Granted Patent B2

US 10,103,956 · App. 14/501,820 · Granted Oct 16, 2018

Abstract

Claims (30)

1. A computer implemented method comprising:

receiving virtual machine utilization data per unit time in a time window for at least two virtual machines, the at least two virtual machines in a first configuration on at least two servers in a server cluster;

determining a CPU utilization metric and a memory utilization metric for each of the at least two virtual machines in the first configuration, wherein the CPU utilization metric comprises a prime utilization of CPU utilization, wherein the prime utilization comprises a mean utilization per unit time within the time window plus one standard deviation of utilization for each corresponding unit time;

simulating a second configuration of the at least two virtual machines on the at least two servers in the server cluster, wherein the second configuration comprises a movement of not more than a predetermined number of the at least two virtual machines between the at least two servers in the cluster;

determining a CPU utilization metric and a memory utilization metric for the at least two virtual machines corresponding to the simulated second configuration, wherein the memory utilization metric comprises a prime utilization of memory utilization;

determining, by a processing device, a first net utilization metric based on a first combination of the CPU utilization metric and the memory utilization metric for each of the at least two servers of the first configuration and determining a second net utilization metric based on a second combination of the CPU utilization metric and the memory utilization metric for each of the at least two servers of the simulated second configuration, wherein configurable weighting values are applied to the CPU utilization metric and the memory utilization metric in the first and second combinations to determine the first and second net utilization metrics;

comparing the first net utilization metric to the second net utilization metric; and

responsive to the second net utilization metric having a lower value that the first net utilization metric, reconfiguring the at least two virtual machines on the at least two servers according to the second configuration.

2. The method of claim 1 , wherein the CPU utilization metric further comprises a peak utilization of at least one of CPU utilization and the memory utilization metric comprises a peak utilization of memory utilization, wherein the peak utilization comprises a highest combined utilization for the at least two virtual machines in the time window.

3. The method of claim 1 , further comprising providing a recommendation of one of the first configuration and the second configuration.

4. The method of claim 1 , wherein the second configuration is determined by executing at least one swap function.

5. The method of claim 4 , wherein the at least one swap function comprises an N to N swap moving a number N of the at least two virtual machines from a first randomly selected server to a second randomly selected server and moving the number N of the at least two virtual machines from the second randomly selected server to the first randomly selected server.

6. The method of claim 4 , wherein the at least one swap function comprises a problem reduction swap for separating a first virtual machine on a first server from a correlated second virtual machine on the first server.

7. The method of claim 4 , wherein the at least one swap function comprises a chunk swap for exchanging at least one of the two or more virtual machines from a first server to a second server and at least one of the two or more virtual machines from the second server to the first server, the exchanged virtual machines of the first server and the second server having approximately the same resource utilization.

8. The method of claim 4 , wherein the at least one swap function comprises a friend swap for moving a first virtual machine having a first resource utilization pattern from a first server to a second server, the second server having a second virtual machine with a second resource utilization pattern that destructively interferes with the first utilization pattern of the first virtual machine.

9. A non-transitory computer-readable storage medium that includes instructions that, when loaded into memory, cause a processor to:

receive virtual machine utilization data per unit time in a time window for at least two virtual machines, the at least two virtual machines in a first configuration on at least two servers in a server cluster;

determine a CPU utilization metric and a memory utilization metric for each of the at least two virtual machines in the first configuration, wherein the CPU utilization metric comprises a prime utilization of CPU utilization, wherein the prime utilization comprises a mean utilization per unit time within the time window plus one standard deviation of utilization for each corresponding unit time;

simulate a second configuration of the at least two virtual machines on the at least two servers in the server cluster, wherein the second configuration comprises a movement of not more than a predetermined number of the at least two virtual machines between the at least two servers in the cluster;

determine a CPU utilization metric and a memory utilization metric for the at least two virtual machines corresponding to the simulated second configuration, wherein the memory utilization metric comprises a prime utilization of memory utilization;

determine a first net utilization metric based on a first combination of the CPU utilization metric and the memory utilization metric for each of the at least two servers of the first configuration and determining a second net utilization metric based on a second combination of the CPU utilization metric and the memory utilization metric for each of the at least two servers of the simulated second configuration, wherein configurable weighting values are applied to the CPU utilization metric and the memory utilization metric in the first and second combinations to determine the first and second net utilization metrics;

compare the first net utilization metric to the second net utilization metric; and

responsive to the second net utilization metric having a lower value that the first net utilization metric, reconfigure the at least two virtual machines on the at least two servers according to the second configuration.

10. The non-transitory computer-readable storage medium of claim 9 , wherein the CPU utilization metric further comprises a peak utilization of at least one of CPU utilization and the memory utilization metric comprises a peak utilization of memory utilization, wherein the peak utilization comprises a highest combined utilization for the at least two virtual machines in the time window.

11. The non-transitory computer-readable storage medium of claim 9 , wherein the instructions further cause the processor to provide a recommendation of one of the first configuration and the second configuration.

12. The non-transitory computer-readable storage medium of claim 9 , wherein the second configuration is determined by executing at least one swap function.

13. The non-transitory computer-readable storage medium of claim 12 , wherein the at least one swap function comprises an N to N swap moving a number N of the at least two virtual machines from a first randomly selected server to a second randomly selected server and moving the number N of the at least two virtual machines from the second randomly selected server to the first randomly selected server.

14. The non-transitory computer-readable storage medium of claim 12 , wherein the at least one swap function comprises a problem reduction swap for separating a first virtual machine on a first server from a correlated second virtual machine on the first server.

15. The non-transitory computer-readable storage medium of claim 12 , wherein the at least one swap function comprises a chunk swap for exchanging at least one of the two or more virtual machines from a first server to a second server and at least one of the two or more virtual machines from the second server to the first server, the exchanged virtual machines of the first server and the second server having approximately the same resource utilization.

16. The non-transitory computer-readable storage medium of claim 12 , wherein the at least one swap function comprises a friend swap for moving a first virtual machine having a first resource utilization pattern from a first server to a second server, the second server having a second virtual machine with a second resource utilization pattern that destructively interferes with the first utilization pattern of the first virtual machine.

Assignments (8)