Technologies for managing cache quality of service

One or more non-transitory machine-readable storage media comprising a plurality of instructions stored thereon that, in response to being executed, cause a network interface controller (NIC) at a compute node to:

• identify an amount of available cache ways of a last level cache (LLC) of a processor at the compute node;
• determine a workload type associated with each of a plurality of virtual machines (VMs) managed by the compute node based on received network traffic to be processed by each of the plurality of VMs;
• identify an amount of LLC cache ways for each workload type;
• and
• update a at least one register of the NIC to indicate the identified amount of LLC cache ways for each workload type.

A method comprising:

• identifying, by a network interface controller (NIC) at a compute node, an amount of available cache ways of a last level cache (LLC) of a processor at the compute node;
• determining, by the NIC, a workload type associated with each of a plurality of virtual machines (VMs) managed by the compute node based on received network traffic to be processed by each of the plurality of VMs;
• identifying, by the NIC, an amount of LLC cache ways for each workload type;
• and
• updating, by the NIC, at least one register of the NIC to indicate the identified amount of LLC cache ways for each workload type.

A network interface controller (NIC) at a compute node, the NIC comprising:

• an input/output (1/0) interface to couple with a last level cache (LLC) of a processor at the compute node;
• and
• circuitry to:
• identify an amount of LLC cache ways of the processor, the identified amount of LLC cache ways for use to process packets, wherein the LLC cache ways are accessible via the I/O interface;
• and
• indicate the identified amount of LLC cache ways to cause an allocation of LLC cache ways for use to process packets.