15.3.2 Network Address Translation (NAT)

The rapid growth of the Internet eventually exhausted the pool of publicly available IPv4 addresses. Although approximately 4.3 billion addresses are theoretically possible, many are reserved for special purposes, and the explosive growth of personal computers, mobile devices, and Internet of Things (IoT) equipment soon exceeded the available address space.

One technique that significantly delayed address exhaustion was Network Address Translation (NAT). Rather than assigning every device a globally unique public address, NAT allows many devices within a private network to share a single public IPv4 address.

In a typical home or office network, computers, smartphones, printers, and other devices use private IP addresses that are valid only within the local network. When these devices communicate with the Internet, the broadband router replaces the private source address with its own public address before forwarding the packet. Replies are translated back to the appropriate internal device using a translation table maintained by the router.

This approach conserves public IPv4 addresses while providing a simple degree of isolation between private networks and the public Internet. Today, NAT is employed by the vast majority of residential and small-business Internet connections.

Although highly successful, NAT introduces additional complexity because the end-to-end transparency originally intended for the Internet is partially lost. This was one of the motivations for developing IPv6, which provides sufficient addresses for virtually every Internet-connected device to possess its own globally unique address.