Apparatus and method for storing data in a proxy cache in a network

• US 7,171,469 B2
• Filed: 09/16/2002
• Issued: 01/30/2007
• Est. Priority Date: 09/16/2002
• Status: Expired due to Fees

• Current Assignee: NetApp, Inc.
• Original Assignee: Network Appliance Incorporated (NetApp, Inc.)
• Inventors: Grunwald, David, Ayyar, Kartik, Prakash, Ashish, Ackaouy, Emmanuel, Quirion, Brian, Amdur, Matthew
• Primary Examiner(s): Wiley; David
• Assistant Examiner(s): BOUTAH, ALINA A
• Application Number:
  US10/245,828
• Publication Number:
  US 20040054777A1
• Time in Patent Office: 1,597 Days
• Field of Search: 709/225, 709213-216, 709/202, 709/231, 711/130, 711/122, 711/218, 711/219, 707/10
• US Class Current: 709/225
• CPC Class Codes: G06F 16/9574 of access to content, e.g. by caching
• CPC Class Codes: H04L 67/5682 Policies or rules for updating, deleting or replacing the stored data

Abstract

In one embodiment, the invention provides an apparatus for caching data in a network, with the apparatus including a proxy cache configured to receive request for an object from a client and to fetch data blocks from a server. The proxy cache may be configured to cache the data blocks in a hierarchical relationship within the object. The object may be, for example, a data file or a directory. The data blocks that are cached in the proxy cache define an active data set which is based upon a request from a client.

Claims

1. An apparatus for caching data in a network, the apparatus comprising:

a proxy cache configured to receive request for an object from a client and to fetch data blocks from a server, wherein the proxy cache includes a memory; and

the proxy cache configured to cache the data blocks in a hierarchical relationship within the object and as an active data set, wherein the hierarchical relationship of data blocks are stored in the memory in the proxy cache;

wherein in response to a cache miss where the proxy cache does not contain a requested data block, the proxy cache determines which server stores the requested data block by using a translation table of N values and FS_id values in the proxy cache, where an N value is a hash computation based on an address of a server that has the requested data block, a volume name the contains the data block, and a file name of the data block, and an FS_id value is a file system identifier of a server-side file handle from the server, and wherein the proxy cache sends a file handle associated with the requested data block to the server only if the address of the server is identified by the N value, and the proxy cache will fetch the data block from the server and cache the fetched data block in the active data set so that the active data will include a content in a first time period and will include a different content in a second time period and wherein the data blocks in the active data set is updated or a data block in the active data set is flushed, in response to a cache miss.

2. The apparatus of claim 1, wherein the data blocks are in a buffer tree in the proxy cache to permit the hierarchical relationship within a file.

3. The apparatus of claim 1, wherein a data block is placed in the active data set in response to a request from a client.

4. The apparatus of claim 1, wherein the active data set comprises a portion of data stored in at least the server.

5. The apparatus of claim 1, wherein the active data set is defined by a particular time period.

6. The apparatus of claim 1, wherein the proxy cache is configured to fetch a data block from the server in response to a request from the client, if a file requested by the server is missing the data block.

7. The apparatus of claim 1, wherein the proxy cache is configured to fetch a data block from the server in response to a request to the server from the client, if a file requested by the server is absent from the proxy cache.

8. The apparatus of claim 1, wherein the proxy cache and client communicate by use of an open standard protocol.

9. The apparatus of claim 8, wherein the protocol comprises an Internet Protocol.

10. The apparatus of claim 8, wherein the protocol comprises Network File System (NFS) protocol.

11. The apparatus of claim 8, wherein the protocol comprises Common Internet File System (CIFS) protocol.

12. The apparatus of claim 1, wherein the proxy cache computes a unique name as part of a computation to determine the presence or absence of a data block in a data file requested by a client, wherein the unique name is computed based on a hash of an address of the server.

13. The apparatus of claim 1, wherein a portion of the active data set can be flushed from the proxy cache based upon a replacement policy.

14. The apparatus of claim 1, wherein the object comprises a data file.

15. The apparatus of claim 1, wherein an object comprises a part of a file.

48. The apparatus of claim 1, wherein the proxy cache computes a unique file name in response to a request from a client, and based upon the unique file name, obtains a matching cache file identifier that determines if each data block of the requested object is stored in the proxy cache.

49. The apparatus of claim 48, wherein the matching local cache file identifier is indexed into an mode file that points to an appropriate inode record.

50. The apparatus of claim 49, wherein the inode record includes inode numbers that indicate if the data blocks of the requested object are stored or not stored in the proxy cache.

51. The apparatus of claim 50, wherein an inode number will include a special value X if an associated data block in the requested object is not stored in the proxy cache, and wherein the special value X is one of the values comprising −

1, −

3, or other values.

52. The apparatus of claim 50, wherein an inode number will not include a special value X if an associated data block in the requested object is stored in the proxy cache.

53. The apparatus of claim 48, wherein the inode file also points to a second inode record that contains server-side attributes for objects.

54. The apparatus of claim 48, wherein a data block of the requested object is not stored in the proxy cache if a matching cache file identifier based upon the unique file name is not in the proxy cache.

62. The apparatus of claim 1, wherein the data blocks in the active data set include contents that are determined based upon a given time period.

63. The apparatus of claim 1, wherein the data blocks in the active data set can differ depending on the particular time period.

64. The apparatus of claim 1, wherein the hierarchical relationship comprises a hierarchical structure of directories and files stored in the memory.

72. The method of claim 48, wherein the unique file name is computed based on a hash of an address of the server.

16. A method for caching data in a network, the method comprising:

receiving request for an object from a client;

if each data block in the requested object is in the cache, then sending the requested object to the client; and

if a data block in the requested object is missing, then fetching the data block from a server, filling the data block in a hierarchical relationship and as part of an active data set, and sending the requested object to the client, wherein the hierarchical relationship including the data block are stored in a memory in a proxy cache;

wherein in response to a cache miss where the proxy cache does not contain a requested data block, determining, by the proxy cache, which server stores the requested data block by using a translation table of N values and FS_id values in the proxy cache, where an N value is a hash computation based on an address of a server that has the requested data block, a volume name the contains the data block, and a file name of the data block, and an FS_id value is a file system identifier of a server-side file handle from the server, and sending, by the proxy cache, a file handle associated with the requested data block to the server only if the address of the server is identified by the N value, and fetching a data block from the server and caching the fetched data block in the active data set so that the active data will include a content in a first time period and will include a different content in a second time period and wherein the data blocks in the active data set is updated or a data block in the active data set is flushed, in response to a cache miss.

17. The method of claim 16, further comprising:

fetching a data block from the server in response to a request to the server from the client, if a file requested by the server is absent.

18. The method of claim 16, wherein the data blocks are in a buffer tree to permit the hierarchical relationship within a file.

19. The method of claim 16, wherein the active data set comprises a portion of data stored in at least the server.

20. The method of claim 16, wherein the active data set is defined by a particular time period.

21. The method of claim 16, wherein the request for the object is transmitted by use of an open standard protocol.

22. The method of claim 21, wherein the protocol comprises an Internet Protocol.

23. The method of claim 21, wherein the protocol comprises Network File System (NFS) protocol.

24. The method of claim 21, wherein the protocol comprises Common Internet File System (CIFS) protocol.

25. The method of claim 16, further comprising computing a unique name as part of a computation to determine the presence or absence of a data block in a data file requested by a client, wherein the unique name is computed based on a hash of an address of the server.

26. The method of claim 16, further comprising flushing a portion of the active data set based upon a replacement policy.

55. The method of claim 16, further comprising:

computing a unique file name in response to the request from the client; and

based upon the unique file name, obtaining a matching cache file identifier that determines if each data block of the requested object is stored in a proxy cache.

56. The method of claim 55, wherein the matching local cache file identifier is indexed into an inode file that points to an appropriate inode record.

57. The method of claim 56, wherein the inode record includes inode numbers that indicate if the data blocks of the requested object are stored or not stored in the proxy cache.

58. The method of claim 57, wherein an inode number will include a special value X if an associated data block in the requested object is not stored in the proxy cache, and wherein the special value X is one of the values comprising −

1, −

3, or other values.

59. The method of claim 57, wherein an inode number will not include a special value X if an associated data block in the requested object is stored in the proxy cache.

60. The method of claim 55, wherein the inode file also points to a second inode record that contains server-side attributes for objects.

61. The method of claim 55, wherein a data block of the requested object is not stored in the proxy cache if a matching cache file identifier based upon the unique file name is not in the proxy cache.

65. The method of claim 16 wherein the data blocks in the active data set include contents that are determined based upon a given time period.

66. The method of claim 16, wherein the data blocks in the active data set can differ depending on the particular time period.

67. The method of claim 16, wherein the hierarchical relationship comprises a hierarchical structure of directories and files stored in the memory.

68. The method of claim 16, wherein contents in the active data set were previously served by the proxy cache in response to a previous request by a client.

69. The method of claim 16, wherein contents in the active data set were previously retrieved from the server by the proxy cache in response to a previous request by a client.

71. The method of claim 68, wherein the hierarchical relationship of data blocks are stored in the memory in the cache.

73. The method of claim 55, wherein the unique file name is computed based on a hash of an address of the server.

27. An apparatus for caching data in a network, the apparatus comprising:

means for receiving a request for an object from a client;

coupled to the receiving means, means for sending the requested object to the client, if each data block in the requested object is in the cache; and

coupled to the receiving means, means for fetching the data block from a server, for filling the data block in a hierarchical relationship and as part of an active data set, and for sending the requested object to the client, if a data block in the requested object is missing, wherein the hierarchical relationship including the data block are stored in a memory in a proxy cache;

wherein in response to a cache miss where the proxy cache does not contain a requested data block, the proxy cache determines which server stores the requested data block by using a translation table of N values and FS_id values in the proxy cache, where an N value is a hash computation based on an address of a server that has the requested data block, a volume name the contains the data block, and a file name of the data block, and an FS_id value is a file system identifier of a server-side file handle from the server, and wherein the proxy cache sends a file handle associated with the requested data block to the server only if the address of the server is identified by the N value, and the proxy cache will fetch a data block from the server and cache the fetched data block in the active data set so that the active data will include a content in a first time period and will include a different content in a second time period and wherein the data blocks in the active data set is updated or a data block in the active data set is flushed, in response to a cache miss.

28. An apparatus for caching data in a network, the apparatus comprising:

a proxy cache configured to receive request for an object from a client and to fetch data blocks from a server, wherein the proxy cache includes a memory; and

the proxy cache configured to cache the data blocks in a hierarchical relationship within the object and as an active data set which is defined based upon a request from a client, wherein the hierarchical relationship of data blocks are stored in the memory in the proxy cache;

wherein in response to a cache miss where the proxy cache does not contain a requested data block, the proxy cache determines which server stores the requested data block by using a translation table of N values and FS_id values in the proxy cache, where an N value is a hash computation based on an address of a server that has the requested data block, a volume name the contains the data block, and a file name of the data block, and an FS_id value is a file system identifier of a server-side file handle from the server, and wherein the proxy cache sends a file handle associated with the requested data block to the server only if the address of the server is identified by the N value, and the proxy cache will fetch a data block from the server and cache the fetched data block in the active data set so that the active data will include a content in a first time period and will include a different content in a second time period and wherein the data blocks in the active data set is updated or a data block in the active data set is flushed, in response to a cache miss.

29. The apparatus of claim 28, wherein the active data set is modified by use of a data replacement strategy.

30. The apparatus of claim 28, wherein the data replacement strategy comprises one of:

a first-in-first-out method, a least recently-used method;

a data importance method, or a data replacement method based upon a user input.

31. The apparatus of claim 28, wherein the active data set is available to at least another client that is coupled to the proxy cache in order to permit collaborative sharing of the active data set for all clients coupled to the proxy cache.

32. The apparatus of claim 28, wherein the proxy cache is configured to translate a file handle to indicate that the file handle maps to a particular server.

33. The apparatus of claim 28, wherein the file handle is translated by use of a hash algorithm based upon a file system identifier and a server Internet Protocol address.

34. A method for caching data in a network, the method comprising:

receiving request for an object from a client;

determining if at least one data block in the requested object is stored in a cache; and

if each data block in the requested object is in the cache, then sending the requested object to the client; and

if a data block in the requested object is missing, then fetching the data block from a server, allocating the data block in the cache, wherein the data block is allocated in an active data set that is dependent upon a given time period, and sending the requested object to the client;

wherein in response to a cache miss where the proxy cache does not contain a requested data block, determining, by the proxy cache, which server stores the requested data block by using a translation table of N values and FS_id values in the proxy cache, where an N value is a hash computation based on an address of a server that has the requested data block, a volume name the contains the data block, and a file name of the data block, and an FS_id value is a file system identifier of a server-side file handle from the server, and sending, by the proxy cache, a file handle associated with the requested data block to the server only if the address of the server is identified by the N value, and fetching a data block from the server and caching the fetched data block in the active data set so that the active data will include a content in a first time period and will include a different content in a second time period and wherein the data blocks in the active data set is updated or a data block in the active data set is flushed, in response to a cache miss.

35. The method of claim 34, wherein the each data block is in a hierarchical relationship within an object.

36. The method of claim 34, wherein the active data set comprises a portion of data stored in at least the server.

37. The method of claim 34, wherein the active data set is defined by a particular time period.

38. The method of claim 34, wherein the request for the object is transmitted by use of an open standard protocol.

39. The method of claim 38, wherein the protocol comprises an Internet Protocol.

40. The method of claim 39, wherein the protocol comprises Network File System (NFS) protocol.

41. The method of claim 39, wherein the protocol comprises Common Internet File System (CIFS) protocol.

42. The method of claim 34, further comprising computing a unique name as part of a computation to determine the presence or absence of a data block in a data file requested by a client, wherein the unique name is computed based on a hash of an address of the server.

43. The method of claim 34, further comprising flushing a portion of the active data set based upon a replacement policy.

44. The method of claim 43, wherein the data replacement strategy comprises one of:

a first-in-first-out method, a least recently-used method;

a data importance method, or a data replacement method based upon a user input.

45. The method of claim 34, wherein the active data set is available to at least another client in order to permit collaborative sharing of the active data set for all clients that can access the active data set.

46. The method of claim 34, wherein the proxy cache is configured to translate a file handle to indicate that the file handle maps to a particular server.

47. The method of claim 46, wherein the file handle is translated by use of a hash algorithm based upon a file system identifier and a server Internet Protocol address.

70. The method of claim 34, wherein the data blocks in the active data set are in hierarchical relationship and are stored in the memory in the cache.