DYNAMIC OPTIMIZATION OF SIMULATION RESOURCES

• US 20180048532A1
• Filed: 08/11/2016
• Published: 02/15/2018
• Est. Priority Date: 08/11/2016
• Status: Active Grant

• Current Assignee: Rescale, Inc.
• Original Assignee: Rescale, Inc.
• Inventors: Poort, Gregorius E., Poort, Mulyanto W., Kaneshiro, Ryan A., McKenzie, Adam C.
• Granted Patent: US 10,193,762 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 9/5066 Algorithms for mapping a plurality of inter-dependent sub-tasks onto a plurality of physical CPUs mappping at compile time, see G06F8/451
• CPC Class Codes: H04L 41/147 for predicting network behaviour
• CPC Class Codes: H04L 47/70 Admission control; Resource allocation
• CPC Class Codes: H04L 47/83 based on usage prediction
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46

Abstract

The present invention dynamically optimizes computing resources allocated to a simulation task while it is running. It satisfies application-imposed constraints and enables the simulation application performing the simulation task to resolve inter-instance (including inter-server) dependencies inherent in executing the simulation task in a parallel processing or other HPC environment. An intermediary server platform, between the user of the simulation task and the hardware providers on which the simulation task is executed, includes a cluster service that provisions computing resources on hardware provider platforms, an application service that configures the simulation application in accordance with application-imposed constraints, an application monitoring service that monitors execution of the simulation task for computing resource change indicators (including computing resource utilization and application-specific information extracted from output files generated by the simulation application) as well as restart files, and a computing resource evaluation engine that determines when a change in computing resources is warranted.

Claims

1. A system that dynamically optimizes the provisioning of computing resources for execution of an application that performs a task having inter-instance dependencies, wherein the application, in order to execute properly and resolve the inter-instance dependencies, includes one or more application-imposed constraints requiring a pre-configuration specification of at least one of the computing resources allocated to the application, the system comprising:

(a) a cluster service that provisions a cluster of computing resources on a hardware provider platform for executing the application;

(b) an application service that configures the application in accordance with the application-imposed constraints, and initiates execution of the application on the provisioned cluster;

(c) an application monitoring service that monitors execution of the application for (i) a plurality of computing resource change indicators and (ii) one or more restart files generated by the application; and

(d) a computing resource evaluation engine that (i) continuously analyzes the computing resource change indicators and restart files to determine whether modification of the current computing resources is warranted and, if so, (ii) implements a change in the current computing resources while the application is running.

2. The system of claim 1, wherein:

(a) the cluster includes a plurality of CPU cores for executing independent sub-task calculations of the application in parallel, including a first sub-task calculation performed by a first instance of the application on a first CPU core, and a second sub-task calculation performed by a second instance of the application on a second CPU core; and

(b) a third sub-task calculation performed by the first instance is dependent upon the results of the second sub-task calculation performed by the second instance, and thereby constitutes an inter-instance dependency.

3. The system of claim 1, wherein the plurality of computing resource change indicators include:

(a) computing resource utilization data that reflect the extent of the application'"'"'s current usage of the computing resources; and

(b) application-specific data, including data extracted from output files generated by the application during its execution, that facilitate a prediction of the application'"'"'s future usage of computing resources.

4. The system of claim 1, wherein the computing resource change indicators include indicators of one or more of the following:

(a) a change in the number of cells in a simulation data model provided as input to the application, (b) one or more repeating iteration steps in the task performed by the application;

(c) a change in the extent of inter-core communication exhibited by the application; and

(d) a change to a more or less computing resource-intensive phase of the task performed by the application.

5. The system of claim 1, wherein the computing resource evaluation engine, based on its analysis of the computing resource change indicators, generates a prediction regarding the extent of the application'"'"'s future usage of computing resources.

6. The system of claim 1, wherein the computing resource evaluation engine bases its determination of whether modification of the current computing resources is warranted on historical information monitored by the application monitoring service over time.

7. The system of claim 1, wherein the application-imposed constraints include a pre-configuration specification of the number of CPU cores in the provisioned cluster.

8. The system of claim 1, wherein the change in computing resources includes an increase in one computing resource and a decrease in another computing resource.

9. The system of claim 1, wherein each restart file contains data reflecting the application'"'"'s current intermediate state, thereby enabling the application, if restarted in the future, to resume execution from that intermediate state.

10. The system of claim 1, wherein the computing resource evaluation engine implements a change in the current computing resources while the application is running by:

(a) directing the cluster service to provision a new cluster of computing resources on a selected hardware provider platform;

(b) terminating the current cluster on which the application is executing; and

(c) directing the application service to reconfigure the application in accordance with the application-imposed constraints, and to initiate execution of the application on the provisioned new cluster from the intermediate state contained within a selected one of the restart files.

11. The system of claim 1, further comprising a license service that authenticates each user of the application with a corresponding license server before permitting that user to initiate execution of the application on the provisioned cluster.

12. The system of claim 1, further comprising a license service that authenticates each user of a component of the application before permitting that user to initiate execution of the component of the application on the provisioned cluster.

13. The system of claim 1, further comprising:

(a) a hardware metering module that monitors the total time the provisioned computing resources are in use during execution of the task; and

(b) a software metering module that monitors the total time the application is in use during execution of the task.

14. The system of claim 13, wherein the software metering module monitors the total time a component of the application is in use during execution of the task.

15. The system of claim 13, wherein the application is terminated on the provisioned cluster and restarted on a new cluster having different computing resources, and wherein the total time monitored by the software metering module includes the sum of the times the application is in use during execution of the task on the provisioned cluster and the new cluster.

16. The system of claim 13, wherein the software metering module monitors a plurality of applications performing the task.

17. A method for dynamically optimizing the provisioning of computing resources for execution of an application that performs a task having inter-instance dependencies, wherein the application, in order to execute properly and resolve the inter-instance dependencies, includes one or more application-imposed constraints requiring a pre-configuration specification of at least one of the computing resources allocated to the application, the method comprising the following steps:

(a) provisioning a cluster of computing resources on a hardware provider platform for executing the application;

(b) configuring the application in accordance with the application-imposed constraints, and initiating execution of the application on the provisioned cluster;

(c) monitoring execution of the application for (i) a plurality of computing resource change indicators and (ii) one or more restart files generated by the application; and

(d) continuously analyzing the computing resource change indicators and restart files to determine whether modification of the current computing resources is warranted and, if so, implementing a change in the current computing resources while the application is running.

18. The method of claim 17, wherein:

(a) the cluster includes a plurality of CPU cores for executing independent sub-task calculations of the application in parallel, including a first sub-task calculation performed by a first instance of the application on a first CPU core, and a second sub-task calculation performed by a second instance of the application on a second CPU core; and

(b) a third sub-task calculation performed by the first instance is dependent upon the results of the second sub-task calculation performed by the second instance, and thereby constitutes an inter-instance dependency.

19. The method of claim 17, wherein the plurality of computing resource change indicators include:

(a) computing resource utilization data that reflect the extent of the application'"'"'s current usage of the computing resources; and

(b) application-specific data, including data extracted from output files generated by the application during its execution, that facilitate a prediction of the application'"'"'s future usage of computing resources.

20. The method of claim 17, wherein the computing resource change indicators include indicators of one or more of the following:

(a) a change in the number of cells in a simulation data model provided as input to the application, (b) one or more repeating iteration steps in the task performed by the application;

(c) a change in the extent of inter-core communication exhibited by the application; and

(d) a change to a more or less computing resource-intensive phase of the task performed by the application.

21. The method of claim 17, further comprising the step of generating, based on the analysis of the computing resource change indicators, a prediction regarding the extent of the application'"'"'s future usage of computing resources.

22. The method of claim 17, wherein the determination of whether modification of the current computing resources is warranted is based on historical information monitored over time.

23. The method of claim 17, wherein the application-imposed constraints include a pre-configuration specification of the number of CPU cores in the provisioned cluster.

24. The method of claim 17, wherein the change in computing resources includes an increase in one computing resource and a decrease in another computing resource.

25. The method of claim 17, wherein each restart file contains data reflecting the application'"'"'s current intermediate state, thereby enabling the application, if restarted in the future, to resume execution from that intermediate state.

26. The method of claim 17, wherein the implementation of a change in the current computing resources while the application is running includes the following steps:

(a) provisioning a new cluster of computing resources on a selected hardware provider platform;

(b) terminating the current cluster on which the application is executing;

(c) reconfiguring the application in accordance with the application-imposed constraints, and

(d) initiating execution of the application on the provisioned new cluster from the intermediate state contained within a selected one of the restart files.

27. The method of claim 17, further comprising the step of authenticating each user of the application with a corresponding license server before permitting that user to initiate execution of the application on the provisioned cluster.

28. The method of claim 17, further comprising the step of authenticating each user of a component of the application before permitting that user to initiate execution of the component of the application on the provisioned cluster.

29. The method of claim 17, further comprising the steps of:

(a) monitoring the total time the provisioned computing resources are in use during execution of the task; and

(b) monitoring the total time the application is in use during execution of the task.

30. The method of claim 29, further comprising the step of monitoring the total time a component of the application is in use during execution of the task.

31. The method of claim 29, wherein the application is terminated on the provisioned cluster and restarted on a new cluster having different computing resources, and wherein the total time the application is in use during execution of the task includes the sum of the times the application is in use on the provisioned cluster and the new cluster.

32. The method of claim 29, further comprising the step of monitoring a plurality of applications performing the task.

33. A system that facilitates hardware and software metering of tasks performed by a plurality of applications on behalf of a plurality of users, the system comprising:

(a) a cluster service that provisions a first cluster of computing resources on one of a plurality of hardware provider platforms for executing an application on behalf of a user and, in the event a change in computing resources is warranted while the application is running, provisions a second cluster of computing resources and terminates the application on the first cluster;

(b) an application service that configures the application and initiates execution of the application on the provisioned first cluster and, in the event the second cluster is provisioned by the cluster service, reconfigures the application and resumes execution of the application on the second cluster;

(c) a license service that authenticates each user of the application with a corresponding license server before permitting that user to initiate execution of the application on the provisioned first and second cluster;

(d) a hardware metering module that monitors the total hardware time the provisioned computing resources are in use during execution of the task, wherein the total hardware time includes the sum of the times the provisioned computing resources are in use on the first cluster and the second cluster; and

(e) a software metering module that monitors the total software time the application is in use during execution of the task, wherein the total software time includes the sum of the times the application is in use on the first cluster and the second cluster.

34. The system of claim 33, wherein the license service authenticates each user of a component of the application before permitting that user to initiate execution of the component of the application on the provisioned first and second cluster.

35. The system of claim 33, wherein the software metering module monitors the total software time a component of the application is in use during execution of the task.

36. The system of claim 33, wherein the software metering module monitors a plurality of applications performing the task.

37. The system of claim 33, wherein the application performs a task having inter-instance dependencies, and the application, in order to execute properly and resolve the inter-instance dependencies, includes one or more application-imposed constraints requiring a pre-configuration specification of at least one of the computing resources allocated to the application, the system further comprising:

(a) an application monitoring service that monitors execution of the application for (i) a plurality of computing resource change indicators and (ii) one or more restart files generated by the application; and

(b) a computing resource evaluation engine that (i) continuously analyzes the computing resource change indicators and restart files to determine whether modification of the current computing resources is warranted and, if so, (ii) implements a change in the current computing resources while the application is running.

38. The system of claim 37, wherein:

(a) the cluster includes a plurality of CPU cores for executing independent sub-task calculations of the application in parallel, including a first sub-task calculation performed by a first instance of the application on a first CPU core, and a second sub-task calculation performed by a second instance of the application on a second CPU core; and

(b) a third sub-task calculation performed by the first instance is dependent upon the results of the second sub-task calculation performed by the second instance, and thereby constitutes an inter-instance dependency.

39. The system of claim 37, wherein the plurality of computing resource change indicators include:

(a) computing resource utilization data that reflect the extent of the application'"'"'s current usage of the computing resources; and

(b) application-specific data, including data extracted from output files generated by the application during its execution, that facilitate a prediction of the application'"'"'s future usage of computing resources.

40. The system of claim 37, wherein the computing resource change indicators include indicators of one or more of the following:

(a) a change in the number of cells in a simulation data model provided as input to the application, (b) one or more repeating iteration steps in the task performed by the application;

(c) a change in the extent of inter-core communication exhibited by the application; and

(d) a change to a more or less computing resource-intensive phase of the task performed by the application.

41. The system of claim 37, wherein the computing resource evaluation engine, based on its analysis of the computing resource change indicators, generates a prediction regarding the extent of the application'"'"'s future usage of computing resources.

42. The system of claim 37, wherein the computing resource evaluation engine bases its determination of whether modification of the current computing resources is warranted on historical information monitored by the application monitoring service over time.

43. The system of claim 37, wherein the application-imposed constraints include a pre-configuration specification of the number of CPU cores in the provisioned first and second cluster.

44. The system of claim 37, wherein the change in computing resources includes an increase in one computing resource and a decrease in another computing resource.

45. The system of claim 37, wherein each restart file contains data reflecting the application'"'"'s current intermediate state, thereby enabling the application, if restarted in the future, to resume execution from that intermediate state.

46. The system of claim 37, wherein the computing resource evaluation engine implements a change in the current computing resources while the application is running by:

(a) directing the cluster service to provision the second cluster of computing resources on a selected hardware provider platform;

(b) terminating the first cluster on which the application is executing; and

(c) reconfiguring the application in accordance with the application-imposed constraints, and resuming execution of the application on the second cluster from the intermediate state contained within a selected one of the restart files.

47. A method for facilitating hardware and software metering of tasks performed by a plurality of applications on behalf of a plurality of users, the method comprising the following steps:

(a) provisioning a first cluster of computing resources on one of a plurality of hardware provider platforms for executing an application on behalf of a user and, in the event a change in computing resources is warranted while the application is running, provisioning a second cluster of computing resources and terminating the application on the first cluster;

(b) configuring the application and initiating execution of the application on the provisioned first cluster and, in the event the second cluster is provisioned by the cluster service, reconfigures the application and resumes execution of the application on the second cluster;

(c) authenticating each user of the application with a corresponding license server before permitting that user to initiate execution of the application on the provisioned first and second cluster;

(d) monitoring the total hardware time the provisioned computing resources are in use during execution of the task, wherein the total hardware time includes the sum of the times the provisioned computing resources are in use on the first cluster and the second cluster; and

(e) monitoring the total time the application is in use during execution of the task, wherein the total software time includes the sum of the times the application is in use on the first cluster and the second cluster.

48. The method of claim 47, further comprising the step of authenticating each user of a component of the application before permitting that user to initiate execution of the component of the application on the provisioned first and second cluster.

49. The method of claim 47, further comprising the step of monitoring the total software time a component of the application is in use during execution of the task.

50. The system of claim 47, further comprising the step of monitoring a plurality of applications performing the task.

51. The method of claim 47, wherein the application performs a task having inter-instance dependencies, and the application, in order to execute properly and resolve the inter-instance dependencies, includes one or more application-imposed constraints requiring a pre-configuration specification of at least one of the computing resources allocated to the application, the method further comprising the following steps:

(a) monitoring execution of the application for (i) a plurality of computing resource change indicators and (ii) one or more restart files generated by the application; and

(b) continuously analyzing the computing resource change indicators and restart files to determine whether modification of the current computing resources is warranted and, if so, implementing a change in the current computing resources while the application is running.

52. The method of claim 51, wherein:

(a) the cluster includes a plurality of CPU cores for executing independent sub-task calculations of the application in parallel, including a first sub-task calculation performed by a first instance of the application on a first CPU core, and a second sub-task calculation performed by a second instance of the application on a second CPU core; and

(b) a third sub-task calculation performed by the first instance is dependent upon the results of the second sub-task calculation performed by the second instance, and thereby constitutes an inter-instance dependency.

53. The method of claim 51, wherein the plurality of computing resource change indicators include:

(a) computing resource utilization data that reflect the extent of the application'"'"'s current usage of the computing resources; and

(b) application-specific data, including data extracted from output files generated by the application during its execution, that facilitate a prediction of the application'"'"'s future usage of computing resources.

54. The method of claim 51, wherein the computing resource change indicators include indicators of one or more of the following:

(a) a change in the number of cells in a simulation data model provided as input to the application, (b) one or more repeating iteration steps in the task performed by the application;

(c) a change in the extent of inter-core communication exhibited by the application; and

(d) a change to a more or less computing resource-intensive phase of the task performed by the application.

55. The method of claim 51, further comprising the step of generating, based on the analysis of the computing resource change indicators, a prediction regarding the extent of the application'"'"'s future usage of computing resources.

56. The method of claim 51, wherein the determination of whether modification of the current computing resources is warranted is based on historical information monitored over time.

57. The method of claim 51, wherein the application-imposed constraints include a pre-configuration specification of the number of CPU cores in the provisioned first and second cluster.

58. The method of claim 51, wherein the change in computing resources includes an increase in one computing resource and a decrease in another computing resource.

59. The method of claim 51, wherein each restart file contains data reflecting the application'"'"'s current intermediate state, thereby enabling the application, if restarted in the future, to resume execution from that intermediate state.

60. The method of claim 51, wherein the implementation of a change in the current computing resources while the application is running includes the following steps:

(a) provisioning the second cluster of computing resources on a selected hardware provider platform;

(b) terminating the first cluster on which the application is executing;

(c) reconfiguring the application in accordance with the application-imposed constraints; and

(d) resuming execution of the application on the second cluster from the intermediate state contained within a selected one of the restart files.