Reducing write amplification in a flash memory

• US 8,892,811 B2
• Filed: 03/01/2012
• Issued: 11/18/2014
• Est. Priority Date: 03/01/2012
• Status: Active Grant

• Current Assignee: Avago Technologies International Sales Pte Limited (Broadcom, Inc.)
• Original Assignee: Avago Technologies General IP PTE Limited (Broadcom, Inc.)
• Inventors: Ish, Mark, Panda, Siddhartha K.
• Primary Examiner(s): CHOE, YONG J
• Application Number:
  US13/409,458
• Publication Number:
  US 20130232290A1
• Time in Patent Office: 992 Days
• Field of Search: 711/103, 711/E12.008
• US Class Current: 711/103
• CPC Class Codes: G06F 12/0246 in block erasable memory, e.g. flash memory
• CPC Class Codes: G06F 2212/7205 Cleaning, compaction, garbage collection, erase control

Abstract

An apparatus having a memory circuit and a manager is disclosed. The memory circuit generally has (i) one or more Flash memories and (ii) a memory space that spans a plurality of memory addresses. The manager may be configured to (i) receive data items in a random order from one or more applications, (ii) write the data items in an active one of a plurality of regions in a memory circuit and (iii) mark the memory addresses in the active region that store the data items as used. Each data item generally has a respective host address. The applications may be executed in one or more computers. The memory addresses in the active region may be accessed in a sequential order while writing the data items to minimize a write amplification. The random order is generally preserved between the data items while writing in the active region.

Claims

1. An apparatus comprising:

a memory circuit comprising (i) one or more Flash memories and (ii) a memory space that spans a plurality of memory addresses; and

a manager configured to (i) receive a plurality of data items and respective host addresses in a random order from one or more applications, (ii) transform said respective host addresses to said memory addresses, (iii) write said data items in an active one of a plurality of regions in said memory circuit and (iv) mark said memory addresses in said active region that store said data items as used, wherein (a) said applications are executed in one or more computers, (b) said memory addresses in said active region are accessed in a sequential order while writing said data items to minimize a write amplification and (c) said random order is preserved between said data items while writing in said active region.

2. The apparatus according to claim 1, wherein said manager is further configured to (i) read one or more of said data items (a) having valid data and (b) are currently stored in an inactive one of said regions in response to a condition and (ii) write said data items with said valid data in said active region, wherein said writing of said data items read from said inactive region is intermixed with said writing of said data items received from said applications.

3. The apparatus according to claim 2, wherein said condition is true in response to (i) at least a threshold number of said regions having been written to and (ii) said inactive region having a largest number of memory addresses marked as free.

4. The apparatus according to claim 1, wherein said manager is further configured to mark said memory addresses corresponding to said data items stored in an inactive one of said regions as free in response to said data items being copied from said inactive region to said active region.

5. The apparatus according to claim 1, wherein said manager is further configured to mark said memory address corresponding to a given one of said data items stored in said memory circuit as free in response to writing a new one of said data items in said memory circuit where said new data item has a same host address as said given data item.

6. The apparatus according to claim 1, wherein said manager is further configured to mark said memory address corresponding to a given one of said data items stored in said memory circuit as free in response to a trim command received from said applications, wherein said trim command identifies said given data item by said host address.

7. The apparatus according to claim 1, wherein said manager is further configured to (i) receive a read request having a given one of said host addresses from said applications, (ii) transform said given host address into a given one of said memory addresses and (iii) read said data item stored at said given memory address from said memory circuit.

8. The apparatus according to claim 1, wherein (i) a start and an end of each of said regions align to a respective boundary of a plurality of erasable memory blocks of said memory circuit, (ii) each of said erasable memory blocks comprises a plurality of said regions and (iii) said memory circuit implements a solid-state drive.

9. The apparatus according to claim 1, wherein said manager comprises software executed in at least one of (i) said computers and (ii) said memory circuit.

10. The apparatus according to claim 1, wherein said apparatus is implemented as one or more integrated circuits.

11. A method for reducing write amplification in a Flash memory, comprising the steps of:

(A) receiving a plurality of data items and respective host addresses in a random order at a manager from one or more applications executed in one or more computers;

(B) transforming said respective host addresses to said memory addresses;

(C) writing said data items in an active one of a plurality of regions in a memory circuit, wherein (i) said memory circuit comprises (a) one or more of said Flash memories and (b) a memory space that spans a plurality of memory addresses, (ii) said memory addresses in said active region are accessed in a sequential order while writing said data items to minimize said write amplification and (iii) said random order is preserved between said data items while writing in said active region; and

(D) marking said memory addresses in said active region that store said data items as used.

12. The method according to claim 11, further comprising the steps of:

reading one or more of said data items (i) having valid data and (ii) are currently stored in an inactive one of said regions in response to a condition; and

writing said data items with said valid data in said active region, wherein said writing of said data items read from said inactive region is intermixed with said writing of said data items received from said applications.

13. The method according to claim 12, wherein said condition is true in response to (i) at least a threshold number of said regions having been written to and (ii) said inactive region having a largest number of memory addresses marked as free.

14. The method according to claim 11, further comprising the step of:

marking said memory addresses corresponding to said data items stored in an inactive one of said regions as free in response to said data items being copied from said inactive region to said active region.

15. The method according to claim 11, further comprising the step of:

marking said memory address corresponding to a given one of said data items stored in said memory circuit as free in response to writing a new one of said data items in said memory circuit where said new data item has a same host address as said given data item.

16. The method according to claim 11, further comprising the step of:

marking said memory address corresponding to a given one of said data items stored in said memory circuit as free in response to a trim command received by said manager, wherein said trim command identifies said given data item by said host address.

17. The method according to claim 11, further comprising the steps of:

receiving a read request having a given one of said host addresses at said manager from said applications;

transforming said given host address into a given one of said memory addresses; and

reading said data item stored at said given memory address from said memory circuit.

18. The method according to claim 11, wherein (i) a start and an end of each of said regions align to a respective boundary of a plurality of erasable memory blocks of said memory circuit, (ii) each of said erasable memory blocks comprises a plurality of said regions and (iii) said memory circuit implements a solid-state drive.

19. The method according to claim 11, wherein said manager comprises software executed in at least one of (i) said computers and (ii) said memory circuit.

20. A non-transitory computer-readable storage medium containing processor executable instructions configured to perform the steps of claim 11.