Symmetrical multiprocessing in multiprocessor systems

• US 7,219,121 B2
• Filed: 03/29/2002
• Issued: 05/15/2007
• Est. Priority Date: 03/29/2002
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Sethi, Bhupinder S., Dabagh, Alireza, Kaniyar, Sanjay N., Murching, Arvind, Srinivas, N K
• Primary Examiner(s): Vu; Thong
• Assistant Examiner(s): Nguyen; Hai V.
• Application Number:
  US10/112,812
• Publication Number:
  US 20030187914A1
• Time in Patent Office: 1,873 Days
• Field of Search: 709/201, 709/250, 370/235, 707/10
• US Class Current: 709/201
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/10015 Access to distributed or replicated servers, e.g. using brokers

Abstract

A new method and framework for scheduling receive-side processing of data streams received from a remote requesting client by a multiprocessor system computer is disclosed. The method receives data packets from the remote requesting client via a network and, for each data packet, applies a mapping algorithm to portions of the received data packet yielding a mapping value. The method further applies the map value to a processor selection policy to identify a processor in the multiprocessor system as a selected processor to perform receive-side processing of the data packet. The method queues the received data packet for processing by the selected processor and invokes a procedure call to initiate processing of the data packet.

Claims

1. A method for scheduling receive-side processing of data packets in a multiprocessor computer system comprising a network interface card communicably coupled to a network to receive data streams, memory and at least two local processors, the method comprising the steps of:

receiving a data packet from a remote requesting client;

storing the data packet in memory;

first applying, by a scheduling processor of the at least two local processors, a mapping algorithm to the data packet, the mapping algorithm yielding a map value;

second applying the map value to a processor selection policy to identify the local processor as a selected processor to perform receive-side processing of the data packet;

directing the received data packet to the selected processor; and

processing the data packet,

wherein the scheduling processor and the network interface card are separate and distinct components of the computer system.

2. The method of claim 1 wherein the receiving step comprises receiving, by the network interface card, the data packet from the network, and wherein the method further comprises:

storing the data packet and a data packet descriptor in memory.

3. The method of claim 2 wherein the receiving step further comprises selecting, by the network interface card, the scheduling processor and sending a message to the scheduling processor requesting handling of the data packet.

4. The method of claim 3 wherein the selecting step comprises utilizing, by the network interface card, a load-balancing algorithm to select the scheduling processor.

5. The method of claim 4 wherein the load-balancing algorithm determines the least active processor in the multiprocessor system.

6. The method of claim 3 wherein the first applying step comprises reading, by the scheduling processor, the data packet descriptor and the data packet from memory and applying the mapping algorithm to the data packet yielding the map value.

7. The method of claim 6 wherein the second applying step comprises utilizing, by the scheduling processor, a map table to resolve the selected processor.

8. The method of claim 6 wherein the directing step comprises:

queuing, by the scheduling processor, the data packet to run on the selected processor; and

sending, by the scheduling processor, a message to the selected processor requesting processing of the data packet.

9. The method of claim 8 wherein the queuing step comprises building a Network Driver Interface Specification data packet and placing the Network Driver Interface Specification data packet in a receive queue belonging to the selected processor.

10. The method of claim 8 wherein the processing step comprises:

processing, by the selected processor, the data packet queued to the selected processor.

11. The method of claim 1 wherein the first applying step comprises:

storing, by the network interface card, the data packet in memory.

12. The method of claim 11 wherein the second applying step comprises:

utilizing, by the scheduling processor, a map table to resolve the selected processor; and

storing a data packet descriptor including the map value in memory.

13. The method of claim 12 wherein the directing step comprises:

selecting, by the network interface card, a scheduling processor; and

sending, by the network interface card, a message to the scheduling processor requesting handling of the data packet.

14. The method of claim 13 wherein the selecting step comprises utilizing, by the network interface card, a load-balancing algorithm to select the scheduling processor.

15. The method of claim 13 wherein the directing step comprises:

reading, by the scheduling processor, the data packet descriptor from memory;

queuing, by the scheduling processor, the data packet to run on the selected processor; and

sending, by the scheduling processor, a message to the selected processor requesting processing of the data packet.

16. The method of claim 15 wherein the reading step comprises reading the map value from the data packet descriptor.

17. The method of claim 15 wherein the queuing step comprises building a Network Driver Interface Specification data packet and placing the Network Driver Interface Specification data packet in a receive queue belonging to the selected processor.

18. The method of claim 15 wherein the processing step comprises processing, by the selected processor, the data packet queued to the selected processor.

19. The method of claim 1 further comprising:

establishing a network connection between a remote requesting client and the multiprocessor computer system,

wherein the establishing step comprises identifying a type of network connection initiated by the remote requesting client.

20. The method of claim 19 wherein the network connection is a TCP connection.

21. The method of claim 1 wherein the mapping algorithm is a hashing function.

22. The method of claim 21 wherein the map value is a hash value.

23. A computer-readable medium having computer-executable instructions for facilitating scheduling receive-side processing of data packets in a multiprocessor system computer comprising a network interface card communicably coupled to a network to receive data streams, memory and at least two local processors, the computer-readable medium having computer-executable instructions facilitating performing the steps of:

receiving a data packet from the network;

first applying a mapping algorithm to the data packet, the mapping algorithm yielding a map value;

second applying, by a scheduling processor of the at least two local processors, the map value to a processor selection policy to identify the local processor as a selected processor to perform receive-side processing of the data packet;

directing the received data packet to the selected processor; and

processing the data packet,

wherein the scheduling processor and the network interface card are separate and distinct components of the computer system.

24. A method for scheduling receive-side processing of data packets in a multiprocessor system computer comprising a network interface card communicably coupled to a network to receive data streams, memory and at least two local processors, the method comprising the steps of:

issuing, by the network interface card, a select local processor request call having a plurality of call parameters comprising a data packet header and a local processor selection policy; and

receiving, by the network interface card, a select local processor request response having a plurality of response parameters comprising a local processor ID,

wherein the one of the at least two local processors and the network interface card are separate and distinct components of the computer system.

25. The method of claim 24 wherein the step of issuing the select local processor request call the call parameters further comprise a hashing function.

26. The method of scheduling receive-side processing of data packets in a multiprocessor computer system comprising a network interface card communicably coupled to a network to receive data streams, memory and at least two local processors, the method comprising the steps of:

receiving a data packet from the network;

selecting, by one of the at least two local processors, one of the at least two local processors to perform receive-side processing of the data packet; and

directing the received data packet to the selected processors,

wherein the one of the at least two local processors and the network interface card are separate and distinct components of the computer system.

27. The method of claim 26, wherein the selecting includes supplying a mapping algorithm to the data packet to yield a map value.

28. The method of claim 27, wherein the selecting includes applying the map value to a processor selection policy to identify the local processor as the selected processor to perform receive-side processing of the data packet.