Regulating power consumption

• US 7,856,549 B2
• Filed: 01/24/2007
• Issued: 12/21/2010
• Est. Priority Date: 01/24/2007
• Status: Active Grant

• Current Assignee: Valtrus Innovations Limited (f/k/a Dolya Holdco 9 Limited) (Key Patent Innovations Limited)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Wheeler, Andrew R.
• Primary Examiner(s): ELAMIN, ABDELMONIEM I
• Application Number:
  US11/657,144
• Publication Number:
  US 20080177424A1
• Time in Patent Office: 1,427 Days
• Field of Search: 713/1, 713/2, 709/223, 718/1
• US Class Current: 713/1
• CPC Class Codes: G06F 1/3203 Power management, i.e. event-based initiation of a power-saving mode
• CPC Class Codes: G06F 9/5077 Logical partitioning of resources; Management or configuration of virtualized resources specific details on emulation or internal functioning of virtual machines G06F9/455
• CPC Class Codes: G06F 9/5094 where the allocation takes into account power or heat criteria power management in computers in general G06F1/3203; thermal management in computers in general G06F1/206
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

A system for dynamically regulating power consumption in an information technology (IT) infrastructure having a plurality of compute nodes interconnected over a network is provided. The system includes at least one virtual machine (VM) host deployed at each of the plurality of compute nodes, the at least one VM host is operable to host at least one VM guest, and the VM hosts on different ones of the plurality of compute nodes are version compatible to enable migration of the VM guests among the VM hosts. The system further includes a management module connected to the plurality of compute nodes over the network to receive a native measurement of a performance metric of a computing resource in each of the plurality of compute nodes, the management module is operable to dynamically regulate power consumption of the plurality of compute nodes by migrating the VM guests among the VM hosts based at least on the received performance metrics of the plurality of compute nodes.

Claims

1. A system for dynamically regulating power consumption in an information technology (IT) infrastructure having a plurality of compute nodes interconnected over a network, the system comprising:

at least one virtual machine (VM) host deployed at each of the plurality of compute nodes, the at least one VM host is operable to host at least one VM guest, and the VM hosts on different ones of the plurality of compute nodes are version compatible to enable migration of the VM guests among the VM hosts; and

a management module connected to the plurality of compute nodes over the network to receive a native measurement of a performance metric of a computing resource in each of the plurality of compute nodes, the management module is operable to dynamically regulate power consumption of the plurality of compute nodes by migrating the VM guests among the VM hosts based at least on the received performance metrics of the plurality of compute nodes,

wherein the management module is operable to determine whether there is sufficient capacity at another compute node of the plurality of compute nodes to accommodate the at least one VM guest from the one compute node and upon the determining that there is sufficient capacity at the another compute node, migrating the at least one VM guest from the one compute node to the another compute node.

2. The system of claim 1, wherein each of the compute nodes comprises a diagnostic unit operable to measure the performance metric from a performance counter native to the computing resource in the each compute node in order to provide the native measurement of the performance metric of the each compute node.

3. The system of claim 2, wherein each of the computer nodes further comprises a utilization agent operable to monitor the native measurement of the performance metric of the each compute node measured by the diagnostic unit.

4. The system of claim 2, wherein the computing resource in the each compute node includes the diagnostic unit therein.

5. The system of claim 2, wherein the management module includes the utilization agent therein.

6. The system of claim 3, wherein the diagnostic unit and the utilization agent are incorporated into a single module.

7. The system of claim 1, wherein at least one of the plurality of compute nodes is a physical computing platform with a different hardware configuration from that of a physical computing platform of another one of the plurality of compute nodes.

8. The system of claim 1, wherein the IT infrastructure comprises a blade server system and at least some of the plurality of compute nodes are server blades in the blade server system.

9. The system of claim 1, wherein each of the plurality of compute nodes is operable to switch among multiple power modes independent of other ones of the plurality of compute nodes.

10. The system of claim 1, wherein the management module is operable to switch one of the plurality of compute nodes to a power down state to conserve power upon a migration of all the VM guests from the one compute node to the another compute node of the plurality of compute nodes.

11. The system of claim 1, wherein the management module is external to the IT infrastructure of which power consumption it dynamically manages.

16. The method of claim 1, wherein providing migration of the at least one VM guest further comprises:

upon the determining that there is not sufficient capacity at the another one compute node, maintaining the at least one VM guest at the one compute node.

12. A method for dynamically regulating power consumption in a system, the method comprising:

identifying a plurality of compute nodes in the system;

implementing at least one virtual machine (VM) host at each of the plurality of compute nodes;

instantiating at least one VM guest on the at least one VM host at each of the plurality of compute nodes;

monitoring a workload demand of the at least one VM guest at each of the plurality of compute nodes and natively-measured performance metrics at the each compute node; and

determining whether the workload demand of the at least one VM guest at one of the plurality of compute nodes exceeds or underutilizes a capacity of the one compute node; and

upon the determining that the workload demand of the at least one VM guest underutilizes the capacity of the one compute node,

a) providing migration of the at least one VM guest to another one of the identified plurality of nodes in the system; and

b) powering down the one underutilized compute node to conserve power consumption once the migration is provided,

wherein providing migration of the at least one VM guest comprises;

determining whether there is sufficient capacity at the another one compute node to accommodate the at least one VM guest from the one compute node; and

upon the determining that there is sufficient capacity at the another one compute node, migrating the at least one VM guest from the one compute node to the another one compute node.

13. The method of claim 12, wherein the additional compute node is another one of the identified plurality of compute nodes in the system.

14. The method of claim 12, wherein the additional compute node is a new compute node additional to the identified plurality of nodes in the system.

15. The method of claim 12, wherein upon the determining that the workload demand of the at least one VM guest exceeds the capacity of the one compute node, powering up an additional compute node to provide additional capacity to accommodate the workload demand of the at least one VM guest.

17. The method of claim 12, wherein powering down the one underutilized compute node comprises one of:

shutting down a power of the one underutilized compute node; and

placing the underutilized compute node in a standby power mode.

18. The method of claim 12, wherein migrating the at least one VM guest from the one compute node to the another one compute node comprises:

migrating the at least one VM guest from the one compute node to the another one compute node based on at least a latency of the one compute node to power back up once in a power-down mode.

19. A computer readable medium on which is encoded program code for dynamically regulating power consumption in a system having a plurality of compute nodes, the encoded program code comprising:

computer program code executed to implement at least one virtual machine (VM) host at each of the plurality of compute nodes;

computer program code executed to instantiate at least one VM guest on the at least one VM host at each of the plurality of compute nodes;

computer program code executed to monitor a workload demand of the at least one VM guest at each of the plurality of compute nodes and natively-measured performance metrics at the each compute node; and

computer program code executed to determine whether the workload demand of the at least one VM guest at one of the plurality of compute nodes exceeds or underutilizes a capacity of the one compute node; and

computer program code executed to, upon the determining that the workload demand of the at least one VM guest underutilizes the capacity of the one compute node,

a) provide migration of the at least one VM guest to another one of the identified plurality of nodes in the system; and

b) power down the one underutilized compute node to conserve power consumption once the migration is provided,

the computer program code executed to provide migration of the at least one VM guest comprises;

computer program code executed to determine whether there is sufficient capacity at the another one compute node to accommodate the at least one VM guest from the one compute node; and

computer program code executed to, upon the determining that there is sufficient capacity at the another one compute node, migrate the at least one VM guest from the one compute node to the another one compute node.