METHOD AND APPARATUS FOR DETERMINING BANDWIDTH-CONSUMING FRAME FLOWS IN A NETWORK

• US 20100202319A1
• Filed: 04/21/2010
• Published: 08/12/2010
• Est. Priority Date: 07/25/2007
• Status: Active Grant

• Current Assignee: Avago Technologies International Sales Pte Limited (Broadcom, Inc.)
• Original Assignee: Brocade Communications Systems, Inc. (Broadcom, Inc.)
• Inventors: Kanda, Amit, Gnanasekaran, Sathish Kumar
• Granted Patent: US 8,582,432 B2
• Time in Patent Office: Days
• US Class Current: 370/253
• CPC Class Codes: H04L 43/026 using flow identification
• CPC Class Codes: H04L 43/0894 Packet rate
• CPC Class Codes: Y02D 30/50 in wire-line communication networks, e.g. low power modes or reduced link rate

Abstract

A port monitor monitors network traffic that passes through a port of a switch. Frame flows routed through the switch are counted by hardware frame counting logic, which includes a content-addressable memory (CAM) and counters corresponding to various frame flows. Port monitor software includes logical containers (“buckets”), which are used to record and sort counts of the frame flows from hardware based on activity of the frame flows. Frame flow sorting is based on confidence building algorithms such that the thrashing of frame flows between buckets is reduced. Storage and sorting of the frame flows in software allows the activity level of any number of frame flows associated with the switch to be counted, regardless of hardware limitations on how many frame flows can be counted. This allows a real-time analysis of frame flows and a determination of frame flows that are major users of Fibre Channel bandwidth.

Claims

1. A method for determining activity in a network, the method comprising:

counting every frame of a first set of a plurality of flows and a portion of a second set of said plurality of flows over a time period;

selecting flows for said first set and said second set; and

changing said portion of said second set counted every frame so that each flow in said second set is periodically counted every frame over a time period.

2. The method of claim 1, wherein said first set of flows has higher flow rates than said second set of flows.

3. The method of claim 2, wherein said selecting is based on flow rates of frames.

4. A computer readable medium storing instructions, the instructions when executed causing a processor to:

count every frame of a first set of a plurality of flows and a portion of a second set of said plurality of flows over a time period;

select flows for said first set and said second set; and

change said portion of said second set counted every frame so that each flow in said second set is periodically counted every frame over a time period.

5. The computer readable medium of claim 4, wherein said first set of flows has higher flow rates than said second set of flows.

6. The computer readable medium of claim 5, wherein said selecting is based on flow rates of frames.

7. A network device comprising:

a processor;

memory operatively connected to said processor to store software programs and data;

detection logic configured to detect selected values in portions of each frame in each of a plurality of frame flows;

a plurality of hardware counters operatively coupled to said processor and operatively connected to said detection logic and configured to count every frame of selected frame flows; and

a software program stored in said memory to cause said processor to;

configure said plurality of hardware counters to count every frame of a first set of said plurality of flows and a portion of a second set of said plurality of flows over a time period;

select flows for said first set and said second set; and

configure said plurality of hardware counters with said portion of said second set counted every frame so that each flow in said second set is periodically counted every frame over a time period.

8. The network device of claim 7, wherein said first set of flows has higher flow rates than said second set of flows.

9. The network device of claim 8, wherein said selecting is based on flow rates of frames.

10. A network device comprising:

a processor;

memory operatively connected to said processor to store software programs and data;

detection logic configured to detect selected values in portions of each frame in each of a plurality of frame flows;

a plurality of hardware counters operatively coupled to said processor and operatively connected to said detection logic and configured to count frame flow activity; and

a software program stored in said memory to cause said processor to;

logically partition said plurality of frame flows into a first set and a second set based on frame flow activity, each frame flow having an associated count value,

wherein said first set has associated an equivalent number of hardware counters for counting frame flow activity as there are frame flows in said first set, and

wherein said second set has associated a smaller number of hardware counters for counting frame flow activity than there are frame flows in said second set;

periodically determine frame flow activity for frame flows having associated hardware counters and update the associated count values for said frame flows;

assign frame flows with the highest count values to the first set and configure the hardware counters as required; and

assign the remaining frame flows to said second set, change which frame flows of said second set have associated hardware counters and configure the hardware counters as required so that all frame flows in said second set are periodically associated with hardware counters to determine count values.

11. The network device of claim 10, wherein the software program is further configured to:

partition said plurality of frame flows into a third set, wherein said third set has associated an equivalent number of hardware counters for counting frame flow activity as there are frame flows in said third set such that all of the frame flows in said third set are counted by hardware counters, and

assign frame flows comprising lower count values than the frame flows in said first set but higher count values than the frame flows in said second set to said third set and configure the hardware counters as required.

12. The network device of claim 11, wherein said third set comprises a larger number of frame flows than said first set and a smaller number of frame flows than said second set.

13. The network device of claim 10, wherein the first set is smaller than the second set.

14. The network device of claim 10, wherein the network device is a Fibre Channel network device.

15. The network device of claim 10, wherein said detection logic includes a content-addressable memory configured to associate each frame flow having an associated hardware counter to the associated hardware counter.

16. The network device of claim 10, wherein frame flow activity of the first set is determined more frequently than the second set.

17. A method of determining activity in a network, comprising:

determining a plurality of frame flows to be monitored;

partitioning said plurality of frame flows into a first set and a second set, each frame flow having an associated count value,

wherein said first set has associated an equivalent number of hardware counters for counting frame flow activity as there are frame flows in said first set, and

wherein said second set has associated a smaller number of hardware counters for counting frame flow activity than there are frame flows in said second set;

periodically determining frame flow activity for the frame flows having associated hardware counters and updating the associated count values for said frame flows;

assigning frame flows with the highest count values to said first set and configuring the hardware counters as required; and

assigning the remaining frame flows to said second set, changing which frame flows of said second set have associated hardware counters, and configuring the hardware counters as required so that all frame flows in said second set are periodically associated with hardware counters to determine count values.

18. The method of claim 17, further comprising:

partitioning said plurality of frame flows into a third set, wherein said third set has associated an equivalent number of hardware counters for counting frame flow activity as there are frame flows in said third set, and

assigning frame flows comprising lower count values than the frame flows in said first set but higher count values than the frame flows in said second set to said third set and configuring the hardware counters as required.

19. The method of claim 18, wherein said third set comprises a larger number of frame flows than said first set and a smaller number of frame flows than said second set.

20. The method of claim 17, wherein the first set is smaller than the second set.

21. The method of claim 17, wherein the network is a Fibre Channel network.

22. The method of claim 17, wherein a content-addressable memory associates each frame flow having an associated hardware counter to the associated hardware counter.

23. The method of claim 17, wherein frame flow activity of the first set is determined more frequently than the second set.

24. A computer readable medium storing instructions, the instructions when executed causing a processor to:

determine a plurality of frame flows to be monitored;

logically partition said plurality of frame flows into a first set and a second set, each frame flow having an associated count value,

wherein said first set has associated an equivalent number of hardware counters for counting frame flow activity as there are frame flows in said first set, and

wherein said second set has associated a smaller number of hardware counters for counting frame flow activity than there are frame flows in said second set;

periodically determine frame flow activity for the frame flows having associated hardware counters and updating the associated count values for said frame flows;

assign frame flows with the highest count values to said first set and configure the hardware counters as required; and

assign the remaining frame flows to said second set and change which frame flows of said second set have associated hardware counters so that all frame flows in said second set are periodically associated with hardware counters to update count values.

25. The computer readable medium of claim 24, further storing instructions that when executed cause a processor to:

partition said plurality of frame flows into a third set, wherein said third set has associated an equivalent number of hardware counters for counting frame flow activity as there are frame flows in said third set, and

assign frame flows comprising lower count values than the frame flows in said first set but higher count values than the frame flows in said second set to said third set and configure the hardware counters as required.

26. The computer readable medium of claim 24, wherein said third set comprises a larger number of frame flows than said first set and a smaller number of frame flows than said second set.

27. The computer readable medium of claim 26, wherein the first set is smaller than the second set.

28. The computer readable medium of claim 24, wherein the network is a Fibre Channel network.

29. The computer readable medium of claim 24, wherein a content-addressable memory associates each frame flow having an associated hardware counter to the associated hardware counter.

30. The computer readable medium of claim 24, wherein frame flow activity of the first set is determined more frequently than the second set.