System, method, and computer program product for controlling output port utilization

• US 8,547,843 B2
• Filed: 01/20/2006
• Issued: 10/01/2013
• Est. Priority Date: 01/20/2006
• Status: Active Grant

• Current Assignee: Saisei Networks, Inc.
• Original Assignee: Saisei Networks Pte Ltd.
• Inventors: Roberts, Lawrence G., Henderson, Alex E.
• Primary Examiner(s): Rinehart, Mark
• Assistant Examiner(s): ANWAR, MOHAMMAD S
• Application Number:
  US11/335,973
• Publication Number:
  US 20070171826A1
• Time in Patent Office: 2,811 Days
• Field of Search: 370/232, 370/233, 370/235
• US Class Current: 370/235
• CPC Class Codes: H04L 47/10 Flow control; Congestion control
• CPC Class Codes: H04L 47/11 Identifying congestion
• CPC Class Codes: H04L 47/125 by balancing the load, e.g. traffic engineering
• CPC Class Codes: H04L 47/22 Traffic shaping
• CPC Class Codes: H04L 47/2416 Real-time traffic
• CPC Class Codes: H04L 47/2441 relying on flow classification, e.g. using integrated services [IntServ]
• CPC Class Codes: H04L 47/32 by discarding or delaying data units, e.g. packets or frames

Abstract

A system, method and computer program product are provided. In use, a plurality of flows associated with packets destined for an output port is identified. A utilization associated with the output port is further measured. Thus, rates of a plurality of the flows destined for the output port may be individually controlled at an input port thereof, based on the utilization to ensure that the utilization remains less than 99.9% and avoid buffering more than 400 packets with a correspondingly low delay.

Claims

1. A method, comprising:

identifying a plurality of flows associated with internet protocol packets destined for an output port, each of the flows including a collection of packets with a common destination address, source address, protocol, destination port, and source port;

measuring a utilization associated with the output port utilizing an output flow manager;

forwarding, from the output flow manager, the measured utilization to a network processing unit (NPU) including at least one processor for use in routing the packets to the common destination address;

receiving the measured utilization from the NPU at an input flow manager; and

individually controlling rates of each of the flows destined for the output port, at an input port thereof and utilizing the input flow manager, based on the utilization to ensure that the utilization remains less than 99.9%, ensure packet loss is not harmful, and avoid buffering more than 400 packets with a correspondingly low delay;

wherein individually controlling the rates includes controlling a rate of each individual flow by only dropping a single packet from the individual flow such that the rate of the individual flow is forced to slow down by a Transmission Control Protocol (TCP) process utilized at a sending computer including the output port;

wherein a decision process used to determine if the individual flow is be slowed down by dropping the single packet predicts a composite output rate for each output port in advance of any output port overload such that the TCP process has sufficient time to reduce the rate to hold the utilization below 99.9%;

wherein the identifying, the measuring and the controlling are performed by hardware or software facilitating the communication of packets from one point in a network to another point in the network.

2. The method of claim 1, wherein the rates of the plurality of flows destined for the output port are each individually controlled in a different manner, at the input port thereof.

3. The method of claim 1, wherein the rates of the plurality of flows destined for the output port are individually controlled, at the input port thereof, based on the utilization to ensure that the utilization remains less than 95%.

4. The method of claim 1, wherein the rates of the plurality of flows destined for the output port are individually controlled, at the input port thereof, based on the utilization to avoid use of more than 1 MB of buffer capacity.

5. The method of claim 1, wherein the rates of the plurality of flows destined for the output port are individually controlled, at the input port thereof, based on the utilization to avoid use of more than 100 KB of buffer capacity.

6. The method of claim 1, wherein the packets together comprise streaming media.

7. The method of claim 6, wherein the streaming media includes audio streaming media.

8. The method of claim 6, wherein the streaming media includes video streaming media.

9. The method of claim 1, wherein the utilization is measured periodically.

10. The method of claim 9, wherein the utilization is measured periodically every 50 milliseconds.

11. The method of claim 1, wherein the controlling includes rejecting a new flow.

12. The method of claim 1, wherein the controlling includes discarding a new flow.

13. The method of claim 1, wherein the identifying, the measuring and the controlling are performed by a router.

14. The method of claim 1, wherein the rates of the plurality of flows destined for the output port are individually controlled, at the input port thereof, based on the utilization to ensure that the utilization remains less than 90%.

15. The method of claim 1, wherein the rates of the plurality of flows destined for the output port are individually controlled, at the input port thereof, based on the utilization to ensure that the utilization remains less than 80%.

16. A computer program product embodied on a non-transitory computer readable medium, comprising:

computer code for identifying a plurality of flows associated with internet protocol packets destined for an output port, each of the flows including a collection of packets with a common destination address, source address, protocol, destination port, and source port;

computer code for measuring a utilization associated with the output port utilizing an output flow manager;

computer code for forwarding, from the output flow manager, the measured utilization to a network processing unit (NPU) including at least one processor for use in routing the packets to the common destination address;

computer code for receiving the measured utilization from the NPU at an input flow manager; and

computer code for individually controlling rates of each of the flows destined for the output port, at an input port thereof and utilizing the input flow manager, based on the utilization to ensure that the utilization remains less than or equal to 99.9%, ensure packet loss is not harmful, and avoid buffering more than 400 packets with a correspondingly low delay;

wherein the computer program product is operable such that individually controlling the rates includes controlling a rate of each individual flow by only dropping a single packet from the individual flow such that the rate of the individual flow is forced to slow down by a Transmission Control Protocol (TCP) process utilized at a sending computer including the output port;

wherein the computer program product is operable such that a decision process used to determine if the individual flow is be slowed down by dropping the single packet predicts a composite output rate for each output port in advance of any output port overload such that the TCP process has sufficient time to reduce the rate to hold the utilization below 99.9%;

wherein the identifying, the measuring and the controlling are performed by hardware or software facilitating the communication of packets from one point in a network to another point in the network.

17. A system, comprising:

an input flow manager for identifying a plurality of flows associated with internet protocol packets destined for an output port, each of the flows including a collection of packets with a common destination address, source address, protocol, destination port, and source port; and

an output flow manager for measuring a utilization associated with the output port, forwarding the measured utilization to a network processing unit (NPU) including at least one processor for use in routing the packets to the common destination address;

wherein the input flow manager receives the measured utilization from the NPU;

wherein the input flow manager individually controls rates of each of the flows destined for the output port, at an input port thereof, based on the utilization to ensure that the utilization remains less than 99.9%, ensure packet loss is not harmful, and avoid buffering more than 400 packets with a correspondingly low delay;

wherein the input flow manager is operable such that individually controlling the rates includes controlling a rate of each individual flow by only dropping a single packet from the individual flow such that the rate of the individual flow is forced to slow down by a Transmission Control Protocol (TCP) process utilized at a sending computer including the output port;

wherein the input flow manager is operable such that a decision process used to determine if the individual flow is be slowed down by dropping the single packet predicts a composite output rate for each output port in advance of any output port overload such that the TCP process has sufficient time to reduce the rate to hold the utilization below 99.9%;

wherein the identifying, the measuring and the controlling are performed by hardware or software facilitating the communication of packets from one point in a network to another point in the network.