Performance of a storage system

• US 7,979,613 B2
• Filed: 07/30/2008
• Issued: 07/12/2011
• Est. Priority Date: 07/15/2005
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Schwarz, Shemer Shimon, Zeidner, Efraim, Zohar, Ofir
• Primary Examiner(s): Lee; Chun-Kuan
• Assistant Examiner(s): Unelus; Ernest
• Application Number:
  US12/182,567
• Publication Number:
  US 20090037924A1
• Time in Patent Office: 1,077 Days
• Field of Search: 710/74, 710/57, 710/56, 710/52, 711/114, 711/163, 711/154, 718/105
• US Class Current: 710/74
• CPC Class Codes: G06F 11/2056 by mirroring
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3433 for load management allocation of a server based on load conditions G06F9/505; load rebalancing G06F9/5083; redistributing the load in a network by a load balancer H04L67/1029
• CPC Class Codes: G06F 11/3485 for I/O devices
• CPC Class Codes: G06F 2003/0697 device management, e.g. handlers, drivers, I/O schedulers
• CPC Class Codes: G06F 2201/88 Monitoring involving counting
• CPC Class Codes: G06F 2201/885 Monitoring specific for caches
• CPC Class Codes: G06F 2206/1012 Load balancing
• CPC Class Codes: G06F 2209/5022 Workload threshold
• CPC Class Codes: G06F 3/0601 Interfaces specially adapted for storage systems
• CPC Class Codes: G06F 3/0613 in relation to throughput
• CPC Class Codes: G06F 3/0653 Monitoring storage devices or systems
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD
• CPC Class Codes: G06F 9/505 considering the load

Abstract

A method for operating a storage system, including storing data redundantly in the system and measuring respective queue lengths of input/output requests to operational elements of the system. The queue lengths are compared to an average queue length to determine respective performances of the operational elements of the storage system. In response to the average queue lengths and a permitted deviation from the average an under-performing operational element among the operational elements is identified. An indication of the under-performing operational element is provided to host interfaces in the storage system. One of the host interfaces receives requests for specified items of the data directed to the under-performing element, and in response to the indication, some of the requests are diverted from the under-performing operational element to one or more other operational elements of the storage system that are configured to provide the specified items of the data.

Claims

1. A method for operating a storage system, comprising:

storing data redundantly in the storage system;

measuring respective queue lengths of input/output requests to operational elements of the storage system;

determining an average queue length of the operational elements;

determining a performance benchmark for the storage system, the performance benchmark being the average queue length of the operational elements plus a predetermined percentage set by a user;

comparing each of the respective queue lengths to the average of the queue lengths so as to determine respective performances of the operational elements of the storage system;

identifying, in response to the average of the queue lengths and the performance benchmark, an under-performing operational element among the operational elements;

providing an indication of the under-performing operational element to host interfaces comprised in the storage system, the indication identifying that the performance benchmark has been reached;

receiving requests at one of the host interfaces for specified items of the data directed to the under-performing element; and

in response to the indication, diverting a preset percentage of the requests for the specified items of the data directed to the under-performing element from the under-performing operational element to one or more other operational elements of the storage system that are configured to provide the specified items of the data.

2. The method according to claim 1, wherein diverting the preset percentage comprises registering all requests to the under-performing element and diverting the preset percentage of all of the requests to the one or more other operational elements.

3. The method according to claim 2, wherein the preset percentage is selected so that an overall performance of the storage system is optimized.

4. The method according to claim 1, wherein storing the data redundantly comprises storing the data according to at least one of a RAID (Redundant Array of Independent Disks) 0 and a RAID 1 storage scheme.

5. The method according to claim 1, wherein storing the data redundantly comprises storing the data according to at least one of a RAID 3, a RAID 4, a RAID 5, and a RAID 6 storage scheme.

6. The method according to claim 1, wherein the requests comprise a first set of sub-requests and a second set of sub-requests derived from a host-request to the system, and wherein diverting the preset percentage comprises:

diverting the first set of sub-requests from the under-performing element to the one or more other operational elements and receiving a first set of respective sub-responses therefrom; and

forwarding the second set of sub-requests to the one or more other operational elements and receiving a second set of respective sub-responses therefrom,

and further comprising collecting the first and second set of sub-responses so as to provide a host-response to the host-request.

14. The method according to claim 1, wherein the predetermined percentage set by the user is ten percent.

15. The method according to claim 14, wherein the preset percentage is five percent.

7. A method for operating a storage system, comprising:

storing data redundantly in the storage system;

measuring respective queue lengths of input/output requests to operational elements of the storage system;

determining an average queue length of the operational elements;

determining a performance benchmark for the storage system, the performance benchmark being the average queue length of the operational elements plus a predetermined percentage set by a user;

comparing each of the respective queue lengths to the average of the queue lengths so as to determine respective performances of the operational elements of the storage system;

identifying, in response to the average of the queue lengths and the performance benchmark, an over-performing operational element among the operational elements;

providing an indication of the over-performing operational element to host interfaces comprised in the storage system, the indication identifying that the performance benchmark is exceeded;

receiving requests at one of the host interfaces for specified items of the data directed to one or more of the operational elements other than the over-performing element;

in response to the indication, confirming that the over-performing element is configured to provide at least some of the specified items of the data in response to at least one of the requests; and

diverting the at least one of the requests to the over-performing operational element.

8. The method according to claim 7, wherein diverting the at least one of the requests comprises registering all requests to the over-performing element, computing a preset fraction of all the requests, and diverting the preset fraction of the requests received at the one of the host interfaces to the over-performing element.

9. The method according to claim 8, wherein the preset fraction is selected so that an overall performance of the storage system is optimized.

10. The method according to claim 7, wherein storing the data redundantly comprises storing the data according to at least one of a RAID (Redundant Array of Independent Disks) 0 and a RAID 1 storage scheme.

11. The method according to claim 7, wherein storing the data redundantly comprises storing the data according to at least one of a RAID 3, a RAID 4, a RAID 5, and a RAID 6 storage scheme.

12. The method according to claim 7, wherein the requests comprise a first set of sub-requests and a second set of sub-requests derived from a host-request to the system, and wherein diverting the at least one of the requests comprises:

diverting the first set of sub-requests to the over-performing element from the one or more other operational elements and receiving a first set of respective sub-responses therefrom; and

forwarding the second set of sub-requests to the one or more other operational elements and receiving a second set of respective sub-responses therefrom,

and further comprising collecting the first and second set of sub-responses so as to provide a host-response to the host-request.

16. The method according to claim 7, wherein the predetermined percentage set by the user is ten percent.

13. A method for operating a storage system, comprising:

storing data redundantly in the storage system;

measuring respective queue lengths of input/output requests to operational elements of the storage system;

determining an average queue length of the operational elements;

determining a performance benchmark for the storage system, the performance benchmark being the average queue length of the operational elements plus a predetermined percentage set by a user;

comparing each of the respective queue lengths to the average of the queue lengths so as to determine respective performances of the operational elements of the storage system;

identifying, in response to the average of the queue lengths and the performance benchmark, an under-performing operational element among the operational elements;

identifying, in response to the average of the queue lengths and the performance benchmark, an over-performing operational element among the operational elements;

providing a first indication of the under-performing operational element and a second indication of the over-performing operational element to host interfaces comprised in the storage system, the first indication identifying that the performance benchmark has been reached;

receiving requests at one of the host interfaces for specified items of the data directed to the under-performing element;

in response to the first indication, diverting a preset percentage of the requests for the specified items of the data directed to the under-performing element from the under-performing operational element to one or more other operational elements of the storage system that are configured to provide the specified items of the data;

in response to the second indication, confirming that the over-performing element is configured to provide at least some of the specified items of the data in response to at least one of the requests; and

diverting the at least one of the requests to the over-performing operational element.

17. The method according to claim 13, wherein the predetermined percentage set by the user is ten percent.

18. The method according to claim 17, wherein the preset percentage is five percent.