Peer to peer job monitoring and control in grid computing systems

• US 20050015437A1
• Filed: 06/11/2003
• Published: 01/20/2005
• Est. Priority Date: 06/11/2003
• Status: Active Grant

• Current Assignee: Snap, Inc.
• Original Assignee: International Business Machines Corporation
• Inventors: Strait, Gary E.
• Granted Patent: US 8,055,753 B2
• Time in Patent Office: Days
• US Class Current: 709/203
• 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 solution for improved monitoring and control of jobs in grid and batch computing systems provides a centralized server'"'"'s batch manager which is only responsible for workload balancing and job initiation and completion, all other command and status information are communicated directly between the plurality of submitter'"'"'s systems and the plurality of client systems that are processing their respective workloads. The computing system and communication process utilizes event-driven peer to peer communications between submitter'"'"'s systems and client systems and enables more detailed status and control information to be passed without overloading the centralized server, and by avoiding polling also provides more immediate feedback of results. Multiple process threads are employed on both the submitter'"'"'s and client systems, and a user interface consolidates and displays results to the submitter allow commands to be sent to processes running on client batch systems.

Claims

1. A computer system and a communication process, comprising:

a centralized server and a submitter application utilized to provide overall control of a batch or grid computing system for receiving and dispatching requests, in combination with an application process which simultaneously allows clients processing dispatched requests to pass a communication back to that client submitting an original request directly without said communication passing through said centralized server.

7. The computer system and a communication process according to claim 1, wherein said communication process includes a connection process on a submitting client for receiving new connections from processing clients whose identity is not known in advance.

2. A computer system application, comprising a process for communication allowing a client submitting a request to a centralized server for processing on an unknown set of clients to embed instructions in requests distributed through said centralized server to said clients instructing the selected processing clients as to how and where to connect to the submitting client.

3. A user interface process on a submitting client, comprising a step presenting a submitting client submitter with a user interface, and simultaneously receiving and displaying on a user interface a summary of input from a plurality of processing clients, while enabling a step of simultaneously sending commands to one or more of said processing clients.

4. The user interface process of claim 3, further comprising a step for said user interface process of enabling display of additional informational windows displaying additional details of communications with a job step on a selected individual processing client.

5. The user interface process of claim 3, further comprising a step enabling interpretation by said user interface of communications received from processing clients and for summarizing the results and presenting the submitter with selected important results immediately as soon as communicated from said processing clients.

6. A user interface process on a submitting client, comprising a step presenting a submitting client submitter with a user interface, and simultaneously receiving and displaying on a user interface a summary of input from a plurality of processing clients, while enabling a step of simultaneously sending commands to one or more of said processing clients, further including a step of enabling one or more of different instances of user interface processes that may exist on a submitting client submitter simultaneously to detect a preferred communication ports already in use the existence of another instance or instances, and from said detection to select another available communications port, and to also change a visible characteristic of said additional instance, for example the color of the main window and all sub-windows, so that the submitter distinguishes the windows belonging to one instance of the disclosed monitoring system and associated job steps from other instances.

8. A computer system and a communication process, comprising:

a centralized server and a submitter application utilized to provide overall control of a batch or grid computing system for receiving and dispatching requests, in combination with an application process which simultaneously allows clients processing dispatched requests to pass a communication with a computing system, including a system of keys that may be passed through a centralized server for an initial job initiation message, and directly between clients for other connections, to allow the clients to positively identify each other and exclude communications from other clients in said computing system.

9. A computer system communication process comprising, a step on the submitting client per job step for receiving input from an associated processing client, and tagging said input from said associated processing client with identifying information specific to a one job step performed by said processing client and returning said input after tagging to a user interface process for display as input associated with said specific processing client and said tagged input'"'"'s one job step.

10. A computing system communication process comprising:

a command process established on a processing client for the purpose of receiving commands from a user interface process on a submitting client, and a step for forwarding said received commands from a user interface process on a submitting client via a pipeline to a batch job step on said processing client.

11. A computing system communication process comprising:

a response process established on a processing client for the purpose of receiving results from a batch job step on said processing client, and a step forwarding the received results immediately over a communications link to a submitting client for display and analysis.

12. A computing system and communication process comprising:

a computing system coupled to a plurality of processing clients and providing a bi-directional pipe and a communication process application for simultaneously receiving output from a batch job step on a processing client, and for sending input to the same batch job step on one of said processing client.

13. A computing system and communication process comprising:

a computing system coupled via a communications link to a plurality of clients acting as processing and submitting clients and providing a bi-directional pipe and a communication process application for simultaneously receiving output from a batch job step on a processing client, and for sending input to the same batch job step on one of said processing client, wherein in a communication process a submitting client detects the completion of a processing step on a processing client by detection of the termination of a communications link with said processing client.

14. A computing system and communication process comprising, a computing system having a centralized server and communication application wherein a centralized batch manager is only responsible for workload balancing and job initiation and completion while other command and status information are communicated directly between a plurality of submitter'"'"'s systems and a plurality of client systems of said computing system that are processing their respective workloads.

15. The computing system and communication process of claim 14, utilizing event-driven peer to peer communications between the submitter'"'"'s systems and client systems said communication process enables detailed status and control information to be passed without overloading the centralized server and without polling.

16. The computing system and communication process of claim 15, wherein multiple process threads are employed on both submitter systems and client batch systems coupled to said centralized server, and a user interface consolidates and displays results to a submitter and to allow commands to be sent to processes running on client batch systems.