System and method for performing flexible workflow process compensation in a distributed workflow management system

• US 5,870,545 A
• Filed: 04/04/1997
• Issued: 02/09/1999
• Est. Priority Date: 12/05/1996
• Status: Expired due to Term

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Company (HP Inc.)
• Inventors: Davis, James W., Shan, Ming-Chien, Du, Weimin
• Primary Examiner(s): Harrell, Robert B.
• Application Number:
  US08/825,853
• Time in Patent Office: 676 Days
• Field of Search: 364/DIG. 1, 364/DIG. 2, 395/180, 395/181, 395/182.01, 395/182.02, 395/182.03, 395/182.1, 395/183.2, 395/184.01, 395/200.3, 395/200.31-200.51, 395/670, 395/675, 707/103
• US Class Current: 709/201
• 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: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling

Abstract

A system and method for performing flexible workflow process compensation in a distributed workflow management system is described. A computer network includes a plurality of interconnected computers. Each computer includes a processor, memory and input/output facilities. The distributed workflow management system operates over the computer network. A plurality of resources perform the workflow process is performed with each resource operatively coupled to at least one of the computers. A process definition diagram includes computer-readable instructions stored in the memory of at least one of the computers and contains a role specification of process activities for performing the workflow process. A resource manager maps the role specification of process activities to at least one of the resources at runtime. A workflow process engine executes each process activity using the mapped resource and compensates a failed process activity responsive to the process definition diagram.

Claims

1. A method for performing selective compensation of a failed process activity in a distributed workflow management system, the distributed workflow management system operating over the computer network comprising a plurality of interconnected computers, each computer including a processor, memory and input/output facilities, the method comprising the steps of:

defining a process definition diagram comprising computer-readable instructions stored in the memory of at least one of the computers and containing a role specification of process activities for performing a workflow process, at least one of the process activities being formed into a lazy compensation group;

mapping the role specification of process activities using a resource manager to at least one of the resources at runtime;

executing each process activity with the mapped resource using a workflow process engine, including forming a pre image storing a portion of process relevant data provided to the process activity and a post image storing a portion of process relevant data generated by the process activity;

compensating the failed process activity for at least one such process activity preceding the failed process activity if the at least one such process activity is not in the lazy compensation group responsive to the process definition diagram; and

re-executing at least one further such process activity with the mapped resource using a workflow process engine if the at least one further such process activity is in the lazy compensation group responsive to the process definition diagram.

2. A method according to claim 1, wherein the step of compensating the failed process activity further comprises the step of comparing a current post image for a latest execution of the at least one further such process activity to an original post image for a first execution of the at least one further such process activity.

3. A method according to claim 2, further comprising the step of deferring the step of compensating for the at least one further such process activity if the current post image is the same as the original post image.

4. A method according to claim 3, wherein another such process activity contained in the lazy compensation group of the at least one such further process activity generates output parameters, the method further comprising the step of cancelling the step of deferring if the output parameters for the another such process activity are not restored during the step of compensating.

5. A method according to claim 4, further comprising the step of compensating each of the process activities contained in the lazy compensation group of the at least one further such process activity.

6. A method according to claim 4, further comprising the step of compensating each of the process activities contained in the lazy compensation group of the at least one further such process activity.

7. A method according to claim 3, wherein another such process activity contained in the lazy compensation group of the at least one such further process activity receives input parameters, the method further comprising the step of cancelling the step of deferring if the input parameters for the another such process activity during re-execution do not match the input parameters for the another such process activity during original execution.

8. A method according to claim 2, further comprising the step of compensating each of the process activities contained in the lazy compensation group of the at least one further such process activity.

9. A method according to claim 3, wherein another such process activity contained in the lazy compensation group of the at least one such further process activity is executed by the one such further process activity, the method further comprising the step of cancelling the step of deferring if the lazy compensation group of the at least one such further process activity contains other process activities having a deferred compensation.

10. A method according to claim 2, further comprising the step of deferring the step of compensating each of the process activities contained in the lazy compensation group of the at least one further such process activity if the current post image is not the same as the original post image.

11. A method according to claim 2, further comprising the steps of:

cancelling the step of deferring if the lazy compensation group of the at least one such further process activity contains other process activities that have not been re-executed; and

compensating each of the other process activities contained in the lazy compensation group of the at least one further such process activity.

12. A method using a workflow process engine for flexibly compensating a workflow process in a distributed workflow management system, the distributed workflow management system comprising a computer network including a plurality of interconnected computers operatively coupled to a plurality of resources, each computer including a processor, memory and input/output facilities, the method comprising the steps of:

operating the workflow process engine on at least one of the computers, the workflow process comprising a sequence of the process activities represented by a process definition diagram stored in the memory of the computer;

controlling the execution of the workflow process responsive to the process definition diagram on each of the computers using the workflow process engine, including the steps of;

performing a forward activity for performing each process activity in the workflow process using one of the resources;

performing a cancel activity for cancelling such a process activity in the workflow process responsive to a cancel request; and

performing a compensation activity for compensating such a process activity in the workflow process that has failed.

13. A method according to claim 12, further comprising the step of specifying a compensation scope comprising the process activities in the process definition diagram already performed prior to the failed process activity.

14. A method according to claim 13, wherein the step of specifying a compensation scope further comprises the steps of:

designating a start compensation point corresponding to the node in the process definition diagram associated with the failed process activity;

designating an end compensation point corresponding to a node in the process definition diagram associated with such a process activity already performed prior to the failed process activity; and

selecting a scoping strategy for compensation.

15. A method according to claim 14, wherein the step of designating an end compensation point further comprises the step of selecting a specific save point using a name associated with the node in the process definition diagram associated with such a process activity already performed prior to the failed process activity.

16. A method according to claim 14, wherein the workflow process engine further comprises a system log file storing statuses of the process activities, the step of designating an end compensation point further comprises the step of searching the system log file for a last executed save point.

17. A method according to claim 14, wherein the step of designating an end compensation point further comprises the step of following arcs in the process definition diagram from the node in the process definition diagram associated with the failed process activity to a last predecessor save point.

18. A method according to claim 12, further comprising the steps of:

determining application-specific data for the failed process activity;

determining process-relevant data for the failed process activity; and

defining a minimal compensation scope as the compensation scope comprising the process activities in the process definition diagram depending upon the application-specific data or the process-relevant data.

19. A method according to claim 12, further comprising the steps of:

maintaining a system log file in the memory of at least one of the computers using the workflow process engine to store a history of statuses for the process activities; and

defining a temporal-based compensation scope as the compensation scope comprising the process activities in the system log file completed prior to the failed process activity.

20. A method according to claim 12, further comprising the steps of:

determining data for the failed process activity dependent upon execution of prior process activities;

determining process-specific data for the failed process activity;

determining process-relevant data for the failed process activity; and

defining a computable compensation scope as the compensation scope comprising the process activities in the process definition diagram depending upon the execution data, the process-specific data or the process-relevant data.

21. A method according to claim 12, further comprising the steps of:

maintaining a system log file in the memory of at least one of the computers using the workflow process engine to store a history of statuses for the process activities;

identifying the process activities in the process definition diagram reachable from another process activity; and

defining a structural-based compensation scope as the compensation scope by removing the reachable process activities in the process definition diagram not yet started based on the system log file.

22. A method according to claim 12, comprising the steps of:

identifying the process activities in the process definition diagram reachable from another process activity;

determining process-specific data for the failed process activity;

determining process-relevant data for the failed process activity; and

defining an extended compensation scope as the compensation scope comprising the reachable process activities in the process definition diagram and the process activities in the process definition diagram depending upon the process-specific data or the process-relevant data.

23. A method for performing lazy compensation in a distributed workflow management system, the distributed workflow management system operating over a computer network comprising a plurality of interconnected computers, each computer including a processor, memory and input/output facilities, the method comprising the steps of:

defining a process definition diagram comprising computer-readable instructions stored in the memory of at least one of the computers and containing a role specification of process activities for performing a workflow process;

mapping the role specification of process activities using a resource manager to at least one of the resources at runtime;

forming at least one lazy compensation group comprising at least one of the process activities which interacts with only those of the process activities within the at least one lazy compensation group and with only those of the process activities outside the at least one lazy compensation group via process relevant data and whose original input state is restored upon compensation;

executing each process activity with the mapped resource using a workflow process engine, including forming an original pre image of each such process activity before execution and an original post image of each such process activity after execution; and

compensating each such process activity that fails responsive to the process definition diagram by re-executing at least one of the process activities preceding the failed process activity.

24. A method according to claim 23, the step of compensating each such process activity further comprising the step performing lazy compensation for each such process activity in each such lazy compensation group, comprising the steps of:

forming a current post image of each such process activity in the lazy compensation group;

comparing the current post image for the process activity to the original post image for the process activity;

adding the process activity to a list of deferrals and updating process data for the process activity using the original pre image if the current post image for the process activity matches the original post image for the process activity; and

performing the step of compensating each such process activity in the list of deferrals if the current post image for the process activity does not match the original post image for the process activity.

25. A method according to claim 24, further comprising the step of performing extended re-execution of the at least one of the process activities preceding the failed process activity, comprising the steps of:

forming a current pre image of each such process activity in the lazy compensation group;

comparing the current pre image for the process activity to the original pre image for the process activity;

removing the process activity from the list of deferrals and updating the process data for the process activity using the original pre image if the current pre image for the process activity matches the original pre image for the process activity; and

performing the step of compensating each such process activity in the list of deferrals if the current pre image for the process activity does not match the original pre image for the process activity.

26. A method according to claim 25, further comprising the step of terminating the process activity, further comprising the step of performing the step of compensating each such process activity remaining in the list of deferrals.

27. A method for performing flexible compensation of a failed process activity in a distributed workflow management system, the distributed workflow management system operating over the computer network comprising a plurality of interconnected computers, each computer including a processor, memory and input/output facilities, the method comprising the steps of:

defining a process definition diagram comprising computer-readable instructions stored in the memory of at least one of the computers and containing a role specification of process activities for performing a workflow process;

mapping the role specification of process activities using a resource manager to at least one of the resources at runtime; and

executing each process activity with the mapped resource using a workflow process engine; and

compensating the failed process activity responsive to the process definition diagram by re-executing at least one of the process activities preceding the failed process activity.

28. A method according to claim 27, wherein the failed process activity comprises one such process activity which has failed or timed out, the step of compensating the failed process activity further comprising the steps of:

determining whether the role specification for the failed process activity requires compensation;

suspending the step of executing each process activity for all other such process activities which are currently active;

computing an end compensation point for the failed process activity;

computing a compensation scope for the failed process activity beginning from the end compensation point;

cancelling each such process activity associated with a node falling within the compensation scope and removing each associated process activity from the compensation scope;

repeating the step of compensating the failed process activity for each node falling within the compensation scope.

29. A method according to claim 28, further comprising the steps of:

designating the end compensation point as a new start compensation point upon a failure of the end compensation point; and

repeating the step of compensating for each node falling within the compensation scope the failed end compensation point.

30. A method according to claim 27, further comprising the steps of:

placing at least one of the process activities preceding the failed process activity into a compensation queue; and

deferring the re-execution of the at least one of the process activities if input data to the preceding process activity during the re-execution is substantially identical to input data during the step of executing each process activity.

31. A system for performing flexible workflow process compensation in a distributed workflow management system, comprising:

a computer network comprising a plurality of interconnected computers, each computer including a processor, memory and input/output facilities, the distributed workflow management system operating over the computer network;

a plurality of resources upon which the workflow process is performed, each resource being operatively coupled to at least one of the computers;

a process definition diagram comprising computer-readable instructions stored in the memory of at least one of the computers and containing a role specification of process activities for performing the workflow process;

a resource manager mapping the role specification of process activities to at least one of the resources at runtime; and

a workflow process engine executing each process activity using the mapped resource and compensating a failed process activity responsive to the process definition diagram.

32. A system according to claim 31, wherein the process definition diagram comprises a set of work nodes and rule nodes interconnected by unidirectional arcs, each work node being associated with one of the process activities, each rule node being associated with at least one route through the process definition diagram, the workflow process engine evaluating each work node to perform the associated process activity using one of the resources and each rule node to follow the associated at least one route through the process definition diagram.

33. A system according to claim 31, further comprising:

a system log file stored in the memory of at least one of the computers storing a history of statuses for the process activities; and

a compensation scope stored in the memory of at least one of the computers, the workflow process engine further compensating the failed process activity by searching the system log file to select completed process activities falling within the compensation scope and compensating and re-executing the selected process activities.