HIERARCHICAL DEFRAGMENTATION OF RESOURCES IN DATA CENTERS

• US 20120331147A1
• Filed: 06/23/2011
• Published: 12/27/2012
• Est. Priority Date: 06/23/2011
• Status: Active Grant

• Current Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Original Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Inventors: Dutta, Debojyoti, Ganesan, Ashok, Banerjee, Subrata
• Granted Patent: US 8,914,513 B2
• Time in Patent Office: Days
• US Class Current: 709/226
• CPC Class Codes: G06F 2209/5022 Workload threshold
• CPC Class Codes: G06F 9/06 using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
• CPC Class Codes: G06F 9/5088 involving task migration
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

Techniques are provided herein for defragmenting resources within a cloud computing system. The cloud computing system includes a plurality of servers deployed in a plurality of respective racks, wherein the respective racks are deployed in a pod of a data center. An element of the cloud computing system determines for each server in a given rack of servers a number of free resource slots available thereon and a number of resource slots in an idle state, and then further determines whether the number of free resource slots on a first server in the plurality of servers is greater than a predetermined threshold. When the number of free resource slots in the first server is greater than the predetermined threshold, a second server in the plurality of servers is identified with sufficient resource slots thereon to accommodate the number of resource slots in the idle state on the first server, and the resource slots in the idle state on the first server are caused to be migrated to the second server.

Claims

1. A method comprising:

at a device in a cloud computing system, the cloud computing system comprising a plurality of servers deployed in a plurality of respective racks, wherein the respective racks are deployed in a pod of a data center, determining for each server in a given rack of servers a number of free resource slots available thereon and a number of resource slots in an idle state;

determining whether the number of free resource slots on a first server in the plurality of servers is greater than a predetermined threshold;

when the number of free resource slots on the first server is greater than the predetermined threshold, identifying a second server in the plurality of servers with sufficient resource slots thereon to accommodate the number of resource slots in the idle state on the first server; and

causing the resource slots in the idle state on the first server to be migrated to the second server.

2. The method of claim 1, wherein identifying the second server comprises identifying the second server within a same rack as the first server.

3. The method of claim 1, further comprising taking a snapshot of a state of the first server before resource slots are migrated from the first server to the second server.

4. The method of claim 1, further comprising before causing the resource slots in the idle state on the first server to be migrated to the second server, determining whether migrating the resource slots in the idle state on the first server violates any policy governing the operation of the resource slots in the idle state on the first server.

5. The method of claim 1, further comprising, after resource slots in the idle state on the first server are migrated to the second server, causing a rack in which the first server is located to be powered down.

6. The method of claim 1, further comprising, after resource slots in the idle state on the first server are migrated to the second server, causing a pod in which the first server is located to be powered down.

7. The method of claim 1, further comprising determining a utilization bandwidth for a first rack in which the first server is located, and migrating resources from the first rack to a second rack only when an available bandwidth in the second rack can accommodate the utilization bandwidth for the first rack.

8. The method of claim 1, wherein the free resource slots or the resource slots in an idle state comprises storage resources.

9. One or more computer-readable storage media storing instructions that, when executed by a processor, cause the processor to:

at a device in a cloud computing system, the cloud computing system comprising a plurality of servers deployed in a plurality of respective racks, wherein the respective racks are deployed in a pod of a data center, determine for each server in the plurality of servers a number of free resource slots available thereon and a number of resource slots in an idle state;

determine whether the number of free resource slots on a first server in the plurality of servers is greater than a predetermined threshold;

when the number of free resource slots in the first server is greater than the predetermined threshold, identify a second server in the plurality of servers with sufficient resource slots thereon to accommodate the number of resource slots in the idle state on the first server; and

cause the resource slots in the idle state on the first server to be migrated to the second server.

10. The computer-readable storage media of claim 9, wherein the instructions, when executed, cause the processor to identify the second server within a same rack as the first server.

11. The computer-readable storage media of claim 9, wherein the instructions, when executed, cause the processor to take a snapshot of a state of the first server before resource slots are migrated from the first server to the second server.

12. The computer-readable storage media of claim 9, wherein the instructions, when executed, cause the processor to before causing the resource slots in the idle state on the first server to be migrated to the second server, determine whether migrating the resource slots in the idle state on the first server violates any policy governing the operation of the resource slots in the idle state on the first server.

13. The computer-readable storage media of claim 9, wherein the instructions, when executed, cause the processor to add to a list of free slots each free resource slot available on the first server, and add to a list of candidate move slots each resource slot in the idle state.

14. The computer-readable storage media of claim 9, wherein the instructions, when executed, cause the processor to, after resource slots in the idle state on the first server are migrated to the second server, cause a rack in which the first server is located to be powered down.

15. The computer-readable storage media of claim 9, wherein the instructions, when executed, cause the processor to, after resource slots in the idle state on the first server are migrated to the second server, cause a pod in which the first server is located to be powered down.

16. The computer-readable storage media of claim 9, wherein the instructions, when executed, cause the processor to determine a utilization bandwidth for a first rack in which the first server is located, and migrate resources from the first rack to a second rack only when an available bandwidth in the second rack can accommodate the utilization bandwidth for the first rack.

17. An apparatus comprising:

a network interface unit configured to perform communications over a network;

a processor configured to be coupled to the network interface unit, wherein the processor is configured to;

determine for each server in a plurality of servers a number of free resource slots available thereon and a number of resource slots in an idle state;

determine whether the number of free resource slots on a first server in the plurality of servers is greater than a predetermined threshold;

when the number of free resource slots in the first server is greater than the predetermined threshold, identify a second server in the plurality of servers with sufficient resource slots thereon to accommodate the number of resource slots in the idle state on the first server; and

cause the resource slots in the idle state on the first server to be migrated to the second server.

18. The apparatus of claim 17, wherein the processor is further configured to identify the second server within a same rack as the first server.

19. The apparatus of claim 17, wherein the processor is further configured to before causing the resource slots in the idle state on the first server to be migrated to the second server, determine whether migrating the resource slots in the idle state on the first server violates any policy governing the operation of the resource slots in the idle state on the first server.

20. The apparatus of claim 17, wherein the processor is further configured to determine a utilization bandwidth for a first rack in which the first server is located, and migrate resources from the first rack to a second rack only when an available bandwidth in the second rack can accommodate the utilization bandwidth for the first rack.