Job Scheduling - Database Design

[brucevde]

I am trying to automate the scheduling of "Jobs" for multiple applications and am having trouble with the database design. Most of the "Jobs" will execute 1 or more "Steps". A Step is usually a Crystal Report but can also be a custom process (and allow for future expansion). "Jobs" can have multiple "Schedules". I used SQL Server's SysJob tables as a starting point and got this far.

Applications have
-> Jobs have
--> Schedules
--> Steps have
---> Input Parameters (typical Crystal Report parameters)
---> Output Locations (printers, emails, files)

The problem I am having is designing the Input/Output tables because even though Steps are the same for each Schedule, the Input and Output could be different. A single Step could be executed X number of times within a Schedule because there could be multiple Inputs for the Step. There could also be multiple Outputs for each Input. Sometimes there are no Input Parameters but again 1 or more Outputs. Obviously a Report has multiple parameters (Crystal allows a Parameter to have "multiple values").

Maybe I am overthinking the design and should just turn a Step into a "single execution of a report". So instead of a Job having 1 Step with 10 different sets of Input Parameters I would have 10 Steps with 1 Input parameter and multiple Output Locations.

I need some ideas on how to handle several tables all within a many to many to many situation (entity relationship tables aside because I don't think they will work in this case)?