Parallel computing system, method and architecture

• US 7,174,381 B2
• Filed: 12/04/2002
• Issued: 02/06/2007
• Est. Priority Date: 12/04/2001
• Status: Expired due to Fees

• Current Assignee: Simtone Corporation
• Original Assignee: Aspeed Software
• Inventors: Gulko, Abraham, Mellor, David
• Primary Examiner(s): Follansbee, John
• Assistant Examiner(s): Nguyen, Dustin
• Application Number:
  US10/310,115
• Publication Number:
  US 20030177240A1
• Time in Patent Office: 1,525 Days
• Field of Search: 717/119, 717/148, 717/151, 717/149, 718/105, 709/201, 709/223
• US Class Current: 709/226
• CPC Class Codes: G06F 2209/5017 Task decomposition
• CPC Class Codes: G06F 2209/508 Monitor
• CPC Class Codes: G06F 8/451 Code distribution considering CPU load at run-time G06F9/505; load rebalancing G06F9/5083
• CPC Class Codes: G06F 8/456 Parallelism detection
• CPC Class Codes: G06F 9/5038 considering the execution order of a plurality of tasks, e.g. taking priority or time dependency constraints into consideration scheduling strategies G06F9/4881 and subgroups
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: G06F 9/5072 Grid computing

Abstract

A parallel or computing system and method make use of a plurality of adapters each affording a specific type of processing algorithm. At each point in an application, an adapter is identified, to parallelize that portion of the application. The process involves associating an appropriate adapter with the application portion, parsing the application portion to define tasks that may be distributed over the available computers. Combined with the parallelized program is a software server that is designed to cooperate with the adapters to control and supervise distributed processing functions. Such functions include mapping, load balancing, and error detection and correction. Also included are functions that cause the results of the computing to be coordinated in real-time and returned appropriately, according to the application. In addition to the parallelized program and server, an application created by the present invention is provided with the functionality of virtual middleware, which interfaces with and controls existing forms of middleware.

Claims

1. A method for parallelizing an application, comprising the steps of:

providing an adapter with a plurality of processing algorithm type functionalities;

identifying an algorithm within the application to adapt such that the adapter will parallelize the execution of such algorithm;

associating the adapter with the algorithm within the application, wherein the association between the adapter and the algorithm is based on correlation between the algorithm and one of the plurality of processing algorithm type functionalities;

adapting the application to conform to a plurality of logical components, at least one of the plurality of logical components being encapsulated by the adapter and at least one of the plurality of logical components including non-parallelized or non-adapted code;

creating multiple instances of each of the plurality of logical components;

distributing the multiple instances of each of the plurality of logical components for parallel execution across a plurality of discreet systems or a plurality of processors in a single system; and

monitoring and balancing the parallel execution;

wherein the above steps are executed at run-time.

2. The method of claim 1 further comprising:

using one of the multiple instances with respect to one of the plurality of logical components as a master and the remaining multiple instances as slaves;

using the master to communicate with the slaves in order to retrieve execution information from the slaves, the execution information including information specific to application progress; and

using the master to provide adaptive commands to the slaves based on the execution information in order to optimize execution performance of the slaves.

3. The method of claim 1 further comprising:

including a software engine in at least one of the plurality of logical components to support the adapter, wherein the software engine is configured to cooperate with the associated adapter to control and supervise distributed processing functions; and

including computational and execution environment information in at least one of the plurality of logical components.

4. The method of claim 3, wherein the distributed processing functions include at least one of data mapping, serialization, data and range distribution, and stall and error detection and correction.

5. The method of claim 3, wherein the distributed processing functions coordinate results of the parallel execution with respect to the distributed multiple instances corresponding to the plurality of logical components in real-time and return said results.

6. A method for running an application, comprising the steps of:

identifying a portion of the application for parallelization;

associating an adapter with the identified portion, wherein the adapter is configured to provide a plurality of processing algorithm type functionalities;

creating a master including the adapter and remainder portions of the application not identified for parallelization;

using the master that operates in a master mode to instantiate a plurality of stateless slaves for parallel execution, each stateless slave including a clone of the master;

reporting to the master progress of the parallel execution with respect to the plurality of stateless slaves, the progress including application progress information with respect to the clones; and

directing the master to distribute adaptive commands to the plurality of stateless slaves to adjust the parallel execution based on the reported progress;

wherein the above steps are executed at run-time.

7. The method of claim 6 wherein the adaptive commands include data and range information.

8. The method of claim 7 further comprising:

monitoring the progress of the parallel execution with respect to the plurality of stateless slaves after the adaptive commands have been received by the plurality of stateless slaves;

adjusting and re-distributing the adaptive commands including the data and range information to the plurality of stateless slaves based on the monitored progress;

committing the adaptive commands to an optimal level for the parallel execution with respect to the plurality of stateless slaves once an optimal performance level has been reached.

9. The method of claim 6 wherein each stateless slave further includes a software engine configured to cooperate with the clone of the adapter to control and supervise distributed processing functions;

and

wherein each stateless slave further includes computational and execution environment information.

10. The method of claim 6 further comprising:

distributing the plurality of stateless slaves for parallel execution across a plurality of discreet systems or a plurality of processors in a single system.