Measuring IPv6

This web page describes a measurement activity being undertaken by APNIC, the Asia Pacific Network Information Centre. The aim of the activity is to conduct a broad based long term measurement of the level of uptake of IPv6 across the Internet.

Why are we doing this?

We are running out of IP addresses for the prevalent version of the IP protocol, IPv4. Indeed, in some parts of the world, namely the Asia Pacific region, there are no more IPv4 addresses to be obtained. We expect that in the coming months this will also occur in Europe and the Middle East, and a little later we will also see the Americas and Africa also run out of IPv4 addresses.

The technical community has developed a new version of the IP protocol, namely IPv6, that is intended to address this. It was originally anticipated that we would transition the Internet to use IPv6, and complete this activity before we needed to reach for that last available IPv4 address. However, this has not happened, and it now appears that we will have to cope with the twin issues of address exhaustion in IPv4 and IPv6 deployment simultaneously.

This is a huge challenge, and will require the coordinated actions of many actors within the Internet, including content providers, equipment vendors, application developers, Internet Service Providers, policy makers and many others. One of the major inputs to this coordination process is good, reliable data. How well are we doing with IPv6 deployment? Which sectors are going well? What are our expectations about the time frame of transition? How can we help the Internet to transition to IPv6? We would like to inform these questions, and inform the broader conversation about the transition to IPv6, with data.

This is why we are undertaking this IPv6 measurement activity.

Our objective is to publish aggregate information that shows the uptake of IPv6 on a global basis, and on a country-by-country basis, over time across as much of the Internet as we can encompass with these measurement tests.

The Measurement Test

The manner in which this is being undertaken in this measurement program is to use web browsers, such as Internet Explorer, Chrome, and Safari, and we set the browser a small set of tests that expose the client's ability to successfully use IPv6.

To illustrate this approach, we have used three images in the table below.

[This image is only accessible over IPv4] [Image accessible over IPv4 and IPv6] [This image is only accessible over IPv6]

The image on the left should load if your browser is capable of using IPv4.

The centre image is loadable using IPv4 or IPv6, and the image will show which of these protocols your browser prefers.

The image on the right is only loadable only your browser is capable of using IPv6.

With the browser measurement test we use the same approach of loading a number of images, but in the test the images are just one pixel in size, so the data easily fits within a single packet. In the measurement script there are also web directives associated with the image to direct the browser not to display the image. The measurement script also waits for all visible objects on the page to complete loading, so that this measurement test does not impair your normal browsing experience in any way. We have also tried to suppress any visible error messages associated with these tests. The intent is that the measurement test happens in the background, and without disturbing or altering your browsing experience in any way.

The measurement test itself attempts to load five images. The first is an image that is only accessible using IPv4. The second is an image that is accessible over IPv4 or IPv6. The third is an image only accessible using IPv6. The fourth is also an IPv6 image, but in this case the URL does not rely on the Domain Name System to access the image (with certain operating systems and browsers this test generates a different response from the browser). The final image is a "result" load, which is delayed by 10 seconds. This image contains the timer measurements that the browser contains to report back which for the first four images were loaded, and the amount of time takes to load each one.

Data Collection and Privacy

Respecting your privacy is important for APNIC. We do not collect any personal data in this activity. The data we collect and retain includes the combination of success and failure for each of the four tests, the time taken for each of the four tests, and the round trip time for packets to be exchanged between your browser and the server.

In this process we obtain the source IP address of your browser. We retain information relating to the country of registration of this source address and the originating Autonomous System number for this source address.

We do not publish any of the source addresses collected in this activity.

Acknowledgements

This measurement activity is being undertaken by APNIC, the Asia Pacific Network Information Centre, as a service to the Internet community at large.

We would like to acknowledge the generous support that has been provided by Google, the Internet Software Consortium and the RIPE Network Coordination Centre in undertaking this activity.

Contact Us

If you have any questions about this measurement activity please contact us at: research (a) apnic.net

Additional Resources