Dynamic quality of service for storage I/O port allocation

• US 10,216,441 B2
• Filed: 11/25/2015
• Issued: 02/26/2019
• Est. Priority Date: 11/25/2015
• Status: Active Grant

• Current Assignee: Kyndryl Incorporated
• Original Assignee: International Business Machines Corporation
• Inventors: Chitti, Mark V., Freimuth, Douglas M., Hollingsworth, John F., Mandalia, Baiju D.
• Primary Examiner(s): Bertram, Ryan
• Assistant Examiner(s): Kwong, Edmund H
• Application Number:
  US14/952,463
• Publication Number:
  US 20170147248A1
• Time in Patent Office: 1,189 Days
• Field of Search: None
• CPC Class Codes: G06F 3/061 Improving I/O performance
• CPC Class Codes: G06F 3/0635 by changing the path, e.g. traffic rerouting, path reconfiguration
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]

Abstract

Dynamic quality of service for storage I/O port allocation is provided by obtaining indications of anticipated demand on a plurality of storage ports of a storage device of a data center, the anticipated demand being at least in part for storage input/output supporting transfer of data to or from a network, and dynamically adjusting storage port allocations based on the indications of anticipated demand and on real-time evaluation of workloads being serviced by the plurality of storage ports.

Claims

1. A computer-implemented method comprising:

receiving by a storage device of a data center, from an application using the storage device, a notification of a change in elastic network bandwidth allocation provisioned to the data center from an elastic network service provider of an elastic network;

based on the notification, obtaining indications of, and determining, anticipated demand on a plurality of storage ports of the storage device, the anticipated demand being at least in part for storage input/output supporting transfer of data to or from the elastic network, wherein an indication of the anticipated demand indicates an amount of the change in the elastic network bandwidth allocation, and the anticipated demand is based at least in part on the change in the elastic network bandwidth allocation and the indicated amount of that change; and

the storage device dynamically adjusting storage port allocations based on the determined anticipated demand and on real-time evaluation of workloads being serviced by the plurality of storage ports, the dynamically adjusting comprising adjusting respective individual storage input/output bandwidth or workload for at least one storage port of the plurality of storage ports, wherein the dynamically adjusting is performed based on the notification of the change in elastic network bandwidth allocation provisioned to the data center.

2. The method of claim 1, wherein an indication of anticipated demand comprises a signal of a change in upcoming demand from an application of the data center, the upcoming demand being upcoming demand on a first storage port set comprising one or more first storage ports of the plurality of storage ports.

3. The method of claim 2, further comprising predicting, based on the signal of the change in upcoming demand from the application of the data center, and using a statistical model, a level of the upcoming demand on the first storage port set, and predicting a duration of the upcoming demand, wherein the dynamically adjusting storage port allocations comprises allocating bandwidth of a second storage port set of one or more second storage ports of the plurality of storage ports to service at least some of the upcoming demand.

4. The method of claim 3, wherein the allocating bandwidth of the second storage port set comprises one or more of:

(i) releasing bandwidth of the second storage port set to the first storage port set for use by the application of the data center, or (ii) utilizing the second storage port set to service the at least some of the upcoming demand.

5. The method of claim 1, wherein the anticipated demand comprises a predicted near-future demand on one or more storage ports of the plurality of storage ports, the predicted near-future demand being predicted based on a current demand on the one or more storage ports.

6. The method of claim 1, wherein the real-time evaluation of the workloads being serviced by the plurality of storage ports comprises identifying an over-utilized storage port of the plurality of storage ports based on an individual workload currently being serviced by the over-utilized storage port exceeding an overutilization threshold, and wherein the dynamically adjusting storage port allocations comprises repurposing one or more unused or underutilized storage ports of the plurality of storage ports to service at least a portion of the workload currently being serviced by the over-utilized storage port.

7. The method of claim 1, wherein the real-time evaluation of the workloads being serviced by the plurality of storage ports comprises identifying that a range in individual workload levels of the plurality of storage ports exceeds a rebalancing threshold, and wherein the dynamically adjusting storage port allocations comprises rebalancing the individual workloads to within the rebalancing threshold.

8. The method of claim 1, wherein the real-time evaluation of the workloads being serviced by the plurality of storage ports comprises identifying traffic type flags of traffic of the workload, and wherein the dynamically adjusting storage port allocations comprises queuing the traffic according to a prioritization schedule that prioritizes traffic based on traffic type.

9. The method of claim 1, wherein the real-time evaluation of the workloads being serviced by the plurality of storage ports comprises evaluating priority of incoming storage commits, and wherein the dynamically adjusting storage port allocations comprises queuing the incoming storage commits according to a prioritization schedule that prioritizes storage commits based on storage commit type.

10. A computer program product comprising:

a computer readable storage medium readable by a processing circuit and storing instructions for execution by the processing circuit for performing a method comprising;

receiving by a storage device of a data center, from an application using the storage device, a notification of a change in elastic network bandwidth allocation provisioned to the data center from an elastic network service provider of an elastic network;

based on the notification, obtaining indications of, and determining, anticipated demand on a plurality of storage ports of the storage device, the anticipated demand being at least in part for storage input/output supporting transfer of data to or from the elastic network, wherein and indication of the anticipated demand indicates an amount of the change in the elastic network bandwidth allocation, and the anticipated demand is based at least in part on the change in the elastic network bandwidth allocation and the indicated amount of that chance; and

the storage device dynamically adjusting storage port allocations based on the determined anticipated demand and on real-time evaluation of workloads being serviced by the plurality of storage ports, the dynamically adjusting comprising adjusting respective individual storage input/output bandwidth or workload for at least one storage port of the plurality of storage ports, wherein the dynamically adjusting is performed based on the notification of the change in elastic network bandwidth allocation provisioned to the data center.

11. The computer program product of claim 10, wherein an indication of anticipated demand comprises a signal of a change in upcoming demand from an application of the data center, the upcoming demand being upcoming demand on a first storage port set comprising one or more first storage ports of the plurality of storage ports, and wherein the method further comprises predicting, based on the signal of the change in upcoming demand from the application of the data center, and using a statistical model, a level of the upcoming demand on the first storage port set, and predicting a duration of the upcoming demand, wherein the dynamically adjusting storage port allocations comprises allocating bandwidth of a second storage port set of one or more second storage ports of the plurality of storage ports to service at least some of the upcoming demand.

12. The computer program product of claim 10, wherein the anticipated demand comprises a predicted near-future demand on one or more storage ports of the plurality of storage ports, the predicted near-future demand being predicted based on a current demand on the one or more storage ports.

13. The computer program product of claim 10, wherein the real-time evaluation of the workloads being serviced by the plurality of storage ports comprises identifying an over-utilized storage port of the plurality of storage ports based on an individual workload currently being serviced by the over-utilized storage port exceeding an overutilization threshold, and wherein the dynamically adjusting storage port allocations comprises repurposing one or more unused or underutilized storage ports of the plurality of storage ports to service at least a portion of the workload currently being serviced by the over-utilized storage port.

14. The computer program product of claim 10, wherein the real-time evaluation of the workloads being serviced by the plurality of storage ports comprises identifying that a range in individual workload levels of the plurality of storage ports exceeds a rebalancing threshold, and wherein the dynamically adjusting storage port allocations comprises rebalancing the individual workloads to within the rebalancing threshold.

15. The computer program product of claim 10, wherein the real-time evaluation of the workloads being serviced by the plurality of storage ports comprises identifying traffic type flags of traffic of the workload, and wherein the dynamically adjusting storage port allocations comprises queuing the traffic according to a prioritization schedule that prioritizes traffic based on traffic type.

16. The computer program product of claim 10, wherein the real-time evaluation of the workloads being serviced by the plurality of storage ports comprises evaluating priority of incoming storage commits, and wherein the dynamically adjusting storage port allocations comprises queuing the incoming storage commits according to a prioritization schedule that prioritizes storage commits based on storage commit type.

17. A computer system comprising:

a memory; and

a processor in communication with the memory, wherein the computer system is configured to perform a method, the method comprising;

receiving by a storage device of a data center, from an application using the storage device, a notification of a change in elastic network bandwidth allocation provisioned to the data center from an elastic network service provider of an elastic network;

based on the notification, obtaining indications of, and determining, anticipated demand on a plurality of storage ports of the storage device, the anticipated demand being at least in part for storage input/output supporting transfer of data to or from the elastic network, wherein an indication of the anticipated demand indicates an amount of the change in the elastic network bandwidth allocation, and the anticipated demand is based at least in part on the change in the elastic network bandwidth allocation and the indicated amount of that change; and

the storage device dynamically adjusting storage port allocations based on the determined anticipated demand and on real-time evaluation of workloads being serviced by the plurality of storage ports, the dynamically adjusting comprising adjusting respective individual storage input/output bandwidth or workload for at least one storage port of the plurality of storage ports, wherein the dynamically adjusting is performed based on the notification of the change in elastic network bandwidth allocation provisioned to the data center.

18. The computer system of claim 17, wherein an indication of anticipated demand comprises a signal of a change in upcoming demand from an application of the data center, the upcoming demand being upcoming demand on a first storage port set comprising one or more first storage ports of the plurality of storage ports, and wherein the method further comprises predicting, based on the signal of the change in upcoming demand from the application of the data center, and using a statistical model, a level of the upcoming demand on the first storage port set, and predicting a duration of the upcoming demand, wherein the dynamically adjusting storage port allocations comprises allocating bandwidth of a second storage port set of one or more second storage ports of the plurality of storage ports to service at least some of the upcoming demand.

19. The computer system of claim 17, wherein the anticipated demand comprises a predicted near-future demand on one or more storage ports of the plurality of storage ports, the predicted near-future demand being predicted based on a current demand on the one or more storage ports.