The UNICORE Workflow engine is used to control the execution of multiple UNICORE jobs at one or multiple execution sites, dealing with dependencies between jobs and handling any required data movement.

The workflow engine offers a wide range of control constructs and other features: groups, if-else blocks, while, repeat and for-each loops are supported, which can be nested to any depth.  Workflow variables can be defined, modified using scripts and used in jobs, if-else conditions or for loop control. Workflows can be halted at used-defined points and continued later, allowing to integrate user-made decisions and user modifications of workflow variables.

UNICORE stores the input and output data for a single workflow execution in a single storage location, simplifying data management and helping users keep track of their data.

Full-featured graphical workflow editing, workflow submission and monitoring is part of the UNICORE Rich Client (URC). Workflows are written in an XML dialect and can also be submitted and monitored via the UNICORE command line client or the REST API.

More information can be found at https://www.unicore.eu/documentation/manuals/unicore/files/workflow/workflow-manual.html