Computer workload redistribution based on prediction from analysis of local resource utilization chronology data

• US 8,104,041 B2
• Filed: 04/24/2006
• Issued: 01/24/2012
• Est. Priority Date: 04/24/2006
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Belady, Christian L., Vaden, Thomas L, Herington, Daniel Edward, Turicchi, Thomas Edwin Jr.
• Primary Examiner(s): An, Meng
• Assistant Examiner(s): Kawsar, Abdullah Al
• Application Number:
  US11/409,853
• Publication Number:
  US 20070250838A1
• Time in Patent Office: 2,101 Days
• Field of Search: None
• US Class Current: 718/105
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3433 for load management allocation of a server based on load conditions G06F9/505; load rebalancing G06F9/5083; redistributing the load in a network by a load balancer H04L67/1029
• CPC Class Codes: G06F 11/3452 Performance evaluation by statistical analysis
• CPC Class Codes: G06F 11/3466 Performance evaluation by tracing or monitoring

Abstract

The present invention provides for redistributing workloads among computers to optimize resource utilization. Utilization by software workloads of computer resources is monitored to yield utilization data. A utilization chronology is updated using the utilization data. The chronology is analyzed to yield resource utilization predictions. The workloads are redistributed among the resources at least in part as function of said predictions.

Claims

1. A computer-implemented method comprising:

a computer system monitoring utilization of computer resources by software workloads to yield local utilization data;

said computer system updating a utilization chronology using said local utilization data;

said computer system analyzing said chronology to yield resource utilization predictions; and

said computer system redistributing said workloads among said resources at least in part as a function of said predictions, said redistributing involving migrating at least a first workload to a resource that is also serving a second workload wherein said first and second workloads have periodic peaks of the same periodicity and different phases.

2. A method as recited in claim 1 wherein said local utilization data is time-stamped, said chronology associating hourly, daily, weekly, monthly, and yearly utilization data with server and workload configuration data.

3. A method as recited in claim 1 further comprising powering down or causing to enter a low-power state a computing resource from which a workload has been removed by said redistributing.

4. A method as recited in claim 3 wherein the monitoring, updating, analyzing, and redistributing are iterated so that further redistributing involves causing said computer resource to enter a high-power state and migrating a workload to said computer resource.

5. A method as recited in claim 1 further comprising issuing an alarm as a function of said local utilization data, said redistributing being performed in response to said alarm.

6. A method as recited in claim 1 wherein an orchestrator program in non-transitory computer readable media, when executed, coordinates the monitoring, analyzing, and redistributing.

7. A system comprising:

hardware resources; and

non-transitory computer readable media including

workloads distributed for utilizing respective combinations of said hardware resources;

a monitoring program for monitoring utilization of said computer resources by said workloads to yield local utilization data; and

a workload manager for updating a utilization chronology using said local utilization data, for analyzing said chronology to yield resource utilization predictions, and for redistributing said workloads among said resources at least in part as a function of said predictions, said redistributing involving migrating at least a first workload to a resource that is also serving a second workload wherein said first and second workloads have periodic leaks of the same periodicity and different phases.

8. A system as recited in claim 7 wherein said local utilization data is time-stamped, said chronology associating hourly, daily, weekly, monthly, and yearly utilization data with server and workload configuration data.

9. A system as recited in claim 7 wherein said workload manager is configured to power down or cause to enter a low-power state a computing resource from which a workload has been removed by said redistributing.

10. A system as recited in claim 9 wherein said workload manager includes means for causing said computing resource to enter a high-performance state and migrating a workload to said computing resource.

11. A system as recited in claim 7 wherein said workload manager provides for issuing an alarm when said local utilization data indicates a criterion has been met, said workload manager providing for redistribution of said workloads in response to said alarm.

12. A computer product comprising non-transitory computer-readable storage media encoded with code defining:

a resource-utilization chronology; and

a workload manager program configured to, when executed by a processor,

monitor utilization of computer resources by software workloads to yield local utilization data;

update said chronology using said local utilization data;

analyze said chronology to yield resource utilization predictions; and

redistribute said workloads among said resources at least in part as a function of said predictions, the redistributing involving migrating at least a first workload to a resource that is also serving a second workload wherein said first and second workloads have periodic peaks of the same periodicity and different phases.

13. A computer product as recited in claim 12 wherein said local utilization data is time-stamped, said chronology associating hourly, daily, weekly, monthly, and yearly utilization data with server and workload configuration data.

14. A computer product as recited in claim 12 wherein said workload manager program is further configured to power down or cause to enter a low-power state a computing resource from which a workload has been removed by the redistributing.

15. A computer product as recited in claim 14 wherein said workload manager program is further configured so that the monitoring, updating, analyzing, and redistributing are iterated so that further redistributing involves causing said computer resource to enter a high-performance state.

16. A computer product as recited in claim 12 wherein said workload manager program is configured to perform said redistribution in response to a received alarm.