US 12,093,714 B2
System and method for estimation of performance impact upon a hypervisor in a cloud environment
Achintya Guchhait, Bengaluru (IN); Adithya Shreedhar, Bengaluru (IN); and Sriram Gummuluru, Bengaluru (IN)
Assigned to ORACLE INTERNATIONAL CORPORATION, Redwood Shores, CA (US)
Filed by ORACLE INTERNATIONAL CORPORATION, Redwood Shores, CA (US)
Filed on Jul. 20, 2021, as Appl. No. 17/380,750.
Prior Publication US 2023/0031963 A1, Feb. 2, 2023
Int. Cl. G06F 9/455 (2018.01); G06F 9/48 (2006.01); G06F 9/50 (2006.01); G06F 11/34 (2006.01); G06N 20/00 (2019.01)
CPC G06F 9/45558 (2013.01) [G06F 9/455 (2013.01); G06F 9/48 (2013.01); G06F 9/50 (2013.01); G06F 9/5072 (2013.01); G06F 9/5077 (2013.01); G06F 9/5083 (2013.01); G06F 9/5088 (2013.01); G06F 11/3409 (2013.01); G06F 11/3447 (2013.01); G06N 20/00 (2019.01); G06F 2009/4557 (2013.01); G06F 2009/45583 (2013.01); G06F 2009/45591 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system for use with a cloud or other computing environment, for estimation of performance impact upon a hypervisor socket provided within such environment, and use thereof, comprising:
a computer including one or more processors, and a cloud computing environment having hardware and software resources, and a hypervisor operating thereon, said cloud computing environment comprising a provisioning service that clients use to provision and manage compute instances that are launched to meet compute and application requirements;
wherein the system includes a data defining a machine learning model, said machine learning model generated by the system by:
determining a set of virtual machine configurations and workloads;
providing to a workload processor an indication of a particular virtual machine and hypervisor configuration for workload testing over a number of test runs;
executing the workloads over the test runs comprising one or multiple virtual machine neighbors having a same configuration and sharing a socket of the hypervisor; and
during each test run, collecting performance data from a set of performance counters associated with a measured drop in performance, which performance data is used to train the machine learning model indicative of a predicted performance degradation for virtual machines due to neighboring virtual machines;
wherein said machine learning model is adapted for use by the system to;
generate a score value that provides a predicted measure of performance drop that affected virtual machines may experience at a particular point in time, due to activity of neighboring virtual machines on one or more sockets of the hypervisor; and
based on the score value, determining a deployment or placement of virtual machines within the cloud environment to facilitate sharing of resources.