Automated data relocation among storage tiers based on storage load

• US 8,429,346 B1
• Filed: 12/28/2009
• Issued: 04/23/2013
• Est. Priority Date: 12/28/2009
• Status: Active Grant

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Original Assignee: EMC Corporation (Dell Technologies Inc.)
• Inventors: Chen, Xiangping, Can, Khang, Madhukar, Manish, Harvey, David, Ku, Mark, Throop, Dean
• Primary Examiner(s): Dudek, Jr., Edward
• Assistant Examiner(s): BIRKHIMER, CHRISTOPHER D
• Application Number:
  US12/647,867
• Time in Patent Office: 1,212 Days
• Field of Search: None
• US Class Current: 711/114
• CPC Class Codes: G06F 3/0613 in relation to throughput
• CPC Class Codes: G06F 3/0649 Lifecycle management
• 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

Methods and systems are disclosed for relocating data in a physical storage pool comprising a plurality of storage tiers having differing performance characteristics, the physical storage pool being mapped into one or more logical units, each logical unit comprising a plurality of logical slices of data storage. The methods and systems can involve receiving a relocation list, the relocation list including, for each of a plurality of logical slices, a slice identifier, a temperature value, and a current physical location, determining a destination tier for each logical slice on the relocation list, evaluating for each logical slice on the relocation list a performance gain expected to be achieved by moving the logical slice from its current physical location to a new location in the destination tier, and relocating data in a logical slice from its current physical location to the new location based on the evaluation.

Claims

1. A method for relocating data in a physical storage pool comprising a plurality of storage tiers having differing performance characteristics, the physical storage pool being mapped into one or more logical units, each logical unit comprising a plurality of logical slices of data storage, the method comprising:

receiving a relocation list, the relocation list including, for each of a plurality of logical slices, a slice identifier, a temperature value, and a current physical location;

determining a destination tier for each logical slice on the relocation list;

evaluating for each logical slice on the relocation list a performance gain expected to be achieved by moving the logical slice from its current physical location to a new location in the destination tier; and

relocating data in a logical slice from its current physical location to the new location based on the evaluation;

wherein the destination tier comprises a plurality of RAID groups, each RAID group having an expected temperature value and a performance capability value, the performance capability value being proportional to the throughput of the RAID group; and

wherein the new location is in a destination RAID group selected based on a ratio of expected temperature value to performance capability value.

2. The method of claim 1, wherein the temperature value of each logical slice corresponds to a historical frequency at which the data represented by the logical slice is accessed by a user.

3. The method of claim 1, wherein the temperature value of each logical slice corresponds to a predicted frequency at which the data represented by the logical slice will be accessed by a user.

4. The method of claim 1, wherein the performance capability value is calculated based on a response time of the RAID group or a number of input and output operations that the RAID group can process in one second.

5. The method of claim 1, wherein the destination RAID group for each logical slice is chosen such that the RAID groups in the destination tier have balanced workloads.

6. The method of claim 1, wherein the destination RAID group has, among the plurality of RAID groups, the lowest ratio of expected temperature value to performance capability value.

7. The method of claim 1, wherein the destination RAID group for each logical slice is chosen based on one or more constraints specified by a user.

8. The method of claim 1, wherein relocating data based on the evaluation includes giving priority to data relocations having the greatest expected performance gain.

9. The method of claim 1, wherein evaluating a performance gain further comprises, for each logical unit, generating a list of logical slices corresponding thereto and then sorting each generated list of logical slices in order of the temperature value of each logical slice.

10. The method of claim 9, wherein evaluating a performance gain further comprises generating an execution list comprising a plurality of entries, each entry corresponding to the logical slice in each generated list of logical slices with the highest temperature value.

11. The method of claim 10, wherein evaluating a performance gain further comprises sorting the execution list in order of the expected performance gain for each logical slice.

12. The method of claim 11, wherein relocating data further comprises relocating data in each logical slice in the execution list.

13. The method of claim 1, wherein each logical slice corresponds to one gigabyte of data.

14. The method of claim 1, wherein data relocation is only performed during a time period specified by a user.

15. The method of claim 1, wherein data relocation is performed without interrupting user access to the data in each logical slice.

16. The method of claim 1, wherein only one logical slice of any given logical unit is relocated at any given time.

17. The method of claim 1, wherein relocating data comprises copying the data from its current physical location to the new location and then, if successful, marking the current physical location as empty.

18. A storage system comprising:

a physical storage pool including a plurality of storage tiers having differing performance characteristics, the physical storage pool being mapped into one or more logical units, each logical unit comprising a plurality of logical slices of data storage;

a processor coupled to memory and in communication with the physical storage pool;

wherein the processor is programmed to relocate data in a physical storage pool by;

receiving a relocation list, the relocation list including, for each of a plurality of logical slices, a slice identifier, a temperature value, and a current location, the current location including a current physical location;

determining a destination tier for each logical slice on the relocation list;

evaluating for each logical slice on the relocation list a performance gain expected to be achieved by moving the logical slice from its current physical location to a new location in the destination tier; and

relocating data in a logical slice from its current physical location to the new location based on the evaluation;

wherein the destination tier comprises a plurality of RAID groups, each RAID group having an expected temperature value and a performance capability value, the performance capability value being proportional to the throughput of the RAID group; and

wherein the new location is in a destination RAID group selected based on a ratio of expected temperature value to performance capability value.

19. A method for relocating data in a physical storage pool comprising a plurality of storage tiers having differing performance characteristics, the physical storage pool being mapped to a plurality of logical units, each logical unit comprising at least one slice, the method comprising:

receiving a pool candidate list including, for each of a plurality of slices, an entry comprising a slice identifier, a current tier, an optimal tier, and a temperature value, wherein the current tier for each listed slice is different from the optimal tier;

dividing the pool candidate list into a plurality of logical unit candidate lists, each logical unit candidate list including only entries corresponding to slices within a single logical unit;

generating an execution list that only includes, for each of the logical unit candidate lists, an entry corresponding to the slice in that logical unit candidate list with the highest temperature value; and

relocating the data corresponding to each slice in the execution list from a first physical location in the slice'"'"'s current tier to a second physical location in the slice'"'"'s optimal tier in order of expected performance gain.