Method and apparatus for allocating resources to applications using a linearized objective function

• US 7,827,557 B2
• Filed: 03/24/2004
• Issued: 11/02/2010
• Est. Priority Date: 03/24/2004
• Status: Expired due to Fees

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Santos, Ciprianos, Zhu, Xiaoyun, Singhal, Sharad, Beyer, Dirk, Drew, Julie Ward
• Primary Examiner(s): Tang; Kenneth
• Application Number:
  US10/808,073
• Publication Number:
  US 20050228850A1
• Time in Patent Office: 2,414 Days
• Field of Search: 718/104, 718/105, 709/217, 709/219, 709/224, 709/226, 370/238
• US Class Current: 718/104
• CPC Class Codes: G06F 9/5005 to service a request

Abstract

A method, system, and apparatus is disclosed for allocating resources to applications. Available resources of a networked computing system may be determined. For each application, required resources of the application may be determined. An assigned subset of the available resources for each application is determined as a function of the required resources of the application and the available resources. The function reduces communication delays between resources of the subset of the available resources in conformance with bandwidth capacity requirements of the application and in conformance with network bandwidth limitations. The applications may then be associated with the assigned subsets of resources.

Claims

1. A method for allocating resources to a plurality of applications, comprising:

determining, by a computer, available resources of a networked computing system, wherein the available resources comprise processing resources, networking resources, and storage resources;

determining, by the computer, for each application, required resources of the application, wherein the required resources of each application specify a type of servers on which the application can be hosted, the type of servers being based on attributes of the servers including attributes relating to processor architecture and processing power, wherein the servers are part of the processing resources;

determining, by the computer, an assigned subset of the available resources for each application as a function of the required resources of the application and the available resources, wherein determining the assigned subset of available resources for each application is based on a linearized objective function that reduces communication delays between resources of the subset of the available resources in conformance with bandwidth capacity requirements of the application and in conformance with network bandwidth limitations, wherein the linearized objective function includes a linear combination of variables; and

associating the applications with the assigned subsets of resources.

2. The method of claim 1, wherein the networking resources comprise network switches.

3. The method of claim 1, wherein the storage resources comprise a storage area network, wherein the storage area network includes at least one pair of redundant core switches coupled to storage devices, the core switches coupled to the processing resources via a plurality of edge switches.

4. The method of claim 1, wherein determining the assigned subset of available resources for each application is performed by solving a mixed-integer programming problem.

5. The method of claim 4, wherein the available resources include network switches, and the mixed-integer programming problem reduces communication delays between resources of each subset of the available resources by reducing data traffic on network links that interconnect the network switches.

6. The method of claim 1, wherein the attributes relating to processor architecture and processing power specify processor type and processor speed.

7. The method of claim 6, wherein the resource requirements further specify storage patterns of files for each application, wherein determining the assigned subset is based on the resource requirements that further specify storage patterns of files.

8. The method of claim 1, wherein the network bandwidth limitations are expressed as linear constraints, and wherein determining the assigned subset of the available resources for each application is a linear optimization problem.

9. The method of claim 8, wherein determining the assigned subset of the available resources for each application is a mixed integer programming problem.

10. The method of claim 1, wherein determining the assigned subset of the available resources for each application is based on the linearized objective function to reduce a number of hops between processing resources in the assigned subset.

11. The method of claim 1, wherein the linearized objective function is provided by substituting products of binary variables in a non-linear objective function with replacement binary variables in the linearized objective function.

12. A system comprising:

a computer having a processor and configured to;

determine available resources of a networked computing system, wherein the available resources comprise servers, networking resources, and storage resources;

determine required resources for each application of a plurality of applications, wherein the required resources for each application specify a type of the servers on which the application can be hosted, the type of servers being based on attributes relating to processing architecture and processing power;

determine an assigned subset of the available resources for each application as a function of the required resources of the application and the available resources, wherein determining the assigned subset of available resources for each application is based on a linearized objective function that reduces communication delays between resources of the subset of the available resources in conformance with bandwidth capacity requirements of the application and in conformance with network bandwidth limitations, wherein the linearized objective function includes a linear combination of variables; and

associate the applications with the assigned subsets of resources.

13. A system, comprising:

a plurality of network-coupled processing resources;

a plurality of storage resources network-coupled to the processing resources, wherein the processing and storage resources are allocated to a plurality of applications;

a computing arrangement configured to;

determine, for each application of the plurality of applications, required resources of the application, wherein the required resources of each application specify a type of servers on which the application can be hosted, the type of servers being based on attributes of the servers including attributes relating to processor architecture and processing power, wherein the servers are part of the processing resources;

determine an assigned subset of the processing and storage resources for each application as a function of the required resources of the application and the processing and storage resources, wherein determining the assigned subset of processing and storage resources for each application is based on a linearized objective function that reduces communication delays between resources of the subset of the processing and storage resources in conformance with bandwidth capacity requirements of the application and in conformance with network bandwidth limitations, wherein the linearized objective function includes a linear combination of variables; and

associate the applications with the assigned subsets of processing and storage resources.

14. The system of claim 13, wherein the storage resources comprise a storage area network, wherein the storage area network includes at least one pair of redundant core switches coupled to storage devices, the core switches coupled to the network via a plurality of edge switches.

15. The system of claim 14, wherein the computing arrangement is configured to determine each assigned subset by solving a mixed-integer programming problem.

16. The system of claim 15, wherein the processing resources are coupled by network switches, and the mixed-integer programming problem reduces communication delays between resources by reducing data traffic on network links that interconnect the network switches.

17. A non-transitory computer-readable medium configured with instructions for causing a processor of a computer to allocate resources to a plurality of applications, comprising:

determining available resources of a networked computing system, wherein the available resources comprise processing resources, networking resources, and storage resources;

determining, for each application, required resources of the application;

determining an assigned subset of the available resources for each application as a function of the required resources of the application and the available resources, wherein determining the assigned subset of available resources for each application is based on a linearized objective function that reduces communication delays between resources of the subset of the available resources in conformance with bandwidth capacity requirements of the application and in conformance with network bandwidth limitations, wherein the linearized objective function includes a linear combination of variables; and

associating the applications with the assigned subsets of resources,

wherein the required resources of each application is specified in resource requirements that include attributes of the processing resources, wherein the attributes specify processor type and processor speed.

18. The non-transitory computer-readable medium of claim 17, wherein the resource requirements further specify storage patterns of files for each application, wherein determining the assigned subset is based on the resource requirements that further specify storage patterns of files.

19. The non-transitory computer-readable media of claim 17, wherein the network bandwidth limitations are expressed as linear constraints, and wherein determining the assigned subset of the available resources for each application is a linear optimization problem.

20. The non-transitory computer-readable media of claim 19, wherein determining the assigned subset of the available resources for each application is a mixed integer programming problem.

21. The non-transitory computer-readable media of claim 17, wherein determining the assigned subset of the available resources for each application is based on the linearized objective function to reduce a number of hops between processing resources in the assigned subset.

22. The non-transitory computer-readable media of claim 17, wherein the linearized objective function is provided by substituting products of binary variables in a non-linear objective function with replacement binary variables in the linearized objective function.

23. The non-transitory computer-readable medium of claim 17, wherein the processing resources comprise servers each having at least one processor.

24. The non-transitory computer-readable medium of claim 17, wherein the networking resources comprise network switches.

25. The non-transitory computer-readable medium of claim 17, wherein the storage resources comprise a storage area network, wherein the storage area network includes at least one pair of redundant core switches coupled to storage devices, the core switches coupled to the processing resources via a plurality of edge switches.

26. The non-transitory computer-readable medium of claim 25, wherein reducing the communication delays between resources comprises solving a mixed-integer programming problem.

27. The non-transitory computer-readable medium of claim 26, wherein the available resources include network switches coupled with the processing resources, and the mixed-integer programming problem reduces communication delays between resources of each subset of the available resources by reducing data traffic on network links that interconnect the network switches.