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

• US 6,041,306 A
• Filed: 03/21/1997
• Issued: 03/21/2000
• Est. Priority Date: 12/05/1996
• Status: Expired due to Term

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Hewlett-Packard Company (HP Inc.)
• Inventors: Pfeifer, Clemens, Sheard, Nicolas, Davis, James W., Shan, Ming-Chien, Du, Weimin
• Primary Examiner(s): Swann, Tod R.
• Assistant Examiner(s): Smithers, Matthew
• Application Number:
  US08/821,940
• Time in Patent Office: 1,096 Days
• Field of Search: 705/8, 705/7, 705/9, 395/200.31, 395/200.32, 395/200.33, 395/182.13
• US Class Current: 705/7.26
• CPC Class Codes: G06F 9/4843 by program, e.g. task dispatcher, supervisor, operating system
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
• CPC Class Codes: G06Q 10/06316 Sequencing of tasks or work

Abstract

A system and method for performing flexible workflow process execution in a distributed workflow management system is described. The distributed workflow management system is formed by a computer network comprising a plurality of computers. Each computer has a processor, memory and input/output facilities. A workflow process management system operates on one or more of the computers to control the computer network in executing the workflow process. The workflow process includes at least one sequence of multiple actions. A plurality of resources is coupled to respective ones of the computers to carry out the multiple actions. A plurality of state machines are stored as computer-operable code in at least one memory and include a plurality of states interconnected by arcs logically forming a directed graph. The workflow management system further includes logic for instantiating each action with one state and logic for executing the logical sequence of the action as state transitions in each state machine.

Claims

1. A system for performing flexible workflow process in a distributed workflow management system that includes multiple computers, comprising:

a workflow process management system operating on at least one of the computers to control execution of the workflow process which includes process instances;

a plurality of resources coupled to respective ones of the computers to carry out the process instances;

a plurality of state machines that comprise

a process instance state machine that includes a plurality of states including (1) a compensation state that allows a failed process instance at a work node to be compensated such that resumption of execution of the failed process instance can be from a specified end compensation point, and (2) a suspended compensation state that allows the process instance in the compensation state to move to the suspended compensation state when the process instance becomes suspended;

a work node instance state machine for a work node instance manager that manages work nodes of the workflow process, wherein the work node instance state machine includes a plurality of states including the compensation state and the suspended compensation state; and

a rule node instance state machine for a rule node instance manager that manages rule nodes of the workflow process.

2. A system according to claim 1, wherein the work node instance state machine, further comprises:

an initial state logically coupled to an active state via a traverse inward arc and to a suspended initial state via a suspend arc;

the suspended initial state logically coupled to the initial state via a resume arc;

the active state logically coupled to a completed state via a complete arc and a terminate arc, and to the suspended active state via a suspend arc;

the suspended active state logically coupled to the active state via a resume arc and to the completed state via a terminate arc;

the completed state logically coupled to the compensation state via a compensate node arc;

the compensation state logically coupled to the completed state via a terminate arc, to a suspended compensation state via a suspend node and to the initial state via a complete arc; and

the suspended compensation state logically coupled to the compensation state via a resume state to the completed state via a terminate arc.

3. A system according to claim 2, further comprising:

the active state logically coupled to the initial state via a reset arc;

the suspended active state logically coupled to the suspended initial state via a reset arc;

the suspended compensation state logically coupled to the suspended initial state via a reset arc;

the completed state logically coupled to the initial state via a reset arc; and

the compensation state logically coupled to the initial state via a reset arc.

4. A system according to claim 1, wherein at least one of the state machines comprises an activity instance state machine, comprising:

a start business object (BO) state logically coupled to a quiescent state via a BO failed arc and a BO completed arc and to a get resource manager (RM) state via a delegate BO arc;

the quiescent state logically coupled to the get RM state via a start activity arc;

the get RM state logically coupled to the quiescent state via a get RM failed arc and to a get BO state via a get RM okay arc; and

the get business object state logically coupled to the quiescent state via a get BO failed arc, to the start BO state via a get BO okay arc and to itself via a redirect RM arc.

5. A system according to claim 1, wherein the rule node instance state machine, further comprises:

an initial state logically coupled to an active state via a traverse inward arc;

the active state logically coupled to the initial state via a terminate arc, to a suspended state via a suspend arc and to itself via a process event arc; and

the suspended state logically coupled to the initial state via a terminate arc and to the active state via a resume arc.

6. The system of claim 1, wherein the process instance state machine is started and managed by a process instance manager.

7. The system of claim 6, wherein when the process instance state machine for the process instance is in the compensation state, the process instance manager

saves all currently active work nodes of the process instance;

sends a message and relevant information to a compensation manager.