Method, apparatus and computer program product for discovering and prioritizing patterns of component usage in a virtual application container for enhancing prefetching

• US 7,849,156 B2
• Filed: 06/04/2007
• Issued: 12/07/2010
• Est. Priority Date: 06/04/2007
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Mummert, Todd W., Alpern, Bowen L., Bala, Vasanth, Ammons, Glenn, Reimer, Darrell Christopher
• Primary Examiner(s): Lin; Kenny S
• Assistant Examiner(s): HOANG, HIEU T
• Application Number:
  US11/757,477
• Publication Number:
  US 20080301316A1
• Time in Patent Office: 1,282 Days
• Field of Search: 709/224, 709/227, 709/228, 709/229, 709/217, 709/218
• US Class Current: 709/217
• CPC Class Codes: G06F 9/4843 by program, e.g. task dispatcher, supervisor, operating system
• CPC Class Codes: H04L 67/131 Protocols for games, networked simulations or virtual reality

Abstract

Disclosed is a method, computer program product and a server unit of a virtual execution environment having at least one client. The server unit includes a network interface for conducting bidirectional communications with the at least one client and a control unit configured to determine, for individual ones of positions in a virtualized execution of an application, those components that are required at each position; to establish a pseudo-distance between the positions in the virtualized execution of the application; and to identify components that are candidates for prefetching as those components required at one or more positions that are proximate to a current position. The pseudo-distance between positions is indicative of an urgency of prefetching candidate components for regulating an amount of resources devoted to the prefetching of components.

Claims

1. A method, comprising:

for individual ones of positions in a virtualized execution of an application, determining, by a data processor, components that are required at each position, where a position is a point at which the virtualized execution of the application requires at least one new component, and where a component is a decomposition of the application;

the data processor establishing a pseudo-distance between the positions in the virtualized execution of the application by hierarchical clustering of requests for components into a binary tree, where distances between leaves of the binary tree approximate pseudo-distances; and

the data processor, using at least a portion of the binary tree, identifying components that are candidates for prefetching as those components required at one or more positions that are proximate to a current position;

where the pseudo-distance is a measure of a likelihood of moving from one position in the virtualized execution of the application to another position in the virtualized execution of the application within some period of time.

2. The method of claim 1, where the pseudo-distance between the positions is indicative of an urgency of prefetching candidate components, and further comprising regulating, by the data processor, an amount of resources devoted to the prefetching of components in accordance with the pseudo-distance.

3. The method of claim 1, where determining and establishing, by the data processor, are used for building a model of application component usage, and further comprising sending the model or a portion of the model to a client that executes the application for use in generating component prefetch requests prior to a time when at least a first component of a group of components is required.

4. The method of claim 1, where determining and establishing, by the data processor, are used for building a model of application component usage, and further comprising pushing components to a client that executes the application prior to a time when at least a first component of a group of components is required.

5. The method of claim 1, where determining and establishing, by the data processor, are used for building a model of application component usage, the model comprising a tree structure constructed having positions as leaves and internal nodes that represent a hierarchy of phases of execution of the application, where descendants of a given node represent positions or sub-phases within the phase, where the pseudo-distance between positions is approximated by closeness in the tree, and where each node of the tree comprises a set of components required by its descendants for specifying components as candidates for prefetching.

6. A tangible computer-readable storage medium storing data representing program instructions, the execution of which by at least one processor results in operations comprising:

for individual ones of positions in a virtualized execution of an application, determining components that are required at each position, where a position is a point at which the virtualized execution of the application requires at least one new component, and where a component is a decomposition of the application;

establishing a pseudo-distance between the positions in the virtualized execution of the application by hierarchical clustering of requests for components into a binary tree, where distances between leaves of the binary tree approximate pseudo-distances; and

identifying, using at least a portion of the binary tree, components that are candidates for prefetching as those components required at one or more positions that are proximate to a current position;

where the pseudo-distance is a measure of a likelihood of moving from one position in the virtualized execution of the application to another position in the virtualized execution of the application within some period of time.

7. The tangible computer-readable storage medium of claim 6, where the pseudo-distance between the positions is indicative of an urgency of prefetching candidate components, and further comprising an operation of regulating an amount of resources devoted to the prefetching of components in accordance with the pseudo-distance.

8. The tangible computer-readable storage medium of claim 6, where the operations of determining and establishing comprise an operation of building a model at a server component of application component usage, and further comprising an operation of sending the model or a portion of the model to at least one client component that executes the application for use in generating component prefetch requests prior to a time when at least a first component of a group of components is required.

9. The tangible computer-readable storage medium of claim 6, where the operations of determining and establishing comprise an operation of building a model at a server component of application component usage, and further comprising an operation of pushing components to a client component that executes the application prior to a time when at least a first component of a group of components is required.

10. The tangible computer-readable storage medium of claim 6, where the operations of determining and establishing comprise an operation of building a model of application component usage, the model comprising a tree structure constructed having positions as leaves and internal nodes that represent a hierarchy of phases of execution of the application, where descendants of a given node represent positions or sub-phases within the phase, where the pseudo-distance between positions is approximated by closeness in the tree, and where each node of the tree comprises a set of components required by its descendants for specifying components as candidates for prefetching.

11. The tangible computer-readable storage medium of claim 6, located at a server component in a virtual execution environment, where the operations of determining and establishing use traces of prefetch requests that are at least one of:

received from at least one client component, and made to a local cache at the client component.

12. A server of a virtual execution environment that comprises at least one client, comprising:

a network interface for conducting bidirectional communications with the at least one client; and

a memory, coupled to the network interface;

a data processor, coupled to the memory, configured to determine, for individual ones of positions in a virtualized execution of an application, components that are required at each position, where a position is a point at which the virtualized execution of the application requires at least one new component, and where a component is a decomposition of the application;

to establish a pseudo-distance between the positions in the virtualized execution of the application by hierarchical clustering of requests for components into a binary tree, where distances between leaves of the binary tree approximate pseudo-distances; and

to identify, using at least a portion of the binary tree, components that are candidates for prefetching as those components required at one or more positions that are proximate to a current position, where the pseudo-distance between positions is indicative of an urgency of prefetching candidate components for regulating an amount of resources devoted to the prefetching of components;

where the pseudo-distance is a measure of a likelihood of moving from one position in the virtualized execution of the application to another position in the virtualized execution of the application within some period of time.

13. The server of claim 12, where said control unit builds a model of application component usage and is further configured to sending the model or a portion of the model to the at least one client for use by the at least one client in generating component prefetch requests to the server prior to a time when at least a first component of a group of components is required.

14. The server of claim 12, where said control unit builds a model of application component usage and is further configured to push components to the at least one client prior to a time when at least a first component of a group of components is required.

15. The server of claim 12, where said control unit builds a model of application component usage, the model comprising a tree structure constructed having positions as leaves and internal nodes that represent a hierarchy of phases of execution of the application, where descendants of a given node represent positions or sub-phases within the phase, where the pseudo-distance between positions is approximated by closeness in the tree, and where each node of the tree comprises a set of components required by its descendants for specifying components as candidates for prefetching.