Method and apparatus for assigning raid levels

• US 6,718,434 B2
• Filed: 05/31/2001
• Issued: 04/06/2004
• Est. Priority Date: 05/31/2001
• Status: Expired due to Fees

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Alvarez, Guillermo, Swaminathan, Ram, Anderson, Eric A., Veitch, Alistair C., Romer, Ted, Golding, Richard
• Primary Examiner(s): Elmore, Reba I.
• Application Number:
  US09/871,043
• Publication Number:
  US 20020184442A1
• Time in Patent Office: 1,041 Days
• Field of Search: 711/114
• US Class Current: 711/114
• CPC Class Codes: G06F 11/1096 Parity calculation or recalculation after configuration or reconfiguration of the system
• CPC Class Codes: G06F 3/0605 by facilitating the interaction with a user or administrator
• CPC Class Codes: G06F 3/0614 Improving the reliability of storage systems
• CPC Class Codes: G06F 3/0629 Configuration or reconfiguration of storage systems
• CPC Class Codes: G06F 3/064 Management of blocks
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD

Abstract

RAID levels are assigned to data prior to loading the data on a disk array or other data storage device. The RAID levels are determined by applying an algorithm (e.g., a rule-based criteria, a utilization model) to data workload specifications and device specifications.

Claims

1. A method of assigning RAID levels to data prior to loading the data in a data storage device, the device allowing multiple RAID levels, the method comprising accessing at least one of a set of device specifications and workload specifications that characterize the data to be loaded in the data storage device;

and applying an algorithm to the at least one set to determine at least one RAID level for the data.

2. The method of claim 1, wherein the algorithm also identifies LUN sizes for the data.

3. The method of claim 1, wherein the data is characterized as at least one store, and wherein the algorithm includes at least one rule that is applied to at least one store.

4. The method of claim 3, wherein the data is further characterized as at least one stream, and wherein the algorithm includes at least one rule that is also applied to at least one stream.

5. The method of claim 4, wherein attributes from multiple streams bound to a store are combined to produce an expected workload for the store.

6. The method of claim 1, wherein the data is characterized by estimated workloads.

7. The method of claim 1, wherein the algorithm includes at least one rule;

and wherein at least one rule is applied to the data, and a RAID level corresponding to a satisfied rule is assigned to the data.

8. The method of claim 7, wherein the RAID level corresponding to the first rule that is satisfied is assigned to the data.

9. The method of claim 7, wherein a default rule is applied to the data if none of the rules fire.

10. The method of claim 7, wherein the RAID level is determined by applying at least one utilization rule to the data.

11. The method of claim 10, wherein the utilization rules include a rule for capacity-bound workloads, and wherein an appropriate RAID level is assigned to capacity-bound workloads.

12. The method of claim 10, wherein at least one of sequence length, I/O request size and rates and overlaps between streams are considered when selecting the RAID level.

13. The method of claim 10, wherein the utilization rules include a rule for bandwidth-bound workloads, and wherein an appropriate RAID level is assigned to bandwidth-bound workloads.

14. The method of claim 7, wherein the algorithm also identifies LUN sizes for the data;

and wherein a further set of rules is applied on an aggregate set of data to minimize the number of LUNs.

15. The method of claim 1, wherein the algorithm includes a model of the device and data workload, which is used to determine the RAID levels.

16. The method of claim 15, wherein the model includes a set of formulas that, given a particular workload, or combination of workloads, determines the utilization of components of the device, and wherein the utilization is calculated for each component as a limit of at least one of the number of operations per second and total bandwidth available.

17. The method of claim 15, wherein the model is used to determine device configuration parameters in addition to RAID levels;

and wherein the RAID levels and device configuration parameters are determined by selecting the RAID level that results in the smallest array utilization.

18. The method of claim 15, wherein the data is characterized as at least one store and stream;

and wherein if a store is not capacity-bound, at least one of the following numbers is estimated and the RAID level resulting in the smallest number is selected;

available bandwidth and the number of I/O operations per second (IOPS) that will be generated by each stream for each RAID level being considered.

19. A method of configuring a data storage device allowing multiple levels of a redundancy hierarchy, the method comprising:

using rule-based criteria to determine redundancy levels and logical unit sizes for the device; and

configuring the data storage device for logical units having the sizes and redundancy levels determined by the criteria.

20. Apparatus for assigning RAID levels to data prior to loading the data in a data storage device, the apparatus comprising a processor for applying an algorithm to at least one of a set of device specifications and workload specifications for the data to determine RAID levels for the data;

and configuring the data storage device according to the RAID levels identified by the algorithm.

21. The apparatus of claim 20, wherein the algorithm also identifies LUN sizes for the data;

and wherein the logical units are also configured according to the LUN sizes identified by the criteria.

22. The apparatus of claim 20, wherein the data is characterized as at least one store, and wherein the algorithm includes at least one rule that is applicable to at least one store.

23. The apparatus of claim 22, wherein the data is further characterized as at least one stream, and wherein the algorithm includes at least one rule that is also applicable to at least one stream.

24. The apparatus of claim 23, wherein the algorithm adds attributes from multiple streams bound to a store to predict expected utilization of the device.

25. The apparatus of claim 20, wherein the algorithm includes at least one rule;

and wherein at least one rule is applied to the data, and a RAID level corresponding to a satisfied rule is assigned to the data.

26. The apparatus of claim 25, wherein the algorithm assigns the RAID level corresponding to the first-firing rule to the data.

27. The apparatus of claim 25, wherein the algorithm applies a default rule to the data if none of the rules fire.

28. The apparatus of claim 25, wherein the algorithm applies at least one utilization rule to the data.

29. The apparatus of claim 28, wherein the utilization rules include a rule for capacity-bound workloads, and wherein an appropriate RAID level is assigned to capacity-bound workloads.

30. The apparatus of claim 28, wherein the utilization rules include a rule for bandwidth-bound workloads, and wherein an appropriate RAID level is assigned to bandwidth-bound workloads.

31. The apparatus of claim 20, wherein the algorithm includes a model of the device and data workload, the model determining the RAID levels.

32. The apparatus of claim 31, wherein the model includes a set of formulas that, given a particular workload, or combination of workloads, determines the utilization of components of the device.

33. The apparatus of claim 31, wherein the model determines device configuration parameters in addition to RAID levels.

34. The apparatus of claim 31, wherein the data is characterized as at least one store and stream;

and wherein if a store is not capacity-bound, the model estimates the number of I/O operations per second (IOPS) that will be generated by each stream for each RAID level being considered, and the algorithm selects the RAID level resulting in the smallest number of per-disk IOPS.

35. A system comprising:

a data storage device having multiple RAID levels;

means for applying an algorithm to at least one of a set of device specifications and workload specifications for data that will be stored in the device, the algorithm identifying RAID levels for the data; and

means for configuring the data storage device according to the identified RAID levels prior to initially storing the data in the data storage device.

36. An article for instructing a processor to assign RAID levels to data prior to loading the data in a data storage device, the data characterized by at least one of device specifications and workload specifications the device having multiple RAID levels, the article comprising:

computer memory; and

a RAID level identifying program encoded in the memory, the program, when executed, instructing the processor to process at least one of the characterizing specifications to identify RAID levels for the data prior to loading the data in the data storage device.