VNF workload placement for optimal performance

This section provides recommendations on where to place VNFs and how to tune platform to get optimal performance from workload.

NUMA (Non-uniform memory access) Locality:

  • vCPUs executing processes are on same NUMA node as the memory used by these processes ​

  • Memory access are local

  • Huge pages access are local. ​

  • PCI devices are directly associated with NUMA nodes for the purposes of DMA.

  • VNFs that use PCI or SR-IOV devices should be placed on the NUMA node associated with the PCIe devices.