System and method for providing dynamic provisioning within a compute environment

• US 9,069,611 B2
• Filed: 10/10/2011
• Issued: 06/30/2015
• Est. Priority Date: 06/18/2004
• Status: Active Grant

• Current Assignee: III Holdings 12, LLC (Intellectual Ventures LLC)
• Original Assignee: Adaptive Computing Enterprises, Inc. (Ala Services LLC)
• Inventors: Jackson, David B.
• Primary Examiner(s): Lee, Tammy
• Application Number:
  US13/269,893
• Publication Number:
  US 20120030685A1
• Time in Patent Office: 1,359 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 2209/5014 Reservation
• CPC Class Codes: G06F 2209/504 Resource capping
• CPC Class Codes: G06F 9/4881 Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
• CPC Class Codes: G06F 9/50 Allocation of resources, e.g. of the central processing unit [CPU]
• CPC Class Codes: G06F 9/5005 to service a request
• CPC Class Codes: G06F 9/5011 the resources being hardware resources other than CPUs, Servers and Terminals
• CPC Class Codes: G06F 9/5022 Mechanisms to release resources
• CPC Class Codes: G06F 9/5027 the resource being a machine, e.g. CPUs, Servers, Terminals
• CPC Class Codes: G06F 9/5072 Grid computing
• CPC Class Codes: H04L 41/00 Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
• CPC Class Codes: H04L 41/50 Network service management, e.g. ensuring proper service fulfilment according to agreements
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

The disclosure relates to systems, methods and computer-readable media for dynamically provisioning resources within a compute environment. The method aspect of the disclosure comprises A method of dynamically provisioning resources within a compute environment, the method comprises analyzing a queue of jobs to determine an availability of compute resources for each job, determining an availability of a scheduler of the compute environment to satisfy all service level agreements (SLAs) and target service levels within a current configuration of the compute resources, determining possible resource provisioning changes to improve SLA fulfillment, determining a cost of provisioning; and if provisioning changes improve overall SLA delivery, then re-provisioning at least one compute resource.

Claims

1. A method comprising:

determining, at a current time, an availability of compute resources within a compute environment for each workload in a queue of workload based on scheduled times of compute resources established for the each workload to yield an analysis, the each workload to consume scheduled compute resources at a future time, which is later than the current time, according to the scheduled times;

based on the analysis, determining an ability of the compute environment to satisfy a service level agreement associated with the each workload in the queue of workload based on a current configuration of the compute environment;

determining possible workload scheduling changes in the compute environment to improve service levels for at least one workload in the queue of workload;

determining a cost of provisioning compute resources associated with the possible workload scheduling changes; and

if the possible workload scheduling changes improve service levels within the compute environment, taking into account the cost, rescheduling the at least one workload in the compute environment prior to all workload in the queue of workload consuming compute resources in the compute environment.

2. The method of claim 1, further comprising determining if the possible workload scheduling changes improve service levels within the compute environment based on provisioning costs in both a long term and a short term.

3. The method of claim 1, further comprising provisioning compute resources as a respective scheduled time arrives for the at least one workload.

4. The method of claim 1, wherein availability is determined in terms of one of:

a required operating system, hardware architecture, network adapters, real memory, virtual memory, internal and external disk storage, and a software application.

5. The method of claim 1, wherein the availability is further constrained by one of:

a resource state, a reservation and a political policy.

6. The method of claim 1, further comprising determining a priority of workload to compute environment objectives and a time frame in which workload must complete to fulfill service level agreements.

7. The method of claim 1, wherein only certain workload in the queue of workload are enabled to receive provisioning services.

8. The method of claim 1, further comprising, if the compute environment will be rescheduled:

analyzing a workload schedule associated with a node to be rescheduled;

selecting at least one node that can be scheduled with a minimal impact; and

when the at least one node becomes available, instructing a scheduling service to schedule the at least one node.

9. The method of claim 8, wherein scheduling further comprises creating a reservation for the at least one node and provisioning the at least one node.

10. The method of claim 9, wherein provisioning the at least one node further comprises provisioning a group of nodes within the compute environment.

11. The method of claim 9, wherein provisioning the at least one node further comprises provisioning individual nodes piecemeal within the compute environment.

12. The method of claim 11, wherein provisioning the individual nodes piecemeal within the compute environment is based on workload within the compute environment.

13. The method of claim 1, further comprising releasing the compute resources back to a general availability.

14. The method of claim 13, wherein a scheduler releases the compute resources back to a general availability according to one of the following time frames:

when a provisioning manager confirms that a provisioning is complete;

if a provisioning scheduled time is set, ahead of the provisioning scheduled time if the provisioning manager confirms that the provisioning is complete; and

if a re-provisioning scheduled time is set, after the re-provisioning scheduled time if the provisioning manager confirms that the provisioning is one of complete and failed.

15. The method of claim 13, wherein reservations are used to control resource access and schedule resource provisioning in time.

16. The method of claim 15, wherein a scheduler will release resources from reservations prior to a reservation end time if a provisioning manager confirms successful completion of provisioning.

17. The method of claim 16, wherein the scheduler will extend the reservation while the provisioning manager confirms failure of provisioning.

18. The method of claim 1, further comprising:

analyzing historical information of the compute environment; and

determining whether to schedule the compute environment based at least in part on the historical information.

19. The method of claim 18, wherein the historical information is one of compute environment performance information, workload-submission information, and failure rate information.

20. A non-transitory computer-storage medium having stored therein instructions which, when executed by a processor, cause the processor to perform operations comprising:

determining, at a current time, an availability of compute resources within a compute environment for each workload in a queue of workload based on scheduled times of compute resources established for the each workload to yield an analysis, the each workload to consume scheduled compute resources at a future time, which is later than the current time, according to the schedules times;

based on the analysis, determining an ability of the compute environment to satisfy a service level agreement associated with the each workload in the queue of workload based on a current configuration of the compute environment;

determining possible workload scheduling changes in the compute environment to improve service levels for at least one workload in the queue of workload;

determining a cost of provisioning compute resources associated with the possible workload scheduling changes; and

if the possible workload scheduling changes improve service levels within the compute environment, taking into account the cost, rescheduling the at least one workload in the compute environment prior to all workload in the queue of workload consuming compute resources in the compute environment.

21. A system comprising:

a processor; and

a computer-readable storage medium having stored therein instructions which, when executed by the processor, cause the processor to perform operations comprising;

determining, at a current time, an availability of compute resources within a compute environment for each workload in a queue of workload based on scheduled times of compute resources established for the each workload to yield an analysis, the each workload to consume scheduled compute resources at a future time, which is later that the current time, according to the schedules times;

based on the analysis, determining an ability of the compute environment to satisfy a service level agreement associated with the each workload in the queue of workload based on a current configuration of the compute environment;

determining possible workload scheduling changes in the compute environment to improve service levels for at least one workload in the queue of workload;

determining a cost of provisioning compute resources associated with the possible workload scheduling changes; and

rescheduling the at least one workload in the compute environment prior to all workload in the queue of workload consuming compute resources in the compute environment, if the possible workload scheduling changes improve service levels within the compute environment, taking into account the cost.