Methods and apparatus for selective workload off-loading across multiple data centers

• US 20060271544A1
• Filed: 05/27/2005
• Published: 11/30/2006
• Est. Priority Date: 05/27/2005
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Rajamani, Nithya, Devarakonda, Murthy V., Naik, Vijay K., Pacifici, Giovanni, Rosu, Daniela, Dias, Daniel Manuel, Dixon, Graeme N.
• Granted Patent: US 7,853,953 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 2209/509 Offload
• CPC Class Codes: G06F 9/5027 the resource being a machine, e.g. CPUs, Servers, Terminals

Abstract

Techniques for selectively off-loading workload across multiple data centers are disclosed. For example, a technique for processing a user request in accordance with a multiple data center environment comprises the following steps/operations. A user request is obtained at a first data center. The user request is classified based on one or more classification criterion. At least a portion of a workload associated with the classified user request is off-loaded to at least a second data center to be processed, wherein the off-loading step/operation is at least based on one or more administrative policies associated with at least one of the first data center and the second data center. The workload is processed such that the user request is at least substantially satisfied in accordance with a service level agreement for the requesting user.

Claims

1. A method of processing a user request in accordance with a multiple data center environment, comprising the steps of:

obtaining, at a first data center, a user request;

classifying the user request based on one or more classification criterion;

off-loading at least a portion of a workload associated with the classified user request to at least a second data center to be processed, wherein the off-loading step is at least based on one or more administrative policies associated with at least one of the first data center and the second data center; and

processing the workload such that the user request is at least substantially satisfied in accordance with a service level agreement for the requesting user.

2. The method of claim 1, wherein the off-loading step further comprises redirecting at least a portion of a workload associated with the classified user request to the second data center based on an admission control policy associated with the first data center.

3. The method of claim 1, wherein the off-loading step further comprises causing at least one resource associated with the second data center to be made available to a resource scheduler associated with the first data center such that at least a portion of a workload associated with the classified user request can be routed directly to the at least one resource associated with the second data center.

4. The method of claim 1, further comprising the step of determining, at the first data center, one or more metric values associated with one or more benefits to off-loading a portion of the workload.

5. The method of claim 4, wherein the metric value determining step further comprises comparing the metric value to a threshold value and using a result of the comparison as a trigger to initiate workload off-loading.

6. The method of claim 5, wherein the metric value determining step further comprises adjusting the threshold value.

7. The method of claim 6, wherein the threshold value adjusting step further comprises applying a constant decay factor to the threshold value.

8. The method of claim 7, wherein the threshold value adjusting step further comprises resetting the threshold value to an original value.

9. The method of claim 1, further comprising the step of determining from among the multiple data centers which one or more data centers are potential data centers for off-loading at least a portion of the workload.

10. The method of claim 9, wherein the potential data center determining step further comprises determining potential data centers on a priority order basis.

11. The method of claim 9, further comprising the step of sending at least one request query to at least one of the data centers determined to be a potential data center.

12. The method of claim 9, further comprising the step of multicasting a request query to multiple data centers.

13. The method of claim 12, further comprising the step of accepting one of the potential data centers as the data center to be used to off-load, wherein the second data center is the accepted data center.

14. The method of claim 13, further comprising the step of selecting one or more resources to be provisioned.

15. The method of claim 14, further comprising the step of provisioning one or more selected resources for use in processing the workload.

16. The method of claim 15, wherein the resource provisioning step further comprises at least one of the first data center and the second data center performing at least a portion of the resource provisioning step.

17. The method of claim 1, further comprising the step of adjustably setting, at the first data center, a resource request threshold value.

18. The method of claim 1, wherein the off-loading step further comprises the second data center determining its own resource availability before agreeing to lend one or more resources for processing the workload.

19. Apparatus for processing a user request in accordance with a multiple data center environment, comprising:

a memory; and

at least one processing device associated with a first data center, coupled to the memory, and operative to;

(i) obtain a user request;

(ii) classify the user request based on one or more classification criterion; and

(iii) cause to be off-loaded at least a portion of a workload associated with the classified user request to at least a second data center to be processed, wherein the off-loading operation is at least based on one or more administrative policies associated with at least one of the first data center and the second data center;

wherein the workload is processed such that the user request is at least substantially satisfied in accordance with a service level agreement for the requesting user.

20. The apparatus of claim 19, wherein the off-loading operation further comprises redirecting at least a portion of a workload associated with the classified user request to the second data center based on an admission control policy associated with the first data center.

21. The apparatus of claim 19, wherein the off-loading operation further comprises causing at least one resource associated with the second data center to be made available to a resource scheduler associated with the first data center such that at least a portion of a workload associated with the classified user request can be routed directly to the at least one resource associated with the second data center.

22. The apparatus of claim 19, wherein the at least one processing device is further operative to determine one or more metric values associated with one or more benefits to off-loading a portion of the workload.

23. The apparatus of claim 19, wherein the at least one processing device is further operative to determine from among the multiple data centers which one or more data centers are potential data centers for off-loading at least a portion of the workload.

24. The apparatus of claim 19, wherein the at least one processing device is further operative to adjustably set a resource request threshold value.

25. An article of manufacture for use at a first data center in processing a user request in accordance with a multiple data center environment, comprising a machine readable medium containing one or more programs which when executed implement the steps of:

obtaining a user request;

classifying the user request based on one or more classification criterion; and

causing to be off-loaded at least a portion of a workload associated with the classified user request to at least a second data center to be processed, wherein the off-loading operation is at least based on one or more administrative policies associated with at least one of the first data center and the second data center;

wherein the workload is processed such that the user request is at least substantially satisfied in accordance with a service level agreement for the requesting user.