DYNAMIC PERFORMANCE VIRTUALIZATION FOR DISK ACCESS

• US 20100122020A1
• Filed: 11/11/2009
• Published: 05/13/2010
• Est. Priority Date: 11/13/2008
• Status: Active Grant

• Current Assignee: Innovations In Memory LLC
• Original Assignee: Gridiron Systems Inc. (Violin Memory Incorporated)
• Inventors: da la Iglesia, Erik, Sikdar, Som
• Granted Patent: US 8,285,961 B2
• Time in Patent Office: Days
• US Class Current: 711/103
• CPC Class Codes: G06F 3/061 Improving I/O performance
• CPC Class Codes: G06F 3/0631 by allocating resources to storage systems
• CPC Class Codes: G06F 3/0653 Monitoring storage devices or systems
• CPC Class Codes: G06F 3/0665 at area level, e.g. provisioning of virtual or logical volumes
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]

Abstract

A storage control system includes performance monitor logic configured to track performance parameters for different volumes in a storage array. Service level enforcement logic is configured to assign target performance parameters to the different volumes and generate metrics for each of the different volumes identifying how much the performance parameters change for the different volumes responsive to changes in the amounts of tiering media allocated to the different volumes. Resource allocation logic is configured to allocate the tiering media to the different volumes according to the performance parameters, target performance parameters, and metrics for the different volumes.

Claims

1. An apparatus, comprising:

a set of tiering media; and

logic circuitry configured to;

identify service level agreements associated with different storage volumes;

monitor storage access performance for the different storage volumes;

compare the storage access performance with the service level agreements associated with the different storage volumes; and

allocate different portions of the tiering media to the different storage volumes according to comparisons between the storage access performance and the service level agreements for the different storage volumes.

2. The apparatus according to claim 1 wherein the logic circuitry is further configured to:

calculate metrics identifying changes in the storage access performance for the different storage volumes responsive to changes in the allocation or deallocation of the tiering media to the different storage volumes; and

allocate the tiering media to the different storage volumes according to the metrics.

3. The apparatus according to claim 2 wherein the tiering media includes both Dynamic Random Access Memory (DRAM) and Flash memory and the logic circuitry is configured to calculate separate RAM metrics and Flash metrics and allocate the RAM and Flash memory to the different storage devices according to the RAM metrics and Flash metrics.

4. The apparatus according to claim 2 wherein the logic circuitry is configured to deallocate the tiering media from the different storage volumes that have metrics indicating no change or negative changes in the storage access performance.

5. The apparatus according to claim 1 wherein the logic circuitry is configured to determine average read latencies for the different storage volumes over associated time intervals and allocate portions of the tiering media to the different storage volumes according to the average read latencies for the different storage volumes.

6. The apparatus according to claim 1 wherein the logic circuitry is configured to determine a number of read operations provided by each of the different storage volumes over associated time intervals and allocate portions of the tiering media to the different storage volumes according to the number of read operations provided by each of the different storage volumes.

7. The apparatus according to claim 1 wherein the logic circuitry is configured to determine a number of blocks provided by the different storage volumes over associated time intervals and allocate portions of the tiering media to the different storage volumes according to the number of blocks provided by the different storage volumes.

8. The apparatus according to claim 1 wherein the service level agreements identify a target number of read operations, target number of read blocks, or a target average read latency for each of the different storage volumes.

9. The apparatus according to claim 1 wherein the logic circuitry is configured to:

generate violation values for each of the different storage volumes indicating how badly each of the different storage volumes violate the associated service level agreements; and

prioritize allocation of the tiering media to the different storage volumes according to the violation values.

11. The method according to claim 9 wherein the memory access operations include a number of read operations provided by a particular disk within a given time period, an amount of time required to complete the read operations to the particular disk, and an amount of data transferred from the particular disk during the given time period.

10. A method, comprising:

monitoring memory access operations between clients and disks in a disk storage array;

calculating performance metrics for tiering media used in conjunction with the memory access operations, wherein the performance metrics identify a change in performance for the memory access operations responsive to allocation of the tiering media to the disks in the disk storage array; and

allocating the tiering media to the disks in the disk storage array according to the performance metrics.

12. The method according to claim 10 further comprising:

identifying performance targets for different disks within the disk storage array;

identifying actual performance values for the different disks;

comparing the actual performance values with the performance targets; and

allocating the tiering media to the different disks according to how much the actual performance values for the disks differ from the performance targets for the disks.

13. The method according to claim 12 further comprising:

allocating an amount of tiering media to a particular disk wherein the actual performance values for the particular disk are below one or more performance targets for the particular disk;

measuring how much the actual performance values change due to the amount of tiering media allocated to the particular disk; and

allocating or deallocating additional amounts of tiering media to the particular disk according to how much the actual performance values change for the particular disk.

14. The method according to claim 13 further comprising:

assigning a maximum limit on the amount of tiering media that can be assigned to the particular disk;

identifying a total amount of tiering media currently assigned to the particular disk; and

assigning additional amounts of tiering media to the particular disk only when the total amount of media is below the maximum limit.

15. The method according to claim 13 further comprising:

storing a metric value for the particular disk identifying a ratio between an amount of change in the actual performance values and the amount of tiering media allocated to the particular disk; and

using the metric value to determine what additional amounts of tiering media to allocate to the particular disk.

16. The method according to claim 15 wherein the metric value identifies a ratio between a change in average read latency for the particular disk for a given time period and a previous amount of the tiering media allocated to the particular disk.

17. The method according to claim 15 wherein the metric value identifies a ratio between a change in a number of read operations provided by the particular disk for a given time period and a previous amount of the tiering media allocated to the particular disk.

18. The method according to claim 15 wherein the metric value identifies a ratio between a change in a number of blocks transferred by the particular disk for a given time period and a previous amount of the tiering media allocated to the particular disk.

19. A storage control system, comprising:

a first interface configured to receive storage access requests from one or more clients;

a second interface configured to transfer the storage access request to a storage array;

local tiering media;

performance monitor logic configured to track performance parameters for different volumes in the storage array, wherein the performance parameters indicate how fast the different volumes respond to the storage access requests, how many storage access requests are serviced by the different volumes over a predetermined time interval, or how much data is transferred by the different volumes over the predetermined time interval; and

resource allocation logic configured to allocate the tiering media to the different volumes according to the performance parameters for the different volumes.

20. The storage control system according to claim 19 further comprising service level enforcement logic configured to assign target performance parameters to the different volumes, wherein the resource allocation logic is configured to allocate the tiering media according to the performance parameters and the target performance parameters assigned to the different volumes.

21. The storage control system according to claim 20 wherein:

the service level enforcement logic is configured to generate metrics for each of the different volumes identifying how much the performance parameters change for the different volumes responsive to how much tiering media is allocated to the different volumes; and

the resource allocation logic is configured to allocate the tiering media according to the performance parameters for the different volumes, the target performance parameters assigned to the different volumes; and

the metrics for the different volumes.