There are several ways of implementing network address and port translation. In some application protocols that use IP address information, the application running on a node in the masqueraded network needs to determine the external address of the NAT, i.e., the address that its communication peers detect, and, furthermore, often needs to examine and categorize the type of mapping in use. Usually this is done because it is desired to set up a direct communications path (either to save the cost of taking the data via a server or to improve performance) between two clients both of which are behind separate NATs.

For this purpose, the Simple traversal of UDP over NATs (STUN) protocol was developed (RFC 3489, March 2003). It classified NAT implementation as full-cone NAT, (address) restricted-cone NAT, port-restricted cone NAT or symmetric NAT and proposed a methodology for testing a device accordingly. However, these procedures have since been deprecated from standards status, as the methods are inadequate to correctly assess many devices. New methods have been standardized in RFC 5389 (October 2008) and the STUN acronym now represents the new title of the specification: Session Traversal Utilities for NAT.