New research conducted by a team of researchers shows that IPv6 adoption has fundamentally left behind its experimental origins. Their research paper will be presented at the upcoming 2014 ACM SIGCOMM conference in Chicago.
We are probably all used to the idea that IPv6 adoption is increasing and that every year we see more IPv6 connected hosts on the Internet. What sets this paper apart is the sheer amount of data the researchers collected and analyzed. In total the researchers investigated 12 different metrics from ten different data-sets. Their goal is to achieve the most accurate view of IPv6 adoption to date.
“To achieve our goal we assemble a set of six publicly-accessible datasets that speak to one or more aspects of IPv6 adoption. We add four additional, previously-unpublished datasets, including a global Internet traffic dataset that includes traffic statistics from 260 providers and represents 16,200 petabytes/month, or approximately 33-50% of all Internet traffic—the largest traffic sample reported in an IPv6 study. In addition to the traffic data, we add DNS query data from several of the largest globally-distributed IPv4-based replicas of the .com and .net top-level domains (TLDs), as well as nearly all native IPv6 replicas for these TLDs.”
Their findings reflect the wealth of data collected. What I particularly find interesting is the application of their data-sets to a topological understanding of IPv6 adoption.
One cool thing they measure is increase in IPv4 paths vs. increase in IPv6 paths. By comparing the topology of the global IPv4 and IPv6 networks they find that, “[the] number of IPv6 paths has a 110-fold increase from January 2004 to January 2014, while there is only an eight-fold increase in the number of IPv4 paths.” Rarely do we get such an elegant understanding of the growth of IPv6 compared to that of the total Internet.
They’re also able to show that the number of Autonomous Systems (ASes) with few neighbors, primarily multi-homed edge ASes, that support IPv6 has increased significantly since 2004. This is interesting because this represents networks that usually lag behind first adopters.
Well connected ASes with many neighbors are going to have the easiest time adopting IPv6 because they have the most resources and it benefits them the most. They typically have a scale advantage, and are much more likely to have an immediate neighbor that has already deployed IPv6. Less connected networks at the edge are those networks you would expect to lag significantly behind in IPv6 adoption, so it’s refreshing to see that their adoption rates have increased significantly since 2004.
There are other gems in this research that I’ll leave for the curious reader to find. The paper can be downloaded from the International Computer Science Institute’s website. Here is the direct PDF link, you can also download their data if you wish to perform your own analysis. The paper’s authors are Jakub Czyz, Scott Iekel-Johnson, Mark Allman, Eric Osterweil, Jing Zhang, and Michael Bailey.
If you would like to get started with IPv6, please visit our IPv6 resources or begin with our “Start Here” page to help find resources most appropriate for your type of organization. If you have an IPv6 case study you think we should consider for inclusion on our site, please contact us – we are always looking for more!