Data center system that accommodates episodic computation

• US 8,849,469 B2
• Filed: 10/28/2010
• Issued: 09/30/2014
• Est. Priority Date: 10/28/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Belady, Christian L., Larus, James R., Reed, Danny A., Borgs, Christian H., Chayes, Jennifer Tour, Lobel, Ilan, Menache, Ishai, Nazerzadeh, Hamid, Jain, Navendu
• Primary Examiner(s): Singh, Gurkanwaljit
• Application Number:
  US12/913,763
• Publication Number:
  US 20120109705A1
• Time in Patent Office: 1,433 Days
• Field of Search: None
• US Class Current: 700/291
• CPC Class Codes: G06F 1/3203 Power management, i.e. event-based initiation of a power-saving mode
• CPC Class Codes: G06F 1/329 by task scheduling
• CPC Class Codes: G06F 9/5027 the resource being a machine, e.g. CPUs, Servers, Terminals
• CPC Class Codes: G06F 9/505 considering the load
• 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: G06Q 10/06312 Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
• CPC Class Codes: G06Q 10/06315 Needs-based resource requirements planning or analysis
• CPC Class Codes: G06Q 30/0283 Price estimation or determination
• CPC Class Codes: G06Q 30/04 Billing or invoicing
• CPC Class Codes: G06Q 40/04 Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
• CPC Class Codes: G06Q 50/06 Energy or water supply
• CPC Class Codes: H02J 3/008 involving trading of energy or energy transmission rights
• CPC Class Codes: H04L 45/00 Routing or path finding of packets in data switching networks routing or path finding in wireless networks H04W40/00
• CPC Class Codes: H04L 67/1004 Server selection for load balancing
• CPC Class Codes: Y02B 70/10 Technologies improving the efficiency by using switched-mode power supplies [SMPS], i.e. efficient power electronics conversion e.g. power factor correction or reduction of losses in power supplies or efficient standby modes
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management
• CPC Class Codes: Y02P 90/82 Energy audits or management systems therefor
• CPC Class Codes: Y02P 90/90 Financial instruments for climate change mitigation, e.g. environmental taxes, subsidies or financing
• CPC Class Codes: Y04S 10/50 Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
• CPC Class Codes: Y04S 50/12 Billing, invoicing, buying or selling transactions or other related activities, e.g. cost or usage evaluation
• CPC Class Codes: Y04S 50/14 : Marketing, i.e. market research and analysis, surveying, promotions, advertising, buyer profiling, customer management or rewards

Abstract

A data center system is described which includes multiple data centers powered by multiple power sources, including any combination of renewable power sources and on-grid utility power sources. The data center system also includes a management system for managing execution of computational tasks by moving data components associated with the computational tasks within the data center system, in lieu of, or in addition to, moving power itself. The movement of data components can involve performing pre-computation or delayed computation on data components within any data center, as well as moving data components between data centers. The management system also includes a price determination module for determining prices for performing the computational tasks based on different pricing models. The data center system also includes a “stripped down” architecture to complement its use in the above-summarized data-centric environment.

Claims

1. A data center system, comprising:

a plurality of data centers configured to receive power from a plurality of power sources, at least one power source comprising a renewable power source that provides power to a first data center of the plurality of data centers;

a management system configured to manage execution of a plurality of computational tasks, an individual computational task from the plurality of computational tasks having multiple data components associated therewith, wherein the management system is further configured to;

monitor power capacities at the plurality of power sources to detect a current or future instance when a power capacity provided by the renewable power source to the first data center is deficient, wherein, in the current or future instance, the first data center has processed at least some of the multiple data components of the individual computational task and other data components of the individual computational task remain unprocessed, and

transfer at least some of the other data components of the individual computational task that remain unprocessed from the first data center that receives power from the renewable power source with deficient power capacity to a second data center that receives power from a second power source with adequate power capacity; and

a processor configured to execute at least the management system.

2. The data center system of claim 1, wherein the renewable power source is co-located with the first data center.

3. The data center system of claim 1, wherein:

the renewable power source is not connected to an established power distribution grid, and

the plurality of power sources also includes at least one on-grid utility power source that is connected to the established power distribution grid.

4. The data center system of claim 1, wherein the management system is further configured to move at least one of the multiple data components by adjusting a time at which the at least one of the multiple data components is to be performed.

5. The data center system of claim 4, wherein said at least one of the multiple data components is originally designated to be performed at time t, and wherein the management system is configured to move said at least one of the multiple data components to a time prior to the time t.

6. The data center system of claim 4, wherein said at least one of the multiple data components is originally designated to be performed at time t, wherein the management system is configured to move said at least one of the multiple data components to a time after the time t.

7. The data center system of claim 4, wherein the management system is configured to suspend execution of said at least one of the multiple data components, and to later resume execution of said at least one of the multiple data components.

8. The data center system of claim 1, wherein the management system is configured to:

analyze the individual computational task to identify separable data components of the multiple data components that can be decoupled to perform the individual computational task;

parse the individual computational task into the separable data components; and

obtain the at least some of the other data components that are transferred to the second data center from the separable data components.

9. The data center system of claim 1, wherein the management systems is configured to transfer the at least some of the other data components of the individual computational task that remain unprocessed based on a consideration of one or more of:

monetary costs of performing computation in different data centers of the plurality of data centers;

environmental costs of performing computation in the different data centers;

power usage effectiveness (PUE) metrics associated with performing computation in the different data centers;

types of resources provided by the different data centers; and

legal constraints associated with performing computation in the different data centers.

10. The data center system of claim 1, wherein:

individual data centers convert input data components of the plurality of computational tasks into output data components,

the individual data centers store intermediate data components, and

the management system is configured to manage the execution of the plurality of computational tasks based on consideration of at least one of;

costs of processing the input data components by different data centers at different times;

costs of transferring the input data components among the different data centers;

costs of storing the intermediate data components by the different data centers; and

costs of transferring the intermediate data components between the different data centers.

11. The data center system of claim 1, wherein the management system includes a price determination module configured to determine prices associated with execution of the plurality of computational tasks within the data center system, the prices including an individual price for the individual computational task and another individual price for another individual computational task having other data components.

12. The data center system of claim 11, wherein the price determination module is configured to determine the individual price based on at least one characteristic of the renewable power source.

13. The data center system of claim 11, wherein the price determination module is configured to determine the another individual price based on whether the another individual computational task involves movement of an individual other data component of the another individual computational task.

14. The data center system of claim 11, wherein the price determination module is configured to determine the another individual price for the another individual computational task according to a defined first level of service, and wherein the management system is configured to satisfy the first level of service but not exceed the first level of service when performing the another individual computational task.

15. The data center system of claim 14, wherein the management system is configured to move an individual other data component of the another individual computational task to accommodate execution of a further computational task having a second price and a corresponding second level of service, wherein the second price is higher than the another individual price and the second level of service is higher than the first level of service.

16. The data center system of claim 11, wherein the price determination module is configured to determine the individual price for the individual computational task based on Service Level Agreement information associated with the individual computational task, the Service Level Agreement information specifying one or more of:

a throughput metric that defines a desired throughput performance of the individual computational task;

a latency metric that defines a desired latency performance of the individual computational task; and

a reliability metric that defines a desired reliability performance of the individual computational task.

17. The data center system of claim 16, wherein the latency metric comprises one or more of:

an intra-center latency metric which defines a desired intra-center latency within the first data center; and

an end-to-end latency metric which defines a desired end-to-end latency between submission of the individual computational task and a response received following execution of the individual computational task.

18. The data center system of claim 11, wherein the price determination module comprises:

a supply determination module configured to determine a supply curve associated with availability of a computational resource within the data center system over a span of time;

a demand determination module configured to determine a demand curve that expresses a demand for the computational resource as a function of price associated with the computational resource, for a particular time; and

a price assessment module configured to determine a price at which to provide the computational resource at the particular time based on the supply curve and the demand curve.

19. The data center system of claim 18, wherein the supply determination module includes a prediction module configured to predict the supply curve based on historical information, and the demand determination module includes another prediction module configured to predict the demand curve based on other historical information.

20. The data center system of claim 11, wherein the price determination module comprises:

an auction module configured to determine the prices based on bids received by entities; and

a bid assistance module configured to assist the entities in submitting bids to the auction module.

21. The data center system of claim 11, wherein the price determination module is configured to determine the another individual price for the another individual computational task based on a consideration of costs associated with execution of subtasks associated with the another individual computational task by different respective data centers.

22. The data center system of claim 11, wherein the price determination module is configured to generate and provide an itemized bill to an entity that identifies costs associated with performing the another individual computational task using one or more of the plurality of data centers.

23. The data center system of claim 1, wherein the first data center and the renewable power source eliminate one or more of:

power backup mechanisms;

redundant components;

power converter components; and

housing components.

24. A method comprising:

monitoring a power capacity of a renewable off-grid power source that is not connected to an established power distribution grid, wherein the renewable off-grid power source supplies power to a first data center;

executing a computational task at the first data center by processing multiple data components associated with the computational task; and

responsive to the monitoring showing that the power capacity of the renewable off-grid power source is deficient, transferring at least some of the multiple data components to a second data center that receives power from another power source with adequate power capacity, wherein the transferring is performed after execution of the computational task begins but before all of the multiple data components have been processed,

wherein at least the transferring is performed by a computing device.

25. The method of claim 24, further comprising selecting the second data center for the transferring based on power cost.

26. The method of claim 24, further comprising:

analyzing the computational task to identify separable data components of the multiple data components that can be decoupled from other data components of the computational task;

parsing the computational task into the separable data components based on the analyzing; and

obtaining the at least some of the multiple data components from the separable data components.

27. A computer readable memory device or storage device storing computer readable instructions that, when executed by one or more processing devices, cause the one or more processing devices to perform acts comprising:

monitoring power capacities of multiple power sources supplying power to multiple data centers in a data center system, wherein the multiple power sources include an episodic power source that supplies power to a first data center in the data center system; and

responsive to the monitoring showing that power capacity of the episodic power source is deficient, transferring execution of an individual computational task from the first data center supplied by the episodic power source to a second data center in the data center system that is not supplied by the episodic power source, wherein;

the transferring comprises transferring at least one partial result associated with the individual computational task from the first data center to the second data center, and

the transferring is performed after the first data center begins execution of the individual computational task.

28. The computer readable memory device or storage device of claim 27, the acts further comprising:

selecting the second data center for the transferring based on one or more characteristics of a second power source available to the second data center, the characteristics comprising one or more of power availability, power type, power cost, and cost of transferring the execution of the individual computational task.

29. The computer readable memory device or storage device of claim 27, the acts further comprising:

ensuring that the transferring maintains conformance with a service level agreement.