Systems And Methods For Managing Congestion

• US 20150188830A1
• Filed: 12/30/2013
• Published: 07/02/2015
• Est. Priority Date: 12/30/2013
• Status: Active Grant

• Current Assignee: Comcast Cable Communications LLC (Comcast Corporation)
• Original Assignee: Comcast Cable Communications LLC (Comcast Corporation)
• Inventors: Zhao, Zhen
• Granted Patent: US 9,559,972 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: H04L 47/12 Avoiding congestion; Recovering from congestion
• CPC Class Codes: H04L 47/193 at the transport layer, e.g. TCP related
• 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 47/28 in relation to timing considerations
• CPC Class Codes: H04L 47/283 in response to processing delays, e.g. caused by jitter or round trip time [RTT]

Abstract

Systems and methods for managing congestion in a network are disclosed. One method can comprise establishing a connection between a first network device and a second network device, wherein the first network device has an associated data limiter value such as a congestion window (CWND). The size of the CWND can be modified, for example, at a first rate, when no congestion is detected. In response to the congestion window size exceeding a threshold value associated with the connection, the size of the CWND can be modified based on round trip time (RTT) associated with the connection. If congestion on the connection is detected, the congestion window size can be modified to a value based on at least in part the bandwidth and RTT of the connection between the first network device and the second network device.

Claims

1. A method comprising:

establishing a connection between a first network device and a second network device, wherein the first network device has an associated congestion window (CWND);

increasing the size of the associated CWND at a first rate when no congestion is detected;

in response to the associated CWND size exceeding a threshold value, increasing the size of the CWND at a second rate based at least in part on round trip time (RTT) associated with the connection;

detecting congestion on the connection; and

decreasing the CWND size to a value based at least in part on bandwidth and the RTT of the connection.

2. The method of claim 1, wherein the establishing the connection between the first network device and the second network device comprises setting the size of CWND to an initial value.

3. The method of claim 2, wherein the initial value of the size of CWND is 1.

4. The method of claim 1, wherein congestion is detected by identifying a predefined number of duplicate acknowledgments (ACKs) or by identifying a mark in one or more ACKs.

5. The method of claim 1, wherein the size of the CWND is increased at the second rate based on

6. The method of claim 1, wherein the decreasing the CWND size to the value is based on cwnd=k·

B_steady·

RTT.

7. The method of claim 1, further comprising setting the threshold value to ssthresh=k·

B_steady·

RTT.

8. The method of claim 1, wherein the first rate is a non-linear rate or an exponential rate.

9. A method comprising:

establishing a connection between a first network device and a second network device, wherein the first network device has an associated congestion window (CWND);

in a first state of the first network device, increasing the size of the CWND at a first rate when no congestion is detected;

in response to the size of the CWND exceeding a threshold value associated with the first state of the first device, transitioning the first network device to a second state;

in the second state of the first network device, increasing the size of the CWND at a second rate at least in part in view of a round trip time (RTT) of the connection;

when congestion is detected, transitioning the first network device to a third state; and

in the third state, decreasing the CWND size to a value based at least in part on the RTT of the connection.

10. The method of claim 9, wherein establishing a connection between a first network device and a second network device comprises setting the size of CWND to an initial value.

11. The method of claim 9, wherein congestion is detected by identifying a predefined number of duplicate acknowledgments (ACKs) or by identifying a mark in one or more ACKs.

12. The method of claim 9, wherein, in the second state, the size of the CWND is increased based on

13. The method of claim 9, wherein in the third state, the size of CWND is set to a value based on cwnd=k·

B_steady·

RTT.

14. The method of claim 9, further comprising in the third state, setting the threshold value associated with the first state to ssthresh=k·

B_steady·

RTT.

15. The method of claim 9, wherein the first rate is a non-linear rate or an exponential rate.

16. The method of claim 9, wherein the first state of the first network device is a state of implementing a slow start protocol of a TCP congestion control mechanism, the second state of the first device is a state of implementing a congestion avoidance protocol of a TCP congestion control mechanism, and the third state of the first network device is a state of implementing a fast retransmission protocol of a TCP congestion control mechanism.

17. A method comprising:

increasing a data limiter value associated with a connection between a first network device and a second network device based at least in part on round trip time (RTT) associated with the connection in a congestion avoidance state of the connection;

detecting congestion on the connection; and

in response to the congestion, decreasing the data limiter value to a value based at least in part on bandwidth and RTT of the connection.

18. The method of claim 17, wherein congestion is detected by identifying a predefined number of duplicate acknowledgments (ACKs) or by identifying a mark in one or more ACKs.

19. The method of claim 17, wherein the size of the data limiter value is increased according to

20. The method of claim 17, wherein the size of data limiter value is decreased to a value according to cwnd=k·

B_steady·

RTT, wherein cwnd represents data limiter value.