and IPv6 together) hosting as a common offering is most of the world hasn’t heard of this yet. Hard to sell something people haven’t heard of. But, with stories emerging like the CNN story listed above, it’ll start to gain some traction. People still won’t get what IPv6 is, but might think of it like this: “There is this new Internet and soon there will be people surfing the new Internet, and if I want my website to be visible by the old Internet and the new Internet, I need dual stack hosting.”
With web hosting, typically users won’t leave their current provider unless they have a problem. You’re not going to see huge swings of hosting accounts from one provider to another, unless one has a serious problem. Well, when word of the need for dual stack hosting goes mainstream, and it will, there could be huge swings from those that don’t have it, to those that do. It’s going to be a game changer. It’ll be interesting to see who, in the list of large hosting providers, comes out with it first, and who is left in the dust!
Once the depletion date hits, aside from some black market trading, there will be no more IPv4 addresses to be had. Imagine a new hosting company coming online. They won’t be able to get any IPv4s. They’ll only be able to offer IPv6 hosting. And the older hosting companies will charge a premium for dual stack. IPv6 hosting will become the standard way to host a web site. As an end user of the Internet, it’s going to be pretty important for you to be able to reach these IPv6 only sites. Some users will come online only able to reach IPv6 enabled sites. But, there will still be users on IPv4 only connections too! So, anyone with a web site is going to want dual stack hosting pretty bad, to be able to reach all Internet users. Dual stack will come as a premium, after all, only the old hosting companies will have IPv4, and only they can offer it. As IPv6 becomes more and more of the standard, the demand for IPv4 will drop, and then so will the prices. Eventually, IPv4 will deprecate. How long will this take? Impossible to predict. Many say decades. I agree IPv4 will exist for decades longer, but when will it stop being vital for a website to be on dual stack to be successful? When will the majority of the most popular sites stop running on IPv4? I think this could occur in as little as 5 years following the depletion date, with a more likely of 10 years, but this is purely subjective.
In any case, the Internet is going to get really interesting here over the next couple years. Prepare yourself for a major event that will likely impact you more than the year 2000!