Packet Guide to Routing and Switching

Chapter 7. Network Address Translation

The world has run out of IPv4 address space. Of course, this is not really news. In 2011 it was announced that some of the last five /8s of publicly available IPv4 address space were being allocated.

Since then, IPv4 address space depletion has been a continuing concern. Trends such as the Internet of Things (IoT) promise to add IP addresses to everything from cars to refrigerators. The simple truth is that there are not enough addresses to go around (and this has been the case for several years).

Many believe that IPv6, with its vast 128-bit addressing, will save the Internet. To date there have been a series of IPv6 days in which organizations are challenged and encouraged to transition from IPv4 to IPv6. Given these factors, the world should be running on IPv6. But it is not. One of the biggest reasons for this is network address translation (NAT).

NAT is described in RFC 1631as a technique by which a single IPv4 public address can be used by several computers. This is because the computers actually use private addresses (also known as RFC 1918 addresses) which are translated into the public address. The effect is that the number of IPv4 public addresses needed is smaller than the number of computers. It works so well that in all likelihood, we will continue to rely on NAT.

NAT describes the process of converting traffic coming from private, inside IPv4 addresses to traffic that appears to have come from a globally unique outside address. The ...