Disclaimer: I work on VPP. The typical usecase are virtual network functions: think virtual switches/routers used to interconnect VMs or containers, or contain

Disclaimer: I work on VPP. The typical usecase are virtual network functions: think virtual switches/routers used to interconnect VMs or containers, or containerized VPN gateways etc. It is also used for high-performance L3-L4 load-balancers etc. As pointed out by others, what is hard is to move small packets. TCP with iperf is not relevant for this kind of workloads. It is easy to max out 100GbE with 1500-bytes packets, but with 200-bytes packets not so much. This is why they communicate about PPS, not bandwidth. There results seems low but it is hard to tell without knowing the platform or configuration. VPP can sustain 20+Mpps / core (2 hyperthreads) on Skylake @2.5GHz (no turboboost).