Generating an idempotent workflow

1. A method for generating an idempotent workflow comprising:

detecting distributed code;

identifying a boundary function in the distributed code that executes at least a first portion of the distributed code on a first computing device and at least a second portion of the distributed code on a second computing device different from the first computing device;

separating the boundary function into a transition function that identifies a transition from the first computing device to the second computing device, a non-idempotent function that modifies parameter values, and a reversion function that terminates a session;

transforming the non-idempotent function into an idempotent function by including code in the non-idempotent function so that the non-idempotent function is executed once;

generating a flag corresponding to the non-idempotent function, wherein the flag indicates if the non-idempotent function has been executed; and

combining the transition function, the idempotent function, and the reversion function to form the idempotent workflow;

wherein the boundary function comprises a transition from a local process to a remote process; and

wherein the idempotent function comprises code to produce a single output value from multiple executions of the idempotent function.

2. The method of claim 1 comprising storing the transition function, the non-idempotent function, and the reversion function in a persistent workflow database.

3. The method of claim 2 comprising storing an input array with the transition function, the non-idempotent function, or the reversion function in the persistent workflow database.

4. The method of claim 3 , wherein the input array comprises a union of output values corresponding to each previously executed function.

5. The method of claim 2 , wherein the persistent workflow database comprises a field that represents an error generated while executing the transition function, the non-idempotent function, or the reversion function.

6. The method of claim 2 , wherein the persistent workflow database comprises a number of instructions value representing the number of instructions in the transition function, the non-idempotent function, or the reversion function.

7. The method of claim 1 , comprising returning an output value from a previous execution of the transition function, the non-idempotent function, or the reversion function if a second flag indicates that the transition function, the non-idempotent function, or the reversion function has been previously executed.

8. One or more computer-readable storage media for generating an idempotent workflow comprising a plurality of instructions that, when executed by a processor, cause the processor to:

detect distributed code;

identify a boundary function in the distributed code that executes at least a first portion of the distributed code on a first computing device and at least a second portion of the distributed code on a second computing device different from the first computing device;

separate the boundary function into additional functions including a non-idempotent function that returns a number of output values from a number of processes executed remotely;

transform the non-idempotent function into an idempotent function by generating a flag indicating whether the non-idempotent function has been previously executed; and

combine the idempotent function with the additional functions to form the idempotent workflow;

wherein the boundary function comprises a transition from a local process to a remote process; and

wherein the idempotent function comprises code to produce a single output value from multiple executions of the idempotent function.

9. The one or more computer-readable storage media of claim 8 , wherein the instructions cause the processor to store a transition function, the non-idempotent function, and a reversion function in a persistent workflow database.

10. The one or more computer-readable storage media of claim 9 , wherein the instructions cause the processor to store an input array with the transition function, the non-idempotent function, and the reversion function in the persistent workflow database.

11. The one or more computer-readable storage media of claim 10 , wherein the input array comprises a union of output values corresponding to each previously executed function.

12. A system for generating an idempotent workflow comprising:

a processor to execute processor executable code;

a storage device that stores processor executable code, wherein the processor executable code, when executed by the processor, causes the processor to:

detect distributed code;

identify a boundary function in the distributed code, the boundary function comprising a transition from a local process to a remote process;

separate the boundary function into a transition function that identifies a transition from the first computing device to the second computing device, a non-idempotent function that modifies parameter values, and a reversion function that terminates a session;

transform the non-idempotent function into an idempotent function including a flag in the non-idempotent function to indicate that the non-idempotent function has been executed only once; and

combine the transition function, the idempotent function, and the reversion function to form the idempotent workflow;

wherein the boundary function comprises a transition from a local process to a remote process; and

wherein the idempotent function comprises code to produce a single output value from multiple executions of the idempotent function.

13. The system of claim 12 , wherein the processor executable code, when executed by the processor, causes the processor to store the transition function, the non-idempotent function, and the reversion function in a persistent workflow database.

14. The system of claim 13 , wherein the processor executable code, when executed by the processor, causes the processor to store an input array with the transition function, the non-idempotent function, and the reversion function in the persistent workflow database.

15. The system of claim 14 , wherein the input array comprises a union of output values corresponding to each previously executed function.