Apparatus and methods for sharing idle workstations

• US 5,978,829 A
• Filed: 11/05/1996
• Issued: 11/02/1999
• Est. Priority Date: 11/05/1996
• Status: Expired due to Term

• Current Assignee: AT&T Corporation (AT&T, Inc.), Lucent Technologies, Inc. (Nokia Corporation)
• Original Assignee: AT&T Corporation (AT&T, Inc.)
• Inventors: Fowler, Glenn Stephen, Wang, Yi-Min, Huang, Yennun, Vo, Kiem-Phong, Chung, Pi-Yu
• Primary Examiner(s): Banankhah, Majid A.
• Application Number:
  US08/740,908
• Time in Patent Office: 1,092 Days
• Field of Search: 395/673, 395/672, 709/100, 709/102, 709/103, 709/104, 709/107
• US Class Current: 718/102
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: G06F 9/5072 Grid computing
• CPC Class Codes: G06F 9/5088 involving task migration

Abstract

The present invention relates to systems for sharing idle workstation computers that are connected together through a network and shared file system. More particularly, a user of a local host workstation may submit jobs for execution on remote workstations. The systems of the present invention select a remote host that is idle in accordance with a decentralized scheduling scheme and then continuously monitor the activity of the remote host on which the job is executing. If the system detects certain activity on the remote host by one of the remote host'"'"'s primary users, the execution of the job is immediately suspended to prevent inconvenience to the primary users. The system also suspends job execution if the remote host'"'"'s load average gets too high. Either way, the suspended job is migrated by selecting another idle remote workstation to resume execution of the suspended job (from the point in time at which the last checkpoint occurred).

Claims

1. A system for sharing computer resources among a plurality of computers connected together by a network, comprising:

a local computer that accepts jobs for remote execution, the local computer being one of the plurality of computers;

a remote computer that is capable of receiving jobs for remote execution, the remote computer being one of the plurality of computers and executing a current status process to collect current activity information about primary users of the remote computer; and

a scheduling computer that chooses a location for remote execution of the accepted jobs based upon at least the remote computer'"'"'s primary user current activity information, the scheduling computer being one of the plurality of computers.

2. The system of claim 1, wherein the scheduling computer and the local computer are the same computer.

3. The system of claim 1, wherein the scheduling computer and the remote computer are the same computer.

4. The system of claim 1, wherein the local computer, the remote computer and the scheduling computer are three different computers of the plurality of computers.

5. The system of claim 1, wherein the scheduling computer capability of choosing a location for remote execution is executed as a decentralized scheduling process among the plurality of computers.

6. The system of claim 1, wherein the local computer executes a submission process in response to a job being submitted for remote execution by a user of the local computer.

7. The system of claim 1, wherein the remote computer also executes:

a monitor process that monitors the activity of primary users of the remote computer.

8. The system of claim 1, wherein the activity information includes time elapsed since a last activity on an input device by a primary user of the remote computer.

9. The system of claim 8, wherein the input device is directly connected to the remote computer.

10. The system of claim 8, wherein the input device is connected to the remote computer through a remote login process.

11. The system of claim 1, wherein the scheduling computer chooses a remote computer as the computer to remotely execute a job only if the time elapsed since a last activity on the input device by a primary user of the remote computer is greater than a predetermined threshold.

12. The system of claim 1, further comprising a file system shared by at least the local and remote computers.

13. The system of claim 12, wherein a primary user of the remote computer is determined by attribute information stored in a central location in the file system.

14. The system of claim 13, wherein the scheduling computer chooses a location based on current information kept in the file system and the current activity status process on each remote computer posts the current information it collects to the central location in the file system.

15. A system for sharing computer resources among a plurality of computers connected together by a network, comprising:

a local computer that accepts jobs that a user has submitted to the local computer for execution on a remote computer; and

a remote computer, selected by a scheduling process, that executes the job submitted by the user, and executes a monitor process that causes the execution of the job to be suspended if the monitor process detects activity on the remote computer by a primary user of the remote computer.

16. The system of claim 15, wherein the remote computer suspends execution of the job if activity by a primary user of the remote computer is detected on an input device of the remote computer.

17. The system of claim 15, wherein the local computer executes a submission process such that if the job is suspended by the remote computer, the job is resubmitted by the submission process to the scheduling process for execution on a second remote computer.

18. The system of claim 15, wherein the monitor process is spawned from a parent process.

19. The system of claim 15, wherein the monitor process is a fine-grain polling process.

20. The system of claim 15, wherein the job is linked with a checkpointing library.

21. The system of claim 15, further comprising a file system shared by at least the local and remote computers.

22. The system of claim 21, wherein a primary user of the remote computer is determined by attribute information stored in a central location in the file system.

23. The system of claim 15, wherein the scheduling process is a decentralized scheduling process.

24. The system of claim 23, wherein the decentralized scheduling process is executed on the local computer that submits a job for remote execution.

25. The system of claim 15, wherein the scheduling process is executed by a single processor for the entire network.

26. A method of sharing computer resources among a plurality of computers connected together by a network comprising the steps of:

accepting jobs for remote execution on a first computer;

executing a scheduling process to schedule the remote execution of the accepted jobs based on current activity information about each of the plurality of primary users of each of the computers that may be remotely accessed; and

running a current status process on at least one second computer of the plurality of computers, the second computer being one of the remotely accessible computers, the current status process operating to collect current activity information about primary users of the second computer.

27. The method of claim 26, further comprising the step of:

executing a submission process on the first computer in response to a job being submitted by a user of the first computer.

28. The method of claim 26, wherein the step of running a current status process collects activity information that includes time elapsed since a last activity on an input device by a primary user of the second computer.

29. The method of claim 26, wherein the step of executing a scheduling process chooses a remote computer for remote execution of the submitted job only if the time elapsed since a last activity on an input device by a primary user of the remote computer is greater than a predetermined threshold, the remote computer being one of the second computers.

30. The method of claim 29, further comprising the steps of:

executing one of the accepted jobs on the remote computer chosen by the scheduling process; and

monitoring the remote computer for activity by a primary user of the remote computer.

31. The method of claim 30, further comprising the step of:

suspending the remote execution of the job when activity by a primary user is sensed during the step of monitoring.

32. The method of claim 31, further comprising the step of:

resubmitting the suspended job to the scheduling process for execution on a different remote computer when activity by a primary user is sensed during the step of monitoring.

33. The method of claim 30, further comprising the step of:

linking the job to a checkpointing library prior to executing the job. scheduling process is executed as a decentralized scheduling process.

34. The method of claim 26, wherein the step of executing a scheduling process is executed as a decentralized scheduling proceess.

35. The method of claim 26, wherein the step of executing a scheduling process is executed as a centralized scheduling process on a single processor.