Allocating of computing resources for applications

Patent metadata

Field	Value
Publication number	US-12566640-B2
Application number	US-202217841552-A
Country	US
Kind code	B2
Filing date	Jun 15, 2022
Priority date	Jun 17, 2021
Publication date	Mar 3, 2026
Grant date	Mar 3, 2026

Abstract

A method for performing scheduling includes extracting information from at least one log file for an application. The method also includes determining an allocation of cloud resources for the application based on the information from the log file(s).

First claim

What is claimed is:

1 . A method, comprising:
extracting information from at least one log file for an application; determining an allocation of cloud resources for the application based on the information from the at least one log file, wherein the at least one log file is from at least one run of the application, wherein the determining of the allocation of cloud resources comprises:
determining a total dead time for the at least one run;
determining a total task time for the at least one run based on the information;
determining a predicted parallelism based on the information from the at least one log file, wherein the predicted parallelism relates to a distribution of tasks based on expected active cores;
determining a predicted total run time for the application based on the total task time, the total dead time, and the predicted parallelism; and
determining a number of cores to be allocated for the application based on the predicted total run time; and
providing the allocation of the cloud resources for the application to a hardware infrastructure.

2 . The method of claim 1 , wherein the extracting information from the at least one log file further includes:
obtaining at least one of task data, cloud settings, hardware information, cloud economic information or cloud reliability information.

3 . The method of claim 1 , wherein the determining the allocation of the cloud resources further includes:
determining a plurality of hardware infrastructures; determining a predicted run time for each of the plurality of hardware infrastructures based on the information in the at least one log file.

4 . The method of claim 3 , wherein the determining the allocation of the cloud resources further includes:
determining a predicted cost for usage of each of the plurality of hardware infrastructures; and determining the predicted cost versus the predicted run time for each of the plurality of hardware infrastructures.

5 . The method of claim 3 ,
wherein the determining the total dead time and the total task time each include distributing the plurality of tasks over a plurality of cores in one hardware infrastructure of the plurality of hardware infrastructures.

6 . The method of claim 1 , wherein the determining the allocation of the cloud resources further includes:
creating a time-based model of memory usage including garbage collection parameters.

7 . The method of claim 1 , further comprising:
determining whether a change in an allocation of resources has occurred in at least one of an application, input data for the application, or cloud resources, the cloud resources including a cluster of cores assigned to the application; and in response to determining that the change has occurred, performing the extracting and determining.

8 . The method of claim 1 , wherein the determining the allocation of the cloud resources further includes:
determining a scheduling of tasks or stages for the application.

9 . The method of claim 1 , further comprising:
extracting additional information from at least one additional log file for the application; and determining a re-allocation of the cloud resources for the application based on the additional information from the at least one additional log file, the at least one additional log file corresponding to at least one additional run of the application, wherein the at least one additional run is different from the at least one run.

10 . A system, comprising:
a processor configured to:
extract information from at least one log file for an application, wherein the at least one log file is from at least one run of the application;
determine an allocation of cloud resources for the application based on the information from the at least one log file, wherein the determining of the allocation of cloud resources comprises to:
determine a total dead time for the at least one run, wherein the at least one run includes a plurality of stages;
determine a total task time for the at least one run based on the information;
determine a predicted parallelism based on the information from the at least one log file, wherein the predicted parallelism relates to a distribution of tasks based on expected active cores;
determine a predicted total run time for the application based on the total task time, the total dead time, and the predicted parallelism; and
determine a number of cores to be allocated for the application based on the predicted total run time; and

provide the allocation of the cloud resources for the application to a hardware infrastructure; and
a memory coupled to the processor and configured to provide the processor with instructions.

11 . The system of claim 10 , wherein to determine the allocation of the cloud resources, the processor is further configured to:
determine a plurality of hardware infrastructures; determine a predicted run time for each of the plurality of hardware infrastructures based on the information in the at least one log file; determine a predicted cost for usage of each of the plurality of hardware infrastructures; and determine the predicted cost versus the predicted run time for the plurality of hardware infrastructures.

12 . The system of claim 11 , wherein the determining the total dead time and the total task time each include distributing the plurality of tasks over a plurality of cores in one hardware infrastructure of the plurality of hardware infrastructures.

13 . A computer program product embodied in a non-transitory computer readable medium and comprising computer instructions for:
extracting information from at least one log file for an application; and determining an allocation of cloud resources for the application based on the information from the at least one log file, wherein the at least one log file is from at least one run of the application, wherein the determining of the allocation of cloud resources comprises:
determining a total dead time for the at least one run, wherein the at least one run includes a plurality of stages;
determining a total task time for the at least one run based on the information;
determining a predicted parallelism based on the information from the at least one log file, wherein the predicted parallelism relates to a distribution of tasks based on expected active cores;
determining a predicted total run time for the application based on the total task time, the total dead time, and the predicted parallelism; and
determining a number of cores to be allocated for the application based on the predicted total run time; and
providing the allocation of the cloud resources for the application to a hardware infrastructure.

14 . The computer program product of claim 13 , wherein the computer instructions for determining the allocation of the cloud resources further include computer instructions for:
determining a plurality of hardware infrastructures; determining a predicted run time for each of the plurality of hardware infrastructures based on the information in the at least one log file; determining a predicted cost for usage of each of the plurality of hardware infrastructures; and determining the predicted cost versus the predicted run time for the plurality of hardware infrastructures.

15 . The computer program product of claim 14 , wherein the determining the total dead time and the total task time each include distributing the plurality of tasks over a plurality of cores in one hardware infrastructure of the plurality of hardware infrastructures.

16 . The computer program product of claim 13

Assignees

• Capital One Services Llc

Inventors

• Hanel Carson
• Gorsky Sean
• Lin Erica
• Bramhavar Suraj
• Chou Jeffrey

Classifications

• G06F9/4887

  involving deadlines, e.g. rate based, periodic · CPC title

• G06F2209/505

  Clust · CPC title

• G06F2209/5019

  Workload prediction · CPC title

• G06F2209/508

  Monitor · CPC title

• G06F9/5055 Primary

  considering software capabilities, i.e. software resources associated or available to the machine · CPC title