IPv6 FAQ

It is clearly not possible for a substantial transition to IPv6 to take place before the exhaustion of the IPv4 free pools, which is expected some time in 2011. Therefore there will be an ongoing demand for IPv4 address space after this time, and those who need it will get it where they can. This is why the APNIC and other RIR communities are spending a lot of time discussing possible address transfer mechanisms, and other mechanisms for maintaining a supply of IPv4 addresses, even in very small blocks.

There is definitely going to be a period of some time, probably years, during which IPv4 addresses will still be needed, but will be obtained through very different framework to that which is in place today.

Failure to adopt IPv6 addressing may directly affect Internet innovation and development in the Asia Pacific. The Internet permeates all aspects of the economy, so future economic development relies heavily on getting this right as quickly and effectively as possible.

On the upside, organizations that move now still have time to plan a dignified migration to IPv6. Those that do plan now can build IPv6 support into normal network upgrade cycles and training budgets rather than risk a lack of IP address availability hampering their growth in the future.

APNIC treads a fine line when it comes to talking about IPv4 exhaustion. We need to promote awareness that the free pool of IPv4 address space is about to run out, because there is no escaping this fact. On the other hand, as awareness grows, particularly in non-technical sectors, there is an increasing risk of alarm - and even panic - which could certainly be unproductive.

The Internet has demonstrated its ability to evolve through difficult times, but those who are building Internet infrastructure need to be aware of this issue. They need to start planning now, and in detail, for the day when there is not enough IPv4 address space to support their network growth.

Those plans need to cover 2 bases: both the contingencies for ongoing IPv4 network growth when IPv4 address supplies could be limited; and the specific plans for deployment of IPv6 and eventual transition of entire networks to the new addressing system.

This is not a uniquely Asian issue, though the rapid growth in Asia Pacific economies is certainly contributing to accelerated exhaustion. Over the last two years, the region has seen 35% growth in requests for IPv4 address allocations.

Let's be absolutely clear, IP addresses are distributed on the basis of demonstrated need and no country or region has greater access or allocation rights than any other. And as a result of recent decisions through the global IP address policy process, each region will get its last allocation of IPv4 address space at exactly the same time.

There are historical imbalances which occurred prior to the current Regional Internet Registry system. But since the five Regional Registries took over management of Internet resources, APNIC has allocated more than 30% of those resources, and in fact more resources in total than any other RIR.

The biggest unique challenge in Asia is that the pace of growth is faster and is already placing huge stresses on the supply of technical expertise in many countries. In this environment, the extra challenges which I have described above will be much more onerous than in countries where Internet infrastructures are more well established and stable.

We have actually been talking about IPv4 exhaustion for more than 10 years. The Internet community recognised long ago that something needed to be done about the limited supply of IPv4 addresses. IPv6 was accepted as the preferred a solution in 1995, giving the industry many years to migrate gracefully.

What was not well recognised at the time was the simple business reality that in a highly competitive environment, a company will always spend its available resources on profit-making activities. Due to the very nature of IPv6, there is no customer demand for IPv6, and therefore no immediate pressing business case.

Sometimes we hear that the Internet industry has been "unable to deploy IPv6" for one reason or another. The truth is, however, that the industry has decided not to deploy it so far. What we hope however is that there is enough planning in place, or that there will be enough in place over the next 2 years, to allow the transition to happen smoothly. Surveys are being done by APNIC and others on precisely these questions, and so far the results are mixed. We hope to publish more information about this in the new year.

The major business benefit it has over IPv4 is that the address space is simply enormous. We are able to make large allocations which allow networks to be designed and to expand much more easily, and without the use of current address management technologies such as Network Address Translation (NAT).

The eventual promise is that the IPv6 Internet will be cheaper and more efficient at many levels - from the basic infrastructure to the applications that we run on it.

Trying to stick with an IPv4 solution, by reclaiming unused or historical address space, using carrier grade or multiple layer NATs, or increasing the percentage of utilisation, might work for the short or even medium term, but the only long-term solution is an address space with considerably more room to move.

IPv6 will also provide huge new business opportunities. The abundance of addresses will allow a huge number of devices to be connected to the Internet, opening a vast range of new applications. By giving every electronic or electrical device an IP address, and allowing direct communications between them, there is a vast potential for new applications and add value in ways we haven't even imagined yet.

Of course it depends how you interpret the statistics and which ones you choose to quote, but the Asia Pacific is certainly deploying IPv6 at a rate comparable to the other regions. In fact, Japan has been quite fast to adopt the technology.

What we need to recognise is that IPv6 is a global challenge, and it makes little sense for any ISP to deploy IPv6 before the networks that it connects to. So although there has been great interest in IPv6 in this part of the world, the actual deployments are not taking place yet, often because of the lack of interest or visible planning in other parts of the world, and particularly we hear in North America.

Although there are specific problems unique to this region, one advantage we have in adopting IPv6 is that we tend to have less legacy infrastructure. China, for example, is developing Internet infrastructure at a very fast rate compared with many western economies, for instance.

The opportunity for new networks in China to skip IPv4 deployment and build their networks end-to-end in IPv6 offers an enormous long-term advantage as they will never have to migrate. But it still relies on a certain level of deployment in other parts of the world, as I explained before.

It is often said that IPv6 is more secure than IPv4, because IPsec security is "built in" to the protocol specification. However this security is still not automatic in IPv6 and it is also readily available in IPv4, which means that the practical difference is minimal.