Usecase Smart Reruns

Use Case Name: Smart Reruns

Version:
Preconditions:

User has designed a workflow that specifies:

1. conditions under which a complete rerun of workflow should occur when user executes workflow (eg "if data sources I have marked have changed, a smart rerun cannot occur")
2. which intermediary derived data products to save for use by smart reruns

Summary:

Users want the ability to rerun certain workflows in a "smart" way, that is, without regenerating intermediary derived data products if the source data (and hence the derived data for non-stochastic models) have not changed. Some intermediate derived data products take a very long time to produce, hence it is beneficial to skip their generation when possible.
Triggers:

User executes a workflow that uses Smart Rerun features and all Smart Rerun criteria have been met.

Basic course of events:

Workflow developer designs a workflow that makes use of Smart Rerun components.

Upon execution of this workflow, Smart Rerun components check to see if their input data sources have changed (possibly the user can also specify conditions which must also be met). If these data sources have changed (or are being used for the first time), components (re)produce derived data products and save those that are marked as Smart Rerun derived data products. If data sources have not changed, and previously saved derived data products exist, a Smart Rerun can occur reusing these existing products. Special considerations for stochastic systems must be taken into account in the design of this usecase (do we allow Smart Reruns for certain stochastic models?).

Alternative paths:

• It is possible a user would also like the ability to do a Smart Rerun even if Smart Rerun criteria are not met, ie they might want to ignore the fact that data sources have changed and reuse existing derived data products to test something else.

Postconditions:

Intermediary derived data products produced by Smart Rerun actors are saved.

Business rules:
Notes:
Author and date: