NUMA-aware scaling for network devices

• US 9,069,722 B2
• Filed: 05/16/2013
• Issued: 06/30/2015
• Est. Priority Date: 12/17/2009
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Li, Yadong
• Primary Examiner(s): Phan, Tri H
• Application Number:
  US13/895,917
• Publication Number:
  US 20130326000A1
• Time in Patent Office: 775 Days
• Field of Search: 370229-240, 370412-429, 709212-216, 709230-245, 711117-125, 711147-153
• US Class Current: 1/1
• CPC Class Codes: G06F 15/17331 Distributed shared memory [DSM], e.g. remote direct memory access [RDMA]
• CPC Class Codes: H04L 45/00 Routing or path finding of packets in data switching networks routing or path finding in wireless networks H04W40/00
• CPC Class Codes: H04L 49/9036 Common buffer combined with individual queues

Abstract

The present disclosure describes a method and apparatus for network traffic processing in a non-uniform memory access architecture system. The method includes allocating a Tx/Rx Queue pair for a node, the Tx/Rx Queue pair allocated in a local memory of the node. The method further includes routing network traffic to the allocated Tx/Rx Queue pair. The method may include designating a core in the node for network traffic processing. Of course, many alternatives, variations and modifications are possible without departing from this embodiment.

Claims

1. A system comprising one or more storage mediums having stored thereon, individually or in combination, instructions that when executed by one or more processors cause the one or processors to:

automatically allocate a single respective receive queue per respective set of multiple sets of multiple processor cores, each single respective receive queue to enqueue respective entries for respective received network packets based on a hash of multiple header fields of the respective received network packets; and

for each respective set of the sets of multiple processor cores, cause designation of a single respective core of the multiple processor cores to receive interrupts and cause an interrupt vector to be set with affinity to the single respective core.

2. The system of claim 1, wherein at least one set of the multiple sets of multiple processor cores comprises a NUMA (Non-Uniform Memory Access) node.

4. The system of claim 2, wherein at least one set of the multiple sets of multiple processor cores comprises a NUMA (Non-Uniform Memory Access) node.

3. A system, comprising:

multiple sets of multiple processor cores; and

a network adapter; and

a network adapter device driver comprising one or more storage mediums having stored thereon, individually or in combination, instructions that when executed by one or more processors cause the one or more processors to;

automatically allocate a single respective receive queue per respective set of multiple sets of multiple processor cores, each single respective receive queue to enqueue respective entries for respective received network packets based on a hash of multiple header fields of the respective received network packets; and

for each respective set of the sets of multiple processor cores, cause designation of a single respective core of the multiple processor cores to receive interrupts and cause an interrupt vector to be set with affinity to the single respective core.

5. The system of claim 3, wherein the network adapter comprises a DMA (Direct Memory Access) engine to write packet data to a memory and wherein the multiple header fields comprise an Internet Protocol source address, Internet Protocol destination address, a Transmission Control Protocol source port, and a Transmission Control Protocol destination port of each respective network packet.

6. A system comprising one or more storage mediums having stored thereon, individually or in combination, instructions that when executed by one or more processors cause the one or more processors to:

determine a same number of receive queues per respective set of multiple sets of multiple processor cores, each respective receive queue to enqueue respective entries for respective received network packets based on a hash of multiple network packet header fields of the respective received network packets; and

allocate the determined same number of receive queues per respective set of multiple processor cores.

7. The system of claim 6, wherein at least one set of the multiple sets of multiple processor cores comprises a NUMA (Non-Uniform Memory Access) node.

8. The system of claim 6, wherein the number of receive queues consists of a single respective receive queue per set of multiple processor cores.

9. The system of claim 6, wherein the instructions that when executed by one or more processors cause the one or more processors to determine a same number of receive queues per respective set of multiple sets of multiple processor cores comprise instructions that when executed by one or more processors cause the one or more processors to determine the same number of receive queues per respective set of multiple sets of multiple processor cores based on system processor architecture.

10. The system of claim 6, wherein the instructions that when executed by one or more processors cause the one or more processors to determine a same number of receive queues per respective set of multiple sets of multiple processor cores comprise instructions that when executed by one or more processors cause the one or more processors to determine the same number of receive queues per respective set of multiple sets of multiple processor cores based on a speed of a communication link.

11. The system of claim 6, wherein the number of receive queues consists of two receive queues per set of multiple processor cores.

12. A system, comprising:

multiple sets of multiple processor cores; and

a network adapter; and

a network adapter device driver comprising one or more storage mediums having stored thereon, individually or in combination, instructions that when executed by one or more processor cause the one or more processors to;

determine a same number of receive queues per respective set of multiple sets of multiple processor cores, each respective receive queue to enqueue respective entries for respective received network packets based on a hash of multiple network packet header fields of the respective received network packets; and

allocate the determined same number of receive queues per respective set of multiple processor cores.

13. The system of claim 12, wherein at least one set of the multiple sets of multiple processor cores comprises a NUMA (Non-Uniform Memory Access) node.

14. The system of claim 12, wherein the number of receive queues consists of a single respective receive queue per set of multiple processor cores.

15. The system of claim 12, wherein the instructions that when executed by one or more processors cause the one or more processors to determine a same number of receive queues per respective set of multiple sets of multiple processor cores comprise instructions that when executed by one or more processors cause the one or more processors to determine the same number of receive queues per respective set of multiple sets of multiple processor cores based on system processor architecture.

16. The system of claim 12, wherein the instructions that when executed by one or more processors cause the one or more processors to determine a same number of receive queues per respective set of multiple sets of multiple processor cores comprise instructions that when executed by one or more processors cause the one or more processors to determine the same number of receive queues per respective set of multiple sets of multiple processor cores based on a speed of a communication link.

17. The system of claim 12, wherein the number of receive queues consists of two receive queues per set of multiple processor cores.