ADAPTIVE STORAGE RELIABILITY MANAGEMENT

• US 20160110249A1
• Filed: 11/03/2014
• Published: 04/21/2016
• Est. Priority Date: 10/20/2014
• Status: Active Grant

• Current Assignee: SanDisk Technologies LLC (Western Digital Corporation)
• Original Assignee: Fusion-IO Incorporated (Western Digital Corporation)
• Inventors: Orme, Evan, Peterson, James G.
• Granted Patent: US 9,952,781 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1016 Error in accessing a memory location, i.e. addressing error
• CPC Class Codes: G06F 11/1044 with specific ECC/EDC distribution
• CPC Class Codes: G06F 11/1048 using arrangements adapted for a specific error detection or correction feature
• CPC Class Codes: G06F 11/1072 in multilevel memories
• CPC Class Codes: G06F 12/0238 Memory management in non-volatile memory, e.g. resistive RAM or ferroelectric memory
• CPC Class Codes: G06F 2212/1032 Reliability improvement, data loss prevention, degraded operation etc
• CPC Class Codes: G06F 2212/152 Virtualized environment, e.g. logically partitioned system
• CPC Class Codes: G06F 2212/202 Non-volatile memory
• CPC Class Codes: G06F 2212/214 Solid state disk
• CPC Class Codes: G06F 2212/7205 Cleaning, compaction, garbage collection, erase control
• CPC Class Codes: G11C 16/16 for erasing blocks, e.g. arrays, words, groups
• CPC Class Codes: G11C 2029/0409 Online test
• CPC Class Codes: G11C 29/52 Protection of memory contents; Detection of errors in memory contents

Abstract

A storage controller identifies a storage location within a storage division that corresponds to a high error rate. In response, the storage controller may refresh data stored on the storage division by relocating data from the storage division and/or initializing (e.g., erasing) the storage division. In some embodiments, the storage division is selectively refreshed by relocating data from the storage location(s) having high error rates, while deferring a full relocation of other data from the storage division. The storage division may be selectively refreshed based on reliability characteristics of the storage division, such as the remaining data retention time calculated for the storage division.

Claims

1. A method, comprising:

determining a reliability metric of a first storage unit within an erase division comprising a plurality of storage units;

relocating data stored on the first storage unit from the erase division in response to the determined reliability metric; and

deferring relocation of data stored on the erase division based on a remaining retention time of the erase division.

2. The method of claim 1, further comprising determining the reliability metric of the first storage unit in response to read operation on the first storage unit, and wherein relocating the data stored on the first storage unit comprises writing data of the read operation within a different erase division.

3. The method of claim 2, further comprising correcting read errors in the data stored on the first storage unit by use of an error-correcting code.

4. The method of claim 1, wherein relocating the data stored on the first storage unit comprises:

identifying data for relocation from the erase division based on reliability metrics of the respective storage units of the erase division; and

selectively relocating the identified data to one or more other erase divisions.

5. The method of claim 1, further comprising determining the remaining retention time of the erase division based on a reliability metric of a second storage unit within the erase division.

6. The method of claim 1, further comprising determining the remaining retention time of the erase division based on a reliability metric of a set of one or more storage units within the erase division excluding the first storage unit.

7. The method of claim 6, wherein determining the remaining retention time of the erase division comprises determining a current error rate of the erase division based on error rates corresponding to read operations performed on the set of storage units within the erase division.

8. The method of claim 1, further comprising:

identifying a subset of the storage units within the erase division that are in use to store valid data, and

determining the remaining retention time of the erase division based on read error metrics of storage units in the identified subset.

9. The method of claim 8, wherein identifying the subset of storage units comprises identifying storage units within the erase division that are associated with logical addresses in a forward map.

10. The method of claim 8, wherein identifying the subset of storage units comprises identifying storage units of the erase division that comprise invalid data.

11. The method of claim 1, wherein the erase division comprises one of an erase block, a group of erase blocks, and a logical erase block.

12. The method of claim 1, wherein deferring relocation of data stored on the erase division comprises postponing a garbage collection operation on the erase division, wherein the garbage collection operation comprises:

identifying valid data stored on the erase division;

relocating the valid data stored on the erase division to another erase division of the storage medium; and

initializing the erase division.

13. An apparatus, comprising:

a reliability monitor configured to identify a physical address within a storage division of a storage medium having an error rate that exceeds an error rate threshold;

a reliability manager configured to determine a reliability metric of cells of the storage division in response to identifying the physical address; and

a partial relocation module configured to write data stored at the identified physical address to a different physical address such that other data is retained on the storage division.

14. The apparatus of claim 13, further comprising a translation layer to associate logical addresses of a logical address space with respective physical addresses, wherein the partial relocation module is configured to remap a logical identifier from the identified physical address to the different physical address by use of the translation layer, such that one or more other logical addresses remain associated with respective physical addresses within the storage division.

15. The apparatus of claim 13, wherein the partial relocation module postpones refresh of the storage division in response to the determined reliability metric exceeding a threshold.

16. The apparatus of claim 13, wherein the reliability manager determines the reliability metric based on one or more of an error rate of a different physical address within the storage division, a reliability projection for the storage division, a reliability model for the storage division, a partial reliability metric of the storage division, an error rate projection, and a data retention guarantee.

17. The apparatus of claim 13, wherein the reliability metric is based on test read operations performed on other portions of the storage division, excluding the identified physical address.

18. The apparatus of claim 13, wherein the storage manager identifies the physical address having the error rate that exceeds the error rate threshold in response to one of servicing a request to read data from the physical address and a reliability scan.

19. A system, comprising:

a flash storage medium comprising a plurality of erase blocks, each erase block comprising a plurality of pages;

means for identifying a page of a particular one of the erase blocks having a bit error rate that exceeds a bit error rate threshold;

means for determining a remaining data retention time for data stored on other portions of the particular erase block; and

means for postponing recovery of the particular erase block in response to the determined remaining data retention time exceeding a duration threshold, wherein postponing recovery comprises writing data stored on the identified page to a page on a different erase block of the flash storage medium, and retaining data stored on other pages of the particular erase block.

20. The system of claim 19, wherein the determined remaining data retention time of the particular erase block is based on one or more of:

time elapsed since the particular erase block was programmed,

bit error rates of other pages of the particular erase block,

a bit error rate projection for the particular erase block,

a bit error rate model for the particular erase block,

a predetermined schedule for grooming operations on erase blocks of the flash storage medium, and

a data retention guarantee pertaining to data stored on the flash storage medium.