Enhancing tiering storage performance

• US 9,632,707 B2
• Filed: 03/26/2015
• Issued: 04/25/2017
• Est. Priority Date: 05/07/2012
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Hyde, II, Joseph S., Roy, Subhojit
• Primary Examiner(s): Yi, David X
• Assistant Examiner(s): PUCHE, TAHILBA O
• Application Number:
  US14/669,623
• Publication Number:
  US 20150205528A1
• Time in Patent Office: 761 Days
• CPC Class Codes: G06F 3/0608 Saving storage space on storage systems
• CPC Class Codes: G06F 3/061 Improving I/O performance
• CPC Class Codes: G06F 3/0611 in relation to response time
• CPC Class Codes: G06F 3/0641 De-duplication techniques
• CPC Class Codes: G06F 3/0649 Lifecycle management
• CPC Class Codes: G06F 3/0665 at area level, e.g. provisioning of virtual or logical volumes
• CPC Class Codes: G06F 3/0685 Hybrid storage combining heterogeneous device types, e.g. hierarchical storage, hybrid arrays
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD

Abstract

A method for improving data storage and retrieval performance, comprising, for a tiered storage environment having levels corresponding to storage performance, using reference count information of at least one data segment maintained in the storage environment to determine which of the levels in the environment the data segment will be assigned; obtaining input/output performance data for input/output operations on the data segment and transforming the data into short term moving averages and long term moving averages; comparing the averages in a performance level to the long term moving averages of the data segment in the performance level and examining the reference count information over time to determine whether a performance trend of the data segment is increasing or decreasing; and promoting the data segment to a higher performance level in the tiered storage environment, if the data segment has an increasing performance trend.

Claims

1. A method for improving data storage and retrieval performance in a computing environment by a processor, comprising:

for a tiered storage environment having levels corresponding to storage performance, using reference count information of at least one data segment maintained in the tiered storage environment to determine which of the levels in the tiered storage environment the at least one data segment will be assigned;

obtaining input/output performance data for input/output operations on the at least one data segment and transforming the input/output performance data into short term moving averages and long term moving averages;

sorting the at least one data segment having higher reference counts such that the at least one data segment having the most activity is at a top of a list;

comparing the short term moving averages of the at least one data segment in a performance level to the long term moving averages of the at least one data segment in the performance level and examining the reference count information of the at least one data segment in the performance level over time to determine whether a performance trend of the at least one data segment is increasing or decreasing; and

promoting the at least one data segment to a higher performance level in the tiered storage environment, if the at least one data segment has an increasing performance trend and a determination that a migration cost of migrating the at least one data segment to the higher performance level is justified;

wherein the migration cost is associated with an input/output count and a latency value.

2. The method of claim 1, wherein those of the at least one data segment having higher reference counts are assigned to a higher performance level than those having lower reference counts.

3. The method of claim 1, further including obtaining the reference count information from a data deduplication engine.

4. The method of claim 1, further including monitoring the tiered storage environment to associate the reference count information with the at least one data segment.

5. The method of claim 4, wherein upon a first access to the at least one data segment, and a determination that the at least one data segment has an associated reference count above a predetermined threshold, immediately migrating the at least one data segment to the higher performance level.

6. The method of claim 1, wherein using the reference count information includes using the reference count information as a weighted factor with additional characteristics of the at least one data segment to determine the which of the levels to assign the at least one data segment.

7. The method of claim 6, further including using an input/output activity, including at least one of a read count and a write count as the additional characteristics to determine the which of the levels to assign the at least one data segment.

8. The method of claim 7, further including subsequently using the input/output activity solely to determine if the at least one data segment should be migrated from a higher performance level to a lower performance level.