As part of ongoing efforts to chart the rise and rise of IPv6, RIPE Labs looked into the difference between average IPv6 capability, per economy over a few days in June 2020, versus the most recent similar averages for 2021.
2021 marks a full ten years since World IPv6 Day, the coordinated test event that led to the IPv6 launch event on 6 June 2012. We’re all aware that IPv6 deployment is, so far, a mixed story: depending on your network, it might either be the dominant network protocol, or it might be an item kicked into the long-grass of a planning document years into the future, or (more likely) it’s somewhere in the middle of that spectrum.
So as usual, let’s ask ourselves: Where are we now, globally, and what’s changed?
Last question first. What’s changed in the last year?
The pandemic has, of course, changed a great many things, including network traffic patterns, such as a shift away from corporate network traffic towards more domestic traffic during weekdays. Let’s take a look at the data generously made available by APNIC.
Since we’re interested in what’s changed, we’ve opted to create a map of deltas per economy, to show us clearly the differences in percentage of IPv6 readiness by economy:
Figure 1 shows the difference between an average (mean) measure of IPv6 capability per economy over a few days in June 2020 versus the most recent similar averages for 2021. Greys are neutral; blues indicate a positive change (a higher proportion of IPv6 readiness); and oranges indicate a negative change (a lower proportion of IPv6 readiness).
There are a couple of notable upticks in this data that stand out.
First, Puerto Rico (PR) appears to be a clear winner by APNIC measurements, though this must be viewed through the filter of a relatively small population and how capable geo-data packs are at identifying users as being located in PR or the US. The major US networks operate here, so cellular deployments include the same IPv6-ready mobile carriers. Granularity of the data and classification of PR versus US affects measurements heavily, leading to wildly different measures for Puerto Rico — Google says 6.48% IPv6 traffic today, while Facebook measures 2.79%. We’ve written about differing IPv6 metrics in the past.
This is followed by Saudi Arabia, with a significant population over 34 million, where various IPv6 measurement metrics agree that IPv6 readiness has risen over 40% in recent months. Also at the top of this list is Israel, which had some IPv6 traffic before June 2020, but has shown a clear uptick to around 25% readiness in the last year. Paraguay, which effectively had no IPv6 until the tail end of 2019, showed a clear uptick at the end of 2020, and is now approaching 20% readiness. Mexico has been on a steady increase over the last few years and is now at around 40% readiness. The figures for each of these economies indicate significant IPv6 deployment efforts.
There are a couple of surprises in the other direction, however.
The US is the most surprising, with a picture that is slightly more complex. The US has had a strong IPv6 deployment program since serious deployments began, and it’s fair to say there are fewer big turn-up events still waiting to happen, as we might see in some other economies. One year ago, stay-at-home orders were being adhered to, and now in mid-2021 the US is months into their rapid vaccination program. We know that corporate networks tend to lag their domestic and mobile ISP counterparts with IPv6 deployment, so this shift might be an indication of people spending time in offices rather than working from home. A similar situation may be true, for example, in Germany, if restrictions are easing and working from home is becoming less common. We see little evidence of many networks significantly reducing their IPv6 footprints, which suggests that IPv6 availability isn’t going away — the traffic is simply moving around again.
Standing starts and other notable jumps
There are of course economies with absolutely no IPv6 deployment, at least beyond extremely small networks and test deployments. That’s a nontrivial list, but it’s a list that gets a little smaller each year.
We can see one of those above in Figure 1. Pakistan appears to have moved from near-zero IPv6 readiness in mid-2020 up to 2.3% in mid 2021. With a population of over 225 million, this is a nontrivial traffic shift. And it’s not just Pakistan. There are movements in other places:
|Economy||IPv6 readiness, 2021||Internet users (≈, from wiki)|
Table 1 — Economies with notable jumps during the sampled period.
Where are we now?
On the interactive map, you can jump between 2020 or 2021 to compare how each year looks. Given that the size of the population that has Internet access will, in part, drive the traffic volume in any given economy or region, the changes across large economies, sorted by Internet population, is shown in Table 2.
|Economy||IPv6 readiness, 2021||Readiness delta from 2020||Internet users (≈, from wiki)|
Table 2 — Changes across large economies, sorted by Internet population.
Each of these economies has significant Internet populations where small deltas can make a real impact on how much traffic is served over IPv6 versus IPv4. The general trend in the last year indicates a greater level of IPv6 readiness, though of course, this may be mediated by working from home and a future partial shift back towards office work (as suggested above for the US).
More work to be done
IPv6 clearly isn’t ‘done’ yet — there’s a lot of work still to do if the protocol is to eventually become ubiquitous. The same story of previous years is perhaps still true —more ASNs are announcing IPv6 than before, and so piece-by-piece we approach some sort of parity.
Deployment is still uneven though. Many economies do not have any apparent IPv6 traffic at all, and although the large content providers (being the primary participants in the IPv6 launch events) have been IPv6-ready for many years, access networks are a mixed bag and corporate networks still seem to be trailing.
Regardless, there is clear progress in multiple regions around the world, in both large economies and small. This is good news!
Stephen Strowes is a senior researcher at RIPE NCC.
This post was originally published at RIPE Labs.
The views expressed by the authors of this blog are their own and do not necessarily reflect the views of APNIC. Please note a Code of Conduct applies to this blog.