Method for the management of local client cache buffers in a clustered computer environment

• US 20050108300A1
• Filed: 11/17/2003
• Published: 05/19/2005
• Est. Priority Date: 11/17/2003
• Status: Abandoned Application

• Current Assignee: Rackable Systems Incorporated (Hewlett-Packard Enterprise Company)
• Original Assignee: Terrascale Technologies, Inc. (Hewlett-Packard Enterprise Company)
• Inventors: Findleton, Iain B., Zhou, Xinliang
• Application Number:
  US10/714,401
• Publication Number:
  US 20050108300A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 12/0815 Cache consistency protocols
• CPC Class Codes: G06F 12/0866 for peripheral storage systems, e.g. disk cache
• CPC Class Codes: G06F 16/9574 of access to content, e.g. by caching

Abstract

A method for retrieving data elements from a shared medium by a client computer is provided. The shared medium maintains a main list of data version information associated with each data element, while the client maintains another locally-stored list containing all previously retrieved data elements associated with their respective data version. The client computer checks whether there is a record for a data element in the locally-stored list and if so, sends the data version associated with it to the shared medium. The shared medium compares the data version received with the data version stored in its main list and if the two match, sends a confirmation to the client computer. If the data versions do not match, the shared medium sends a new copy of the data element and a new data version to the originating client computer, such that the client computer updates its locally-stored list.

Claims

1. A mehod for executing a common task in a clustered computing environment comprising a plurality of computers interconnected to collaborate on said common task, said plurality of computers including at least a client computer and a shared storage medium storing data elements, said shared storage medium maintaining a main list of data version information associated with said data elements, said method comrising:

said client computer maintaining a locally-stored list containing previously retreived data elements associated with their data version;

said client computer reading from said locally-stored list data version associated with said data element and sending a request over a data network including said data version to said shared medium;

if said data version received from said client computer does not match said main list data version associated with said data element, said shared storage medium sending to said client compputer a new copy of said data element and a new data version, said client computer updating said locally-stored list with said new copy of said data element and said new data version;

if said data version received from said client computer matches said main list data version associated with said data element, said shared storage medium sending to said client computer confirmation that said locally-stored data element associated with said data version is valid;

at least one of said plurality of computers modifying said data element stored on said shared storage medium and said client computer using said retrived data element to execute said common task;

whereby transfer of copies of data elements between said shared storage medium and said plurality of computers is reduced and an amount of network load needed to retrieve data elements from said shared storage medium is reduced.

2. The method as claimed in claim 1, wherein said client sending said data version to said shared medium comprises sending a null-value data version in the case in which said data element is not stored in said client memory and said shared medium.

3. The method as claimed in claim 1, wherein said request for said data element contains an address range defining said data element on said shared medium.

4. The method as claimed in claim 3, wherein said address range comprises non-contiguous storage blocks.

5. The method as claimed in claim 1, wherein said client computer communicates with said shared medium through a network block device driver.

6. The method as claimed in claim 1, wherein said shared medium is a server memory storage space.

7. A method for maintaining a main list of data version information associated with data elements on a shared medium, said data version information being used for data retrieval, comprising:

creating a list of data structures identifying data elements on said shared medium and said data version information;

receiving a request on a data network for writing at least one of said data elements;

following modification to said at least one of said data elements, giving a new data version to said at least one of said data elements that was modified.

8. The method as claimed in claim 7, wherein if said data elements being modified are associated with multiple separate data structures containing data version information, creating a new single data structure in siad list associated with said data elements modified and removing said multiple separate data structures from said list.

9. The method as claimed in claim 7, wherein said initial version state is an initial version number and wherein said initial version number is incremented to obtain said new version state.

10. The method as claimed in claim 7, wherein said list of said structures is a double linked binary tree list.

11. A method for managing data version information associated with data elements on a shared storage medium in a clustered computing environment, said data version information being used for data retrieval by a plurality of computers interconnected in said clustered computing environment, comprising:

creating a list of data structures identifying data elements on said shared storage medium and said data version information;

receiving a request on a data network from at least one of said plurality of computers for writing at least one of said data elements;

following modifications to said at least one of said data elements, giving a new data version to said at least one of said data elements that was modified.

12. The method as claimed in claim 11, wherein if said data elements being modified are associated with multiple separate data structures containing data version information, creating a new single data structure in said list associated with said data elements modified and removing said multiple separate data structures from said list.

13. The method as claimed in claim 11, wherein said initial version state is an initial version number and wherein said initial version number is incremented to obtain said new version state.

14. The method as claimed in claim 11, wherein said list of data structures is a double linked binary tree list.