US 11,900,162 B2
Autonomous application management for distributed computing systems
Suresh Mathew, San Ramon, CA (US); Nikhil Gopinath Kurup, Tampa, FL (US); Hari Chandrasekhar, Highlands Ranch, CO (US); and Benjamin Thomas, San Jose, CA (US)
Assigned to SEDAI, INC., Pleasanton, CA (US)
Filed by Sedai Inc., Pleasanton, CA (US)
Filed on Feb. 23, 2022, as Appl. No. 17/678,907.
Application 17/678,907 is a division of application No. 17/387,984, filed on Jul. 28, 2021, granted, now 11,294,723.
Claims priority of provisional application 63/214,783, filed on Jun. 25, 2021.
Claims priority of provisional application 63/214,784, filed on Jun. 25, 2021.
Prior Publication US 2022/0413917 A1, Dec. 29, 2022
Int. Cl. G06F 9/50 (2006.01); G06N 3/08 (2023.01); G06F 11/07 (2006.01); G06F 8/71 (2018.01); G08B 21/18 (2006.01); H04L 43/16 (2022.01); H04L 67/10 (2022.01); H04L 67/00 (2022.01); G06F 11/30 (2006.01); G06F 11/34 (2006.01)
CPC G06F 9/5016 (2013.01) [G06F 8/71 (2013.01); G06F 9/5094 (2013.01); G06F 11/079 (2013.01); G06F 11/0721 (2013.01); G06F 11/0769 (2013.01); G06F 11/3006 (2013.01); G06F 11/34 (2013.01); G06F 11/3452 (2013.01); G06N 3/08 (2013.01); G08B 21/182 (2013.01); H04L 43/16 (2013.01); H04L 67/10 (2013.01); H04L 67/34 (2013.01); G06F 2209/501 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method to manage a computing resource for a distributed computing system, comprising:
executing a first test function using the distributed computing system at a first plurality of allocation setpoints for the computing resource;
based on the execution, obtaining one or more performance metrics for the first test function for each setpoint of the first plurality of allocation setpoints;
training a machine learning model based on the obtained one or more performance metrics, wherein training the machine learning model comprises fitting a function to each of the obtained performance metrics over the first plurality of allocation setpoints;
utilizing the trained machine learning model to manage the computing resource for a second function, wherein utilizing the trained machine learning model to manage the computing resource for the second function comprises:
receiving the second function to be optimized for the computing system;
executing the second function at a second plurality of allocation setpoints for the computing resource, wherein the second plurality of allocation setpoints comprises a fewer number of setpoints compared to the first plurality of allocation setpoints;
obtaining the one or more performance metrics for the second function for each of the second plurality of allocation setpoints;
applying the trained machine learning model to the obtained performance metrics for the second function for each of the second plurality of allocation setpoints;
determining an optimal allocation setpoint for the one or more performance metrics based on the trained machine learning model; and
providing a recommendation of a setpoint for the computing resource for the second function based on the determined optimal allocation setpoint.