Symbiotic wide-area file system and method

• US 20040111441A1
• Filed: 12/09/2002
• Published: 06/10/2004
• Est. Priority Date: 12/09/2002
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Saito, Yasushi, Karamanolis, Christos
• Granted Patent: US 8,315,975 B2
• Time in Patent Office: Days
• US Class Current: 707/200
• CPC Class Codes: G06F 16/1834 implemented based on peer-to-peer networks, e.g. gnutella p2p communication protocols H04L67/104

Abstract

In one embodiment, the invention provides a method and apparatus for a wide area file system, including: creating a peer-to-peer wide-area file system that allows read and write sharing of data. In another embodiment, the invention provides a method and apparatus for a wide area file system, including using per-file replication to achieve high availability and performance in the wide-area distributed file system.

Claims

1. A method for a wide area file system, the method comprising:

creating a peer-to-peer wide-area file system that allows read and write sharing of data.

2. The method of claim 1, wherein the action of creating the peer-to-peer wide-area file system comprises:

maintaining a graph of replicas of files.

3. The method of claim 1, further comprising:

propagating updates on a per-file basis.

4. The method of claim 2, further comprising:

exchanging updates by the replicas in a peer-to-peer fashion.

5. The method of claim 1, further comprising:

creating a replica of a file whenever the file is accessed.

6. The method of claim 5, wherein the file comprises a directory.

7. The method of claim 5, wherein the file comprises data.

8. The method of claim 2, wherein the replicas include core replicas.

9. The method of claim 8, wherein a core replica is related to maintenance of a hierarchical name space.

10. The method of claim 2, wherein the replicas include non-core replicas.

11. The method of claim 1, further comprising:

removing an inactive replica.

12. The method of claim 1, further comprising:

randomly selecting a plurality of replicas.

13. The method of claim 11, wherein a plurality of regions includes at least one of the selected replicas.

14. The method of claim 11, wherein a local region includes at least one of the selected replicas.

15. The method of claim 1, further comprising:

selecting, by a node, m existing replicas for a file, where m is an integer;

selecting an existing replica closest to the node;

selecting m-2 replicas by a random traversal from the from the existing replica closest to the node.

16. The method of claim 1, further comprising:

adding a non-core replica to the graph.

17. An apparatus for a wide area file system, the apparatus comprising:

at least one node configured to create a peer-to-peer wide-area file system that allows read and write sharing of data.

18. The apparatus of claim 17, wherein the at least one node is configured to maintain a graph of replicas of files.

19. The apparatus of claim 17, wherein the node is configured to propagate updates on a per-file basis.

20. The apparatus of claim 18, wherein the replicas is configured to exchanging updates in a peer-to-peer fashion.

21. The apparatus of claim 17, wherein the node is configured to create a replica of a file whenever the file is accessed.

22. The apparatus of claim 21, wherein the file comprises a directory.

23. The apparatus of claim 21, wherein the file comprises data.

24. The apparatus of claim 18, wherein the replicas include core replicas.

25. The apparatus of claim 24 wherein a core replica is related to maintenance of a hierarchical name space.

26. The apparatus of claim 18, wherein the replicas include non-core replicas.

27. The apparatus of claim 17, wherein the node is configured to remove an inactive replica.

28. The apparatus of claim 17, wherein the node is configured to randomly select a plurality of replicas.

29. The apparatus of claim 28, wherein a plurality of regions includes at least one of the selected replicas.

30. The apparatus of claim 28, wherein a local region includes at least one of the selected replicas.

31. The apparatus of claim 17, wherein a node is configured to select m existing replicas for a file, where m is an integer, to select an existing replica closest to the node, and to select m-2 replicas by a random traversal from the from the existing replica closest to the node.

32. The apparatus of claim 17, wherein the node is configured to add a non-core replica to the graph.

33. A method for a wide area file system, the method comprising:

using per-file replication to achieve high availability and performance in a wide-area distributed file system.

34. The method of claim 33, further comprising:

permitting access to any one of replicas of a file, where different regions include at least one replica.

35. The method of claim 33, wherein the action of using per-file replication comprises:

creating replicas of files in a sparse graph in response to an access to a file.

36. An apparatus for a wide area file system, the apparatus comprising:

at least one node configured to use per-file replication to achieve high availability and performance in a wide-area distributed file system.

37. The apparatus of claim 36, wherein at least one node permits access to any one of replicas of a file, where different regions include at least one replica.

38. The apparatus of claim 36, wherein at least one node is configured to create replicas of files in a sparse graph in response to an access to a file.

39. A server, comprising:

a protocol handler configured to receive requests from a client device to access a file;

a replication engine configured to create a replica of a file in response to an initial access to a file;

a log module configured to log replica-update operations to a replica; and

a membership module configured to provide the information to permit the replication engine to make decisions on integration, updates, requests functions.

40. A method for a distributed file system, the method comprising:

creating a replica of a file in response to an initial access to the file.

41. The method of claim 40, further comprising:

removing the replica when the replica becomes inactive.

42. The method of claim 40, further comprising:

removing the replica in response to a lack of disk space in a node.

43. An article of manufacture, comprising:

a machine-readable medium having stored thereon instructions to;

create a peer-to-peer wide-area file system that allows read and write sharing of data.

44. An article of manufacture, comprising:

a machine-readable medium having stored thereon instructions to;

use per-file replication to achieve high availability and performance in a wide-area distributed file system,

45. An article of manufacture, comprising:

a machine-readable medium having stored thereon instructions to;

create a replica of a file in response to an initial access to the file.

46. An apparatus for a wide area file system, the apparatus comprising:

means for creating a peer-to-peer wide-area file system that allows read and write sharing of data.