System for setting congestion avoidance flag at intermediate node to reduce rates of transmission on selected end systems which utilizing above their allocated fair shares

• US 5,675,742 A
• Filed: 06/26/1995
• Issued: 10/07/1997
• Est. Priority Date: 04/22/1988
• Status: Expired due to Term

• Current Assignee: Enterasys Networks Incorporated (Extreme Networks, Inc.)
• Original Assignee: Digital Equipment Corporation (HP Inc.)
• Inventors: Chiu, Dah-Ming, Ramakrishnan, K. K., Jain, Rajendra K.
• Primary Examiner(s): Lee, Thomas C.
• Assistant Examiner(s): Luu, Le Hien
• Application Number:
  US08/494,502
• Time in Patent Office: 834 Days
• Field of Search: 364/200, 364/554, 379/279, 395/200, 395/200.13, 395/200.11, 370/94, 370/60
• US Class Current: 709/226
• CPC Class Codes: H04L 47/10 Flow control; Congestion control
• CPC Class Codes: H04L 47/11 Identifying congestion
• CPC Class Codes: H04L 47/263 Rate modification at the source after receiving feedback
• CPC Class Codes: H04L 47/27 Evaluation or update of window size, e.g. using information derived from acknowledged [ACK] packets
• CPC Class Codes: H04L 47/33 using forward notification
• CPC Class Codes: Y02D 30/50 in wire-line communication networks, e.g. low power modes or reduced link rate

Abstract

Method and apparatus for operating a digital communication network to avoid congestion by detecting load conditions at intermediate stations exceeding an overload condition, and flagging information packets associated with those streams of traffic accounting for more than their fair share of throughput at such overloaded stations.

Claims

1. In a network of end systems communicating by means of transmission and reception of digital information packets routed through at least one intermediate system, a method for indicating when the intermediate system is approaching a state of congestion, said method being performed at the intermediate system and comprising the steps of:

A. determining whether the intermediate system is in an overload condition characterized by operation of the intermediate system above an optimal operating level but below a congestion level;

B. if the intermediate system is in the overload condition;

i. selecting the end systems that are transmitting information packets at rates that are above their respective allocated fair shares; and

ii. signaling one or more of the selected end systems by setting at the intermediate system a congestion avoidance flag in information packets routed through the intermediate system to one or more of the selected end systems.

2. The method of claim 1, wherein said intermediate system temporarily stores packets to create a queue having a queue length equal to the sum of the number of packets in the queue and the number of packets being serviced by the intermediate system, and wherein the overload condition is determined by determining, over a load averaging interval, an average queue length, that is greater than a first preselected queue length.

3. The method of claim 2, wherein operation of the intermediate system includes cycles of activity, each of which includes a busy period and an idle period, and wherein said load averaging interval includes at least one completed previous cycle of activity.

4. The method of claim 2, wherein the first preselected length is approximately equal to one.

5. The method of claim 1, wherein different streams of traffic comprising information packets are routed through said intermediate system, said method further comprising the step of:

A1. identifying those streams of traffic that are causing the overload condition; and

in step (B), setting a congestion avoidance flag only in an information packet that is a constituent of a stream of traffic identified as causing the overload condition.

6. The method of claim 5, wherein a stream of traffic is identified as causing the overload condition if its constituent information packets account for a throughput at said intermediate system, during a traffic measuring interval, that is greater than an allocated share of an estimated throughput capacity of the intermediate system.

7. The method of claim 6, wherein the throughput of the intermediate system as a function of load is characterized by a knee, and the estimated throughput capacity of the intermediate system is set to equal approximately the throughput at the knee.

8. The method of claim 6, wherein the estimated throughput capacity of the intermediate system is determined by setting the estimated capacity of the intermediate system equal to a total intermediate system throughput from all streams of traffic passing through the intermediate system during said traffic measuring interval, and multiplying the estimated capacity by a capacity factor.

9. The method of claim 8, wherein the capacity factor is approximately equal to 0.9.

10. The method of claim 1 wherein the congestion avoidance flag consists of a single bit.

11. In a network of end systems communicating by means of transmission and reception of digital packets routed through at least one intermediate system, a method for indicating when the intermediate system is approaching a state of congestion, said method comprising the steps of:

A. determining at an intermediate system whether the intermediate system is in an overload condition; and

B. if the intermediate system is in the overload condition, the intermediate system i. selecting end systems that are transmitted packets at rates that are above their respective fair shares; and

ii. signaling selected end systems by including at the intermediate system congestion avoidance information in the packets routed through the intermediate system to one or more of the selected end systems.

12. The method of claim 11, wherein said intermediate system temporarily stores packets to create a queue having a queue length equal to the sum of the number of packets in the queue and the number of packets being serviced by the intermediate system, and wherein the overload condition is determined by determining, over a load averaging interval, an average queue length that is greater than a first preselected queue length.

13. The method of claim 12, wherein operation of the intermediate system includes cycles of activity, each of which includes a busy period and an idle period, and wherein said load averaging interval includes at least one completed previous cycle of activity.

14. The method of claim 12, wherein the first preselected length is approximately equal to one.

15. The method of claim 11, wherein different streams of traffic comprising packets are routed through said intermediate system, said method further comprising the steps of:

identifying those streams of traffic that are causing the overload condition;

in step Bi selecting end stations that are associated with the identified streams of traffic; and

in step Bii, including congestion avoidance information only in packets that are constituents of the streams of traffic identified as causing the overload condition.

16. The method of claim 15, wherein the stream of traffic is identified as causing the overload condition if its constituent packets account for a throughput at said intermediate system, during a traffic measuring interval, that is greater than an allocated share of an estimated throughput capacity of the intermediate system.

17. The method of claim 16, wherein the throughput of the intermediate system as a function of load is characterized by a knee, and the estimated throughput capacity of the intermediate system is set to equal approximately the throughput at the knee.

18. The method of claim 16, wherein the estimated throughput capacity of the intermediate system is determined by (i) setting the estimated capacity of the intermediate system equal to a total intermediate system throughput from all streams of traffic passing through the intermediate system during said traffic measuring interval, and (ii) multiplying the estimated capacity by a capacity factor.

19. The method of claim 18, wherein the capacity factor is approximately equal to 0.9.

20. The method of claim 11, wherein the congestion avoidance information consists of a congestion avoidance flag.

21. The method of claim 11, wherein the congestion avoidance information consists of information about one or more desired operating conditions of the intermediate station.

22. A feedback device, for use in a network of end systems communicating by transmission and reception of digital packets routed through at least one intermediate system, comprising,

A. means for determining at an intermediate system whether said intermediate system is in an overload condition characterized by operation of the intermediate system above an optimal operating level and below a congestion operating level;

B. means at said intermediate system for selecting end systems that are operating above their respective allocated fair shares of throughput; and

C. means for signaling one or more of the selected end systems by setting at the intermediate system a congestion avoidance flag in information packets routed through the intermediate system to one or more of the selected end systems when the intermediate system determines that said intermediate system is in said overload condition.

23. The feedback device of claim 22, wherein said intermediate system temporarily stores packets to create a queue having a queue length equal to the sum of the number of packets in the queue and the number of packets being serviced by the intermediate system, and wherein said means for determining an overload condition further comprises means for comparing average queue length at the intermediate system, averaged over a load averaging interval, to a first preselected queue length.

24. The feedback device of claim 23, wherein operation of the intermediate system includes cycles of activity, each of which includes a busy period and an idle period, and wherein said load averaging interval includes at least one completed previous cycle of activity.

25. The feedback device of claim 23, wherein the first preselected length is approximately equal to one.

26. The feedback device of claim 23 further comprising means for determining that all streams of traffic routed through said intermediate system are causing the overload condition, said means comprising means for comparing the average queue length to a second preselected queue length, longer than the first preselected length.

27. The feedback device of claim 26 wherein the second preselected length is approximately equal to two.

28. The feedback device of claim 22, wherein said means for selecting includes means for identifying streams of traffic that are causing the overload condition said means selecting end systems that are part of the identified streams of traffic.

29. The feedback device of claim 28, wherein said means for identifying streams of traffic as causing the overload condition comprises:

A11. means for comparing the number of information packets associated with a stream of traffic passing through the intermediate system during a traffic measuring interval with a share of an estimated throughput capacity of the intermediate system allocated to said stream of traffic.

30. The feedback device of claim 29, wherein said means for identifying streams of traffic as causing the overload condition further comprises means for determining the estimated throughput capacity of the intermediate system by measuring a total intermediate system throughput from all streams of traffic passing through the intermediate system during said traffic measuring interval.

31. The feedback device of claim 29, wherein operation of the intermediate system includes cycles of activity, each of which includes a busy period and an idle period, and wherein said means for identifying streams of traffic as causing the overload condition further comprises means for detecting when such cycles of activity begin and end.

32. The device of claim 22 wherein the congestion avoidance flag consists of a single bit.

33. A network of end systems communicating by means of transmission and reception of digital packets routed through at least one intermediate system, the intermediate system including:

A. means for determining whether the intermediate system is in an overload condition;

B. means for selecting end systems that are operating above their respective allocated fair shares of throughput; and

C. means for signaling one or more of the selected end systems, said means including congestion avoidance information in the packets routed through the intermediate system to the selected end systems.

34. The system of claim 33, wherein the means for signaling sets a congestion avoidance flag.

35. The system of claim 33, wherein the means for signaling includes in the packets information about one or more desired operating conditions of the intermediate station.