Back to all articles
IPv6IPv4Network ArchitectureHistory

IPv6: Why It's Taking 20 Years to Upgrade the Internet

Chris VanceJuly 2, 2024

On February 3, 2011, the Internet Assigned Numbers Authority (IANA) handed out the very last block of available IPv4 addresses. The pool was officially empty. The 4.3 billion addresses were gone.

This was not a surprise. The engineering community saw this coming in the early 1990s. In 1998, they published the official standard for IPv6 (Internet Protocol Version 6), a brilliant, elegant replacement that would give us enough IP addresses to assign one to every atom on the surface of the earth.

The plan was simple: Everyone upgrades their routers, we switch to IPv6, and the crisis is averted.

More than 25 years later, global IPv6 adoption is only hovering around 45%.

Why is it taking so long? Why is the technology industry, which normally obsoletes hardware every three years, completely failing to upgrade the most foundational protocol of the internet?

The Brilliance of IPv6

First, we must acknowledge that IPv6 is an engineering masterpiece. It doesn't just add more addresses; it fixes dozens of subtle flaws in the aging IPv4 protocol.

  • **Infinite Addresses:** An IPv4 address is 32 bits (`192.168.1.1`). An IPv6 address is 128 bits (`2001:0db8:85a3:0000:0000:8a2e:0370:7334`). We will never, ever run out again.
  • **No More NAT:** Because we have infinite addresses, every single device on earth—your phone, your fridge, your smart lightbulb—can have a globally unique, publicly routable IP address. We don't need Network Address Translation (NAT) to hide multiple devices behind one IP anymore. The internet can return to its true "end-to-end" design.
  • **Smarter Routing:** The IPv6 packet header was redesigned to be vastly simpler and more efficient for core internet routers to process, speeding up global traffic.
  • **Auto-Configuration:** IPv6 devices don't necessarily need a DHCP server to get an address. Using a feature called SLAAC (Stateless Address Autoconfiguration), a device can plug into a network, listen to the router, and mathematically generate its own globally unique IP address instantly.

    • The Fatal Flaw: Lack of Backward Compatibility

    If IPv6 is so great, what went wrong?

    The engineers who designed IPv6 made one critical, arguably fatal, decision: IPv6 is not backward compatible with IPv4.

    An IPv4 computer cannot talk to an IPv6 computer. An IPv4 router cannot process an IPv6 packet. They are entirely different languages.

    This meant there was no easy "upgrade path." You couldn't just slowly phase it in. To transition to IPv6, a network administrator had to run both protocols simultaneously (a setup called Dual-Stack). The router needed an IPv4 address to talk to the old internet, and an IPv6 address to talk to the new internet.

    The Economic Problem

    Because there was no backward compatibility, there was no economic incentive to upgrade.

    Imagine you are the CTO of a major corporation in 2010. Your IT team says, "We need 2 million dollars to upgrade our firewalls and routers to support IPv6."

    You ask: "Will this make our website faster?" *No.*

    "Will it give us new features?" *No.*

    "Will the customers notice a difference?" *No.*

    "If we don't upgrade, will our website go offline?" *No, the IPv4 internet will keep working.*

    The CTO denies the budget.

    And because the corporations didn't upgrade their websites, the Internet Service Providers (ISPs) saw no reason to upgrade their home routers. And because the ISPs didn't upgrade, the consumer electronics companies saw no reason to force IPv6 on their devices. It was a massive, global Mexican standoff.

    The Savior and the Villain: NAT

    The other reason the transition stalled was that we invented a Band-Aid: NAT (Network Address Translation).

    As we discussed in a previous blog, NAT allowed millions of home networks and corporate offices to hide thousands of computers behind a single IPv4 address.

    NAT was supposed to be a temporary hack to buy us a few years to deploy IPv6. Instead, it worked so incredibly well that it removed the urgency of the crisis entirely. Why spend millions upgrading to IPv6 when a $50 home router with NAT solves the IP shortage instantly?

    The Tipping Point

    So what finally started pushing the needle toward IPv6 in recent years?

    1. Mobile Networks: Smartphones exploded. A single cellular carrier in India or China might have 500 million subscribers. You cannot use NAT to hide 500 million phones behind a few IPv4 addresses; the overhead is too massive. The cellular networks *had* to switch to IPv6. If you are reading this on 4G or 5G right now, you are almost certainly using an IPv6 address.

    2. The Black Market for IPv4: Because there are no new IPv4 addresses available, a massive black market emerged. Today, a single IPv4 address sells for around $40. If a new cloud hosting company wants to buy a block of 65,000 IPv4 addresses, it will cost them over $2.5 million just for the numbers. IPv6 addresses are virtually free. The economics finally flipped.

    The Future

    We are currently in the messy middle. Major tech giants (Google, Facebook, Netflix) and cellular providers are almost entirely IPv6. But thousands of smaller websites, corporate VPNs, and legacy systems are stubbornly clinging to IPv4.

    We will likely be living in a Dual-Stack world for another decade. The transition to IPv6 is the largest infrastructure upgrade in human history, and it proves that in technology, having the mathematically superior solution is only half the battle; the other half is convincing the accountants to pay for it.