Geert Jan de Groot

Thank you for your analysis. I do have a few points to make though. IPv6 was originally defined to solve address space issues with IPv4. So, on the Internet, we would add IPv6, run dual-stack, and when everybody has IPv6 access, we could stop providing services via IPv4. When we're talking about IPv6 deployment, that should be the goal, and how are we faring? One way to look at this is mobile access, that is access via smartphones. Many, many people have smartphones and for many it is the primary or even the only network access device. Looking at this, the situation is much less favorable. Mobile devices that have wifi access - if the wifi provides IPv6, then many (but not all) devices will use it. In some case IPv4 is preferred, in some cases IPv6 is. Providing and analyzing the guest network of a youth organisation is interesting in this regard. But the real problem is that IPv6 is almost never made available over mobile (4G / 5G) networks. I am not able to buy a mobile subscription that provides IPv6 in the Netherlands, and I think this applies to many EU countries (would be interesting to investigate!) From what I have been told (and please correct me!), providing IPv6 connectivity incurs an additional chargeable connection slot at the mobile network infrastructure, incurring additional cost. If that is true (again, I don't know!) then providing IPv6 makes no business sense since the operator is doubling his connection cost w/o clear benefit. There would be a clear benefit if all providers do this (forced to do this) because in this case all providers incur the same cost and hence the cost would be "zero", at least from an intra-provider competition point of view. This would be of great benefit for one's national economy because we could start working on the next phase, plan to phase-out IPv4. So, I asked. When the auction process for 5G spectrum space was done in the Netherlands, I suggested that the 5G service would be dual-stack as part of the bidding process. This suggestion was denied; the regulator commented that the service should be "technology-neutral" and hence didn't feel like supporting IPv6 as part of the 5G service. I think this would be interesting successive work. One, find out if, where and how IPv6 connectivity is made available in mobile networks (I'm only aware of Slovenia doing this). And two, as long as regulators don't see the advantage of doing IPv6 as a solution to the problems of address space scarcity on one's national infrastructure, I think there is some work here. Perhaps an organisation like ISOC could consider this work, sort of as a next phase of the IPv6 deployment initiative. It is possible (though some providers find it surprisingly difficult) to make one's consumer home-network dual-stack. But without involvement of mobile access IPv6 remains a "server only" thing. And that's not what we had in mind when the IPv6 thing started.