File system adapted for use with a dispersed data storage network

• US 20090094320A1
• Filed: 07/14/2008
• Published: 04/09/2009
• Est. Priority Date: 10/09/2007
• Status: Active Grant

• Current Assignee: Pure Storage, Inc.
• Original Assignee: Imc. Cleversafe
• Inventors: Palthepu, Srinivas, Dhuse, Greg, Volvovski, Ilya, Thornton, Vance, Mark, Zachary, Gladwin, S. Christopher, Baptist, Andrew, Cilfone, Bart, Resch, Jason, Quigley, John
• Granted Patent: US 8,209,363 B2
• Time in Patent Office: Days
• US Class Current: 709/203
• CPC Class Codes: G06F 16/148 File search processing
• CPC Class Codes: G06F 16/182 Distributed file systems

Abstract

A file system interface to a dispersed data storage network is disclosed. The disclosed interface accepts read and write commands from a file system resident on a user'"'"'s computer and generates network commands that are forwarded to slice servers that form the storage component of the dispersed data storage network. The slice servers then fulfill the read and write commands.

Claims

1. A method of reading a data segment from a dispersed data storage network, said dispersed data storage network comprising at least one grid client and a plurality of slice servers, wherein at least some of said plurality of slice servers are located in separate facilities, and wherein a data segment cannot be reconstructed without accessing more than one of said plurality of slice servers, said method comprising the steps of:

i) requesting access to a file object referencing data stored on said dispersed data storage network from an interface operating on one of said plurality of computers;

ii) loading a metadata object associated with said file object into said file system cache;

iii) loading said file object into said file system cache; and

iv) reading data referenced by said file object from said dispersed data storage network.

2. A method of writing a data segment to a dispersed data storage network, said dispersed data storage network comprising at least one grid client and a plurality of slice servers, wherein at least some of said slice servers are located in separate facilities, and wherein a data segment written to said dispersed data storage network cannot be reconstructed without accessing more than one of said plurality of slice servers, said method comprising the steps of:

i) requesting the creation of a file object to be associated with said data to be stored on said dispersed data storage network from an interface operating on said computer;

ii) identifying a first directory object to associate with said file object;

iii) copying said first directory object to a writeable second directory object;

iv) creating a reference to said file object within said writeable second directory object;

v) invaliding said first directory object;

vi) copying said writeable second directory object to a third read only directory object;

vii) writing the contents of said writeable second directory object to said dispersed data storage network; and

viii) destroying said writeable second directory object.

3. A method of writing a collection of data to a dispersed data storage network, said dispersed data storage network comprising at least one grid client and a plurality of slice servers, wherein at least some of said plurality of slice servers are located in separate facilities, wherein a collection of data is segmented into one or more data segments, and wherein a data segment written to said dispersed data storage network cannot be reconstructed without accessing more than one of said plurality of slice servers, said method comprising the steps of:

i) opening a file system object;

ii) writing a collection of data to said file system object;

iii) segmenting said collection of data into a plurality of data segments, wherein a particular data segment is of a different size from one or more other data segments; and

iv) for each data segment, slicing the data segment into a plurality of data slices and writing said plurality of data slices to a plurality of slice servers.

4. The method of claim 3 further comprising the step of incrementing the size of data segments associated with said collection of data so that data segments including data near a beginning of the collection of data is smaller than data segments including data near an end of the collection of data.

5. A grid access computer for use as part of a distributed computer system implementing a dispersed data storage network, said dispersed data storage network including at least a plurality of slice servers coupled to a network, wherein at least some of said plurality of slice servers are located in separate facilities, said grid access computer comprising:

i) a network port for communicating with at least one client computer and said plurality of slice servers over said network;

ii) a network layer for receiving packets from said network port; and

iii) a file object layer in communication with said network layer, said file object layer adapted to parse packets received from said network layer and based on said packets, issue requests to said plurality of slice servers.

6. The grid access computer of claim 5 further comprising a server layer adapted to parse network packets issued by a network file system operating on said client computer, said server layer in communication with said network layer and further in communication with said file object layer.

7. The grid access computer of claim 6 wherein said network file system is selected from a group consisting of NFS, WebDAV, and FTP.