Method and apparatus for accelerating vm-to-vm network traffic using cpu cache

A processor, configured to be installed on a host platform including system memory, the processor comprising:

• a plurality of cores, each having an associated level 1 (L2) cache and a level 2 (L2) cache;
• a last level cache (LLC), communicatively coupled to the plurality of cores;
• a memory controller, communicatively coupled to the plurality of cores, configured to support access to the system memory when the processor is installed in the computer system;
• and
• a virtual queue manager (VQM), communicatively coupled to the LLC;
• wherein each of the L1 caches, L2 caches, and LLC include a plurality of cacheline slots for storing cacheline data and an associated cache agent, wherein during operation the host platform is configured to execute, via the processor, software including a hypervisor and a plurality of VMs, execution of the software further implementing a virtual switch configured to forward packetized traffic between VMs, and wherein the processor is further configured, during operation of the host platform, to:
• receive a packet at the VQM, the packet generated by a first VM and destined for a second VM;
• inspect the packet to determine whether the packet should be forwarded via a VM-to-VM cache fast path;
• and
• if the packet is to be forwarded via the VM-to-VM cache fast path, write packet data for the packet from the VQM to one or more cachelines in a VM-VM shared data buffer in the LLC allocated for cache fast path forwarding between the first VM and the second VM.