Granted Patent B2

US 10,162,553 · App. 13/303,870 · Granted Dec 25, 2018

Abstract

Claims (45)

1. A method for providing object level de-duplication for solid state storage devices (SSDs), comprising:

receiving a block of data at a solid state storage device (SSD), wherein the block of data comprises a header portion and a payload, and wherein the header portion comprises context information;

determining, by the SSD, that the payload should be de-duplicated prior to storage based solely on the context information stored within the header portion without analyzing the payload;

compressing the payload such that the compressed payload includes random data;

building a hash, the hash consisting of a subset of the random data less than an entirety of the compressed payload;

de-duplicating the compressed payload using the hash; and

storing the de-duplicated payload to the solid state storage device.

2. The method of claim 1 , wherein the context information is provided from hierarchical layers above the SSD.

3. The method of claim 2 , wherein the context information is indicative of whether the payload comprises random data.

4. The method claim 1 , wherein payload data comprises a plurality of sub-objects.

5. The method of claim 4 comprising de-duplicating based on a correlation between sub-objects of other payloads.

6. The method of claim 1 , comprising before storing, separating out the header from payload and storing the header and the payload in separate locations on the SSD.

7. The method of claim 1 , further comprising determining that the payload is not to be de-duplicated when the cost of improperly determining to de-duplicate the payload is high.

8. The method of claim 1 , further comprising deriving the context information from the header portion by analyzing the header portion in view of a known structure of the header portion.

9. The method of claim 3 , further comprising de-duplicating the payload only when the payload is non-random.

10. The method of claim 1 , wherein the context information comprises a sensitivity level associated with the payload, indicative of a cost of improperly determining to de-duplicate the payload.

11. The method of claim 1 , wherein the context information is uniquely associated with the payload.

12. A solid state device (SSD) providing object level de-duplication, comprising:

a host interface for communicating to a host device;

a flash memory for storing data; and

a de-duplication engine comprising logic stored in a computer readable medium, the logic configured to:

receive a block of data, wherein the block of data comprises a header portion and a payload, and wherein the header portion comprises context information;

determine the payload should be de-duplicated prior to storage based solely on the context information stored within the header portion without analyzing the payload;

compress the payload such that the compressed payload includes random data;

build a hash, the hash consisting of a subset of the random data less than an entirety of the compressed payload;

de-duplicate the compressed payload using the hash; and

store the de-duplicated payload to the flash memory.

13. The solid state device of claim 12 , wherein the context information is provided from hierarchical layers above the SSD.

14. The solid state device of claim 12 , wherein payload data comprises a plurality of sub-objects.

15. The solid state device of claim 12 , wherein the logic is further configured to determine that the payload is not to be de-duplicated when the cost of improperly determining to de-duplicate the payload is high.

16. A method for providing object level de-duplication for solid state storage devices (SSDs), comprising:

receiving, at a solid state storage device (SSD), a first block of data to be written to the SSD, wherein the first block of data comprises first header information and a first object;

determining, at the SSD based solely on the first header information without analyzing the first object, the first object should be de-duplicated prior to storage;

compressing the first object such that the first object includes random data;

building a hash, the hash consisting of a subset of the random data less than an entirety of the compressed first object;

de-duplicating, at the SSD, the compressed first object using the hash; and

storing the de-duplicated first object to the SSD.

17. The method of claim 16 , wherein the first header information is provided by an application layer that had previously accessed the first block of data.

18. The method of claim 16 , wherein determining, based on the first header information, the first object should be de-duplicated prior to storage, comprises:

deriving, at the SSD, context information based on the first header information by analyzing the header portion in view of a known structure of the header portion; and

determining, at the SSD, based on the context information, whether the first object should be de-duplicated prior to storage.

19. The method of claim 16 , wherein the first block of data further comprises second header information and a second object, wherein the method further comprises determining, at the SSD, based on the first header information and the second header information, whether the second object should be de-duplicated prior to storage.

20. The method of claim 16 , further comprising:

receiving, at the SSD, a second block of data to be written to the SSD, wherein the second block of data comprises second header information and a second object;

determining, at the SSD, based on the first header information and the second header information, whether the second object should be de-duplicated prior to storage.

Assignments (12)