System and method for granular deduplication

• US 10,228,858 B1
• Filed: 08/14/2017
• Issued: 03/12/2019
• Est. Priority Date: 02/11/2015
• Status: Active Grant

• Current Assignee: Innovations In Memory LLC
• Original Assignee: Violin Systems LLC
• Inventors: Stoakes, Timothy, Ratna, Vikas, Garg, Amit
• Primary Examiner(s): Verbrugge, Kevin
• Application Number:
  US15/676,401
• Time in Patent Office: 575 Days
• Field of Search: None
• CPC Class Codes: G06F 3/0608 Saving storage space on storage systems
• CPC Class Codes: G06F 3/0619 in relation to data integrity, e.g. data losses, bit errors
• CPC Class Codes: G06F 3/0641 De-duplication techniques
• CPC Class Codes: G06F 3/065 Replication mechanisms
• CPC Class Codes: G06F 3/0665 at area level, e.g. provisioning of virtual or logical volumes
• 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 3/0688 Non-volatile semiconductor memory arrays

Abstract

A method of managing a memory system is described, the memory system having a processor with a client interface, a buffer memory, and a storage memory. A LUN (logical unit) is exposed to the client, the LUN having a memory size, and QoS specification. When the processor receives a command from the client to write data to the LUN, determining if the QoS specification includes deduplication and: if the QoS does not include deduplication, processing the command by storing the data in the storage memory and creating metadata uniquely referencing a storage location of the data that is written; or if the QoS includes deduplication, determine if deduplication is possible while meeting a latency value as specified in the QoS specification and performing deduplication.

Claims

1. A method of managing a memory system, comprising:

providing a processor and storage memory;

allocating a logical unit (LUN) of a plurality of LUNs to a client;

receiving a command to write data to the allocated LUN;

determining if the allocated LUN is configured for deduplication;

if the allocated LUN is not configured for deduplication, process the command by storing the data in the storage memory and creating metadata referencing a storage location of the data that is written to the storage memory;

if the allocated LUN is configured for deduplication, performing deduplication of the data and creating metadata referencing the location where the deduplicated data is stored;

and

storing the metadata for the deduplicated data for a plurality of LUNs in a single instance repository (SIR), and storing a single instance of the deduplicated data in the storage memory.

2. The method of claim 1, further comprising:

for each LUN of the plurality of LUNs configured for deduplication;

accepting a write command and data from an associated client in an input buffer memory;

maintaining a status buffer memory identifying a logical block address (LBA) of each received write command;

determining if the LBA of the write command is present in the status buffer memory; and

,

if the LBA of the write command is present in the status buffer memory, determining that the write command is an overlapped write command;

or,

if the LBA of the write command is not present in the status buffer memory, determining that the write command is not an overlapped write command.

3. The method of claim 2, further comprising:

maintaining a common deduplication buffer memory for a plurality of LUNs;

processing the non-deduplicated write commands in the common deduplication buffer memory for the plurality of LUNs for deduplication where the order of processing of commands is in accordance with a policy;

processing the overlapped write commands in sequential order; and

deleting the LBA of the write command from the status buffer memory when the deduplication process for the write command has been completed.

4. The method of claim 3, wherein the policy for processing non-deduplicated data does not include processing write commands in sequential order.

5. A method of managing a memory system, comprising:

providing a memory system with a processor having a client interface, buffer memory, and storage memory;

allocating a logical unit (LUN) of a plurality of LUNs to the client, the allocated LUN having a QoS specification;

receiving a command to perform an operation to the allocated LUN;

determining if the QoS specification for the allocated LUN includes deduplication and;

if the QoS specification does not include deduplication, process the command;

or

if the QoS specification includes deduplication, determine if deduplication is possible while meeting the QoS specification and;

if deduplication is possible, performing deduplication of the data and creating metadata referencing the storage location where the deduplicated data is stored;

or,

if deduplication is not possible, storing the data in a backing LUN associated with the allocated LUN, and

draining the data stored in the backing LUN when deduplication is possible by re-ingesting the data from the backing LUN and processing the commands in accordance with the QoS specification.

6. The method of claim 5, further comprising:

storing metadata for deduplicated data for a plurality of LUNs in a single instance repository (SIR), and storing a single instance of the deduplicated data in the storage memory.

7. The method of claim 6, further comprising:

for each LUN of the plurality of LUNs having a QoS that includes deduplication;

accepting a command from the client in an input buffer memory;

maintaining a status buffer memory identifying the LBA of each received command;

determining if the LBA of the command is present in the status buffer memory; and

,

if the LBA of the command is present in the status buffer memory, determining that the command is an overlapped command;

or,

if the LBA of the command is not present in the status buffer memory, determining that the command is not an overlapped command.

8. The method of claim 7, further comprising:

maintaining a first common buffer memory for non-overlapped commands of the plurality of LUNs;

maintaining a second common buffer memory for overlapped commands of the plurality of LUNs;

processing the non-overlapped commands in the first common buffer memory for the plurality of LUNs from the input buffer memories where the order of processing of commands is in accordance with a policy;

processing the overlapped commands in the second common buffer memory in sequential order; and

deleting the LBA of the command from the status buffer when the command has been completed.

9. The method of claim 8, wherein the policy for processing non-overlapped commands does not include processing commands in sequential order.

10. The method of claim 9, wherein the command is at least one of a WRITE, TRIM, UNMAP, WRITE_SAME, COMPARE_AND_WRITE or XCOPY command.

11. The method of claim 10, wherein overlapped commands for a first LUN are processed independently of overlapped commands for a second LUN.

12. The method of 7, wherein when a first command and a second command are determined to be overlapped:

comparing the first command and the second command to determine if the commands result in a same change in the data or metadata for the LBA;

deleting one of the first command or the second command; and

processing the remaining one of the first command and second command as a non-overlapped command.

13. A memory management system, comprising:

a storage memory; and

a processor configured to;

store metadata for deduplicated data for a plurality of logical units (LUNs) in a single instance repository (SIR);

store a single instance of the deduplicated data in the storage memory;

for each LUN of the plurality of LUNs;

accept commands in an input buffer memory;

maintain a status buffer memory identifying the logical block addresses (LBA) of each received command;

determine if the LBA of the command is present in the status buffer memory;

if the LBA of the command is present in the status buffer memory, determine that the command is an overlapped command; and

if the LBA of the command is not present in the status buffer memory, determine that the command is not an overlapped command.

14. The memory management system of claim 13, the processor further configured to:

maintain a first common buffer memory for non-overlapped commands of the plurality of LUNs;

maintain a second common buffer memory for overlapped commands of the plurality of LUNs;

process the non-overlapped commands in the first common buffer memory in non-sequential order;

process the overlapped commands in the second common buffer memory in sequential order; and

delete the LBA of the command from the status buffer when the command has been completed.

15. The memory management system of claim 13, the processor further configured to process overlapped commands for a first LUN independently of overlapped commands for a second LUN.

16. The memory management system of claim 13, the processor further configured to:

determine a first command and a second command as overlapped commands;

delete one of the first command or the second command when the commands result in a same change in the data or metadata for the LBA; and

process the remaining one of the first command and second command as a non-overlapped command.