CLUSTERED COMPUTER SYSTEM

• US 20070255782A1
• Filed: 05/16/2007
• Published: 11/01/2007
• Est. Priority Date: 09/26/2005
• Status: Active Grant

• Current Assignee: Advanced Cluster Systems, Inc.
• Original Assignee: Advanced Cluster Systems, LLC
• Inventors: Frajka, Tamas, Westerhoff, John, Tannenbaum, Zvi, Tannenbaum, Ariela, Minor, Bryan
• Granted Patent: US 8,402,080 B2
• Time in Patent Office: Days
• US Class Current: 709/201
• CPC Class Codes: G06F 2209/509 Offload
• CPC Class Codes: G06F 9/4843 by program, e.g. task dispatcher, supervisor, operating system
• CPC Class Codes: G06F 9/5033 considering data affinity
• CPC Class Codes: G06F 9/5072 Grid computing
• CPC Class Codes: H04L 67/02 based on web technology, e.g. hypertext transfer protocol [HTTP]
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46

Abstract

A clustered computational system comprises a clustered computational engine, a program host, and a queue manager. The clustered computational engine comprises a plurality of clustered computers and is configured to perform computations. The program host stores a plurality of programs that define computations that can be performed by the clustered computational engine and a plurality of inputs to the programs. The queue manager is configured to determine when the program host has stored sufficient inputs to allow the clustered computational engine to perform a computation defined by the program and to schedule the performance of the computation by the clustered computational engine.

Claims

1. A clustered computational system comprising:

a clustered computational engine comprising a plurality of clustered computers and configured to perform computations;

a program host computer that stores, on a computer-accessible storage medium, a plurality of programs and a plurality of inputs to the programs, wherein the programs define computations that can be performed by the clustered computational engine; and

a queue manager configured to periodically poll the program host computer in order to determine when the program host has stored sufficient inputs to allow the clustered computational engine to perform a computation defined by the program and to schedule the performance of the computation by the clustered computational engine.

2. The system of claim 1, wherein the clustered computational engine comprises a clustered math engine.

3. The system of claim 2, wherein the clustered math engine comprises at least one computer executing gridMathematica.

4. The system of claim 1, wherein the program host is configured to communicate with a web server to receive program selections and inputs.

5. The system of claim 1, further comprising a web server configured to receive user selections of computations to be performed on the clustered computational engine, to transmit the user selections to the program host, and to receive results of computations performed by the clustered computational engine.

6. The system of claim 5, wherein the web server is further configured to display the results of computations.

7. The system of claim 6, wherein the web server is in communication with a graphics component that assists the web server to display the results in graphical format.

8. The system of claim 7, wherein the graphics component is further configured to perform computations and the system is able to assign the graphics component to perform computations that do not require the clustered computational engine.

9. The system of claim 8, wherein the graphics component comprises webMathematica.

10. A system for providing a web-based interface to gridMathematica, comprising:

a web server in communication with webMathematica and configured to present to a user a plurality of web pages for selecting computations that can be performed by gridMathematica and for entering inputs for the computations; and

a queue manager configured to schedule computations to be performed by gridMathematica and to communicate with gridMathematica to cause gridMathematica to perform a scheduled computation.

11. The interface of claim 10, wherein the queue manager is further configured to schedule computations in accordance with a priority assigned to at least one computation.

12. The interface of claim 10, wherein the web server is further configured to present to the user at least one status web page showing a status of at least one scheduled computation.

13. The interface of claim 10, wherein the web server is further configured to present to the user at least one results web page that shows results of a computation performed by gridMathematica.

14. The interface of claim 13, wherein webMathematica assists the web server to show results of the computation in graphical format.

15. The interface of claim 14, further configured such that webMathematica performs at least some computations that do not require gridMathematica.

16. A method of performing a computation, the method comprising:

storing at least one program that defines a computation;

receiving a selection of a program;

storing a plurality of inputs to the selected program;

detecting that a complete set of inputs to the selected program has been stored;

scheduling execution of the selected program;

transmitting the selected program and the inputs to a first remote location for execution of the computation defined by the program at the first remote location; and

receiving results of the executed computation and storing the results in a computer-accessible storage medium.

17. The method of claim 16, wherein receiving a selection of a program comprises receiving a user selection of the program from a second remote location that is different from the first remote location.

18. The method of claim 17, wherein receiving a selection of a program further comprises receiving the selection from a user terminal that comprises a web browser.

19. The method of claim 16, further comprising executing the computations at the first remote location on a clustered computational engine.

20. The method of claim 19, wherein the clustered computational engine comprises a clustered math engine.