Characterization of data access using file system

• US 6,442,682 B1
• Filed: 02/18/1999
• Issued: 08/27/2002
• Est. Priority Date: 02/18/1999
• Status: Expired due to Term

• Current Assignee: Network Appliance Incorporated (NetApp, Inc.)
• Original Assignee: Auspex Systems, Inc.
• Inventors: Bhanjois, Bhimsen, Bodas, Amod, Coviello, John S., Narayanaswamy, Lakshman, Pothapragada, Srinivas
• Primary Examiner(s): Breene, John
• Assistant Examiner(s): Le, Debbie M
• Application Number:
  US09/251,753
• Time in Patent Office: 1,286 Days
• Field of Search: 707/1, 707/200, 713/1, 713/2, 713/100, 710/10, 710/104, 709/220, 709/200, 709/221, 709/222
• US Class Current: 713/1
• CPC Class Codes: G06F 16/10 File systems; File servers

Abstract

An apparatus and a method manage data stored on one or more data storage devices using an adaptive file system by characterizing the data on the data storage devices managed by the file system; and tuning the file system by selecting one or more options to configure a kernel during boot-up and an operating system during mount time.

Claims

1. An adaptive file system for managing data stored on one or more data storage devices, comprising:

means for characterizing the data on the data storage devices managed by the file system; and

means for tuning the file system, according to the data characterization, by selecting one or more options to configure a kernel during boot-up and an operating system during mount time.

2. The adaptive file system of claim 1, wherein one of the options optimizes the file system for sequential read/write operations, further comprising:

means for disabling caching of the data; and

means for performing read/write operations directly to the data storage device.

3. The adaptive file system of claim 2, further comprising means for pre-allocating blocks of data.

4. The adaptive file system of claim 1, wherein one of the options optimizes the file system for large file random read operations, further comprising:

means for determining an average block size of the large file; and

means for reading-ahead blocks of data based on the determined average block size.

5. The adaptive file system of claim 1, wherein one of the options optimizes the file system for large file random write operations, further comprising means for writing data directly to the data storage device.

6. The adaptive file system of claim 5, further comprising:

a buffer adapted to store a page to be overwritten by the large file random write operation; and

means for disabling the buffer.

10. The adaptive file system of claim 1, wherein one of the options optimizes the file system for sequential read operations, further comprising:

means for determining all files in a directory; and

means for prefetching each file in the directory into a buffer.

7. An adaptive file system for managing data stored on one or more data storage devices, comprising:

means for characterizing the data on the data storage devices managed by the file system;

means for tuning the file system, according to the data characterization, by selecting one or more options to configure a kernel during boot-up and an operating system during mount time, wherein one of the options optimizes the file system for small file random read/write operations; and

means for performing a delayed read/write operation.

8. An adaptive file system for managing data stored on one or more data storage devices, comprising:

means for characterizing the data on the data storage devices managed by the file system;

means for tuning the file system, according to the data characterization, by selecting one or more options to configure a kernel during boot-up and an operating system during mount time, wherein one of the options optimizes the file system for accessing metadata;

means for generating a search parameter; and

means for performing a search in accordance with the search parameter.

9. The adaptive file system of claim 8, wherein the search parameter compares either left-most or right-most characters of a file name.

11. A method for managing data stored on one or more data storage devices using an adaptive file system, comprising:

characterizing the data on the data storage devices managed by the file system based on access or storage attributes of the data; and

using the data characterization, tuning the file system by selecting one or more options to configure a kernel during boot-up and an operating system during mount time.

12. The method of claim 11, wherein one of the options optimizes the file system for sequential read/write operations, further comprising:

disabling caching of the data; and

performing read/write operations directly to the data storage device.

13. The method of claim 12, further comprising pre-allocating blocks of data.

14. The method of claim 11, wherein one of the options optimizes the file system for large file random read operations, further comprising:

determining an average block size of the large file; and

reading-ahead blocks of data based on the determined average block size.

15. The method of claim 11, wherein one of the options optimizes the file system for large file random write operations, further comprising writing data directly to the data storage device.

16. The method of claim 15, further comprising buffering a page to be overwritten by the large file random write operation.

20. The method of claim 11, wherein one of the options optimizes the file system for sequential read operations;

determining all files in a directory; and

prefetching each file in the directory into a buffer.

17. A method for managing data stored on one or more data storage devices using an adaptive file system, comprising:

characterizing the data on the data storage devices managed by the file system based on access or storage attributes of the data;

using the data characterization, tuning the file system by selecting one or more options to configure a kernel during boot-up and an operating system during mount time, wherein one of the options optimizes the file system for small file random read/write operations; and

performing a delayed read/write operation.

18. A method for managing data stored on one or more data storage devices using an adaptive file system, comprising:

characterizing the data on the data storage devices managed by the file system based on access or storage attributes of the data;

using the data characterization, tuning the file system by selecting one or more options to configure a kernel during boot-up and an operating system during mount time, wherein one of the options optimizes the file system for accessing metadata;

generating a search parameter; and

performing a search in accordance with the search parameter.

19. The method of claim 18, wherein the search parameter compares either left-most or right-most characters of a file name.

21. A computer system, comprising:

an interconnect bus;

one or more processors coupled to the interconnect bus and adapted to be configured for server specific functionalities including network processing, file processing, storage processing and application processing;

a configuration processor coupled to the interconnect bus and to the processors, the configuration processor dynamically assigning processor functionalities upon request;

one or more data storage devices coupled to the processors and managed by a file system;

means for characterizing the data on the data storage devices managed by the file system based on access or storage attributes of the data; and

means for tuning the file system according to the data characterization by selecting one or more options in an operating system kernel and a mount table.