Remediating gaps between usage allocation of hardware resource and capacity allocation of hardware resource

• US 9,519,515 B2
• Filed: 04/17/2015
• Issued: 12/13/2016
• Est. Priority Date: 12/16/2010
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Calcaterra, Jeffrey A., Hintermeister, Gregory R., Williams, Michael D.
• Primary Examiner(s): Wai, Eric C
• Assistant Examiner(s): Lee, James J
• Application Number:
  US14/689,719
• Publication Number:
  US 20150220365A1
• Time in Patent Office: 606 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 11/3024 where the computing system component is a central processing unit [CPU]
• CPC Class Codes: G06F 11/3447 Performance evaluation by modeling
• CPC Class Codes: G06F 2209/5019 Workload prediction
• CPC Class Codes: G06F 9/5011 the resources being hardware resources other than CPUs, Servers and Terminals

Abstract

A usage allocation of a hardware resource to each of a number of workloads over time is determined using a demand model. The usage allocation of the resource includes a current and past actual usage allocation of the resource, a future projected usage allocation of the resource, and current and past actual usage of the resource. A capacity allocation of the resource is determined using a capacity model. The capacity allocation of the resource includes a current and past capacity and a future projected capacity of the resource. Whether a gap exists between the usage allocation and the capacity allocation is determined using a mapping model. Where the gap exists between the usage allocation of the resource and the capacity allocation of the resource, a user is presented with options determined using the mapping model and selectable by the user to implement a remediation strategy to close the gap.

Claims

1. A method comprising:

forecasting, by a processor, a future usage allocation of a hardware resource to each workload of a plurality of workloads over time, using a demand model, the future usage allocation of the hardware resource specifying a future projected usage allocation of the hardware resource, based on at least a current and past actual usage of the hardware resource as monitored, by one of;

the demand model specifying a given workload is to run on dedicated hardware, and the forecasting includes determining the usage allocation of the hardware resource to the given workload as a complete usage allocation of the hardware resource in relation to the dedicated hardware;

the demand model specifying that the given workload is to be fully redundant, and the forecasting includes increasing the usage allocation of the hardware resource to the given workload by at least a factor of two;

forecasting, by the processor, a capacity allocation of the hardware resource, using a capacity model, the capacity allocation of the hardware resource specifying a future projected capacity of the hardware resource allocatable to the workloads at future times based in part on when computing devices that contribute to the hardware resource are online or offline at the future times;

determining, by the processor, whether a gap is predicted to occur that causes the usage allocation of the hardware resource to be greater than the capacity allocation of the hardware resource, using a mapping model;

in response to determining that the gap is predicted to occur, presenting a user with one or more options determined by the processor using the mapping model and selectable by the user, wherein the one or more options including an option to prevent a part of the hardware resource to be offline at a specific future time; and

implementing a remediation strategy to prevent the gap from actually occurring based on a selection of the one or more options by the user.

2. The method of claim 1, wherein determining the usage allocation of the hardware resource to a given workload over time, using the demand model, comprises:

receiving, from the user, the demand model in relation to the given workload, the demand model as received from the user specifying likely usage of the hardware resource by the given workload over time; and

,

determining the usage allocation of the hardware resource to the given workload, based on the likely usage of the hardware resource by the given workload over time.

3. The method of claim 1, wherein determining the usage allocation of the hardware resource to a given workload of the plurality of workloads over time, using the demand model, comprises:

receiving, from the user, the demand model in relation to the given workload, the demand model as received from the user specifying likely demand of the given workload over time, where the likely demand has a unit of measure other than a unit of measure of the usage allocation of the hardware resource;

translating the likely demand of the given workload over time as received from the user to a likely usage of the hardware resource by the given workload over time; and

,

determining the usage allocation of the hardware resource to the given workload, based on the likely usage of the hardware resource by the given workload over time.

4. The method of claim 1, wherein determining the usage allocation of the hardware resource to a given workload of the plurality of workloads over time, using the demand model, comprises:

monitoring the actual usage of the hardware resource by the given workload over time, as specified by the demand model in relation to the given workload;

determining the usage allocation of the hardware resource to the given workload, based on the actual usage of the hardware resource by the given workload over time as has been monitored.

5. The method of claim 1, wherein determining the usage allocation of the hardware resource to a given workload of the plurality of workloads over time, using the demand model, comprises:

where the demand model specifies that the given workload is to run on dedicated hardware, determining the usage allocation of the hardware resource to the given workload as a complete usage allocation of the hardware resource in relation to the dedicated hardware.

6. The method of claim 1, wherein determining the usage allocation of the hardware resource to a given workload of the plurality of workloads over time, using the demand model, comprises:

where the demand model specifies that the given workload is to be fully redundant, increasing the usage allocation of the hardware resource to the given workload by at least a factor of two.

7. The method of claim 1, wherein determining the usage allocation of the hardware resource to a first workload and to a second workload of the plurality of workloads over time, using the demand model, comprises:

where the demand model specifies that the first workload is not to run on a same hardware as the second workload, determining the usage allocation of the hardware resource to the first workload and to the second workload such that the demand model is satisfied.

8. The method of claim 1, wherein determining the usage allocation of the hardware resource to each workload, using the demand model, comprises:

aggregating the usage allocation of the hardware resource to each workload over time, to yield a total usage allocation of the hardware resource to all the workloads over time.

9. The method of claim 1, wherein determining whether the gap exists between the usage allocation of the hardware resource and the capacity allocation of the hardware resource, using the mapping model, comprises:

displaying a graph having a first axis denoting time and a second axis denoting usage of the hardware resource.

10. The method of claim 9, wherein the graph further includes a maximum amount of potential usage of the hardware resource.

11. The method of claim 9, wherein the graph is segmented within the second axis, over time as denoted by the first axis, in correspondence with the usage allocation of the hardware resource to each workload,

wherein a current and past actual usage of the hardware resource by each workload is represented as a line within a segment of the graph corresponding to the current and past usage allocation of the hardware resource to the workload.

12. The method of claim 1, wherein determining whether the gap exists between the usage allocation of the hardware resource and the capacity allocation of the hardware resource, using the mapping model, comprises:

forecasting a likelihood that the usage allocation of the hardware resource to a particular workload of the plurality of workloads will be insufficient to accommodate actual usage of the hardware resource by the particular workload at a specific future time, using the mapping model,

wherein presenting the user with the one or more options comprises;

displaying a graphical user interface as dictated by the mapping model, a user provided with the one or more options within the graphical user interface that will reduce the likelihood that the usage allocation of the hardware resource to the particular workload will be insufficient to accommodate the usage of the hardware resource by the particular workload at the specific future time;

receiving a selection by the user of one or more desired options of the options provided to the user; and

,

implementing the desired options in the relation to the hardware resource to reduce the likelihood that the usage allocation of the hardware resource to the particular workload will be insufficient to accommodate the actual usage of the hardware resource by the particular workload at the specific future time.

13. The method of claim 12, wherein displaying the graphical user interface in which a user is provided with the options that will reduce the likelihood that the usage allocation of the hardware resource to the particular resource will be insufficient to accommodate the usage of the hardware resource by the particular workload at the specific future time comprises:

where the demand model specifies that a given workload is to run on dedicated hardware, displaying the graphical user interface in which the user is provided with a given option of the options, the given option to at least temporarily ignore the demand model,

such that at least temporarily ignoring the demand model results in an increase in available usage allocation of the hardware resource to the particular workload by a constrained usage allocation of the hardware resource,

wherein the constrained usage allocation of the hardware resource is equal to the usage allocation of the hardware resource to the given workload that is attributable to satisfying the demand model.

14. The method of claim 12, wherein displaying the graphical user interface in which a user is provided with the options that will reduce the likelihood that the usage allocation of the hardware resource to the particular resource will be insufficient to accommodate the usage of the hardware resource by the particular workload at the specific future time comprises:

where the demand model specifies that a given workload is to be fully redundant, displaying the graphical user interface in which the user is provided with a given option of the options, the given option to at least temporarily ignore the demand model,

such that at least temporarily ignoring the demand model results in an increase in available usage allocation of the hardware resource to the particular workload by a constrained usage allocation of the hardware resource,

wherein the constrained usage allocation of the hardware resource is equal to the usage allocation of the hardware resource to the given workload that is attributable to satisfying the demand model.

15. The method of claim 12, wherein displaying the graphical user interface in which a user is provided with the options that will reduce the likelihood that the usage allocation of the hardware resource to the particular resource will be insufficient to accommodate the usage of the hardware resource by the particular workload at the specific future time comprises:

where the demand model specifies that a first workload is not to run on a same hardware as a second workload, displaying the graphical user interface in which the user is provided with a given option of the options, the given option to at least temporarily ignore the demand model,

such that at least temporarily ignoring the demand model results in an increase in available usage allocation of the hardware resource to the particular workload by a constrained usage allocation of the hardware resource,

wherein the constrained usage allocation of the hardware resource is equal to the usage allocation of the hardware resource to the first workload and to the second workload that is attributable to satisfying the demand model.

16. The method of claim 12, wherein displaying the graphical user interface in which a user is provided with the options that will reduce the likelihood that the usage allocation of the hardware resource to the particular resource will be insufficient to accommodate the usage of the hardware resource by the particular workload at the specific future time comprises:

where the capacity model specifies a portion of the hardware resource is scheduled to be offline at the specific future time, displaying the graphical user interface in which the user is provided with a given option of the options, the given option to prevent the portion of the hardware resource to be offline at the specific future time,

such that preventing the portion of the hardware resource to be offline at the specific future time results in an increase in available usage allocation of the hardware resource to the particular workload.

17. The method of claim 12, wherein displaying the graphical user interface in which a user is provided with the options that will reduce the likelihood that the usage allocation of the hardware resource to the particular resource will be insufficient to accommodate the usage of the hardware resource by the particular workload at the specific future time comprises:

displaying the graphical user interface in which the user is provided with a given option of the options, the given option to make available an additional amount of the hardware resource at the specific future time,

such that making available the additional amount of the hardware resource at the specific future time results in an increase in available usage allocation of the hardware resource to the particular workload.

18. The method of claim 1, wherein the hardware resource is distributed over the computing devices of one or more hardware systems,

and wherein each workload encompasses one or more application computer programs running on the computing devices.

19. A non-transitory storage device storing computer-readable code executable by a processor, the computer-readable code comprising:

first computer-readable code to forecast a future usage allocation of a hardware resource to each workload of a plurality of workloads over time, using a demand model, the future usage allocation of the hardware resource specifying a future projected usage allocation of the hardware resource, based on at least a current and past actual usage of the hardware resource as monitored, by one of;

the demand model specifying a given workload is to run on dedicated hardware, and the forecasting includes determining the usage allocation of the hardware resource to the given workload as a complete usage allocation of the hardware resource in relation to the dedicated hardware;

the demand model specifying that the given workload is to be fully redundant, and the forecasting includes increasing the usage allocation of the hardware resource to the given workload by at least a factor of two;

second computer-readable code to forecast a capacity allocation of the hardware resource, using a capacity model, the capacity allocation of the hardware resource specifying a future projected capacity of the hardware resource allocable to the workloads at future times based in part on when computing devices that contribute to the hardware resource are online or offline at the future times;

third computer-readable code to determine whether a gap is predicted to occur that causes the usage allocation of the hardware resource to be greater than the capacity allocation of the hardware resource, using a mapping model;

fourth computer-readable code to, in response to the third computer-readable code determining that the gap is predicted to occur, present a user with one or more options determined by the processor using the mapping model and selectable by the user, wherein the one or more options include an option to prevent a portion of the hardware resource to be offline at a specific future time; and

fifth computer-readable code to implement a remediation strategy to prevent the gap from actually occurring based on a selection of the one or more options by the user.

20. A system comprising:

one or more computing devices;

a hardware resource at the one or more computing devices and to be used by a plurality of workloads;

a computer-readable data storage storing a computer program implementing a demand model, a capacity model, and a mapping model; and

,

a processor to execute the computer program,

wherein the computer program is to;

forecast a future usage allocation of a hardware resource to each workload of a plurality of workloads over time, using a demand model, the future usage allocation of the hardware resource specifying a future projected usage allocation of the hardware resource, based on at least a current and past actual usage of the hardware resource as monitored, by one of;

the demand model specifying a given workload is to run on dedicated hardware, and the forecasting includes determining the usage allocation of the hardware resource to the given workload as a complete usage allocation of the hardware resource in relation to the dedicated hardware;

the demand model specifying that the given workload is to be fully redundant, and the forecasting includes increasing the usage allocation of the hardware resource to the given workload by at least a factor of two;

forecast a capacity allocation of the hardware resource, using a capacity model, the capacity allocation of the hardware resource specifying a future projected capacity of the hardware resource allocable to the workloads at future times based in part on when computing devices that contribute to the hardware resource are online or offline at the future times;

determine whether a gap is predicted to occur that causes the usage allocation of the hardware resource to be greater than the capacity allocation of the hardware resource, using a mapping model;

in response to determining that the future gap is predicted to occur, present a user with one or more options determined by the processor using the mapping model and selectable by the user, wherein the one or more options include an option to prevent a portion of the hardware resource to be offline at a specific future time; and

implement a remediation strategy to prevent the gap from actually occurring based on a selection of the one or more options by the user.