The transition to IPv6 is one of the most significant and protracted projects in the history of the Internet. Decided in the 1990s in response to the foreseeable growth in the number of connected devices, the migration is still ongoing in 2025, more than twenty years after the publication of RFC 2460. Understanding why this transition is both urgent and complex is essential to grasping the architecture of modern networks.

1. The depletion of IPv4 addresses

IPv4, designed in 1981, offers an address space of 2³² addresses, or approximately 4.3 billion unique IP addresses. This figure seemed astronomical in 1981, when the Internet comprised just a few hundred computers. Today, it is woefully inadequate in the face of the explosion in connected devices: smartphones, tablets, IoT devices, vehicles and industrial equipment. There are now more than 15 billion connected devices worldwide, yet only 4.3 billion possible IPv4 addresses.

The exhaustion of IPv4 addresses occurred in stages. On 3 February 2011, the IANA (Internet Assigned Numbers Authority) allocated the last five /8 blocks of IPv4 addresses from its global pool, marking the official exhaustion of the central pool.

The regional registries (RIRs) have successively exhausted their reserves:

· APNIC 2011

· RIPE NCC (Europe) in 2012

· ARIN (North America) in 2015

· LACNIC and AFRINIC shortly afterwards.

Since then, there have been no public IPv4 addresses available through standard allocation; only recycled addresses or those from the secondary market can be obtained.

In response to this shortage, two workarounds have been widely implemented. NAT (Network Address Translation) allows multiple devices on a private network to share a single public IP address. This is why your internet router has a public address, whilst all the devices in your home have private addresses (192.168.x.x, 10.x.x.x). NAT works well for outbound connections, but it introduces complications for peer-to-peer protocols, VoIP, online gaming and inbound connections. CGNAT (Carrier-Grade NAT), deployed by ISPs, goes even further by sharing the same public IPv4 address amongst several different subscribers, creating a double NAT that further complicates diagnostics and certain uses.

A secondary market for IPv4 addresses has developed in response to the shortage. Organisations holding large blocks of unused addresses (such as US universities that were allocated full /8 blocks in the 1980s, companies in liquidation) are reselling them at prices that reached $40 to $60 per address in 2025, or $40,000 to $60,000 for a /24 block of 256 addresses. These high costs provide a strong economic incentive to migrate to IPv6.

2. IPv6: Global Deployment Status in 2025

IPv6 offers an address space of 2¹²⁸ addresses, or 340 sextillion addresses, enough to assign several billion addresses to every grain of sand on Earth. Beyond the address space, IPv6 brings significant architectural improvements: the simplified IP header speeds up processing by routers, native multicast replaces IPv4 broadcast more efficiently, stateless autoconfiguration (SLAAC) allows a device to configure itself automatically without a DHCP server, and IPv6 mobility (RFC 6275) keeps connections active during network changes.

The global roll-out of IPv6 is making significant progress, although significant disparities remain between countries and types of network. India has become the world leader in IPv6 adoption, accounting for around 70% of traffic, thanks to the massive roll-out by the operator Jio of a native IPv6 4G/5G network for several hundred million subscribers. The United States has an adoption rate of around 55%, driven by Comcast, T-Mobile and major web services (Google, Netflix and Facebook have had IPv6 enabled for several years). Germany is also approaching 55%, with Deutsche Telekom having rolled out IPv6 on a massive scale across its fixed and mobile networks.

France is expected to reach around 40% adoption by 2025. Free is ahead of the curve, with around 60% of its traffic using IPv6, having been the first French ISP to enable IPv6 by default on its Freeboxes. Orange, SFR and Bouygues Telecom have gradually enabled dual-stack (IPv4 and IPv6 simultaneously) on their FTTH services. The global average stands at around 45% according to APNIC and Google statistics, but this average masks significant disparities: Africa remains below 10% overall, held back by older network equipment and migration costs.

Region / Country

Adoption IPv6 (2025)

Location

Inde

~70 %

Global leader, Jio’s native IPv6 network

United States

~55 %

Comcast, T-Mobile, Google and Netflix have been rolled out

Germany

~55 %

Deutsche Telekom in the lead

France

~40 %

Free is the market leader (~60%), offering dual-stack FTTH services

Japon

~45 %

NTT, KDDI and SoftBank have been deployed

Chine

~30 %

China Mobile, rapid growth

Africa

<10%

Under development, older equipment

Global average

~45 %

Source : APNIC / Google IPv6 Statistics 2025

For corporate networks, the transition to IPv6 is often slower than for public networks, as it involves updating numerous internal systems: server addressing, firewall rules, automation scripts, monitoring systems and LDAP directories. Dual-stack, where each device has both an IPv4 and an IPv6 address, is the recommended transition strategy: it allows IPv6 to be adopted gradually without disrupting existing IPv4 services.

💡 Preparing your network for IPv6

Any new network deployment in 2025 must be designed to be natively dual-stack. Equipment purchased today will still be in use in 10 years’ time, by which point IPv4 will likely have been phased out in many regions. Check the IPv6 compatibility of all equipment (firewalls, monitoring sensors, directories, backup systems) before deployment. Managing IPv6 prefixes within an organisation requires careful consideration of the addressing plan: unlike with IPv4, there is no need to count addresses, but it is essential to structure prefixes hierarchically to facilitate aggregation in routing tables.

Modifié le: vendredi 9 octobre 2026, 10:03