June marked the 9th anniversary of “IPv6 Launch Day”, a global effort on 6th June 2012 in which leading Internet companies and access providers, including Google and Facebook, switched on IPv6 for their services. The idea was not only to lead by example, but also to make clear the migration to IPv6 was feasible and had limited risks to the stability of the networks and services that make up the Internet.
Although uptake of IPv6 was relatively slow in the years following that day, the subsequent depletion of available IPv4 addresses at the Regional Internet Registries and emergence of a secondary market in used Internet addresses seem to have caused the adoption rate to pick up. Data published by large services such as Google and Facebook as well as dedicated measurements by APNIC now all indicate that somewhere in the region of 25 to 30 percent of global Internet users are using IPv6. While this at first appears a significant proportion of the user population, and in fact it is, this isn’t anywhere close to the goal as it was envisioned 25 years ago when IPv6 was standardised – which was for the transition to be fully complete before the global supply of IPv4 addresses had been depleted.
A lower-than-expected uptake has again started the discussion on whether policy measures and regulatory oversight could help – but what would the consequence be?
Uneven global adoption is cause for concern
When zooming into the national level or even looking at deployment in individual networks, we see a very scattered picture as there are still massive differences in the level of IPv6 deployment. Those 1 in 3 users that use IPv6 are distributed unevenly, with most residing in more developed countries – India being the noteworthy exception to that trend. Additionally, a large portion of those users are connected to only a few networks, such as T-Mobile in the US, Sky Broadband in the UK, and Jio in India, which all have IPv6 adoption rates close to 100 percent.
This uneven adoption is cause for concern. While the Internet continues to grow and IPv6 support often still is lacking, the pressure on the already limited supply of IPv4 addresses will continue to rise. With the result being even higher prices for addresses in the secondary market (which are already commonly exceeding $35 per address). At a global level, this risks increasing the digital divide, where increasingly we see operators and emerging markets struggling to obtain the IPv4 addresses needed to connect their users and services. But even developed markets must start considering the effects of increasing prices on competition and affordability – where ultimately it will be the user who pays the price for those IPv4 addresses.
- Here's our take on the best VPN services available now
EU action could be coming
Of course, a full transition to IPv6 would avoid all this, but the rational expectation is that this will still take years (and quite possibly decades) to complete. Policymakers all over the world are waking up to the importance of reaching this goal. Several factors have caught their attention, such as the concern that we cannot sustain exponential Internet growth with IPv4 alone, as well the negative consequences of sharing IPv4 addresses (such as using Carrier Grade NAT to lessen the burden of the address shortage), which for instance hampers law enforcement.
This is not new, and under the maxim of “stimulate where you can, regulate where you must”, a substantial number of governments have already taken steps to promote IPv6 via policies which mandate the inclusion of IPv6 support for government ICT and network procurement.
Unfortunately, while some success has been seen, for many this is not enough – and the level and pace of IPv6 adoption is still considered too low. As the European Commission described it in its cybersecurity strategy published in December: “[...] accelerate the uptake of key Internet standards including IPv6 [...] not excluding regulatory measures like a European sunset clause for IPv4 to steer the market if there is insufficient progress towards their adoption.”
Now these words were obviously chosen with the best intentions, despite the open question of what “sufficient” progress would be. But they are also slightly worrisome, as it does leave a clear signal to the market that an intervention might be coming. Again, nothing new – with the Internet being such a fundamental part of our lives and economies, this certainly wouldn’t be the first legislative proposal that seeks to steer the industry in the right direction. We can safely say that the days where the Internet could exist as a truly unregulated space are long gone, never to return.
The dangers of intervention
An EU IPv6 mandate could be seen as quite controversial. Until now, the owners and operators of networks and services have been totally free in choosing which protocols, technologies, and standards they support. Nowhere is there any law or binding document that says a service provider must deploy a particular technology – at least not regarding the Internet, and especially not regarding the network layers. Many would argue that this has been a large contributing factor to the Internet’s success – it has always been the case that the market could pick its own winners.
Whether to step over this boundary is becoming a true dilemma. Does the low uptake of IPv6 cause such problems that an intervention is warranted? And if so, would one be fundamentally altering the Internet by removing the free choice and market-driven approach that, over the last 40 years, has made it what it is?
Subsequently, if there is consensus that this is a form of market failure and intervention is warranted, in what form should it come? The Commission’s text already hints that mandating use of IPv6 might not be the only option. You could also take aim at IPv4, which after all might be presented as the root cause. But it will not disappear overnight and a full ban of IPv4 might require a transition period which could last for decades. Although we cannot expect any immediate results, the option of sunsetting IPv4 does have the benefit of leaving all future options open – which is one of the main drawbacks of making IPv6 mandatory. What if the Internet further evolves, and a new and better protocol is identified and developed to standard – would we do it all over again, or are we forever committed to IPv6?
Correcting our course without legislation
While there are no easy and quick answers here, we do need to really consider our options. Not only as policymakers, but most of all as an industry itself. This can all be avoided, but for that to happen we must pick up the pace on IPv6 and show that the market is indeed still able to correct its own course. We, the Internet, have stayed clear of legislation dictating technical design for four decades, and if this comes to an end, it truly would be our own failure.
Meanwhile, policymakers are slowly testing the waters – for instance in France where “IPv6 readiness” has become a requirement for operators bidding for 5G spectrum licenses. On the other end of the spectrum, I recently became aware of the Finnish regulator, which outside of promoting IPv6 adoption, now insists that customers must be able to request their own non-shared IPv4 address from their provider. This is a measure that could increase costs, and which could act as a significant barrier to new entrants who must spend money to obtain IPv4 addresses, even if they fully deploy IPv6.
It is becoming increasingly clear that one way or another, IPv6 will happen. We can only hope that when it does, it happens in the Internet way: bottom-up and driven by users’ needs and market dynamics. Let us try and avoid the failures of the past where many technology choices were mandated in a top-down fashion and consider why the Internet has lived on to survive them all.
Marco Hogewoning, Manager of Public Policy and Internet Governance, RIPE NCC