System and method for allocating resources by examining a system characteristic

• US 7,752,623 B1
• Filed: 09/16/2004
• Issued: 07/06/2010
• Est. Priority Date: 09/16/2004
• Status: Active Grant

• Current Assignee: Valtrus Innovations Limited (f/k/a Dolya Holdco 9 Limited) (Key Patent Innovations Limited)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Crawford, Isom Lawrence Jr.
• Primary Examiner(s): An; Meng-Ai
• Assistant Examiner(s): Huaracha; Willy W
• Application Number:
  US10/942,298
• Time in Patent Office: 2,119 Days
• Field of Search: 718/104, 718/1, 718/102, 718/103, 718/105, 710/244, 711/173, 395/200.11, 712/290, 709/103, 709/220, 709/221, 709/223
• US Class Current: 718/104
• CPC Class Codes: G06F 2209/5019 Workload prediction
• CPC Class Codes: G06F 2209/508 Monitor
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: G06F 9/5077 Logical partitioning of resources; Management or configuration of virtualized resources specific details on emulation or internal functioning of virtual machines G06F9/455

Abstract

In one embodiment, a method for allocating resources in a shared resource domain comprises gathering performance data for multiple applications, determining that an application of the multiple applications is not achieving a service level objective (SLO), calculating an additional amount of a resource for the application estimated to cause the application to achieve the SLO, and examining, before initiating reallocation operations, at least one system characteristic to determine whether the calculated additional amount of the resource would improve performance of the application.

Claims

1. A method for allocating resources in a shared resource domain, comprising:

gathering performance data for multiple applications;

determining that an application of said multiple applications is not achieving a service level objective (SLO);

calculating an additional amount of resources for said application estimated to cause said application to achieve said SLO;

examining, before initiating reallocation operations, at least one system characteristic, by performing a system call to examine the queue length of a run queue of a computing domain associated with said application and by comparing said additional amount to a level of allocation granularity associated with said resources, to determine whether said calculated additional amount of said resources would improve performance of said application; and

requesting said additional amount only when said additional amount is determined to improve performance of said application.

2. The method of claim 1, wherein said examining comprises analyzing an application characteristic data structure associated with said application.

3. The method of claim 2, further comprising:

updating a value of said application characteristic data structure in response to repetitive allocation and deallocation operations.

4. The method of claim 1, wherein said examining further comprises:

adjusting said additional amount in response to said comparing.

5. The method of claim 4, wherein said resources comprise a plurality of processors.

6. The method of claim 5, wherein said multiple applications are executed in respective virtual partitions and each processor is assigned to only one partition.

7. The method of claim 1, further comprises:

requesting said additional amount of said resources from a software process responsible for reallocation of resources between said multiple applications.

8. The method of claim 7, wherein said software process reallocates said additional amount of said resources after determining whether said additional amount of said resources is available.

9. The method of claim 8, wherein said determining whether said additional amount of said resources is available comprises:

identifying idle resources.

10. The method of claim 8, wherein said determining whether said additional amount of said resources is available comprises:

identifying resources assigned to applications of lower priority than said application.

11. A computing system, comprising:

a plurality of computing resources;

computer readable storage medium including executable instructions for facilitating reallocation of said plurality of computing resources in said computing system;

a plurality of computing domains, wherein a respective application, a manager process, and a respective performance monitoring process are executed in each of said plurality of computing domains; and

an executable allocation process for allocating said plurality of computing resources between said plurality of computing domains;

wherein when a manager process determines that performance data from its associated performance monitor does not satisfy a service level objective (SLO), said manager process is operable to (i) calculate an additional amount of said plurality of computing resources appropriate to achieve said SLO;

(ii) to examine at least one system characteristic to determine whether said additional amount will improve performance of its associated application by performing a system function call to examine the queue length of a run queue of a computing domain associated with said application and by comparing said additional amount to a level of allocation granularity associated with said plurality of computing resources; and

(iii) request said additional amount from said allocation process only when said additional amount is determined to improve performance of its application.

12. The computing system of claim 11, wherein said manager process performs said examining by analyzing application characteristic information statically encoded in a data structure.

13. The computing system of claim 11, wherein said plurality of computing resources comprise processors.

14. The computing system of claim 13, wherein each processor is assigned to only one of said plurality of computing domains and said manager process modifies said additional amount to equal an integer number of processors.

15. The computing system of claim 13, wherein said executable allocation process reallocates said additional amount by identifying idle resources.

16. The computing system of claim 13, wherein said executable allocation process reallocates said additional amount by identifying resources associated with applications of lower priority than the application associated with said requesting manager process.

17. A computer readable storage medium including executable instructions for facilitating reallocation of resources in a computing system, said computer readable medium comprising:

code for generating performance data for a plurality of applications;

code for detecting that an application is not achieving a service level objective using performance data from said code for generating;

code for calculating an additional amount of resources using performance data and a current allocation of said resources associated with said application;

code for executing a system function call to examine the queue length of a run queue of a computing domain associated with said application;

code for comparing said additional amount to a level of allocation granularity associated with said resources;

code for examining said application to determine whether improved performance of said application will result from allocation of said additional amount of said resources; and

code for initiating reallocation operations only when said additional amount of said resources will improve performance of said application.

18. The computer readable storage medium of claim 17, wherein said plurality of applications are executed in respective partitions.