Load balancing data access in virtualized storage nodes

• US 9,009,724 B2
• Filed: 09/24/2010
• Issued: 04/14/2015
• Est. Priority Date: 09/24/2010
• 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: Gold, Stephen, DiCorpo, Jeffrey, Pelly, Simon
• Primary Examiner(s): An, Meng
• Assistant Examiner(s): Rashid, Wissam
• Application Number:
  US12/889,727
• Publication Number:
  US 20120079499A1
• Time in Patent Office: 1,663 Days
• US Class Current: 718/105
• CPC Class Codes: G06F 3/0607 by facilitating the process of upgrading existing storage systems, e.g. for improving compatibility between host and storage device
• CPC Class Codes: G06F 3/0611 in relation to response time
• CPC Class Codes: G06F 3/0635 by changing the path, e.g. traffic rerouting, path reconfiguration
• CPC Class Codes: G06F 3/0664 at device level, e.g. emulation of a storage device or system
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]

Abstract

Systems and methods of load balancing data access in virtualized storage nodes are disclosed. An embodiment of a method includes receiving a data access request from a client for data on a plurality of the virtualized storage nodes. The method also includes connecting the client to one of the plurality of virtualized storage nodes having data for the data access request. The method also includes reconnecting the client to another one of the plurality of virtualized storage nodes to continue accessing data in the data access request.

Claims

1. A method of load balancing data access in virtualized storage nodes, comprising:

receiving a data access request from a client for data on a plurality of the virtualized storage nodes;

connecting the client to one of the plurality of virtualized storage nodes by moving a virtual drive to one of a plurality of virtualized storage nodes that contain a virtual cartridge having portions of the data for the data access request;

reconnecting the client to another one of the plurality of virtualized storage nodes, by moving the virtual drive to another virtualized storage node that contains another virtual cartridge to continue accessing remaining data in the data access request; and

directing new data write requests to a deduplication store already having duplicate data that matches the new data based on initial hash signatures.

2. The method of claim 1, further comprising signaling the client that the connection to one of the plurality of virtualized storage nodes is busy while reconnecting the client to another one of the plurality of virtualized storage nodes.

3. The method of claim 1, further comprising interrupting a connection with the client within a timeout window of a data access application.

4. The method of claim 3, wherein interrupting the connection with the client is transparent to the client.

5. The method of claim 3, wherein the connection with the client is restored before a data access application registers an error in data access operations.

6. The method of claim 1, further comprising reducing backend network traffic between the plurality of virtualized storage nodes.

7. The method of claim 1, further comprising balancing bandwidth for the plurality of virtualized storage nodes.

8. The method of claim 1, further comprising moving data among the plurality of virtualized storage nodes on a backend network during times of low backup operation traffic.

16. The method of claim 1, further comprising a virtual drive moving to different nodes as a read progresses through a virtual cartridge stored across the plurality of the virtualized storage nodes.

17. The method of claim 1, further comprising writing an initial portion of the new data to a wrong node, and locating the appropriate deduplication store based on the initial portion of the data written to the wrong node, and then moving the initial portion of the data written to the wrong node to a correct node.

18. The method of claim 1, further comprising presenting each of the plurality of the virtualized storage nodes as a unique port to a backup application executing on the client.

19. The method of claim 18, further comprising disabling a physical port of a virtual drive while enabling presentation of the virtual drive to a backup application executing on the client from a different physical port on a different one of the plurality of the virtualized storage nodes.

20. The method of claim 1, wherein pieces of a virtual cartridge are contained in different deduplication stores based on references to other duplicate data in other virtual cartridges of the virtualized storage nodes.

9. A system for load balancing data access in virtualized storage nodes, the system embodied at least in part in program code stored on a non-transitory computer readable storage and executable by a processor and comprising:

a processor;

an interface between the virtualized storage nodes and a client, the interface configured to receive data access requests from the client for data on the plurality of the virtualized storage nodes; and

a connection manager operatively associated with the interface, the connection manager configured to connect the client to one of the plurality of virtualized storage nodes by moving a virtual drive to one of a plurality of virtualized storage nodes that contain a virtual cartridge having portions of data for a data access request, and then reconnect the client to another one of the plurality of virtualized storage nodes by moving the virtual drive to another virtualized storage node that contains another virtual cartridge to provide access to data in the data access request, the data being stored across multiple of the plurality of virtualized storage nodes, and the connection manager directing new data write requests to a deduplication store already having duplicate data that matches the new data based on initial hash signatures.

10. The system of claim 9, wherein the interface signals the client that the connection to one of the plurality of virtualized storage nodes is busy when the connection manager reconnects the client to another one of the plurality of virtualized storage nodes.

11. The system of claim 9, wherein the connection manager reconnects the client to another one of the plurality of virtualized storage nodes within a timeout window of a data access application executing for the client.

12. The system of claim 9, wherein the connection manager reconnects the client to another one of the plurality of virtualized storage nodes transparently to the client.

13. The system of claim 9, wherein the connection manager reconnects the client to another one of the plurality of virtualized storage nodes before a data access application registers an error in data access operations.

14. A data access system embodied at least in part in program code stored on a non-transitory computer readable storage and executable by a processor and comprising:

a processor;

an interface for a plurality of virtualized storage nodes, the interface configured to receive data access requests from a client for data on the plurality of the virtualized storage nodes; and

a connection manager operatively associated with the interface, the connection manager configured to connect the client to one of the plurality of virtualized storage nodes by moving a virtual drive to one of a plurality of virtualized storage nodes that contain a virtual cartridge having portions of the data for the data access request, and then reconnect the client to another one of the plurality of virtualized storage nodes by moving the virtual drive to another virtualized storage node that contains another virtual cartridge to provide access to data stored across the plurality of virtualized storage nodes, and the connection manager directing new data write requests to a deduplication store already having duplicate data that matches the new data based on initial hash signatures.

15. The system of claim 14, wherein:

the interface signals the client that one of the plurality of virtualized storage nodes is busy while the connection manager reconnects the client to another one of the plurality of virtualized storage nodes; and

the connection manager reconnects the client to another one of the plurality of virtualized storage nodes within a timeout window of a data access application executing for the client before the data access application registers an error so that the connection/reconnection is transparent to the client.