Equivalence and difference of the dual device under test setup and the single device under test setup of RFC 8219

Summary RFC 8219 has defined a comprehensive benchmarking methodology for the IPv6 transition technologies. It recommends two kinds of measurement setups: The dual device under test (DUT) setup facilitates the benchmarking of the customer edge (CE) and provider edge (PE) devices together using a legacy RFC 2544 or RFC 5180 Network Performance Tester, whereas the single DUT setup requires a separate technology‐specific tester for the benchmarking of each device. As such, special‐purpose testers do not exist for the vast majority of the IPv6 transition technologies; the only viable solution can be the usage of the dual DUT setup. In this paper, we investigate if the two kinds of measurement setups provide the same or different results; moreover, we examine how the single DUT measurement results can be estimated from the dual DUT measurement results. To that end, we make theoretical considerations and also perform IPv4 packet forwarding and stateless IP/ICMP translation (SIIT) measurements using both measurement setups and analyze the results of the throughput and latency measurements. It was found that the throughput results of the dual DUT setup could approximate well those of the single DUT setup and their differences followed the predictions of our theoretical considerations. However, the latency results did not always follow the theoretical expectations.