Loosely coupled mass storage computer cluster having a set of data storage controllers interoperable for reading and writing data objects

• US 8,019,941 B2
• Filed: 02/11/2008
• Issued: 09/13/2011
• Est. Priority Date: 10/24/1995
• Status: Expired due to Fees

• Current Assignee: SeaChange International Incorporated
• Original Assignee: SeaChange International Incorporated
• Inventors: Ywoskus, John A., McGrath, Edward J., Carlozzi, Michael D., Trasatti, Philip J., Mann, Bruce E.
• Primary Examiner(s): Thai; Tuan V.
• Application Number:
  US12/029,278
• Publication Number:
  US 20090055401A1
• Time in Patent Office: 1,310 Days
• Field of Search: 711/114, 711/100, 711/112, 711/154
• US Class Current: 711/114
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 11/2028 eliminating a faulty processor or activating a spare
• CPC Class Codes: G06F 11/2038 with a single idle spare processing component
• CPC Class Codes: G06F 2211/1028 Distributed, i.e. distributed RAID systems with parity
• CPC Class Codes: G06F 2211/1045 Nested RAID, i.e. implementing a RAID scheme in another RAID scheme
• CPC Class Codes: H04N 7/17336 Handling of requests in head-ends

Abstract

A method and apparatus redundantly store data, in particular video data objects, in a distributed computer system having at least three processor systems, each processor system being connected in point to point two way channel interconnection with each other processor system. The data is stored in a redundant fashion both at the computer system level as well as the processor system level. Accordingly, the failure of a single processor does not adversely affect the integrity of the data. The computer system can also overlay a switching system connected in a ring fashion for providing a fault tolerance to the failure of a single connected processor system at the switch level. Accordingly, there results a fault tolerant data distribution system.

Claims

1. A data cluster, comprising:

three or more cluster members that store a plurality of data objects, each cluster member comprising data storage and a processor that executes a data storage controller, wherein each of the plurality of data objects comprise data blocks that are stored across the cluster members according to a distributed, redundant storage process;

the data storage controller of at least one of the cluster members receiving individual data service requests for a data object and interoperating with other data storage controllers within a set of the three or more cluster members to access the corresponding data storage to read or write data blocks comprising the data object according to the distributed, redundant storage process.

2. The data cluster of claim 1 wherein the data storage controller responds to the individual data service requests from a client.

3. The data cluster of claim 2 wherein the data storage controller responds to the individual data service requests from an external client over an external link coupled to any one of the three or more cluster members.

4. The data cluster of claim 2 wherein each cluster member further comprises a data service application module that communicates the individual data service requests from the client to the data storage controller.

5. The data cluster of claim 1 wherein:

the data storage controller receives a request to store the data object in the data cluster and interoperates with the other data storage controllers to write the data blocks comprising the data object across the data storage within the set of the three or more cluster members according to the distributed, redundant storage process.

6. The data cluster of claim 5 wherein the data storage within the set of the three or more cluster members comprise a shared cluster volume, one of the data storage controllers within the set of three or more cluster members is designated to control the writing of the data blocks comprising the data object across all of the data storage of the cluster volume.

7. The data cluster of claim 5 wherein the data storage for each cluster member is private, the data storage controller and each of the other data storage controllers individually control the writing of the data blocks to the corresponding data storage.

8. The data cluster of claim 5 wherein each data storage controller comprises:

an interface module that allocates the data blocks comprising the data object for storage across the data storage within the set of the three or more cluster members according to the distributed, redundant storage process.

9. The data cluster of claim 5 wherein each data storage controller comprises:

a local file provider and a remote file provider that write the data blocks comprising the data object across the data storage within the set of the three or more cluster members according to the distributed, redundant storage process.

10. The data cluster of claim 9 wherein the remote file provider communicates with the other data storage controllers within the set of the three or more cluster members to write data blocks in the corresponding data storage according to the distributed, redundant storage process.

11. The data cluster of claim 9 wherein the local file provider writes data blocks to the data storage of the local cluster member.

12. The data cluster of claim 11 wherein the local file provider writes data blocks to the data storage of the local cluster member through a private file system.

13. The data cluster of claim 1 wherein:

the data storage controller receives a request to retrieve the data object stored in the data cluster by interoperating with the other data storage controllers to read the data blocks comprising the data object stored across the data storage within the set of the three or more cluster members according to the distributed, redundant storage process.

14. The data cluster of claim 13 wherein each data storage controller comprises:

an interface module that assembles the data blocks read from across the data storage within the set of the three or more cluster members according to the distributed, redundant storage process into the data object and delivers the requested data to a requesting device.

15. The data cluster of claim 13 wherein each data storage controller comprises:

a local file provider and a remote file provider that read the data blocks comprising the data object from the data storage within the set of the three or more cluster members according to the distributed, redundant storage process.

16. The data cluster of claim 15 wherein the remote file provider communicates with the other data storage controllers within the set of the three or more cluster members to read data blocks in the corresponding data storage according to the distributed, redundant storage process.

17. The data cluster of claim 15 wherein the local file provider reads data blocks from the data storage of the local cluster member.

18. The data cluster of claim 17 wherein the local file provider obtains data blocks from the data storage of the local cluster member through a private file system.

19. The data cluster of claim 13 wherein:

in the presence of a failure of one of the three or more cluster members, the data storage controller responding to the request reads data blocks from active cluster members in the set, reads redundant data blocks from the active cluster members in place of missing data blocks stored at the failed cluster member, and reconstructs the missing data blocks using the data blocks and redundant data blocks.

20. The data cluster of claim 1 wherein the data object is video data.

21. The data cluster of claim 1 wherein the data storage controller receives individual data service requests for the data object and interoperates with the other data storage controllers within the set of the three or more cluster members in parallel to access the corresponding data storage to read and write data blocks comprising the data object according to a distributed, redundant storage process.

22. The data cluster of claim 1 wherein the distributed, redundant storage process is a Redundant Array of Independent Disks (RAID) storage process.

23. The data cluster of claim 22 wherein the RAID storage process is a RAID-5 storage process.

24. The data cluster of claim 1 wherein the data blocks are stored with metadata for accessing the data object.

25. The data cluster of claim 24 wherein the metadata comprises information for determining locations of individual data blocks.

26. The data cluster of claim 24 wherein the metadata comprises one or more fragment headers, each of the one or more fragment headers being associated with at least a portion of the data object.

27. The data cluster of claim 1 wherein locations of constituent blocks of the data object are determined based on an individual block number to be located, fragment size, and total number of the constituent data blocks.

28. The data cluster of claim 27 wherein the fragment size equals the cluster size.

29. The data cluster of claim 1, wherein the data storage controller interoperates with the other data storage controllers to access a corresponding set of files from the data storage of the set of cluster members through which to read and write the data blocks comprising the data object according to a distributed, redundant storage process.

30. A method of accessing data storage within a data cluster, the data cluster comprising three or more cluster members that store a plurality of data objects, each cluster member comprising data storage and a processor that executes a data storage controller, wherein each of the plurality of data objects comprise data blocks that are stored across the cluster members according to a distributed, redundant storage process, the method comprising at the data storage controller of at least one of the cluster members:

receiving individual data service requests for a data object; and

interoperating with other data storage controllers within a set of the three or more cluster members to access the corresponding data storage to read or write data blocks comprising the data object according to the distributed, redundant storage process.

31. The method of claim 30, further comprising, at the data storage controller:

interoperating with the other data storage controllers to access a corresponding set of files from the data storage of the set of cluster members through which to read and write the data blocks comprising the data object according to a distributed, redundant storage process.

32. A data cluster, comprising:

three or more cluster members that store a plurality of data objects, each cluster member comprising data storage and a processor that executes a data storage controller, wherein each of the plurality of data objects comprise data blocks that are stored across the cluster members according to a distributed, redundant storage process; and

the data storage controller of at least one of the cluster members receiving individual data service requests for a data object and interoperating with other data storage controllers within a set of the three or more cluster members to access the corresponding data storage to read or write data blocks comprising the data object according to the distributed, redundant storage process, the data blocks being stored with metadata for accessing the data object and the metadata comprising information for determining locations of individual data blocks.

33. A method of accessing data storage within a data cluster, the data cluster comprising three or more cluster members that store a plurality of data objects, each cluster member comprising data storage and a processor that executes a data storage controller, wherein each of the plurality of data objects comprise data blocks that are stored across the cluster members according to a distributed, redundant storage process, the method comprising at the data storage controller of at least one of the cluster members:

receiving individual data service requests for a data object; and

interoperating with other data storage controllers within a set of the three or more cluster members to access the corresponding data storage to read or write data blocks comprising the data object according to the distributed, redundant storage process, the data blocks being stored with metadata for accessing the data object and the metadata comprising information for determining locations of individual data blocks.

34. A data cluster, comprising:

three or more cluster members that store a plurality of data objects, each cluster member comprising data storage and a processor that executes a data storage controller, wherein each of the plurality of data objects comprise data blocks that are stored across the cluster members according to a distributed, redundant storage process; and

the data storage controller of at least one of the cluster members receiving individual data service requests for a data object and interoperating with other data storage controllers within a set of the three or more cluster members to access the corresponding data storage to read or write data blocks comprising the data object according to the distributed, redundant storage process;

the data storage controller of the at least one cluster member determining locations of the data blocks comprising the data object based on an individual block number to be located, fragment size, and total number of the data blocks comprising the data object.

35. A method of accessing data storage within a data cluster, the data cluster comprising three or more cluster members that store a plurality of data objects, each cluster member comprising data storage and a processor that executes a data storage controller, wherein each of the plurality of data objects comprise data blocks that are stored across the cluster members according to a distributed, redundant storage process, the method comprising at the data storage controller of at least one of the cluster members:

receiving individual data service requests for a data object; and

interoperating with other data storage controllers within a set of the three or more cluster members to access the corresponding data storage to read or write data blocks comprising the data object according to the distributed, redundant storage process;

determining locations of the data blocks based on an individual block number to be located, fragment size, and total number of the data blocks comprising the data object.