Minimizing write operations to a flash memory-based object store

• US 8,700,842 B2
• Filed: 01/03/2011
• Issued: 04/15/2014
• Est. Priority Date: 04/12/2010
• Status: Active Grant

• Current Assignee: SanDisk Technologies LLC (Western Digital Corporation)
• Original Assignee: SanDisk Enterprise IP LLC (Western Digital Corporation)
• Inventors: Dinker, Darpan
• Primary Examiner(s): NGUYEN, HIEP T
• Application Number:
  US12/983,762
• Publication Number:
  US 20110252186A1
• Time in Patent Office: 1,198 Days
• US Class Current: 711/103
• CPC Class Codes: G06F 11/1446 Point-in-time backing up or restoration of persistent data
• CPC Class Codes: G06F 11/1474 in transactions G06F16/20 takes precedence
• CPC Class Codes: G06F 11/2023 Failover techniques
• CPC Class Codes: G06F 11/2071 using a plurality of controllers
• CPC Class Codes: G06F 11/2082 Data synchronisation
• CPC Class Codes: G06F 12/08 in hierarchically structured memory systems, e.g. virtual memory systems
• CPC Class Codes: G06F 16/23 Updating
• CPC Class Codes: G06F 16/24569 Query processing with adaptation to specific hardware, e.g. adapted for using GPUs or SSDs
• CPC Class Codes: G06F 3/0613 in relation to throughput
• CPC Class Codes: G06F 3/0667 at data level, e.g. file, record or object virtualisation
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
• CPC Class Codes: G06F 3/0679 Non-volatile semiconductor memory device, e.g. flash memory, one time programmable memory [OTP]
• CPC Class Codes: G06F 3/0685 Hybrid storage combining heterogeneous device types, e.g. hierarchical storage, hybrid arrays

Abstract

Approaches for minimizing the amount of write transactions issued to an object store maintained on a solid state device (SSD). Transactions requested against an object store maintained on a SSD may be committed once transaction information for the transaction is durably stored in a non-volatile dynamic random access memory (DRAM), which may be maintained in a HDD controller. Further, data blocks stored in a volatile cache of a database server that issues write requests to an object store maintained on one or more SSDs may be considered persistent stored once confirmation is received that the data blocks are written to a double-write buffer stored on a non-volatile medium, such as NV RAM in a HDD controller. Additionally, any data blocks that are to be written over in a non-volatile DRAM are first ensured to be no longer present within the volatile write cache maintained a the solid state device.

Claims

1. A method for prolonging the service life of a solid state device, comprising:

in response to receiving a first request to commit a transaction against an object store maintained on one or more solid state devices, sending, to a non-volatile dynamic random access memory (DRAM), a second request to store transaction information for the transaction in the non-volatile DRAM; and

in response to receiving a first response indicating that the transaction information has been persistently stored in a transaction log on the non-volatile DRAM, sending a second response indicating that the transaction, made against the object store maintained on the one or more solid state devices, has been committed.

2. The method of claim 1, wherein the non-volatile DRAM is comprised within a hard-disk drive (HDD) controller.

3. The method of claim 1, wherein transaction information for multiple transactions is stored sequentially in the transaction log maintained on the non-volatile DRAM, wherein each of the multiple transactions is made against and updates information in the object store maintained on the one or more solid state devices.

4. The method of claim 1, wherein at least one of the one or more solid state devices is a flash memory chip.

5. A method for prolonging the service life of a solid state device, comprising:

in response to receiving a request to persistently store one or more blocks stored in a volatile cache of a database server that issues write requests to an object store maintained on one or more solid state devices, writing the one or more blocks to a double-write buffer stored on a non-volatile storage medium, wherein the non-volatile storage medium is not one of the one or more solid state devices; and

in response to receiving a first message indicating that the one or more blocks have been written to the double-write buffer stored on the non-volatile storage medium, sending a second message indicating that the one or more blocks have been persistently stored in the object store.

6. The method of claim 5, wherein the non-volatile storage medium is non-volatile dynamic random access memory (DRAM) of a hard-disk drive (HDD) controller.

7. The method of claim 5, further comprising:

identifying, within the double-write buffer, a plurality of data blocks to be written to the one or more solid state devices; and

subsequent to sending the second message, writing the plurality of data blocks to the one or more solid state devices in a single operation.

8. The method of claim 5, further comprising:

identifying a configurable number of dirty blocks within the volatile cache of the database server, wherein a dirty data block is a data block that has changes made to the data block which have not been persistently stored to a solid state device or to the non-volatile storage medium of the HDD.

9. A method for maintaining a double-write buffer for a database stored on one or more solid state devices, comprising:

writing a first set of data blocks to a volatile write cache maintained on a solid state device; and

in response to receiving a request to write a second set of data blocks to the double-write buffer maintained in a non-volatile dynamic random access memory (DRAM), ensuring that any data blocks that are to be written over in the non-volatile DRAM as a result of writing the second set of data blocks to the non-volatile DRAM are no longer present within the volatile write cache maintained on the solid state device.

10. The method of claim 9, wherein ensuring comprises receiving confirmation that the entire contents of the volatile write cache have been persistently stored on the solid state device.

11. The method of claim 9, wherein the non-volatile RAM resides within a hard-disk drive (HDD) controller.

12. A non-transitory machine readable storage medium storing one or more sequences of instructions for prolonging the service life of a solid state device, wherein execution of the one or more sequences of instructions by one or more processors causes:

in response to receiving a first request to commit a transaction against an object store maintained on one or more solid state devices, sending, to a non-volatile dynamic random access memory (DRAM), a second request to store transaction information for the transaction in the non-volatile DRAM; and

in response to receiving a first response indicating that the transaction information has been persistently stored in a transaction log on the non-volatile DRAM, sending a second response indicating that the transaction, made against the object store maintained on the one or more solid state devices, has been committed.

13. The non-transitory machine readable storage medium of claim 12, wherein the non-volatile DRAM is comprised within a hard-disk drive (HDD) controller.

14. The non-transitory machine readable storage medium of claim 12, wherein transaction information for multiple transactions is stored sequentially in the transaction log maintained on the non-volatile DRAM, wherein each of the multiple transactions is made against and updates information in the object store maintained on the one or more solid state devices.

15. The non-transitory machine readable storage medium of claim 12, wherein at least one of the one or more solid state devices is a flash memory chip.

16. A non-transitory machine readable storage medium storing one or more sequences of instructions for prolonging the service life of a solid state device, wherein execution of the one or more sequences of instructions by one or more processors causes:

in response to receiving a request to persistently store one or more blocks stored in a volatile cache of a database server that issues write requests to an object store maintained on one or more solid state devices, writing the one or more blocks to a double-write buffer stored on a non-volatile storage medium, wherein the non-volatile storage medium is not one of the one or more solid state devices; and

in response to receiving a first message indicating that the one or more blocks have been written to the double-write buffer stored on the non-volatile storage medium, sending a second message indicating that the one or more blocks have been persistently stored in the object store.

17. The non-transitory machine readable storage medium of claim 16, wherein the non-volatile storage medium is non-volatile dynamic random access memory (DRAM) of a hard-disk drive (HDD) controller.

18. The non-transitory machine readable storage medium of claim 16, wherein execution of the one or more sequences of instructions by one or more processors further causes:

identifying, within the double-write buffer, a plurality of data blocks to be written to the one or more solid state devices; and

subsequent to sending the second message, writing the plurality of data blocks to the one or more solid state devices in a single operation.

19. The non-transitory machine readable storage medium of claim 16, wherein execution of the one or more sequences of instructions by one or more processors further causes:

identifying a configurable number of dirty blocks within the volatile cache of the database server, wherein a dirty data block is a data block that has changes made to the data block which have not been persistently stored to a solid state device or to the non-volatile storage medium of the HDD.

20. A non-transitory machine readable storage medium storing one or more sequences of instructions for maintaining a double-write buffer for a database stored on one or more solid state devices, wherein execution of the one or more sequences of instructions by one or more processors causes:

writing a first set of data blocks to a volatile write cache maintained on a solid state device; and

in response to receiving a request to write a second set of data blocks to the double-write buffer maintained in a non-volatile dynamic random access memory (DRAM), ensuring that any data blocks that are to be written over in the non-volatile DRAM as a result of writing the second set of data blocks to the non-volatile DRAM are no longer present within the volatile write cache maintained on the solid state device.

21. The non-transitory machine readable storage medium of claim 20, wherein ensuring comprises receiving confirmation that the entire contents of the volatile write cache have been persistently stored on the solid state device.

22. The non-transitory machine readable storage medium of claim 20, wherein the non-volatile RAM resides within a hard-disk drive (HDD) controller.