Detecting and correcting errors in data

1. A method for use in detecting corruption in data written to a storage device, the method comprising:

receiving one or more blocks of data to be written to the storage device;

selecting from at least one of the blocks only a portion of the data in the block;

applying a rule to derive information characterizing the selected portion of data;

storing this information; and

after the data has been written to the storage device, using the stored information in deciding whether the data has been corrupted.

2. The method of claim 1 , where using the stored information in deciding whether the data has been corrupted includes:

reading the selected portion of data from the storage device;

again applying the rule to derive information characterizing the selected portion of data as read from the storage device;

assessing whether the stored information matches the information derived from the data as read from the storage device; and

if the two pieces of information do not match, concluding that the data has been corrupted.

3. The method of claim 1 , where selecting a portion of the data includes selecting a portion of the data in each block to be written during a write operation.

4. The method of claim 3 , where applying a rule to derive information includes calculating a checksum of the data selected from more than one of the blocks.

5. The method of claim 1 , where applying a rule to derive information includes calculating a checksum for the selected portion of data.

6. The method of claim 1 , where applying a rule to derive information includes reading the contents of the selected portion of data.

7. The method of claim 1 , where the selected portion of data includes a single byte of data.

8. The method of claim 1 , where the selected portion of data includes multiple bytes of data.

9. The method of claim 1 , where the selected portion of data includes a single word of data.

10. The method of claim 1 , where the selected portion of data includes multiple words of data.

11. A computer system comprising:

a storage device; and

a data-integrity component configured to:

receive one or more blocks of data to be written to the storage device;

select from at least one of the blocks only a portion of the data in the block;

apply a rule to derive information characterizing the selected portion of data;

effect storage of the information; and

after the data has been written to the storage device, use the stored information in deciding whether the data has been corrupted.

12. The system of claim 11 , where the data-integrity component, in using the stored information in deciding whether the data has been corrupted, is configured to:

read the selected portion of data from the storage device;

again apply the rule to derive information characterizing the selected portion of data as read from the storage device;

assess whether the stored information matches the information derived from the data as read from the storage device; and

if the two pieces of information do not match, conclude that the data has been corrupted.

13. The system of claim 11 , where the data-integrity component, in selecting a portion of the data, is configured to select a portion of the data in each block to be written during a write operation.

14. The system of claim 13 , where the data-integrity component, in applying a rule to derive information, is configured to calculate a checksum of the data selected from more than one of the blocks.

15. The system of claim 11 , where the data-integrity component, in applying a rule to derive information, is configured to calculate a checksum for the selected portion of data.

16. The system of claim 11 , where the data-integrity component, in applying a nile to derive information, is configured to read the contents of the selected portion of data.

17. An article comprising at least one computer-readable storage medium containing program code that, when executed by a computer, causes the computer to:

receive one or more blocks of data to be written to the storage device;

select from at least one of the blocks only a portion of the data in the block;

apply a rule to derive information characterizing the selected portion of data;

effect storage of the information; and

after the data has been written to the storage device, use the stored information in deciding whether the data has been corrupted.

18. The article of claim 17 , where, in using the stored information in deciding whether the data has been corrupted, the computer is configured to:

read the selected portion of data from the storage device;

again apply the rule to derive information characterizing the selected portion of data as read from the storage device;

assess whether the stored information matches the information derived from the data as read from the storage device; and

if the two pieces of information do not match, conclude that the data has been corrupted.

19. The article of claim 17 , where, in selecting a portion of the data, the computer is configured to select a portion of the data in each block to be written during a write operation.

20. The article of claim 19 , where, in applying a rule to derive information, the computer is configured to calculate a checksum of the data selected from more than one of the blocks.

21. The article of claim 17 , where, in applying a rule to derive information, the computer is configured to calculate a checksum for the selected portion of data.

22. The article of claim 17 , where, in applying a rule to derive information, the computer is configured to read the contents of the selected portion of data.

23. A method for use in correcting corruption found in data written to a storage device, the method comprising:

receiving a request to read data from the storage device;

reading the requested data from the storage device;

applying a rule to assess whether the data has been corrupted; and

if corruption has occurred:

reading a duplicate copy of the requested data from another storage device; and

delivering the duplicate copy of the requested data to the requesting device.

24. The method of claim 23 , also comprising writing the duplicate copy of data over the corrupted data on the storage device.

25. A computer system comprising:

a first storage device on which data is stored;

a second storage device that stores a duplicate copy of at least some of the data stored on the first storage device; and

a data-integrity component configured to:

receive a request to read certain data from the first storage device;

read the requested data from the first storage device;

apply a rule to assess whether the data has been corrupted; and

if corruption has occurred:

read a duplicate copy of the requested data from the second storage device; and

deliver the duplicate copy of the requested data to the requesting device.

26. The system of claim 25 , where the data-integrity component is also configured to write the duplicate copy of data over the corrupted data on the first storage device.

27. An article comprising at least one computer-readable storage medium containing program code that, when executed by a computer, causes the computer to:

receive a request to read data from the storage device;

read the requested data from the storage device;

apply a rule to assess whether the data has been corrupted; and

if corruption has occurred:

read a duplicate copy of the requested data from another storage device; and

deliver the duplicate copy of the requested data to the requesting device.

28. The article of claim 27 , where the program code also causes the computer to write the duplicate copy of data over the corrupted data on the storage device.

29. A method for use in detecting and correcting corruption in data written to a storage device, the method comprising:

receiving one or more blocks of data to be written to the storage device;

selecting a portion of the data in the blocks;

applying a rule to derive information characterizing the selected portion of data;

storing this information; and

after the data has been written to the storage device:

receiving a request to read the data from the storage device;

reading the requested data;

using the stored information in deciding whether the data has been corrupted; and

if corruption has occurred:

reading a duplicate copy of the requested data from another storage device; and

delivering the duplicate copy of the requested data to the requesting device.

30. The method of claim 29 , where using the stored information in deciding whether the data has been corrupted includes:

reading the selected portion of data from the storage device;

again applying the nile to derive information characterizing the selected portion of data as read from the storage device;

assessing whether the stored information matches the information derived from the data as read from the storage device; and

if the two pieces of information do not match, concluding that the data has been corrupted.

31. The method of claim 29 , also comprising writing the duplicate copy of data over the corrupted data on the storage device.