Right-sizing resource requests by applications in dynamically scalable computing environments

Patent metadata

Field	Value
Publication number	US-12596574-B2
Application number	US-202217893864-A
Country	US
Kind code	B2
Filing date	Aug 23, 2022
Priority date	Aug 27, 2021
Publication date	Apr 7, 2026
Grant date	Apr 7, 2026

Abstract

Methods, systems, and techniques for right-sizing resource requests for applications in a dynamically scalable computing environment. In one aspect, a method comprises monitoring resource usage of at least one computer resource by an application executing on a computer system, and monitoring resource requests for the computer resource(s) associated with the application. The method further comprises determining, for the computer resource(s), a resource usage upper bound associated with the application, testing the resource usage upper bound against at least one threshold, determining, from the testing, a resource request adjustment, and dynamically applying the resource request adjustment to the resource requests for the computer resource(s) associated with the application.

First claim

The invention claimed is:

1 . A computer-implemented method for right-sizing resource requests for applications in a dynamically scalable computing environment, the method comprising:
monitoring, by a computer system, resource usage of at least one computer resource by an application executing in the dynamically scalable computing environment; monitoring, by the computer system, resource requests for the at least one computer resource associated with the application; determining, by the computer system, for the at least one computer resource, a resource usage upper bound associated with the application; testing, by the computer system, the resource usage upper bound against at least one threshold; determining, by the computer system, from the testing, a resource request adjustment; and the computer system dynamically applying the resource request adjustment to the resource requests for the at least one computer resource associated with the application; wherein determining, for the at least one computer resource, a resource usage upper bound associated with the application comprises:
testing whether, during a predetermined time period, there have been a predetermined number of instances where the resource usage has exceeded a mean of the resource usage of the computer resource by the application plus a predetermined number of standard deviations of the resource usage of the computer resource by the application;
responsive to a determination that there have been a predetermined number of instances where the resource usage has exceeded a mean of the resource usage of the computer resource by the application plus a predetermined number of standard deviations of the resource usage of the computer resource by the application, setting the resource usage upper bound to a first predetermined percentage of a maximum resource usage of the computer resource by the application; and
responsive to a determination that there have not been a predetermined number of instances where the resource usage of the computer resource by the application has exceeded a mean of the resource usage of the computer resource by the application plus a predetermined number of standard deviations of the resource usage of the computer resource by the application, setting the resource usage upper bound to a second predetermined percentage of the mean of the resource usage of the computer resource by the application.

2 . The method of claim 1 , wherein the resource request adjustment is applied automatically to the resource requests absent human intervention to reduce latency.

3 . The method of claim 1 , further comprising:
evaluating performance of the resource request adjustment as applied to the resource requests; and dynamically modifying at least one of the threshold and the resource usage upper bound based on the performance of the resource request adjustment as applied to the resource requests.

4 . The method of claim 3 , wherein modifying at least one of the threshold and the resource usage upper bound based on the performance of the resource request adjustment as applied to the resource requests is performed by machine learning.

5 . The method of claim 4 , wherein evaluating performance of the resource request adjustment as applied to the resource requests and modifying at least one of the threshold and the resource usage upper bound based on the performance of the resource request adjustment as applied to the resource requests is carried out recursively.

6 . The method of claim 1 , wherein the threshold is determined by machine learning.

7 . The method of claim 1 , wherein the at least one computer resource includes at least one of CPU resource, memory resource, network utilization, disk I/O, graphical processing unit (GPU) uses and database usage.

8 . The method of claim 1 , wherein the application is a containerized application.

9 . The method of claim 1 , wherein the application is a virtual machine application.

10 . The method of claim 1 , wherein the resource usage upper bound is different from an actual maximum resource usage of the application.

11 . The method of claim 1 , wherein the resource requests include a limit value for the at least one computer resource, and wherein the resource usage upper bound differs from the limit value for the at least one computer resource.

12 . The method of claim 1 , wherein testing the resource usage upper bound against at least one threshold derived from the resource requests comprises:
testing whether the resource usage upper bound divided by a mean of the resource requests exceeds a first predetermined value; responsive to a determination that the resource usage upper bound divided by the mean of the resource requests exceeds the first predetermined value, testing whether the resource usage upper bound divided by the mean of the resource requests exceeds a second predetermined value that is greater than the first predetermined value by more than a third predetermined value; and responsive to a determination that the resource usage upper bound divided by the mean of the resource requests fails to exceed the first predetermined value, testing whether subtracting the resource usage upper bound divided by the mean of the resource requests from a fourth predetermined value yields a result that exceeds a fifth predetermined value.

13 . The method of claim 12 , wherein determining, from the testing, the resource request adjustment comprises:
responsive to a determination that the resource usage upper bound divided by the mean of the resource requests exceeds the second predetermined value by more than the third predetermined value, setting the resource request adjustment to increase the resource requests; responsive to a determination that subtracting the resource usage upper bound divided by the mean of the resource requests from the fourth predetermined value yields a result that exceeds the fifth predetermined value, setting the resource request adjustment to decrease the resource requests; responsive to a determination that the resource usage upper bound divided by the mean of the resource requests fails to exceed the second predetermined value by more than the third predetermined value, setting the resource request adjustment to leave the resource requests unchanged; and responsive to a determination that subtracting the resource usage upper bound divided by the mean of the resource requests from the fourth predetermined value yields a result that fails to exceed the fifth predetermined value, setting the resource request adjustment to leave the resource requests unchanged.

14 . The method of claim 13 , wherein setting the resource request adjustment to increase the resource requests comprises setting the resource request adjustment to increase the resource requests by a percentage equal to 100% multiplied by the lesser of (a) a sixth predetermined value and (b) a seventh predetermined value subtracted from the resource usage upper bound divided by the mean of the resource requests.

15 . The method of claim 13 , wherein setting the resource request adjustment to decrease the resource requests comprises setting the resource request adjustment to decrease the resource requests by a percentage equal to 100% multiplied by the lesser of (a) an eighth predetermined value and (b) the resource usage upper bound divided by the mean of the resource requests subtracted from a ninth predetermined value.

16 . A data processing system comprising at least one processor and memory coupled to the at least one processor, wherein the memory contains instruct

Assignees

• Royal Bank Of Canada

Inventors

• Stringer Matthew
• Li Merissa
• Alemi Kaveh
• Aminu Ore
• Packiriswamy Venkatesan
• Agrawal Manoj
• Mahajan Vishal

Classifications

• G06F9/45558 Primary

  Hypervisor-specific management and integration aspects · CPC title

• G06F2209/504

  Resource capping · CPC title

• G06F2209/508

  Monitor · CPC title

• G06F9/5011

  the resources being hardware resources other than CPUs, Servers and Terminals · CPC title

• G06F2009/4557

  Distribution of virtual machine instances; Migration and load balancing · CPC title