Benchmark for simulation with/without caching

[eileen-kuehn]

For presentations by @tfesenbecker we need some plots visualising the difference regarding performance when caching is enabled/disabled. The performance might be twofold: 1) performance of the simulation content, aka what is the internal duration of simulation, and 2) time to run the actual simulation.

For this, we should set up a demonstration that defines an input dataset defining the jobs, pools, and input files. The dataset should have the characteristic that caching actually improves performance.

Tasks for this ticket:

• definition of dataset including job input, pool input, and input file input,
• script for plotting the monitoring data that are produced by the simulation, and
• documentation of the workflow (what is the intention having this demonstration, what is expected, which parameters to set for simulation, etc.).