Method and system for fast incremental backup using comparison of descriptors

1. A method for combining file-level and block-level backups for a fast incremental backup of a storage device, the method comprising:

(a) creating a bitmap of blocks of the storage device with all blocks that require backing up being marked in the bitmap;

(b) extracting descriptors of the files and directories from a descriptors area of the storage device;

(c) discarding at least some of the descriptors from a set of descriptors associated with each file and then generating a common hash value of a resulting set of descriptors;

(d) recognizing identical hash values related to a prior backup of the storage device and to a current state of the storage device;

(e) checking coincidence of blocks that form a single file, wherein the descriptors of the files have identical hash values, and, for at least some of the blocks that form the single file, comparing contents of the blocks;

(f) unmarking, in the bitmap, the coincident blocks;

(g) backing up contents of the blocks that are marked in the bitmap; and

(h) storing pointers for coincident blocks, wherein the pointers point to blocks from the prior backup.

2. The method of claim 1 , wherein the prior backup is a backup of the storage device at a previous point in time.

3. The method of claim 1 , wherein the prior backup is a backup of a different storage device.

4. The method of claim 1 , wherein the prior backup includes multiple backups of the storage device.

5. The method of claim 1 , wherein the discarding includes resetting selected data access attributes of the descriptors.

6. The method of claim 1 , wherein the discarding includes resetting selected attributes of the descriptors.

7. The method of claim 1 , wherein the descriptors are MFT entries.

8. The method of claim 1 , wherein the descriptors are inodes entries.

9. The method of claim 1 , wherein the descriptors are directory entries and numbers of all clusters, according to which the directory entries are ordered.

10. The method of claim 1 , wherein step (c) further comprises generating hash values for blocks and comparing the hash values of the blocks of the storage device with the hash values of previously archived blocks, that have the same addresses.

11. The method of claim 1 , wherein step (e) further comprises bit-wise comparison of the content of the blocks.

12. The method of claim 1 , wherein step (b) further comprises reading the descriptors from a disk area of the data storage device.

13. The method of claim 1 , further comprising elimination of information unnecessary for the descriptors comparison from the descriptors area.

14. The method of claim 1 , wherein the descriptors include data access attributes of any of “Update sequence” in MFT file record header and time read in “$STANDARD_INFORMATION” from attributes of MFT file record.

15. The method of claim 1 , wherein MFT contents from a start of a file record to an end marker are converted into hash values.

16. The method of claim 1 , wherein the descriptors include data access attribute “last access date” in operating systems with FAT file system.

17. The method of claim 1 , wherein the descriptors include data access attribute “access time” in operating system with Ext2/Ext3 file system.

18. The method of claim 1 , wherein inodes contents, indirect blocks, double indirect blocks and triple indirect blocks with file data are converted into hash values.

19. A method for combining file-level and block-level backups for a fast incremental backup, the method comprising:

(a) creating a bitmap of blocks of a first storage device with all blocks that require backing up being marked in the bitmap;

(b) extracting descriptors of the files and directories corresponding to blocks of the first storage device and a second storage device;

(c) generating a common hash value of the descriptors;

(d) recognizing identical hash values related to a backup of the second storage device and to a current state of the first storage device;

(e) checking coincidence of blocks that form a single file, wherein the descriptors have identical hash values, and, for at least some of the blocks that form the single file, comparing contents of the blocks;

(f) unmarking, in the bitmap, the coincident blocks;

(g) backing up contents of the units blocks that are marked in the bitmap; and

(h) storing pointers for coincident blocks, wherein the pointers point to blocks from the backup of the second storage device.

20. The method of claim 19 , further comprising discarding at least some of the descriptors prior to the generating step.

21. A system for combining file-level and block-level backups for a fast incremental backup of a storage device, the system comprising:

(a) a bitmap of blocks of the storage device with all blocks that require backing up being marked in the bitmap;

(b) a plurality of descriptors of the files and directories stored in a descriptors area of the storage device;

(c) wherein at least some of the descriptors from a set of descriptors associated with each file are discarded and then a common hash value of a resulting set of descriptors is generated;

(d) a set of identical hash values related to a prior backup of the storage device and to a current state of the storage device;

(e) means for checking coincidence of blocks that form a single file, wherein the descriptors have identical hash values, and, for at least some of the blocks that form the single file, comparing contents of the blocks;

(f) wherein, in the bitmap, the coincident blocks are unmarked;

(g) a backup of contents of the blocks that are marked in the bitmap; and

(h) a plurality of pointers for coincident blocks, wherein the pointers point to blocks from the prior backup.

22. The system of claim 21 , wherein the discarding comprises resetting selected data access attributes of the descriptors.

23. The system of claim 21 , wherein the discarding includes resetting selected attributes of the descriptors.

24. The system of claim 21 , wherein the common hash value comprises hash values for blocks and comparing the hash values of the blocks of the storage device with the hash values of previously archived blocks, that have the same addresses.

25. The system of claim 21 , wherein the descriptors include data access attributes, including any of the “Update sequence” in MFT file record header and time read in “$STANDARD_INFORMATION” from attributes of MFT file record.

26. The system of claim 21 , wherein MFT contents from a start of a file record to an end marker are converted into hash values.

27. The system of claim 21 , wherein inodes contents, indirect blocks, double indirect blocks and triple indirect blocks with file data are converted into hash values.

28. A system for combining file-level and block-level backups for a fast incremental backup, comprising:

a main storage that includes blocks and files;

a bitmap of the blocks with all blocks that require backing up being marked in the bitmap;

a block-based archive of a previous state of the main storage; and

a plurality of descriptors of the files,

wherein for those files whose control sum values of the descriptors are not identical to control sum values of the descriptors of the previously archived files, an incremental block-based backup of corresponding blocks is performed, wherein for at least some of the blocks that form the single file, contents of the blocks are compared to determine if these blocks need to be backed up and unmarking, in the bitmap, coincident blocks that correspond to the archived logical storage units with coincident descriptors and backing up contents of the blocks that are marked in the bitmap.

29. The method of claim 1 , wherein, for those files that contain additional blocks, the bitmap only resets those bits that correspond to the additional blocks.

30. The method of claim 1 , wherein, for those files that contain additional metadata, but no new contents, the bitmap only resets those bits that correspond to the new metadata.

31. The method of claim 1 , further comprising, for those blocks whose hashes are not identical, comparing the blocks on a cluster by cluster basis, a cluster being a group of blocks.