Manager and host-based integrated power saving policy in virtualization systems
A mechanism for a manager and host-based integrated power saving policy in virtualization systems is disclosed. A method of the invention includes receiving configuration and power information of a host machine from a management agent on the host machine, performing a macro-level power saving scheduling algorithm that takes into consideration the received configuration and power information of the host machine, and requesting that the host machine alter a number of active running CPU cores as part of the macro-level power saving scheduling algorithm.
1. A method, comprising:
receiving, by a host controller machine, configuration and power information of a host machine from a management agent on the host machine, the configuration information comprising a total number of CPU cores on the host machine and a number of active running CPU cores on the host machine;
performing, by the host controller machine, a macro-level power saving scheduling algorithm that takes into consideration the received configuration and power information of the host machine; and
requesting, by the host controller machine, that the host machine alter the number of active running CPU cores as part of the macro-level power saving scheduling algorithm.
2. The method of claim 1 , wherein taking into consideration the received configuration and power information of the host machine further comprises considering the total number of CPU cores when at least one of load balancing the host machine or considering the host machine as part of a host selection algorithm.
3. The method of claim 2 , further comprising denoting one or more loads as having a lower priority to the host machine so that the denoted one or more loads run on CPU cores having a lower power consumption.
4. The method of claim 1 , wherein the power information comprises an actual power consumption of each CPU core on the host machine.
5. The method of claim 1 , wherein the requesting that the host machine alter a number of active running CPU cores further comprises requesting the host machine to activate one or more of its inactive CPU cores.
6. The method of claim 1 , wherein the requesting that the host machine alter a number of active running CPU cores further comprises requesting the host machine to shut down one or more of its active running CPU cores.
7. The method of claim 1 , further comprising scheduling one or more virtual machines (VMs) to the host machine based on the macro-level power saving scheduling algorithm.
8. A system, comprising:
a memory;
a processing device communicably coupled to the memory; and
a power policy agent executable from the memory by the processing device, the power policy agent to:
receive configuration and power information of a host machine from a management agent on the host machine, the configuration information comprising a total number of CPU cores on the host machine and a number of active running CPU cores on the host machine;
perform a macro-level power saving scheduling algorithm that takes into consideration the received configuration and power information of the host machine; and
request that the host machine alter a number of active running CPU cores as part of the macro-level power saving scheduling algorithm.
9. The system of claim 8 , wherein taking into consideration the received configuration and power information of the host machine further comprises considering the total number of CPU cores when at least one of load balancing the host machine or considering the host machine as part of a host selection algorithm.
10. The system of claim 9 , wherein the power policy agent further to denote one or more loads as having a lower priority to the host machine so that the denoted one or more loads run on CPU cores having a lower power consumption.
11. The system of claim 8 , wherein the power information comprises an actual power consumption of each CPU core on the host machine.
12. The system of claim 8 , wherein the requesting that the host machine alter a number of active running CPU cores further comprises at least one of requesting the host machine to activate one or more of its inactive CPU cores or requesting the host machine to shut down one or more of its active running CPU cores.
13. The system of claim 8 , further comprising scheduling one or more virtual machines (VMs) to the host machine based on the macro-level power saving scheduling algorithm.
14. A non-transitory machine-readable storage medium including data that, when accessed by a processing device, cause the processing device to perform operations comprising:
receiving, by the processing device, configuration and power information of a host machine comprising the processing device from a management agent on the host machine, the configuration information comprising a total number of CPU cores on the host machine and a number of active running CPU cores on the host machine;
performing a macro-level power saving scheduling algorithm that takes into
consideration the received configuration and power information of the host machine; and
requesting that the host machine alter a number of active running CPU cores as part of the macro-level power saving scheduling algorithm.
15. The non-transitory machine-readable storage medium of claim 14 , wherein taking into consideration the received configuration and power information of the host machine further comprises considering the total number of CPU cores when at least one of load balancing the host machine or considering the host machine as part of a host selection algorithm.
16. The non-transitory machine-readable storage medium of claim 15 , wherein the operations further comprise denoting one or more loads as having a lower priority to the host machine so that the denoted one or more loads run on CPU cores having a lower power consumption.
17. The non-transitory machine-readable storage medium of claim 14 , wherein the requesting that the host machine alter a number of active running CPU cores further comprises at least one of requesting the host machine to activate one or more of its inactive CPU cores or requesting the host machine to shut down one or more of its active running CPU cores.
18. The non-transitory machine-readable storage medium of claim 14 , wherein the power information comprises an actual power consumption of each CPU core on the host machine.
19. The non-transitory machine-readable storage medium of claim 14 , wherein the operations further comprise scheduling one or more virtual machines (VMs) to the host machine based on the macro-level power saving scheduling algorithm.