Network Address Translation

Network Address Translation (NAT) is a mechanism that maps multiple private IP addresses to a smaller number of public IP addresses, typically a single public address assigned to a router or firewall. Developed as a response to the exhaustion of IPv4 addresses, NAT has become so ubiquitous that it is now a defining feature of the consumer internet: most home networks, corporate LANs, and mobile devices sit behind NAT devices that rewrite the source or destination addresses of packets as they cross the boundary between private and public networks.

NAT breaks the end-to-end principle of internet design, which held that any host should be able to communicate directly with any other host. A device behind a NAT cannot receive unsolicited inbound connections unless the NAT device is explicitly configured to forward specific ports — a process called port forwarding that is notoriously difficult for non-technical users. This asymmetry — outbound connections work, inbound connections do not — has shaped the architecture of internet applications. Peer-to-peer systems, voice-over-IP, and online games all require complex NAT traversal techniques (STUN, TURN, ICE) to establish direct connections between hosts that cannot see each other.

NAT was sold as a temporary workaround for IPv4 address scarcity. It has persisted for three decades, becoming so deeply embedded in network infrastructure that removing it would require redesigning not just protocols but the economics of internet service provision. NAT is not a temporary measure; it is a permanent deformation of the internet's architecture, born from a failure to deploy IPv6 and sustained by the inertia of billions of deployed devices. The internet was designed to be a network of peers. NAT made it a network of clients.