An agent-to-agent test permits ThousandEyes users to monitor two endpoints and visualize the network path in both directions, which can mean faster, more accurate identification of problems than with an agent-to-server test. However, unlike an agent-to-server test, where the test target is typically intended to receive network traffic initiated by other hosts, in an agent-to-agent test often one or both agents cannot receive network traffic initiated by other hosts, either from the public Internet or from other network locations. Changes to network address translation (NAT), port address translation (PAT) and/or stateful packet filter (SPF, i.e. firewall) rules would normally be required to allow the test traffic.