Setting delay precedence on queues before a bottleneck link based on flow characteristics

Patent metadata

Field	Value
Publication number	US-2016308769-A1
Application number	US-201514686575-A
Country	US
Kind code	A1
Filing date	Apr 14, 2015
Priority date	Apr 14, 2015
Publication date	Oct 20, 2016
Grant date	—

Abstract

Embodiments include detecting an increase in delay of a flow assigned to a first queue of a network device, where the increase is sufficient to cause the flow rate of the flow to decrease if the flow is delay-sensitive. Embodiments further include determining whether an amount of bandwidth consumed by the flow decreases sufficiently after the increase is detected, and assigning the flow to a second queue based, at least in part, on determining the amount of bandwidth consumed by the flow does not decrease sufficiently. Specific embodiments include evaluating a series of two or more bandwidth measurements of the flow according to a bandwidth reduction measure to determine whether the amount of bandwidth consumed by the flow decreases by sufficiently. More specific embodiments include the first queue being configured to receive delay-sensitive flows and the second queue being configured to receive delay-insensitive nice flows.

First claim

What is claimed is:

1 . A method comprising:
detecting an increase in delay of a flow assigned to a first queue of a network device, wherein the increase is sufficient to cause the flow rate of the flow to decrease if the flow is delay-sensitive; determining whether an amount of bandwidth consumed by the flow decreases sufficiently after the increase is detected; and assigning the flow to a second queue based, at least in part, on determining the amount of bandwidth consumed by the flow does not decrease sufficiently.

2 . The method of claim 1 , wherein the assigning the flow to the second queue causes subsequent packets of the flow to be stored in the second queue.

3 . The method of claim 1 , further comprising:
evaluating a series of two or more bandwidth measurements of the flow according to a bandwidth reduction measure to determine whether the amount of bandwidth consumed by the flow decreases sufficiently.

4 . The method of claim 3 , wherein the two or more bandwidth measurements represent amounts of bandwidth consumed by the flow in the network device over sequential time periods.

5 . The method of claim 1 , wherein the flow comprises a set of flows having a same classification.

6 . The method of claim 1 , further comprising:
determining respective delay periods for two or more packets of the flow received by the network device; and evaluating the respective delay periods based, at least in part, on a delay increase measure to determine whether the increase is sufficient to cause the flow rate to decrease if the flow is delay-sensitive.

7 . The method of claim 6 , further comprising:
determining a delay period for a particular packet of the two or more packets received by the network device, wherein the delay period is calculated based on a time difference between an egress time of the particular packet at the network device and a timestamp from a source node in the particular packet.

8 . The method of claim 1 , wherein the second queue is configured to receive delay-insensitive nice flows.

9 . The method of claim 1 , wherein the first queue is configured to receive delay-sensitive flows.

10 . The method of claim 9 , further comprising:
prior to detecting the increase in delay of the flow, assigning the flow to the first queue in response to the flow asserting itself as a delay-sensitive flow.

11 . The method of claim 1 , wherein the subsequent packets of the flow are not stored in the second queue until after a propagation delay period of the flow expires, wherein a timer for the propagation delay period begins when the increase is determined to be sufficient to cause the flow rate of the flow to decrease if the flow is delay-sensitive.

12 . At least one machine readable storage medium having instructions stored therein, and when executed by at least one processor the instructions cause the at least one processor to:
detect an increase in delay of a flow assigned to a first queue of a network device, wherein the increase is sufficient to cause the flow rate of the flow to decrease if the flow is delay-sensitive; determine whether an amount of bandwidth consumed by the flow decreases sufficiently after the increase is detected; and assign the flow to a second queue based, at least in part, on determining the amount of bandwidth consumed by the flow does not decrease sufficiently.

13 . The at least one machine readable storage medium of claim 12 , wherein the instructions when executed by the at least one processor cause the at least one processor to:
evaluate a series of two or more bandwidth measurements of the flow according to a bandwidth reduction measure to determine whether the amount of bandwidth consumed by the flow is decreases sufficiently.

14 . The at least one machine readable storage medium of claim 13 , wherein the two or more bandwidth measurements represent amounts of bandwidth consumed by the flow in the network device over sequential time periods.

15 . The at least one machine readable storage medium of claim 12 , wherein the instructions when executed by the at least one processor cause the at least one processor to:
determine respective delay periods for two or more packets of the flow received by the network device; and evaluate the respective delay periods based, at least in part, on a delay increase measure to determine whether the increase is sufficient to cause the flow rate to decrease if the flow is delay-sensitive.

16 . The at least one machine readable storage medium of claim 15 , wherein the instructions when executed by the at least one processor cause the at least one processor to:
determine a delay period for a particular packet of the two or more packets received by the apparatus by at least one of:
calculating a time difference between an egress time of the particular packet at the network device and an ingress time of the particular packet at the network device; and
calculating a time difference between an ingress time of the particular packet at the network device and a timestamp from a source node in the particular packet.

17 . An apparatus comprising:
at least one processor; and at least one memory element comprising instructions that when executed by the at least one processor cause the apparatus to:
detect an increase in delay of a flow assigned to a first queue of a network device, wherein the increase is sufficient to cause the flow rate of the flow to decrease if the flow is delay-sensitive;
determine whether an amount of bandwidth consumed by the flow decreases sufficiently after the increase is detected; and
assign the flow to a second queue based, at least in part, on determining the amount of bandwidth consumed by the flow does not decrease sufficiently.

18 . The apparatus of claim 17 , wherein the instructions when executed by the at least one processor cause the apparatus to:
evaluate a series of two or more bandwidth measurements of the flow according to a bandwidth reduction measure to determine whether the amount of bandwidth consumed by the flow is decreases sufficiently.

19 . The apparatus of claim 17 , wherein the instructions when executed by the at least one processor cause the apparatus to:
prior to detecting the increase in delay of the flow, assign the flow to the first queue in response to the flow asserting itself as a delay-sensitive flow.

20 . The apparatus of claim 17 , wherein the flow comprises a set of flows having a same classification.

Assignees

• Cisco Tech Inc

Inventors

• Versteeg William C
• Ramalho Michael A

Classifications

• H04L47/12 Primary

  Avoiding congestion; Recovering from congestion · CPC title

• H04L47/522 Primary

  Dynamic queue service slot or variable bandwidth allocation · CPC title

• H04L43/0882

  Utilisation of link capacity · CPC title

• H04L47/801

  Real time traffic · CPC title

• H04L43/0852

  Delays · CPC title