1. Proposed title of this feature request

Linux has RPS and RFS to distribute packet receiving process.

• https://docs.kernel.org/networking/scaling.html
  ===> RPS: Receive Packet Steering
  ===> RFS: Receive Flow Steering

OCP documentation has the following section to enable RFS:

• 9.3.1. Use the Machine Config Operator (MCO) to activate RFS
  https://docs.redhat.com/en/documentation/openshift_container_platform/4.18/html/scalability_and_performance/ibm-z-recommended-host-practices#use-the-mco-to-activate-rfs_ibm-z-recommended-host-practices

However, there is no OCP document to enable RPS.
Could we enhance our documentation to enable RPS?

2. What is the nature and description of the request?

This is a request to enhance OCP documentation.

3. Why does the customer need this? (List the business requirements here)

My customer is hitting packet receiving performance issue.
Linux has some featires to distribute the packet receiving process to multiple CPU cores: RSS, RFS and RPS.
In the customer's case, RSS cannot be enabled due to the virtualization mechanism limitation, RFS didn't show enough performance, and RPS showed the best performance.
The customer has to enable RFS to keep good network performance for network latency/drop sensitive application pods.

However, OCP doesn't have any official procedure to enable RPS,

I'm aware that "PerformanceProfile" of Tuned Operator has "performance.openshift.io/enable-physical-dev-rps: true" annotation, but this doesn't match because PerformanceProfile requires CPU isolation.
In this case, we just want to distribute the packet receive process to all CPU cores without CPU isolation.

4. List any affected packages or components.

documetation