Method for fault handling in a co-operative workflow environment

• US 20070101179A1
• Filed: 10/27/2005
• Published: 05/03/2007
• Est. Priority Date: 10/27/2005
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Kankar, Pankaj, Mann, Vijay, Chandra, Sunil, Chafle, Girish
• Granted Patent: US 7,475,275 B2
• Time in Patent Office: Days
• US Class Current: 714/4
• CPC Class Codes: G06F 11/0709 in a distributed system consisting of a plurality of standalone computer nodes, e.g. clusters, client-server systems
• CPC Class Codes: G06F 11/0784 Routing of error reports, e.g. with a specific transmission path or data flow
• CPC Class Codes: G06F 11/0793 Remedial or corrective actions recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40
• CPC Class Codes: G06F 11/1474 in transactions G06F16/20 takes precedence

Abstract

Embodiments herein provide a fault-handling scheme based on forward recovery for cooperative workflow environments. The fault handling scheme relies on the correct placement of transaction scopes and their associated fault and compensation handlers for maintaining correct application semantics, a fault propagation scheme for forwarding faults to a workflow component that has the corresponding fault handler, and a distributed mechanism for collecting data of completed workflow components to facilitate recovering from faults. The fault handling scheme makes use of control flow messages to facilitate compensation of nested transaction scopes (residing in different components). The workflow components are also modified with additional code for aiding with fault propagation and fault recovery.

Claims

1. A method for fault handling in a co-operative workflow environment comprising:

employing a placement strategy for a transaction scope and associated fault handlers and compensation handlers across different workflow components;

empoying a fault propagation scheme for forwarding a fault occurring in any of said workflow components to a fault handler;

employing a distributed data collection mechanism to facilitate recovery from a fault occurring in any said workflow component;

employing a fault recovery mechanism using compensation,

wherein said transaction scope defines a transaction boundary and provides a fault handling and a fault recovery context for a set of activities and said transaction scope may be nested inside another transaction scope to create outer and inner scopes, and

wherein said cooperative workflow environment comprises a plurality of said workflow components deployed at different locations in a distributed infrastructure and working co-operatively with other components to complete an overall global task.

2. The method according to claim 1, wherein a plurality of said workflow components comprise said transaction scope, and wherein said method further comprise identifying a workflow component that has an end of said transaction scope as a root workflow component for a given transaction scope and identifying other workflow components as non-root workflow components.

3. The method according to claim 1, wherein, when employing said placement strategy, end of said transaction scope is placed in the said root workflow component,

wherein said fault handlers for a given said transaction scope reside in said root workflow component, and

wherein said compensation handlers for a given said transaction scope reside in said root workflow component.

4. The method according to claim 1, further comprising augmenting said workflow components with additional fault handling code to implement said fault propagation scheme according to which said workflow component forwards a fault, that has either occurred within said workflow component or has been received from another said workflow component as part of said fault propagation scheme, to one of a following workflow said components whichever comes first in the control flow path:

the root workflow component of the given said transaction scope;

next join workflow component; and

next fork workflow component,

wherein said join workflow component comprises a workflow component that has more than one incoming links and said fork workflow component comprises a workflow component that has more than one outgoing link.

5. The method according to claim 1, further comprising sending, using a root workflow component, a completion control message in case of normal completion.

6. The method according to claim 5, further comprising exiting normally by said workflow components upon receipt of said completion control message.

7. The method according to claim 1, further comprising initiating said distributed data collection mechanism by a root workflow component by sending a data collection control message once said fault occurring in any said workflow component in a scope has been propagated to a root workflow component of that said scope.

8. The method according to claim 7, further comprising appending data variables of said workflow components to said data collection control message and sending said data collection control message to said workflow components which are next in sequence in the control flow upon receipt of said data collection control message, and performing one of the following steps:

exiting if said workflow components are non-root workflow components, or

waiting for one of a compensate control message or a no compensate control message if said workflow components are said root workflow components.

9. The method according to claim 1, wherein, when employing said fault recovery mechanism using said compensation, the root workflow component of a scope sends a compensate control message to root workflow components of all said inner scopes for which fault recovery has to be initiated, and said root workflow component of said scope sends a no compensate control message to root workflow components of all said inner scopes for which no said fault recovery has to be done.

10. The fault recovery mechanism according to claim 9, further comprising exiting normally by said root workflow components upon receipt of said no compensate control message, and initiating said fault recovery using said compensation handlers upon receipt of said compensate control message from said root workflow component of an immediate outer scope.

11. A system for fault handling in a co-operative workflow environment comprising multiple workflow engines that host workflow components wherein said host workflow components are adapted to implement:

a placement strategy for a transaction scope and associated fault handlers and compensation handlers across different said workflow components;

a fault propagation scheme for forwarding a fault occurring in any said workflow component to a fault handler;

a distributed data collection mechanism to facilitate recovery from a fault occurring in any said workflow component;

a fault recovery mechanism using compensation,

wherein said transaction scope defines a transaction boundary and provides a fault handling and a fault recovery context for a set of activities and said transaction scope may be nested inside another said transaction scope resulting in outer and inner scopes, and

wherein said cooperative workflow environment comprises a plurality of said workflow components deployed at different locations in a distributed infrastructure and working co-operatively with other components to complete an overall global task.

12. The system according to claim 11, further comprising a plurality of said workflow components that comprise said transaction scope, out of which a workflow component that has an end of said transaction scope is identified as a root workflow component for a given said transaction scope and other said workflow components are identified as non-root workflow components.

13. The system according to claim 11, wherein said workflow components implement said placement strategy such that an end of a said transaction scope resides in a root workflow component;

said fault handlers for a given said transaction scope reside in said root workflow component; and

said compensation handlers for a given said transaction scope reside in said root workflow component.

14. The system according to claim 11, wherein said workflow components are augmented with additional fault handling code to implement said fault propagation scheme, wherein said workflow component is adapted to forward a fault, that has either occurred within said workflow component or has been received from another said workflow component as part of said fault propagation scheme, to one of a following said workflow components whichever comes first in a control flow path:

the root workflow component of a given said transaction scope;

next join workflow component; and

next fork workflow component,

wherein said join workflow component comprises a workflow component that has more than one incoming links and said fork workflow component comprises a workflow component that has more than one outgoing links.

15. The system according to claim 11, further comprising root workflow components that send a completion control message in case of normal completion.

16. The system according to claim 11, further comprising workflow components that implement said distributed data collection mechanism, wherein said distributed data collection mechanism is initiated by the root workflow component by sending a data collection control message once a fault occurring in any said workflow component in a scope has been propagated to said root workflow component of said scope.

17. The system according to claim 11, wherein said workflow components are adapted to perform one of:

exiting normally upon receipt of a completion control message; and

implementing said distributed data collection mechanism by appending data variables to a data collection message and sending said message to said workflow components which are next in sequence in a control flow, wherein, upon receipt of a data collection message, said workflow components are adapted to perform one of;

exiting if said workflow components are non-root workflow components;

or

waiting for a compensate control message or a no compensate control message if said workflow components are root workflow components.

18. The system according to claim 11, wherein workflow components implement said fault recovery mechanism using said compensation, wherein the root workflow component of a scope is adapted to send a compensate control message to root workflow components of all said inner scopes for which compensation has to be initiated, and wherein said root workflow component of said scope is adapted to send a no compensate control message to said root workflow components of all said inner scopes for which no said fault recovery has to be done.

19. The system according to claim 18, wherein said root workflow components are adapted to exit normally upon receipt of said no compensate control message, and wherein said root workflow components are adapted to initiate said fault recovery using said compensation handlers upon receipt of said compensate control message from a root workflow component of an immediate outer scope.

20. A computer program product comprising a computer usable medium having a computer readable program, wherein the computer readable program, when executed on a computer, causes the computer to perform a method for fault handling in a co-operative workflow environment, comprising:

employing a placement strategy for a transaction scope and associated fault handlers and compensation handlers across different workflow components;

employing a fault propagation scheme for forwarding a fault occurring in any of said workflow components to a fault handler;

employing a distributed data collection mechanism to facilitate recovery from a fault occurring in any said workflow component;

employing a fault recovery mechanism using compensation,

wherein said transaction scope defines a transaction boundary and provides a fault handling and fault recovery context for a set of activities and a said transaction scope may be nested inside another transaction scope resulting in outer and inner scopes, and

wherein said cooperative workflow environment comprises a plurality of said workflow components deployed at different locations in a distributed infrastructure and working co-operatively with other components to complete an overall global task.