Replay to reconstruct program state

• US 9,170,915 B1
• Filed: 12/06/2011
• Issued: 10/27/2015
• Est. Priority Date: 12/06/2011
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Fateev, Maxim E.
• Primary Examiner(s): Kang, Insun
• Application Number:
  US13/311,960
• Time in Patent Office: 1,421 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 11/3414 Workload generation, e.g. scripts, playback
• CPC Class Codes: G06F 9/3836 Instruction issuing, e.g. dynamic instruction scheduling or out of order instruction execution
• CPC Class Codes: G06F 9/3871 Asynchronous instruction pipeline, e.g. using handshake signals between stages
• 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

The state of a workflow application in a distributed computing environment can be reconstructed by replaying previously executed portions of the workflow application. Embodiments maintain non-serialized event data relating to workflow history and use the data to determine subsequent actions to take in order to advance the workflow. In some embodiments, workflow instances can be serialized in order to create a checkpoint.

Claims

1. A system for reconstructing the state of an application, the system comprising one or more hardware computing devices configured to execute:

one or more activity modules configured to execute an action of a workflow definition; and

a workflow decision module configured to;

receive a first workflow history record from a data store, wherein the first workflow history record comprises data related to an event of a workflow application execution;

determine a first activity module to execute based at least in part on the first workflow history record, wherein determining the first activity module alters the program state of the workflow decision module;

create a first command to execute the first activity module;

issue the first command;

determine whether the first command completed successfully;

if the first command completes successfully;

replay the determination of the first activity module to execute; and

recreate the first command;

receive a second workflow history record from the data store; and

in response to receiving the second workflow history record;

disregard the recreated first command, wherein disregarding the recreated first command comprises preventing re-execution of the first command;

determine a second activity module to execute based at least in part on the altered program state;

create a second command to execute the second activity module; and

issue the second command.

2. The system of claim 1, wherein disregarding the recreated first command comprises deleting the recreated first command.

3. The system of claim 1, wherein replaying the determination of the first activity module to execute comprises:

receiving again the first workflow history record from the data store; and

determining again the first activity module to execute based at least in part on the first workflow history record, wherein determining again the first activity alters the program state of the workflow decision module.

4. The system of claim 1, further comprising a plurality of servers configured to process a purchase.

5. The system of claim 4, wherein the plurality of servers comprises a first server and a second server, and wherein the first server receives a workflow history record representing an action taken by the second server.

6. A computer-implemented method of reconstructing the state of an application, the method comprising:

receiving, from an electronic data store, a first event record comprising data related to an event of a workflow application at a previous time;

executing a first portion of the workflow application, wherein the execution is based at least in part on the first event record received from the electronic data store and wherein the execution alters a state of the workflow application;

creating a first command, wherein the first command is based at least in part on the execution of the first portion of the workflow application, and wherein completion of the first command advances the workflow application;

issuing the first command;

determining whether the first command completed successfully;

if the first command completes successfully;

replaying the execution of the first portion of the workflow application; and

recreating the first command without re-execution of the first portion of the workflow application;

receiving, from the electronic data store, a second event record comprising data related to the completion of the first command;

executing a second portion of the workflow application, wherein the execution of the second portion is based at least in part on the altered state of the workflow application;

creating a second command, wherein the second command is based at least in part on the execution of the second portion, and wherein completion of the second command further advances the workflow application; and

issuing the second command;

wherein said replaying is implemented by a computer system comprising computer hardware configured with specific executable instructions.

7. The computer-implemented method of claim 6, wherein replaying the execution of the first portion of the workflow application comprises:

receiving again the first event record from the data store; and

executing again the first portion of the workflow application, wherein the executing again is based at least in part on the first event record received again from the electronic data store, and wherein execution alters again a state of the workflow application.

8. The computer-implemented method of claim 6, wherein the computer system comprises a plurality of servers.

9. The computer-implemented method of claim 8, wherein the plurality of servers comprises a first server and a second server, and wherein the first server receives the first event record comprising data related to an event of a workflow application at a previous time, the event occurring on the second server.

10. The computer-implemented method of claim 6, further comprising disregarding the recreated first command.

11. The computer-implemented method of claim 6, further comprising deleting the recreated first command.

12. The computer-implemented method of claim 6, further comprising saving, to the electronic data store, a third event record comprising data related to the second command.

13. The computer-implemented method of claim 12, wherein data of the third event relates to one of the scheduling of the second command, the execution of the second command, or the completion of the second command.

14. The computer-implemented method of claim 6, wherein issuing the first command or second command comprises causing the first command or second command to be scheduled for execution.

15. The computer-implemented method of claim 6, wherein the first command or the second command comprises processing of payment information.

16. The computer-implemented method of claim 6, wherein the first command or the second command comprises executing a workflow application.

17. Non-transitory physical computer storage comprising instructions stored therein for implementing, in one or more processors, operations for reconstructing the state of an application, the operations comprising:

receiving, from an electronic data store, a first event record comprising data related to a command initiated by an application at a previous time;

determining a first activity of the application, wherein the determining is based at least in part on the first event record received from the electronic data store, and wherein the determining alters a state of the application;

creating a first command to execute the first activity, wherein completion of the first command allows advancement of the application;

causing the first command to be scheduled for execution;

determining whether the first command completed successfully; and

if the first command completes successfully;

replaying the determining of the first activity; and

recreating the first command without causing re-execution of the first activity;

receiving, from the electronic data store, a second event record comprising data related to the completion of the first command;

determining a second activity of the application, wherein the determination of the second activity is based at least in part on the altered state of the workflow application;

creating a second command to execute the second activity, wherein completion of the second command allows further advancement of the application; and

issuing the second command.

18. The non-transitory physical computer storage of claim 17, the operations further comprising saving, to the electronic data store, a third event record comprising data related to the second command.

19. The non-transitory physical computer storage of claim 17, the operations further comprising disregarding the recreated first command.

20. The non-transitory physical computer storage of claim 17, the operations further comprising deleting the recreated first command.

21. The non-transitory physical computer storage of claim 17, wherein at least one of the first activity or the second activity is execution of a workflow application.

22. The non-transitory physical computer storage of claim 17, wherein at least one of the first activity or the second activity is related to processing purchases.

23. The non-transitory physical computer storage of claim 17, wherein the application is a distributed application executing on a plurality of servers.

24. The non-transitory physical computer storage of claim 23, wherein the application is a distributed application, wherein the plurality of servers comprises a first server and a second server, and wherein the first server receives the first event record comprising data related to an event of the distributed application, the event occurring at a previous time on the second server.

25. The non-transitory physical computer storage of claim 17, the operations further comprising serializing the altered state.

26. The non-transitory physical computer storage of claim 25, the operations further comprising deserializing the altered state.