Method for data maintenance

• US 9,305,012 B2
• Filed: 12/05/2013
• Issued: 04/05/2016
• Est. Priority Date: 09/02/2011
• Status: Active Grant

• Current Assignee: Compuverde AB (Pure Storage, Inc.)
• Original Assignee: Compuverde AB (Pure Storage, Inc.)
• Inventors: Bernbo, Stefan, Melander, Christian, Persson, Rodger, Petersson, Gustav
• Primary Examiner(s): Truong, Lechi
• Application Number:
  US14/098,021
• Publication Number:
  US 20140095559A1
• Time in Patent Office: 852 Days
• Field of Search: 719/310, 719/318, 709/223
• US Class Current: 1/1
• CPC Class Codes: G06F 16/182 Distributed file systems
• CPC Class Codes: G06F 16/1834 implemented based on peer-to-peer networks, e.g. gnutella p2p communication protocols H04L67/104
• CPC Class Codes: G06F 16/184 implemented as replicated file system
• CPC Class Codes: G06F 16/22 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 9/54 Interprogram communication
• CPC Class Codes: H04L 41/0213 Standardised network management protocols, e.g. simple network management protocol [SNMP]

Abstract

A method for data storage implemented in a data storage system is disclosed. Data storage nodes may be interconnected by means of a communications network. The method may include sending a request for a first data item to a plurality of storage nodes. The first data item may include a reference to a second data item stored in the storage system. The method may include receiving the first data item from at least one storage node, and sending a request for the second data item to the plurality of storage nodes based on the reference included in the first data item.

Claims

1. A method for implementing a distributed directory structure in a data storage system, the method comprising:

identifying a first data item in a data storage system, wherein the first data item comprises a first collection object, the first collection object comprises one or more references to one or more other data items in the data storage system, the one or more references comprise a reference to a second data item stored in the data storage system, and the first collection object corresponds to a first subdirectory in a distributed directory system;

determining a first multicast address to use for requesting the first collection object;

sending a first request for the first collection object using the first multicast address, the first request including a unique identifier for the first collection object;

receiving at least one copy of the first collection object based on sending the first request using the first multicast address;

determining a second multicast address used to request the second data item in a cluster of the data storage system, wherein the reference to the second data item included in the first collection object comprises the second multicast address, the second data item comprises a second collection object, the second collection object corresponds to a second subdirectory, and the first subdirectory is a parent directory of the second subdirectory;

sending a second request for the second data item to the second multicast address included in the reference to the second data item included in the first collection object; and

receiving the second data item via a unicast message based on the second request.

2. The method as in claim 1, wherein multiple copies of the second data item are received based on the second request, each of the multiple copies being received from a different data storage node of the cluster.

3. The method as in claim 1, wherein the first data item and the second data item are stored in different clusters of data storage nodes.

4. The method as in claim 1, wherein the second collection object comprises at least one reference to at least one other data item in the data storage system, and the at least one reference comprises a reference to a third data item stored in the data storage system.

5. The method as in claim 4, wherein each of the first data item, the second data item, and the third data item are stored on different data storage nodes in the data storage system.

6. The method as in claim 1, wherein a collection object is a type of file utilized to implement the distributed directory structure.

7. The method as in claim 6, wherein collection objects in the data storage system each comprise at least one of a reference to another collection object or a reference to a data file.

8. The method as in claim 7, wherein the collection objects in the data storage system each represent a folder or subfolder of the distributed directory structure.

9. The method as in claim 1, wherein the first collection object also includes a reference to a data file, and the data file comprises payload data.

10. A method for locating a data file within a distributed directory system, the method comprising:

determining a path for accessing a first data item, wherein the first data item comprises a data file, the path for accessing the first data item includes one or more subdirectories, and each of the one or more subdirectories is represented in the distributed directory system using a corresponding collection object;

determining a first multicast address to use for requesting a first collection object that represents a first subdirectory in the path;

sending a first request for the first collection object using the first multicast address, the first request including a unique identifier for the first collection object;

receiving at least one copy of the first collection object based on sending the first request using the first multicast address, wherein the first collection object comprises a second multicast address for a second collection object representing a second subdirectory, the first subdirectory being a parent directory of the second subdirectory;

sending a second request for the second object using the second multicast address, wherein sending multicast requests for subsequent collection objects representing further subdirectories are continued until at least a subdirectory that includes the first data items is located; and

receiving the first data item via a unicast message.

11. The method as in claim 10, wherein a copy of each collection object is stored on multiple data storage nodes.

12. The method as in claim 11, further comprising determining which copy of a identified collection object is to be requested based at least in part on load information for the data storage nodes that store the identified collection object.

13. The method as in claim 10, further comprising using a root key to identify a collection object representing a root directory for the path, and requesting the collection object representing the root directory based on the root key.

14. A server implemented at least in part using hardware, the server comprising a processor configured to:

identify a first data item in a data storage system, wherein the first data item comprises a first collection object, the first collection object comprises one or more references to one or more other data items in the data storage system, the one or more references comprise a reference to a second data item stored in the data storage system, and the first collection object corresponds to a first subdirectory in a distributed directory system;

determine a first multicast address to use for requesting the first collection object;

send a first request for the first collection object using the first multicast address, the first request including a unique identifier for the first collection object;

receive at least one copy of the first collection object based on sending the first request using the first multicast address;

determine a second multicast address used to request the second data item in a cluster of the data storage system, wherein the reference to the second data item included in the first collection object comprises the second multicast address, the second data item comprises a second collection object, the second collection object corresponds to a second subdirectory, and the first subdirectory is a parent directory of the second subdirectory;

send a second request for the second data item to the second multicast address included in the reference to the second data item included in the first collection object; and

receive the second data item via a unicast message based on the second request.

15. A data storage node comprising a processor configured to:

identify a first data item in a data storage system, wherein the first data item comprises a first collection object, the first collection object comprises one or more references to one or more other data items in the data storage system, the one or more references comprise a reference to a second data item stored in the data storage system, and the first collection object corresponds to a first subdirectory in a distributed directory system;

determine a first multicast address to use for requesting the first collection object;

send a first request for the first collection object using the first multicast address, the first request including a unique identifier for the first collection object;

receive at least one copy of the first collection object based on sending the first request using the first multicast address;

determine a second multicast address used to request the second data item in a cluster of the data storage system, wherein the reference to the second data item included in the first collection object comprises the second multicast address, the second data item comprises a second collection object, the second collection object corresponds to a second subdirectory, and the first subdirectory is a parent directory of the second subdirectory;

send a second request for the second data item to the second multicast address included in the reference to the second data item included in the first collection object; and

receive the second data item via a unicast message based on the second request.

16. The data storage node as in claim 15, wherein the processor is configured to add a new data file to the second subdirectory corresponding to the second collection object by adding a reference to the new data file to the second collection object.

17. The data storage node as in claim 15, wherein the processor is configured to add a new subdirectory to the first subdirectory corresponding to the first collection object by adding a reference to a third collection object to the first collection object, the third collection object corresponding to the new subdirectory.

18. The data storage node as in claim 15, wherein the processor is configured to delete the second subdirectory corresponding to the second collection object by removing the reference to the second collection object from the first collection object.