Network device architecture for consolidating input/output and reducing latency

• US 20060087989A1
• Filed: 03/30/2005
• Published: 04/27/2006
• Est. Priority Date: 10/22/2004
• Status: Active Grant

• Current Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Original Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Inventors: Edsall, Thomas, Bergamasco, Davide, Gai, Silvano, Bonomi, Flavio, Dutt, Dinesh
• Granted Patent: US 7,830,793 B2
• Time in Patent Office: Days
• US Class Current: 370/299
• CPC Class Codes: H04L 12/413 with random access, e.g. carrier-sense multiple-access with collision detection [CSMA-CD]
• CPC Class Codes: H04L 47/10 Flow control; Congestion control
• CPC Class Codes: H04L 47/32 by discarding or delaying data units, e.g. packets or frames
• CPC Class Codes: H04L 47/33 using forward notification
• CPC Class Codes: H04L 47/39 Credit based
• CPC Class Codes: H04L 49/90 Buffering arrangements

Abstract

The present invention provides methods and devices for implementing a Low Latency Ethernet (“LLE”) solution, also referred to herein as a Data Center Ethernet (“DCE”) solution, which simplifies the connectivity of data centers and provides a high bandwidth, low latency network for carrying Ethernet and storage traffic. Some aspects of the invention involve transforming FC frames into a format suitable for transport on an Ethernet. Some preferred implementations of the invention implement multiple virtual lanes (“VLs”) in a single physical connection of a data center or similar network. Some VLs are “drop” VLs, with Ethernet-like behavior, and others are “no-drop” lanes with FC-like behavior. Some preferred implementations of the invention provide guaranteed bandwidth based on credits and VL. Active buffer management allows for both high reliability and low latency while using small frame buffers. Preferably, the rules for active buffer management are different for drop and no drop VLs.

Claims

1. A method for processing more than one type of network traffic in a single network device, the method comprising:

receiving first frames and second frames on physical links that are logically partitioned into a plurality of virtual lanes;

partitioning buffers of the network device into first buffer spaces for first frames received on a first virtual lane and second buffer spaces for second frames received on a second virtual lane;

storing the first frames in the first buffer spaces;

storing the second frames in the second buffer spaces;

applying a first set of rules to the first frames; and

applying a second set of rules to the second frames.

2. The method of claim 1, further comprising the step of assigning a guaranteed minimum amount of buffer space per virtual lane.

3. The method of claim 1, wherein the first frames comprise Ethernet frames.

4. The method of claim 1, wherein the first set of rules causes first frames to be dropped in response to latency.

5. The method of claim 1, wherein the second set of rules does not cause second frames to be dropped in response to latency.

6. The method of claim 1, wherein the first set of rules causes an explicit congestion notification to be transmitted from the network device in response to latency of second frames.

7. The method of claim 1, wherein the second set of rules causes an explicit congestion notification to be transmitted from the network device in response to latency of second frames.

8. The method of claim 1, wherein the partitioning step comprises dynamically partitioning buffers according to one or more factors from the group of factors consisting of buffer occupancy, time of day, traffic loads, congestion, guaranteed minimum bandwidth allocation, known tasks requiring greater bandwidth and maximum bandwidth allocation.

9. The method of claim 1, wherein the second set of rules causes frames to be dropped to avoid deadlocks.

10. The method of claim 1, wherein the first set of rules applies a probabilistic drop function in response to latency.

11. The method of claim 1, wherein the second set of rules comprises implementing a buffer-to-buffer crediting scheme for the second frames.

12. The method of claim 1, further comprising the step of storing the first frames and the second frames in virtual output queues (“

VOQs”

), wherein each VOQ is associated with a destination port/virtual lane pair.

13. The method of claim 1, further comprising the step of performing buffer management in response to at least one of buffer occupancy per virtual lane and packet age.

14. The method of claim 6, wherein the explicit congestion notification is sent to one of a source device or an edge device.

15. The method of claim 6, wherein the explicit congestion notification is sent via one of a data frame or a control frame.

16. The method of claim 11, wherein the buffer-to-buffer crediting scheme comprises crediting according to frame size.

17. The method of claim 11, wherein buffer-to-buffer credits are managed by an arbiter.

18. The method of claim 11, wherein the second set of rules comprises implementing the buffer-to-buffer crediting scheme within a network device and using PAUSE frames for flow control on the second virtual lane.

19. The method of claim 12, further comprising the step of performing buffer management in response to at least one of VOQ length, buffer occupancy per virtual lane, overall buffer occupancy and packet age.

20. A network device, comprising:

means for receiving first frames and second frames on physical links that are logically partitioned into a plurality of virtual lanes;

means for partitioning buffers of the network device into first buffer spaces for first frames received on a first virtual lane and second buffer spaces for second frames received on a second virtual lane;

means for storing the first frames in the first buffer spaces;

means for storing the second frames in the second buffer spaces;

means for applying a first set of rules to the first frames; and

means for applying a second set of rules to the second frames.

21. A network device, comprising:

a plurality of ports configured for receiving frames on a plurality of physical links;

a plurality of line cards, each line card in communication with one of the plurality of ports and configured to do the following;

receive frames from one of the plurality of ports;

identify first frames received on a first virtual lane and second frames received on a second virtual lane;

partition buffers into first buffer spaces for the first frames and second buffer spaces for the second frames;

store the first frames in the first buffer spaces;

store the second frames in the second buffer spaces;

apply a first set of rules to the first frames; and

apply a second set of rules to the second frames.

22. A method for carrying more than one type of traffic in a single network device, the method comprising:

identifying first frames received on first virtual lanes and second frames received on second virtual lanes;

dynamically partitioning buffers of the network device into first buffer spaces for the first frames and second buffer spaces for the second frames, wherein the buffers are dynamically partitioned according to one or more factors from the group of factors consisting of overall buffer occupancy, buffer occupancy per virtual lane, time of day, traffic loads, congestion, guaranteed minimum bandwidth allocation, known tasks requiring greater bandwidth and maximum bandwidth allocation;

storing the first frames in first virtual output queues (“

VOQs”

) of the first buffer spaces;

storing the second frames in second VOQs of the second buffer spaces;

applying a first set of rules to the first frames; and

applying a second set of rules to the second frames.

23. A network device, comprising:

means for identifying first frames received on first virtual lanes and second frames received on second virtual lanes;

means for dynamically partitioning buffers of the network device into first buffer spaces for the first frames and second buffer spaces for the second frames, wherein the buffers are dynamically partitioned according to one or more factors from the group of factors consisting of overall buffer occupancy, buffer occupancy per virtual lane, time of day, traffic loads, congestion, guaranteed minimum bandwidth allocation, known tasks requiring greater bandwidth and maximum bandwidth allocation;

means for storing the first frames in first virtual output queues (“

VOQs”

) of the first buffer spaces;

means for storing the second frames in second VOQs of the second buffer spaces;

means for applying a first set of rules to the first frames; and

means for applying a second set of rules to the second frames.

24. A network device, comprising:

a plurality of ports configured for receiving frames on a plurality of physical links;

a plurality of line cards, each line card in communication with one of the plurality of ports and configured to do the following;

identify first frames received on first virtual lanes and second frames received on second virtual lanes;

dynamically partition buffers of the network device into first buffer spaces for the first frames and second buffer spaces for the second frames, wherein the buffers are dynamically partitioned according to one or more factors from the group of factors consisting of overall buffer occupancy, buffer occupancy per virtual lane, time of day, traffic loads, congestion, guaranteed minimum bandwidth allocation, known tasks requiring greater bandwidth and maximum bandwidth allocation;

store the first frames in first virtual output queues (“

VOQs”

) of the first buffer spaces;

store the second frames in second VOQs of the second buffer spaces;

apply a first set of rules to the first frames; and

apply a second set of rules to the second frames.