The Internet Ran Out of Addresses on February 3, 2011. Nobody Noticed Because of a Workaround Invented in 1994.

Why IPv4 Was Designed with 4.3 Billion Addresses (and Why That Was Not Enough)

Every device that communicates on the internet needs an IP address, a number that identifies it on the network. IPv4, the version of the Internet Protocol introduced in 1981, uses 32-bit addresses. A 32-bit number has 2 to the power of 32 possible values: exactly 4,294,967,296 or about 4.3 billion unique addresses.

In 1981, 4.3 billion addresses seemed impossible to exhaust. The internet was a research network used by universities, defense contractors, and government agencies. The team that designed IPv4, led by Vint Cerf and Bob Kahn, was solving a specific communication problem for a specific community. The idea that every person on earth would one day own multiple networked devices was not part of the planning scenario.

The address space was allocated generously in the early years. Class A allocations gave entire blocks of 16 million addresses to single organizations. MIT received 18.9.0.0/8, a block of over 16 million addresses, in 1981. Apple received 17.0.0.0/8. The US Department of Defense received several entire /8 blocks. These early allocations consumed enormous portions of the address space for organizations that, even at their largest, would never have tens of millions of devices.

In 1992, a network engineer named Vint Cerf, now working on internet policy rather than protocol design, published a warning through the Internet Society that IPv4 addresses would run out within years if usage growth continued. The warning was correct about the problem but optimistic about the timeline. Usage grew faster than projected.

NAT: The Workaround That Bought the Internet 15 Extra Years

The technology that delayed IPv4 exhaustion by roughly fifteen years is Network Address Translation, or NAT. Designed by a team including Paul Francis and Kjeld Bøgh Andersen and described in RFC 1631 in 1994, NAT allows many devices to share a single public IP address by mapping private internal addresses to the single external address through the router.

When you connect to the internet through a home router, your devices have private IP addresses: numbers in the ranges 192.168.x.x or 10.x.x.x. These ranges are defined in RFC 1918 as reserved for private use and never routable on the public internet. Your router has one public IP address assigned by your ISP. When your phone sends traffic to a server, the router replaces the private source address with the public address and records the mapping. When the server responds, the router looks up the mapping and delivers the response to the correct internal device.

NAT allowed a household of six devices to consume only one public IP address. Mobile carriers used the same trick for millions of customers, a technique called Carrier-Grade NAT or CGNAT. Internet cafes, corporate networks, and university campuses all ran NAT. The technology was effective enough that the 2011 exhaustion event passed largely without public notice. The addresses ran out, and the internet kept working.

The cost of NAT is complexity. Applications that need a device to be directly addressable, certain peer-to-peer applications, voice and video calling, and some gaming services, face challenges when operating behind NAT. Techniques like STUN and TURN, documented in RFCs 5389 and 5766, allow peer-to-peer applications to establish connections through NAT layers, but they add latency and require intermediary servers.

IPv6: The Solution That Has Been Arriving Since 1998

IPv6, defined in RFC 2460 in 1998, is the designed solution to IPv4 exhaustion. IPv6 uses 128-bit addresses, providing 2 to the power of 128 possible values: approximately 340 undecillion unique addresses. Written in full, this number has 39 digits. Every person on earth could have more addresses than there are grains of sand on all beaches, with addresses left over. Address exhaustion under IPv6 is not a practical concern.

IPv6 deployment has been slower than expected for twenty-five years despite the clear need. The problem is backwards compatibility: IPv4 and IPv6 are not interoperable. A device using only IPv6 cannot communicate with a server that has only an IPv4 address. This requires running dual-stack deployments, where networks and servers support both protocols simultaneously, which doubles operational complexity.

As of 2024, Google's IPv6 statistics show that approximately 45 percent of users reach Google's services over IPv6, up from near-zero a decade ago. The transition is happening but has taken decades rather than the years optimists projected.

Your current IP address, the one visible to websites you visit, is almost certainly an IPv4 address allocated from your ISP's pool, which the ISP itself shares among customers using NAT. Unless you are on a modern mobile carrier or corporate network with IPv6 deployment, you are using an address space that was theoretically exhausted fourteen years ago, sustained by translation technology that was designed as a temporary fix.

Conclusion

The internet that was designed for 4.3 billion addresses now has roughly 5 billion connected devices, sustained by NAT and slowly transitioning to IPv6. The address your device shows when you check your IP is one small part of this infrastructure.

The What Is My IP tool at ToolHQ shows your current public IP address, location, and ISP information as seen by the websites you visit. The address displayed is the one that identifies your connection to the public internet.

Frequently Asked Questions

Try These Free Tools