Analysis of energy-related factors for selecting computational job locations

• US 8,086,544 B2
• Filed: 09/03/2008
• Issued: 12/27/2011
• Est. Priority Date: 09/03/2008
• Status: Active Grant

• Current Assignee: Kyndryl Incorporated
• Original Assignee: International Business Machines Corporation
• Inventors: Boss, Gregory J., Doran, James R., Hamilton, Rick A. II, Jackson, Steven V., Sand, Anne R.
• Primary Examiner(s): BORISSOV, IGOR N
• Application Number:
  US12/203,345
• Publication Number:
  US 20100057641A1
• Time in Patent Office: 1,210 Days
• Field of Search: 713/320, 713/300, 713/321, 713/323, 713/330, 713/340, 705/400, 705/412, 705/37.R, 705/7.11, 705/7.12, 705/7.13, 705/7.22, 705/7.23, 705/7.24, 705/7.25, 705/7.26, 705/7.29, 705/7.31, 705/7.32, 705/7.35, 705/7.36, 705/7.37, 705/7.38, 709/229, 709223-226
• US Class Current: 705/400
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
• CPC Class Codes: G06Q 10/063 Operations research, analysis or management
• CPC Class Codes: G06Q 30/0283 Price estimation or determination
• CPC Class Codes: G06Q 50/06 Energy or water supply
• CPC Class Codes: Y02P 90/82 Energy audits or management systems therefor

Abstract

The selection of an optimal data center location for running a computational workload is based on multiple energy criteria. The location is chosen based on multivariate and predictive analysis of total direct and indirect energy costs, and other user-defined factors. Among the direct and indirect costs are power costs and cooling costs as well as structural and other details of a given data center. Among the other factors to be considered that can have an impact on present and future costs are weather patterns, data and forecasts, availability of energy providers, and energy attributes. A forecaster factors these direct and indirect costs along with extrinsic information such as historical trends and predictive sources into a forecast which is then input to a decision engine along with user defined criteria and with anticipated compute tasks and requirements to select a final location or locations for handling the workload.

Claims

1. A system, comprising:

a processing unit, computer readable memory and a computer readable storage medium;

first program instructions to collect energy-related data from each of a plurality of different data center sources, the energy-related data comprising power costs, weather data, energy attributes, and cooling costs for each of the plurality of different data center sources;

second program instructions to predict projected energy consumption forecasts for processing a compute job at each of the different data center sources as a function of a workload requirement of the compute job and the collected energy-related data, wherein the predicted projected energy consumption forecasts are short-term forecasts having a term of prediction spanning only a life of the compute job;

third program instructions to select a data center source location for performing the compute job from the plurality of different data center sources as a function of the predicted projected energy consumption forecasts, feedback data indicative of effectiveness of the predicted projected energy consumption forecasts, and user-defined criteria; and

fourth program instructions to calculate actual costs for the selected data center source at a time that the compute job is run, and to provide additional input data to the feedback data indicative of the effectiveness of the predicted projected energy consumption forecasts by comparing the calculated actual costs to the predicted projected energy consumption forecast for the selected data center source;

wherein the first, second, third and fourth program instructions are stored on the computer readable storage medium for execution by the processing unit via the computer readable memory.

2. The system according to claim 1 wherein the second program instructions are further to predict the projected energy consumption forecasts by simulating a model using a Monte Carlo analysis to randomly generate values for uncertain variables.

3. The system according to claim 1 wherein the second program instructions are further to predict the projected energy consumption forecasts for processing the computational job as a function of a news report of a pending strike in a location of the data center sources.

4. The system according to claim 3 wherein the energy-related data further comprises computer processing speed, data center charges, environmental and business risks, labor rates, age of equipment, supply availability and processing overhead data for each of the plurality of different data center sources;

and

wherein the user-defined criteria comprises a third party data center efficiency rating criteria, a method used to generate energy criteria and a speed of processing criteria.

5. A method for determining a physical location for running a computational workload comprising:

collecting energy-related data from each of a plurality of different data center sources, the collected energy-related data comprising power costs, weather data, energy attributes, and cooling costs for each of the plurality of different data center sources;

predicting by a processing unit a projected energy consumption forecasts for processing a compute job at each of the different data center sources as a function of a workload requirement of the compute job and the collected energy-related data, wherein the predicted projected energy consumption forecasts are short-term forecasts having a term of prediction spanning only a life of the compute job;

selecting by the processing unit a data center source location for performing the compute job from the plurality of different data center sources as a function of the predicted projected energy consumption forecasts, feedback data indicative of effectiveness of the predicted projected energy consumption forecasts, and user-defined criteria; and

the processing unit calculating actual costs for the selected data center source at a time that the compute job is run, and providing additional input data to the feedback data indicative of the effectiveness of the predicted projected energy consumption forecasts by comparing the calculated actual costs to the predicted projected energy consumption forecast for the selected data center source.

6. The method according to claim 5 further comprising:

predicting by the processing unit the projected energy consumption forecast by simulating a model using a Monte Carlo analysis to randomly generate values for uncertain variables.

7. The method according to claim 5 wherein the processing unit predicts the projected energy consumption forecasts for processing the computational job as a function of a news report of a pending strike in a location of the data center sources.

8. The method according to claim 7 wherein the energy-related data further comprises computer processing speed, data center charges, environmental and business risks, labor rates, age of equipment, supply availability and processing overhead data of the plurality of different data center sources;

and

wherein the user-defined criteria comprises a third party data center efficiency rating criteria, a method used to generate energy criteria and a speed of processing criteria.

9. An article of manufacture, comprising:

a computer-readable storage device having computer readable program code embodied therewith, the computer readable program code containing program instructions that, when implemented on a computer processing unit, cause the computer processing unit to;

collect energy-related data from each of a plurality of different data center sources, the collected energy-related data comprising power costs, weather data, energy attributes, and cooling costs for each of the plurality of different data center sources;

predict projected energy consumption forecasts for each of the data center sources as a function of a workload requirement of the compute job and the collected energy-related data, wherein the predicted projected energy consumption forecasts are short-term forecasts having a term of prediction spanning only a life of the compute job;

select a data center source location for performing the compute job from the plurality of different data center sources as a function of the predicted projected energy consumption forecasts, feedback data indicative of effectiveness of the predicted projected energy consumption forecasts, and user-defined criteria; and

calculate actual costs for the selected data center source at a time that the compute job is run, and provide additional input data to the feedback data indicative of the effectiveness of the predicted projected energy consumption forecasts by comparing the calculated actual costs to the predicted projected energy consumption forecast for the selected data center source.

10. The article of manufacture according to claim 9 wherein the program instructions are further to cause the computer processing unit to predict the projected energy consumption forecasts by simulating a model using a Monte Carlo analysis to randomly generate values for uncertain variables.

11. The article of manufacture according to claim 9 wherein the program instructions are further to predict the projected energy consumption forecasts for processing the computational job as a function of a news report of a pending strike in a location of the data center sources.

12. The article of manufacture according to claim 11 wherein the energy-related data further comprises computer processing speed, data center charges, environmental and business risks, labor rates, age of equipment, supply availability and processing overhead data for each of the plurality of different data center sources;

and

wherein the user-defined criteria comprises a third party data center efficiency rating criteria, a method used to generate energy criteria and a speed of processing criteria.