The Freename Forum

Imagine the Internet is running out of IPv4 addresses and nobody cares. Scientist Geoff Huston has gotten to the bottom of the phenomenon of the gradual transition to the successor protocol IPv6 - and questions it.

Since 2002, we have been reporting here on the Internet of the future, based on the IPv6 Internet protocol. IP addresses form the basis of the Domain Name System (DNS), converting combinations of numbers into domains that are easier to remember. In the currently widespread IPv4 version, however, the available addresses are becoming scarce, as “only” 4.2 billion alternatives are conceivable in mathematical terms. Since the early 1990s, the engineers of the Internet Engineering Task Force (IETF) have therefore been working on the successor protocol IPv6, which has long been available to the public and theoretically makes 3.4 X 10 to the power of 38 addresses possible; however, its breakthrough is still pending. The address shortage has led to lively trading with prices of between US$ 30 and US$ 40 per IPv4 address, but despite regular warnings from the five Regional Internet Registries (RIRs) that the Internet is running out of addresses, there is no sign of IPv6 being used across the board. Back in May 2022, Geoff Huston of the Asia-Pacific Network Information Centre (APIC), one of the five RIRs, commented that if the prospect of running out of IPv4 addresses had given any urgency to the RIRs' efforts, they had been living with it for a decade and were now jaded. However, he at least hoped that the transition had come closer.

Around two and a half years later, Huston has also given up this hope. For the year 2024, he has calculated that just over a third of internet users will be able to access a pure IPv6 service; everyone else will still only be using IPv4 - even though the addresses should have long since run out, as the number of devices with an internet connection is constantly increasing. By the end of 2024, it is estimated that around 20 billion devices will be using the internet. This is possible because the internet shares each individual IPv4 address with an average of seven devices. “If end-to-end was the sustaining principle of the Internet architecture then as far as the users of IPv4-based access and services are concerned, then it's all over!” said Huston. At the same time, however, he was also self-critical: IPv6 is not faster, more versatile or more secure than IPv4; nor does it offer any other tangible benefits in the form of lower costs, higher revenues or greater market share, but merely provides more addresses.

For Huston, the solution is therefore to consider the transition from IPv4 to IPv6 complete when IPv4 is no longer required, i.e. when an Internet service provider can operate a viable service that uses only IPv6 addresses and has no supported IPv4 access mechanisms at all. The responsibility therefore lies with the networks and network operators to first invest in the transition to a dual-stack platform with the aim of phasing out IPv4 support. It is also important to move everything from the network to applications. Huston poses the almost heretical question: “Is universal unique endpoint addressing a 1980s concept whose time has come and gone? If network transactions are localized, then what is the residual role of a unique global endpoint addressing clients or services?” He even questions the definition of the Internet as a common transmission structure and a common protocol address pool, but without giving a conclusive answer. But it doesn't need to be, because one thing is clear: despite the shortage of IPv4 addresses, there are no immediate technical limits to the growth of the Internet - and certainly not to the registration of domain names.

You can find the article by Geoff Huston at:
https://blog.apnic.net/2024/10/22/the-ipv6-transition/