Reducing network congestion using a load balancer

Patent metadata

Field	Value
Publication number	US-12388754-B2
Application number	US-202218051218-A
Country	US
Kind code	B2
Filing date	Oct 31, 2022
Priority date	Oct 31, 2022
Publication date	Aug 12, 2025
Grant date	Aug 12, 2025

Abstract

Reducing network congestion using a load balancer is described. The load balancer receives network traffic for processing work for a delay-sensitive application. The load balancer is coupled with multiple application instances executing on compute nodes that are each capable of processing the work. The load balancer determines from the network traffic whether there is an indication of network congestion. If there is an indication of network congestion, the load balancer selects one of the application instances for processing the work based at least in part on reducing the network congestion. The load balancer causes the work to be processed at the selected application instance.

First claim

What is claimed is:

1. A method for reducing network congestion using a load balancer, the method comprising:
receiving, at the load balancer, network traffic for processing work for a delay-sensitive application, wherein the load balancer is coupled with a plurality of application instances executing on a plurality of compute nodes that are each capable of processing the work;
determining, from the received network traffic, an indication of network congestion;
responsive to the determining the indication of network congestion, selecting one of the plurality of application instances for processing the work, wherein the selecting is based at least in part on reducing the network congestion without degrading output quality through a more efficient compression mechanism used by the selected one of the plurality of application instances compared to other ones of the plurality of application instances to lessen bandwidth that will be used for carrying network traffic responsive to processing the work, and wherein the selected one of the plurality of application instances is capable of processing the work such that less bandwidth is used for carrying network traffic responsive to the processed work without degrading output quality as compared to other ones of the plurality of application instances due to the more efficient compression mechanism; and
causing the work to be processed at the selected one of the plurality of application instances.

2. The method of claim 1 , wherein determining the indication of network congestion from the network traffic includes analyzing Explicit Congestion Notification (ECN) data included within the network traffic.

3. The method of claim 1 , wherein a first set of one or more of the plurality of compute nodes has a first hardware capability, wherein a second set of one or more of the plurality of compute nodes has a second hardware capability, wherein the first hardware capability is capable of processing work such that less bandwidth is used for carrying network traffic responsive to the processed work as compared to the second hardware capability, and wherein selecting one of the plurality of application instances for processing the work includes selecting one of the plurality of application instances at the first set of the plurality of compute nodes.

4. The method of claim 3 , wherein the work is related to video processing, and wherein the first hardware capability includes a graphics processing unit (GPU) or hardware accelerator that more efficiently encodes video including a more efficient compression mechanism compared to the second hardware capability that does not include a GPU or hardware accelerator.

5. The method of claim 1 , further comprising:
interrupting a non-delay-sensitive task being executed on the selected one of the plurality of application instances; and
resuming the interrupted non-delay-sensitive task after the work is finished executing.

6. The method of claim 1 , wherein the selected one of the plurality of application instances for processing the work has more processing cycles available for processing the work including compressing compared to other ones of the plurality of application instances.

7. A non-transitory machine-readable storage medium that provides instructions that, if executed by a processor of a load balancer, will cause said load balancer to perform operations for reducing network congestion comprising:
receiving, at the load balancer, network traffic for processing work for a delay-sensitive application, wherein the load balancer is coupled with a plurality of application instances executing on a plurality of compute nodes that are each capable of processing the work;
determining, from the received network traffic, an indication of network congestion;
responsive to the determining the indication of network congestion, selecting one of the plurality of application instances for processing the work, wherein the selecting is based at least in part on reducing the network congestion without degrading output quality through a more efficient compression mechanism used by the selected one of the plurality of application instances compared to other ones of the plurality of application instances to lessen bandwidth that will be used for carrying network traffic responsive to processing the work, and wherein the selected one of the plurality of application instances is capable of processing the work such that less bandwidth is used for carrying network traffic responsive to the processed work without degrading output quality as compared to other ones of the plurality of application instances due to the more efficient compression mechanism; and
causing the work to be processed at the selected one of the plurality of application instances.

8. The non-transitory machine-readable storage medium of claim 7 , wherein determining the indication of network congestion from the network traffic includes analyzing Explicit Congestion Notification (ECN) data included within the network traffic.

9. The non-transitory machine-readable storage medium of claim 7 , wherein a first set of one or more of the plurality of compute nodes has a first hardware capability, wherein a second set of one or more of the plurality of compute nodes has a second hardware capability, wherein the first hardware capability is capable of processing work such that less bandwidth is used for carrying network traffic responsive to the processed work as compared to the second hardware capability, and wherein selecting one of the application instances for processing the work includes selecting one of the plurality of application instances at the first set of the plurality of compute nodes.

10. The non-transitory machine-readable storage medium of claim 9 , wherein the work is related to video processing, and wherein the first hardware capability includes a graphics processing unit (GPU) or hardware accelerator that more efficiently encodes video including a more efficient compression mechanism compared to the second hardware capability that does not include a GPU or hardware accelerator.

11. The non-transitory machine-readable storage medium of claim 7 , wherein the operations further comprise:
interrupting a non-delay-sensitive task being executed on the selected one of the plurality of application instances; and
resuming the interrupted non-delay-sensitive task after the work is finished executing.

12. The non-transitory machine-readable storage medium of claim 7 , wherein the selected one of the plurality of application instances for processing the work has more processing cycles available for processing the work including compressing compared to other ones of the plurality of application instances.

13. A load balancer, comprising:
a processor; and
a non-transitory machine-readable storage medium that provides instructions that, if executed by the processor, will cause the processor to perform operations comprising:
receiving, at the load balancer, network traffic for processing work for a delay-sensitive application, wherein the load balancer is coupled with a plurality of application instances executing on a plurality of compute nodes that are each capable of processing the work;
determining, from the received network traffic, an indication of network congestion;
responsive to the determining the indication of network congestion, selecting one of the plurality of application instances for processing the work, wherein the selecting is based at least in part on reducing the network congestion without degrading output quality through a more efficient compression mechanism used by the selected one of the plurality of application i

Assignees

• Ericsson Telefon Ab L M

Inventors

• Svensson Fredrik
• Skarin Per
• Eker Johan
• Millnert Victor

Classifications

• H04L47/122

  by diverting traffic away from congested entities · CPC title

• H04L43/062

  related to network traffic · CPC title

• H04L67/1001

  for accessing one among a plurality of replicated servers · CPC title

• H04L67/101

  based on network conditions · CPC title

• H04L47/38

  by adapting coding or compression rate · CPC title