The Coming IPv6 Evolution

The coming exhaustion of the IPv4 address space has been in the news for a few years now. Various organizations have warned that the end is nigh, and finally, it appears the transition to IPv6 is really starting to really pick up steam. The transition to IPv6 will provide more capability, more opportunity, more performance, and a generally better user (and administrator) experience.

The switch to IPv6 will impact every web site, every server, every device – every Internet user – over the next couple of years. If you haven’t already gotten started with your transition plan, now is the time to do so.

The Problem

IPv4 address are 32 bits, commonly expressed 4 octets in dotted quad notation, like this: x.x.x.x where x is a number between 0 and 254. People have been using and memorizing IP addresses for almost 30 years. For instance, your home network is probably using IP addresses in the range of 192.168.x.x.

Because the IPv4 addresses we all know and love are 32 bits, there are about 4 billion unique addresses. However, some of those addresses are reserved for special use and some for home networks. Also, in the 1980’s, before anyone realized just how big this Internet thing would be, addresses were assigned in a very inefficient manner: for example, MIT, Apple, Prudential Insurance, General Electric and IBM each have 8 million addresses – HP has 16 million. Naturally, because the Internet started in the US, all of these big block allocations were made to US entities. As other countries came online, and continue to come online, and as new technologies like virtualization, cell phone, tablets, TVs, PVRs, cars, even toasters join the Internet, the number of IP addresses available is simply insufficient. The consensus is that the last block of IP addresses will be allocated in 2012.

Doomsday

What happens after the last block of address is allocated? Probably not much – at least immediately. China and India would be affected first, as they have the largest online growth. ISPs will start charging more for service (or they won’t issue IPv4 addresses at all, issue only IPv6 addresses), as IP addresses which were formerly essentially free will now an increasingly high value. Markets will develop to buy and sell IP addresses. To avoid using addresses, ISPs will stop allocating Internet IP addresses to their customers, instead making up their own address space using NAT (like how your home router does this for your home network), essentially creating little Internets.

Over time, this “little Internets” problem will spread, and will occur throughout the world. There are 3 major issues with the “little Internets” “solution” to IPv4 address exhaustion:

  1. Devices on separate little Internets will not be able directly communicate – or at least not without jumping through very tricky hoops. Gaming, VoIP, and P2P all become less reliable, more difficult to develop, and harder to use.
  2. Hosting will become more expensive, crippling innovation. Right now, if you have the next big idea, you pay $10/month (maybe even less) for web hosting, $10/yr for a domain name, and you’re off. Imagine how crippling it would be if there was an additional IP address fee tacked on – perhaps starting at $5/month, then quickly jumping to much more than that.
  3. Because of damage done to P2P technologies and the added expense of hosting, the Internet will more quickly transform into less of a collaborative platform, and more into a consumer/provider platform, becoming more like Cable TV than the free for all open forum it is now.

IPv4 exhaustion will be very bad for the common person, even if he doesn’t understand the problem. However, it will be very good for current established institutions because it will become increasingly more expensive to compete with them. For example, Google started out as two college kids – they ended up taking down a billion dollar company’s product (DEC’s Altavista). With the added costs and challenges that would exist in a post-IPv4 exhaustion world, Google could not have possible become what it is today.

The Solution

The solution is to grow the address space – make more addresses. Sounds simple, right? Well, not really.

The size of the IPv4 address space is very firmly planted in the IPv4 protocol; there is no way to change it. As early as the late 1980’s, people began to realize that the address space was not big enough, so work on the next protocol began. Meanwhile, the world changed: mobile phones came about and become extremely popular, virtualization became popular (again), VoIP was invented, and computer science evolved new ideas.

In 1998, the specification for IPv6 was finalized. IPv6 includes optimizations for mobile devices (such as cell phones), quality of service is built in (so VoIP works better), security is built in, improved efficiency/performance, and the address space is vastly larger (128 bits instead of 32 bits). There are about 4 billion IPv4 address compared to the 3.4×1038 provided by IPv6. Considering as how there are 5×1028 for each human being alive today, IPv6 address space should last us a while.

Today, all modern operating systems (Windows XP, Linux 2.6, MacOS 10.3) support IPv6, and almost all software does.

Who’s Pushing IPv6

The US government realized that IPv4 exhaustion was going to be a problem, so it mandated that all new technology purchases be IPv6 capable, and all networks be upgraded to IPv6 by 2008. As with most government projects, the work is not 100% done, but many federal systems are running IPv6 right now.

China started the “China Next Generation Internet” project to push IPv6 adoption within its borders very hard. The 2008 Olympics websites all other network services were available over IPv6. China of courses realizes that if IPv6 isn’t adopted, and IPv4 doomsday comes along, it will be at a serious disadvantage.

ISPs also realized that without IPv6, service will become more expensive and cut into margin. So ISPs such as Comcast and Verizon are implementing IPv6.

In Australia and Europe, many ISPs have already implemented IPv6.

Mobile phone service providers are pushing IPv6 as mobile phones are major consumer of IP addresses. For this reason, T-Mobile is running an IPv6 test program, and Verizon is mandating that all new phones on its network be IPv6 capable in the near future.

Realizing that many of its customers will soon have IPv6 access, and may not always have IPv4 access, Google is moving its infrastructure to IPv6.

Other companies are realizing the same thing, such as Akamai and Facebook.

Should My Next Project Run On IPv6?

Yes. The benefits include:

  • Future proofing. IPv4 will be around for a long time still, but it won’t always be cheap, and it will become more complicated to administer. If your site will last more than a year, you should provide IPv6 connectivity.
  • Customer experience. If you serve only over IPv4, and a customer using only IPv6 visits your site, your site may not work (if it does work, it will be slower). If you serve only IPv4, and a customer supporting both IPv4 and IPv6 visits your site, the user may experience improved performance.

Supporting IPv6 can be easy. Ask your ISP or hosting provider if they support IPv6 (many backbone providers and large, commercial ISPs do) – if they don’t, go elsewhere. Your operating system already supports it, and chances are, the rest of your software does too.

As Cameron Byrne from T-Mobile USA repeatedly told content providers at Google’s IPv6 Implementor’s Conference:

Our users are going to access your content over IPv6. The only relevant question is “will we make the AAAA record or will you?” Wouldn’t you rather be the one to do it so you have control?

Cross posted to the Isobar blog – please comment there.

CC BY-SA 4.0 The Coming IPv6 Evolution by Craig Andrews is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.