This article describes how eCatt hangs together, in a nutshell of course...
eCatt test scripts are composed of test scripts, system settings (landscape settings), variants (data to be assigned to a test script) and test configurations (where all components are linked together). From the development system an eCatt can be run on any system that is set up as part of the eCatt landscape. Thus triggered from the development system, scripts can perform a variety of actions on the test or QA system.
An elaborate log is maintained on the development system (hence detailed information on the script that was executed is available on your development platform).
How ? By logging on to these systems and executing the scripts on those systems. Thus if your script performs a bit of Abap coding, the coding is placed on the target system before it is executed.
Complex ? Actually, it is. Setting up an eCatt and keeping it up to date is quite a time consumer. But having your test cases covered in eCatt scripts can also be a time-saver....