Benchmarking the Mapping of Address and Port With Encapsulation Border Relay Routers

Mapping of Address and Port with Encapsulation (MAP-E) is considered one of the most important IPv6 transition technologies, providing a high level of scalability thanks to its utterly stateless behavior in the operator network. The primary device executing this technology is the Border Relay (BR), which represents the pivot of its scalability. Thus, benchmarking the performance and scalability of this device using a solution based on existing accredited standards is crucial for service providers before its deployment in their core networks. The Benchmarking Working Group (BMWG) of the Internet Engineering Task Force (IETF) has published a comprehensive guide for such a task in its Request for Comments (RFC) 8219. The research work presented in this paper involves conducting comprehensive benchmarking for the MAP-E BR implementation of the FD.io Vector Packet Processing (VPP), a well-known high-performance open-source networking software, and analyzing its performance and scalability, using two different test systems. To accomplish the research, a MAP-E encapsulation capability has been integrated into “Maptperf,” an RFC 8219 compliant research Tester developed originally to benchmark the Mapping of Address and Port using Translation (MAP-T) BR. The results of benchmarking the MAP-E BR of VPP demonstrated the functionality and efficiency of the Tester encapsulation plugin as well as the high scalability of the tested implementation, even when serving an increasing number of Customer Edge (CE) devices or deploying a varying number of worker threads in the test. Moreover, they showed that the tested implementation achieved high performance levels in throughput, Frame Loss Rate (FLR), latency, and Packet Delay Variation (PDV) tests, despite utilizing a low number of worker threads.