SYSTEM AND METHOD FOR MAXIMIZING PORT BANDWIDTH WITH MULTI-CHANNEL DATA PATHS

• US 20190327178A1
• Filed: 04/18/2018
• Published: 10/24/2019
• Est. Priority Date: 04/18/2018
• Status: Active Grant

• Current Assignee: Avago Technologies International Sales Pte Limited (Broadcom, Inc.)
• Original Assignee: Avago Technologies General IP PTE Limited (Broadcom, Inc.)
• Inventors: Wu, Kenny, Gervasio, JR., Gregorio, Chhabra, Lalit, Shenoy, Ravi
• Granted Patent: US 10,735,340 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 13/385 for adaptation of a particular data processing system to different peripheral devices
• CPC Class Codes: G06F 13/4018 with data-width conversion
• CPC Class Codes: G06F 13/4022 using switching circuits, e.g. switching matrix, connection or expansion network G06F13/4009 takes precedence
• CPC Class Codes: H04L 47/41 by acting on aggregated flows or links
• CPC Class Codes: H04L 47/6225 Fixed service order, e.g. Round Robin
• CPC Class Codes: H04L 47/6255 queue load conditions, e.g. longest queue first
• CPC Class Codes: H04L 49/3027 Output queuing
• CPC Class Codes: H04L 49/3045 Virtual queuing
• CPC Class Codes: H04L 49/90 Buffering arrangements

Abstract

A networking adaptor and method of transferring data are depicted and described herein. One example of the networking adaptor is provided with a host interface and a network interface. The network interface may include a transmit portion and a receive portion. The transmit portion may include a first set of data paths and the receive portion may include a second set of data paths. Both the first set of data paths and second set of data paths are configurable to be aggregated or de-aggregated to support a single port operation that represents a combined bandwidth of the data paths in the first set of data paths or the second set of data paths.

Claims

1. A networking adaptor, comprising:

a host interface; and

a network interface, wherein the network interface includes a transmit portion and a receive portion, wherein the transmit portion comprises a first set of data paths and wherein the receive portion comprises a second set of data paths, wherein both the first set of data paths and second set of data paths are configurable to be aggregated or de-aggregated to support a single port operation that represents a combined bandwidth of the data paths in the first set of data paths or the second set of data paths.

2. The networking adaptor of claim 1, wherein the transmit portion comprises a multiplexer that connects the data paths in the first set of data paths to a number of different data ports that belong to the network interface.

3. The networking adaptor of claim 2, wherein the multiplexer facilitates the aggregation of the data paths in the first set of data paths.

4. The networking adaptor of claim 3, wherein the multiplexer receives inputs from a set of transmission buffers, wherein the set of transmission buffers couple the first set of data paths to the multiplexer.

5. The networking adaptor of claim 4, wherein the multiplexer aggregates the first set of data paths for a command based on an input key included in the command, wherein the input key comprises at least one of a work queue identifier and an exchange identifier along with a modulo operation that is applied to the input key.

6. The networking adaptor of claim 4, wherein the multiplexer further comprises an egress arbiter that is used to select and multiplex one of the data paths from the first set of data paths for frame transmission when a data path in the first set of data paths has at least one frame available to transmit in a round-robin fashion between the first set of data paths.

7. The networking adaptor of claim 1, wherein the receive portion comprises a de-multiplexer that de-aggregates and distributes frames between the second set of data paths.

8. The networking adaptor of claim 7, wherein the de-multiplexer assigns and marks a serial number onto a beginning of each frame before the frame is de-multiplexed into one of the second set of data paths.

9. The networking adaptor of claim 8, wherein the de-multiplexer provides each frame to a set of receive buffers, wherein the set of receive buffers couples the second set of data paths to the de-multiplexer.

10. The networking adaptor of claim 9, further comprising:

a buffer manager that uses the serial number to reorder frames prior to providing the frames to the host interface.

11. A method of transferring commands from a host to a network, the method comprising:

receiving a command at a host interface;

distributing the command among a first set of data paths, wherein each data path in the first set of data paths have substantially a first bandwidth;

multiplexing the command onto one network port when the command requires more than the first bandwidth; and

transmitting the command via the one network port.

12. The method of claim 11, wherein the one network port comprises a bandwidth that is a product of the first bandwidth and a number of data paths in the first set of data paths.

13. The method of claim 11, further comprising:

placing the distributed command onto a plurality of transmission buffers to ensure that the one network port can deliver a frame at the bandwidth provided by the one network port.

14. A method of transferring commands from a network to a host, the method comprising:

receiving a frame at a network port;

utilizing a de-multiplexer to assign and mark a serial number onto a beginning of the frame;

de-multiplexing the frame with the de-multiplexer to create a de-multiplexed frame;

distributing the de-multiplexed frame among a set of receive data paths, wherein each data path in the set of data paths comprises a bandwidth that is less than a bandwidth of the network port;

re-ordering the de-multiplexed frame at an egress of the set of receive data paths; and

providing the re-ordered frame to a host interface.

15. The method of claim 14, further comprising:

providing the de-multiplexed frame to a set of receive buffers.

16. The method of claim 14, further comprising:

distributing frames to multiple virtual output queues.

17. The method of claim 16, wherein an input key is used for the distribution of the frames to the multiple virtual output queues.

18. The method of claim 17, wherein the input key comprises at least one of a work queue identifier and an exchange identifier.

19. The method of claim 18, wherein the input key is obtained from a frame header or a table lookup.

20. The method of claim 14, wherein the host interface comprises a PCIe interface and wherein the network port comprises a plurality of Fibre Channel (FC) ports.