Dynamic application placement based on cost and availability of energy in datacenters

• US 9,207,993 B2
• Filed: 05/13/2010
• Issued: 12/08/2015
• Est. Priority Date: 05/13/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Jain, Navendu
• Primary Examiner(s): SISON, JUNE Y
• Application Number:
  US12/779,059
• Publication Number:
  US 20110282982A1
• Time in Patent Office: 2,035 Days
• Field of Search: 709/223
• US Class Current: 1/1
• 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: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

An optimization framework for hosting sites that dynamically places application instances across multiple hosting sites based on the energy cost and availability of energy at these sites, application SLAs (service level agreements), and cost of network bandwidth between sites, just to name a few. The framework leverages a global network of hosting sites, possibly co-located with renewable and non-renewable energy sources, to dynamically determine the best datacenter (site) suited to place application instances to handle incoming workload at a given point in time. Application instances can be moved between datacenters subject to energy availability and dynamic power pricing, for example, which can vary hourly in day-ahead markets and in a time span of minutes in realtime markets.

Claims

1. A computer-implemented method comprising:

for hosting sites that run application instances to process workload;

tracking realtime energy parameters associated with the hosting sites and energy resources for the hosting sites, and

tracking realtime availability of computing resources associated with the hosting sites;

constructing a framework including;

a hosting sites model that associates the realtime energy parameters and the realtime availability of the computing resources with operation of the hosting sites;

an application instances model that associates the application instances with the computing resources and the energy resources, wherein the application instances model includes power footprints and memory footprints for individual application instances;

solving the framework to output decisions based on the hosting sites model and the application instances model, wherein the solving considers the realtime energy parameters, the realtime availability of the computing resources, and the power footprints and the memory footprints for the individual application instances; and

automatically migrating at least some of the application instances between the hosting sites based on the decisions from the solving the framework.

2. The method of claim 1, wherein the solving the framework further considers service level agreements associated with the application instances.

3. The method of claim 1, wherein the solving the framework further considers costs of the automatically migrating, other costs associated with network bandwidth, and network latencies between the hosting sites.

4. The method of claim 1, wherein the realtime energy parameters include costs and availability of the energy resources, and the energy resources include both renewable and non-renewable energy sources.

5. The method of claim 1, wherein the realtime energy parameters include both temporal and geographical dynamic variations in energy prices of the energy resources.

6. The method of claim 1, further comprising:

simulating future availability and cost of the energy resources based on historical energy pricing and availability data,

wherein the solving the framework further considers the simulated future availability and cost of the energy resources.

7. The method of claim 6, wherein the energy resources include both renewable and non-renewable energy resources.

8. The method of claim 1, further comprising:

redirecting incoming user requests to selected hosting sites based on the decisions from the solving the framework.

9. The method of claim 1, further comprising:

managing power states of the hosting sites based on the decisions.

10. A power control system comprising:

a processor and a storage device;

a data collector stored on the storage device and executable by the processor, the data collector configured to collect computing parameters related to a plurality of datacenter servers, the computing parameters reflecting availability of computing elements;

a simulator stored on the storage device and executable by the processor, the simulator configured to;

collect historical parameters and realtime energy parameters of the datacenter servers reflecting availability of renewable and non-renewable energy sources and dynamic variations in energy prices, and

calculate data points relating to the historical parameters and the realtime energy parameters for designated time intervals;

a data processing engine stored on the storage device and executable by the processor, the data processing engine configured to correlate the computing parameters and the data points for the designated time intervals from the data collector and the simulator to produce correlated data related to the designated time intervals;

a decision component stored on the storage device and executable by the processor, the decision component configured to receive the correlated data from the data processing engine and to make decisions based on a framework including a hosting site model, an application instance model, and a location energy model, the decisions related to power state changes of the hosting sites associated with the designated time intervals; and

a power state actuator stored on the storage device and executable by the processor, the power state actuator configured to receive the decisions from the decision component and to manage power state changes associated with the received decisions within the designated time intervals.

11. The power control system of claim 10, wherein the simulator is further configured to collect the realtime energy parameters by computing the availability of an individual energy source based on predicted power available from the individual energy source for a specific datacenter server for a specific time.

12. The power control system of claim 10, wherein the decisions made by the decision component are based on reducing monetary costs of operating the plurality of datacenter servers.

13. The power control system of claim 10, wherein the decisions made by the decision component are based on reducing impact on user interactiveness.

14. The power control system of claim 10, wherein the decisions made by the decision component are based on increasing long term availability of the computing elements.

15. The power control system of claim 10, wherein the decisions made by the decision component are based on increasing use of the renewable energy sources.

16. The power control system of claim 10, wherein the decisions made by the decision component are based on reducing carbon footprint and pollutant emission due to electricity used in operation of the plurality of datacenter servers.

17. The power control system of claim 10, wherein the decisions made by the decision component are based on balancing bandwidth cost and energy cost.

18. The power control system of claim 10, wherein the application instance model is based on information related to at least one of processor usage and memory footprint size of an individual application instance, number of network connections maintained by the individual application instance, power footprint of the individual application instance, service level agreements on hosting of the individual application instance, or latency threshold for migrating the individual application instance between two sites.

19. A computer-implemented method, comprising:

receiving data related to servers, the data reflecting historical parameters of computing elements and energy sources, realtime energy parameters for the energy sources, and availability of the computing elements, wherein the energy sources comprise both renewable and non-renewable energy sources;

accessing a framework that includes;

a hosting site model associated with operations of the servers,

an application instances model of application instances that are run on the servers, and

a location energy model that estimates power consumption at the servers;

entering the received data into the framework to make server power state decisions based on the received data; and

automatically relocating individual application instances between the servers based on the server power state decisions.

20. The method of claim 19, the data further reflecting availability of an individual energy source, the availability being based on predicted power available from the individual energy source to a specific server for a specific time.

21. The method of claim 19, wherein the server power state decisions are based on at least one of reducing monetary costs for running the application instances, reducing impact on user interactiveness, increasing long term availability of the computing elements, increasing use of the renewable energy sources, reducing carbon footprint and pollutant emission due to the power consumption at the servers, or balancing bandwidth cost and energy cost.