Method of recovering from software failures using replanning

• US 8,239,828 B2
• Filed: 01/08/2008
• Issued: 08/07/2012
• Est. Priority Date: 01/08/2008
• Status: Active Grant

• Current Assignee: Globalfoundries US Incorporated (GlobalFoundries, Inc.)
• Original Assignee: International Business Machines Corporation
• Inventors: Liu, Zhen, Riabov, Anton V.
• Primary Examiner(s): Das, Chameli
• Application Number:
  US11/971,056
• Publication Number:
  US 20090177910A1
• Time in Patent Office: 1,673 Days
• Field of Search: 717/120, 717/124
• US Class Current: 717/120
• CPC Class Codes: G06F 11/3664 Environments for testing or debugging software
• CPC Class Codes: G06F 11/3672 Test management

Abstract

A method for recovering from software failures, includes: receiving failure information that identifies a failing component of a first processing graph; modifying a planning domain that includes a plurality of component descriptions according to the failure information; and composing a second processing graph by using the modified planning domain so that the second processing graph does not include the failing component.

Claims

1. A method for recovering from software failures, comprising:

receiving failure information that identifies a failing component of a first processing graph;

modifying a planning domain that includes a plurality of component descriptions according to the failure information;

composing using a processor of a computer, a second processing graph by using machine code executable by the computer to process the modified planning domain,

wherein the method further comprises;

deploying the second processing graph in an execution environment;

executing the second processing graph; and

providing a result of the executed second processing graph if a failing component is not identified during the execution of the second processing graph.

2. The method of claim 1, wherein if a component of the first processing graph did not fail and it produced an output, the method further comprises:

composing the second processing graph by using the output.

3. The method of claim 1, wherein after receiving the failure information and prior to composing the second processing graph, the method further comprises:

determining an end state of the executed first processing graph;

generating a planning task based on the determined end state, consistent state, recovery action and state modification descriptions for each component in the first processing graph;

generating a recovery plan according to the planning task; and

performing a recovery action included in the recovery plan.

4. The method of claim 1, wherein the first processing graph is composed for a request.

5. The method of claim 4, wherein the request is a set of tags.

6. The method of claim 4, wherein the second processing graph is composed by generating and solving a planning task, the planning task including the modified planning domain and the request, the second processing graph satisfying the request.

7. The method of claim 6, wherein the planning task is expressed in Stream Processing Planning Language (SPPL).

8. A method for recovering from software failures, comprising:

receiving a set of tags specifying a user request;

generating, using a processor of a computer, a processing graph by using machine code executable by the computer to process a domain description that includes a plurality of component descriptions, the processing graph including a plurality of components that satisfy the request;

deploying the processing graph in an execution environment;

storing the user request; and

invoking the processing graph to retrieve a response, wherein if a component of the processing graph fails;

obtaining failure information identifying the failed component;

removing the failed component description from the domain description;

generating an alternative processing graph by using the modified domain description, the alternative processing graph including a plurality of components that satisfy the request;

deploying the alternative processing graph in the execution environment;

invoking the alternative processing graph; and

providing a result of the executed alternative processing graph if a failing component is not identified during the execution of the alternative processing graph.

9. A non-transitory computer readable medium tangibly embodying a program of instructions executable by a processor to perform method steps for recovering from software failures, the method steps comprising:

receiving failure information that identifies a failing component of a first processing graph;

modifying a planning domain that includes a plurality of component descriptions according to the failure information;

composing a second processing graph by using the modified planning domain;

wherein modifying the planning domain includes removing the failing component'"'"'s description from the planning domain,

wherein the method steps further comprise;

deploying the second processing graph in an execution environment;

executing the second processing graph; and

providing a result of the executed second processing graph if a failing component is not identified during the execution of the second processing graph.

10. The computer readable medium of claim 9, wherein if a component of the first processing graph did not fail and it produced an output, the method steps further comprise:

composing the second processing graph by using the output.

11. The computer readable medium of claim 9, wherein after receiving the failure information and prior to composing the second processing graph, the method steps further comprise:

determining an end state of the executed first processing graph;

generating a planning task based on the determined end state, consistent state, recovery action and state modification descriptions for each component in the first processing graph;

generating a recovery plan according to the planning task; and

performing a recovery action included in the recovery plan.

12. The computer readable medium of claim 9, wherein the first processing graph is composed for a request.

13. The computer readable medium of claim 12, wherein the request is a set of tags.

14. The computer readable medium of claim 12, wherein the second processing graph is composed by generating and solving a planning task, the planning task including the modified planning domain and the request, the second processing graph satisfying the request.

15. The computer readable medium of claim 14, wherein the planning task is expressed in Stream Processing Planning Language (SPPL).

16. A non-transitory computer readable medium tangibly embodying a program of instructions executable by a processor to perform method steps for recovering from software failures, the method steps comprising:

receiving a set of tags specifying a user request;

generating a processing graph by using a domain description that includes a plurality of component descriptions, the processing graph including a plurality of components that satisfy the request;

deploying the processing graph in an execution environment;

storing the user request;

invoking the processing graph to retrieve a response, wherein if a component of the processing graph fails;

obtaining failure information identifying the failed component;

removing the failed component description from the domain description;

generating an alternative processing graph by using the modified domain description, the alternative processing graph including a plurality of components that satisfy the request;

deploying the alternative processing graph in the execution environment;

invoking the alternative processing graph; and

providing a result of the executed alternative processing graph if a failing component is not identified during the execution of the alternative processing graph.