TCP CONGESTION CONTROL FOR LARGE LATENCY NETWORKS

• US 20150043339A1
• Filed: 12/27/2012
• Published: 02/12/2015
• Est. Priority Date: 12/28/2011
• Status: Active Grant

• Current Assignee: Cdf K.E. Yuan
• Original Assignee: Cdf K.E. Yuan
• Inventors: Wang, Jingyuan, Wen, Jiangtao, Han, Yuxing, Zhang, Jun
• Granted Patent: US 9,838,318 B2
• Time in Patent Office: Days
• US Class Current: 370/230
• CPC Class Codes: H04L 12/6418 Hybrid transport
• CPC Class Codes: H04L 43/0864 Round trip delays
• CPC Class Codes: H04L 47/11 Identifying congestion
• 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 65/1066 Session management
• CPC Class Codes: H04L 69/163 In-band adaptation of TCP data exchange; In-band control procedures
• CPC Class Codes: H04W 80/06 Transport layer protocols, e.g. TCP [Transport Control Protocol] over wireless transmission control protocol/Internet protocol [TCP/IP] or user datagram protocol [UDP] H04L69/16

Abstract

A congestion control mechanism for TCP communication sessions is described. The congestion control mechanism adjusts the size of the congestion window based on a number of parallel virtual TCP Hybla connections. The number of parallel virtual TCP Hybla connections used to determine the congestion window can be dynamically adjusted. The enhanced TCP Hybla congestion control mechanism provides improved throughput for heterogenous networks having a high latency link and having possible packet loss.

Claims

1. A method for congestion control of a communication session over a heterogeneous network having a high latency link, the method comprising:

determining a round trip time (RTT) for the communication session;

determining a normalized RTT (p) based on an equalizing round trip time (RTT₀);

determining a number, N, of virtual communication sessions; and

setting a congestion window (Cω

nd) size for the communication session based on N TCP virtual communication sessions and p.

2. The method of claim 1, wherein the congestion window is set based on N TCP Hybla virtual communication sessions.

3. The method of claim 2, wherein the congestion window is set according to:

4. The method of claim 3, wherein p is calculated according to:

5. The method of claim 4, wherein the number of virtual communication sessions is determined according to:

6. The method of claim 5, wherein Nis determined as a combination of a number, N₁, of virtual TCP Hybla communication sessions and a number, N₂, of TCP Reno sessions, wherein N₁ and N₂ are set according to:

7. The method of claim 6, further comprising increasing Cω

nd for received ACKs, when operating in a slow start phase, according to;

Cω

nd_i+1=Cω

nd_i+N₁·

(2^p−

1)+N₂

8. The method of claim 7, further comprising increasing Cω

nd for received ACKs, when operating in a congestion avoidance phase, according to;

9. The method of claim 8, further comprising decreasing Cω

nd for packet loss events, according to;

10. The method of claim 9, wherein N₁ and N₂ are determined periodically to balance the aggressiveness of the Hybla portion of the congestion control with the more stable performance of the Reno portion of the congestion control.

11. The method of claim 1, further comprising:

determining if Cω

nd is larger than a slow start threshold (SS_thresh); and

operating in a slow start phase when Cω

nd is less than SS_thresh and operating in a congestion avoidance phase when Cω

nd is larger than SS_thresh.

12. The method of claim 11, further comprising:

updating an average RTT value (RTT_average) using the determined RTT; and

updating a minimum RTT (RTT_min) with the determined RTT when the determined RTT is less than RTT_min,

wherein p is calculated according to;

13. The method of claim 12, where N is determined according to:

14. The method of claim 13, wherein N is determined as a combination of a number, N₁, of virtual TCP Hybla communication sessions and a number, N₂, of TCP Reno sessions, wherein N₁ and N₂ are set according to:

15. The method of claim 14, further comprising increasing Cω

nd for received ACKs, when operating in the slow start phase, according to;

Cω

nd_i+1=Cω

nd_i+N₁·

(1^p−

1)+N₂

16. The method of claim 15, further comprising increasing Cω

nd for received ACKs, when operating in the congestion avoidance phase, according to;

17. The method of claim 16, further comprising decreasing Cω

nd for packet loss events, according to;

18. The method of claim 17, wherein N₁ and N₁ are determined periodically to balance the aggressiveness of the Hybla portion of the congestion control with the more stable performance of the Reno portion of the congestion control.

19. A computing device for communicating over a network, the computing device comprising:

a processor for executing instructions; and

a memory for storing instructions, which when executed by the processor configure the computing device to perform a method according to claim 1.

20. A non-transitory computer readable medium storing instructions for configuring a computing device to perform a method according to claim 1.