IPv6 Essentials

, author of IPv6 Essentials

There are many different ideas and opinions about IPv6 around in the market. There are even people who think IPv6 will never come to life. Despite this perception, major companies around the world have been quietly developing products and services using the next generation protocol. And with the recent release of my book, IPv6 Essentials, O'Reilly has also entered the fray. Even with the enormous investment in IPv6 development, however, there are many misconceptions surrounding IPv6.

This article explains what you need to know in order to put things into perspective. It also discusses some of the myths that exist and explains IPv6's background, so that you can form your own opinion. This article is not designed to explain how IPv6 works in detail. To find out more about IPv6's inner workings, pick up a copy of IPv6 Essentials.

This article also assumes you have a good understanding of IPv4 and general networking principles.

TCP/IP version 4 (IPv4) is probably the most used protocol in our networks and in the Internet. It has proven over the years to be robust, stable, expandable, and reliable. Everyone who checks their email or surfs a Web site uses IPv4. It was developed almost 30 years ago, at a time when the current size and uses of the Internet were beyond imagination. Still, developers managed to create a lasting protocol that will be around for several years to come. About ten years ago, work began on a successor protocol. It was called IPng (next generation) at that time. The main reason for working on a new protocol was the expected IP address exhaustion. This is the reason why many people think that if they currently have enough IPv4 addresses, there is no need to think about IPv6. Additionally, we've developed technologies like Network Address Translation (NAT) and Classless Interdomain Routing (CIDR) to delay the address-space issues.

Why We Need IPv6

However, the perception that there is no need for IPv6 is a false one, for several reasons.

Reason number one: people who say there are enough IPv4 addresses are the ones who have big allocations. The government, universities, and organizations in the U.S. alone hold approximately 74 percent of the total IPv4 address space. One example: Genuity, a provider of IP network services based in the U.S., has three Class A addresses. That amounts to approximately 48 million addresses -- more than twice the address space the whole country of China has (approximately 20 million addresses). In Asia, IPv4 addresses are so rare that providers have no choice but to offer IPv6 commercial services. They are even funded by the governments to do so (Korea and Japan, for example). This is why the number of ISPs offering IPv6 on a commercial level is much higher in Asia than in Europe and the U.S. However, it is only a matter of time before European and American ISPs will need to rethink their stance on IPv6. Gartner Group, for example, estimates that the official IPv4 address space will be exhausted by 2005 or 2006.

Reason number two: NAT is an adequate short-term fix, but not a long-term solution. It doesn't allow for end-to-end security, breaks the peer-to-peer model, and creates a bottleneck in the network. For many current and future applications, especially in the area of portals and eCommerce, end-to-end security will be a requirement. And that, you will only get with IPv6.

Reason number three: In the future, it will not only be personal or network computers that will require an IP address. There are more and more devices that need a permanent IP address. The day is approaching where your mobile phone, car, PDA, TV, and even your refrigerator will have an IP address. With these changes, the need for IP address space will increase exponentially. Future services will require a higher level of security and quality of service (QOS) that, while feasible, will not be economically viable with IPv4.

IPv6 Advantages

The extended address space is certainly not the only reason to consider IPv6. The developers of the new protocol have learned from 25 years of experience with IPv4, and so they've optimized IPv6 for the complex networks of the future. IPv6 has many advantages, such as integrated security, an improved addressing schema, and more efficient routing and autoconfiguration that should reduce administrative costs. The features supporting mobility with IPv6 also offer possibilities for wireless networks, which IPv4 could never handle. And mobility is the future of the Internet.

Another common misconception is that IPv6 can be used only if the network backbone and the Internet backbone have been upgraded to IPv6. So why care about this new protocol today? But with the number of transition mechanisms that have already been defined and that are extended and optimized continously, upgrading is not a requirement. It is possible to create IPv6 islands at the edge of the network (or even allow an isolated IPv6 host) that utilize the IPv4 backbone to connect to other remote IPv6 islands or hosts. The IPv6 packets will travel through the IPv4 infrastructure encapsulated in an IPv4 packet (tunneling). So you can start building new segments or extending your existing segments with IPv6 and wait to upgrade your backbone in the normal course of business.

Who Uses IPv6 Today?

I had planned to write a whole chapter on 6bone for IPv6 Essentials. (I do discuss 6bone in Chapter 1, "IPv6 Versus IPv4.") The 6bone is a global IPv6 test network that was started in 1996, which today has more than 1,000 participating hosts. This shows that IPv6 is alive and can be used. Anyone interested can hook up to the 6bone and get started. When I started working on the 6bone chapter, I suddenly realized how many commercial and production networks already used IPv6: NTT Communications in Japan and Telia in Sweden, to name only two. I called some of the companies and talked to the managers of those IPv6 networks. And their baseline was: yes, there are a few things still missing, but we have what we need to get started and make it work. So this call goes out to all application developers: What the market needs is for you to port your applications to IPv6.

Early adopters have already started using IPv6 commercially, and since 2000, most vendors ship production stacks. Market analysts see the following timelines for IPv6 deployment: ISP and enterprise adoption will start in late 2002. Gartner Group says that by 2006, 50 percent of ISPs will offer IPv6 services. It is expected that by 2005 or 2006, the number of IPv4 hosts will slowly start to decrease and the number of IPv6 hosts will grow substantially and be higher than the number of IPv4 hosts.

At conferences, many people have become accustomed to being able to plug in a wireless card, sit anywhere in the conference area, and surf the Web or check email. We also take for granted having wireless connectivity at hotels, airports, and even outside areas. Did you know that at IETF conferences and IPv6 summits, wireless is used with IPv6? Pretty cool.

Ongoing Research

6net is an interesting three-year research project that was started in January 2002. It will build the world's highest capacity IPv6 research network. The goal is to demonstrate that continued growth of the Internet can be met using new IPv6 technology. The total investment will be 25.6 million Euro (25 million USD). Sixty-five percent of the funds will be provided by the project partners and 35 percent by the European Commission (information taken from the 6net Web site). Partners include companies like Cisco Systems, IBM, Sony, NTT Communications, many national research networks, and a number of universities.

Another project is Euro6IX. It is funded by the European IST (Information Society Technologies) program, and the goal is to support the rapid introduction of IPv6 in Europe. The first objective is to research an appropriate architecture to design and deploy the first Pan- European, noncommercial IPv6 Internet Exchange (IX) Network. It will connect several regional, neutral IPv6 Internet Exchange points across Europe, and achieve the same level of robustness and service quality as is currently offered by IPv4 Internet Exchange networks. The next objective is to use the deployed IPv6 infrastructure to research, test, and validate IPv6-enabled applications and services.

Another interesting development was at the Global IPv6 summit in March 2002 in Madrid, the first large-scale distribution of a congress over IPv6. The summit included remote participation of speakers and other participants. This was done using ISABEL, an advanced piece of software for real-time videoconferencing that has been around for many years and has been ported to IPv6 in reasonable time. This is real-world validation that IPv6 transition scenarios work.

NTT Communications, an ISP in Japan that offers commercial IPv6 services, has over 100 paying IPv6 customers as of March 2002. Toshiba developed an IPv6-enabled refrigerator, Sony an IPv6-enabled PlayStation, and an IPv6-enabled car has been developed as part of the WIDE project. The government in Japan supported the national IPv6 promotion council with 80 million in USD (8 billion Yen) in 2001, and released a tax incentive program for ISPs that buy IPv6-ready routers in 2002.

There are a number of Internet Web servers that are IPv6-accessible. A list can be found at www.ipv6.org/v6-www.html. To get there, you need a configured IPv6 stack. If you don't know how to do this, read Chapter 11, "Get Your Hands Dirty," in IPv6 Essentials and you'll be ready to go. You can also visit my IPv6 Web site (Apache on Linux).

In Germany, there is an interesting pilot project in the works called Guardian Angel System (GANS). It was initiated by doctors, the University Hospital at Tübingen, and Stuttgart and Ericsson Telebit as part of the EU project 6WINIT. In emergency situations, remote medical teams that cannot solve a medical emergency problem are able to contact the Guardian Angel experts via IPv6-based transmission. They use a one-button activation of all necessary systems and transmission of all data (vital data, videostream, and voice) works over that IPv6- based connection. The system is currently being evaluated on a simulator to make sure that no patients are needlessly put in danger. At the end of 2002, a pilot with Universal Mobile Telecommunications System (UMTS), a new wireless protocol, is planned.

And last but not least, a goodie for the gamers among you: two guys from Canada have ported Quake to IPv6 in only two days. Go and get it at www.viagenie.qc.ca/en/ipv6/quake/ipv6-quake.shtml.


I hope this article has dispelled some of the myths about IPv6 and generated some interest on your part. Should you care about IPv6? Yes, you should. IPv6 is inevitable. Although different Internet pundits argue about the time it will take for adoption, they no longer question its inevitability. New applications and services are being developed that leverage the advanced features of IPv6. If you suddenly need these services for your business and have not at least planned and tested for IPv6, the cost of introducing the new service, along with IPv6 at the same time, will be a big hit on your budget, your team, and your network. So, the best course of action is to begin learning about IPv6 today and make it part of your IT strategy. You don't need to tear down your IPv4 networks -- the old saying, "Never touch a running system," is still true. But if you build new segments or if you need to extend existing IPv4 segments because they have hit the limits, consider using IPv6. The sooner you test and plan, the easier your migration will be.

My next article will cover the IPv6 business case and show you how an investment in IPv6 will pay off in the long term.

O'Reilly & Associates recently released (July 2002) IPv6 Essentials.

Return to the O'Reilly Network.