Persistence of experiments

[amaiberg]

Currently we have nothing connecting our execution with the database. We need to implement the logic that interfaces between the controller and the persistence layer.
Some need to be addressed to tackle this issue:

Are we storing intermediate CASes? If yes, how do we decide which ones?
Are we storing every experiment? How/where do we specify we want to use a database instead of an in-memory process? Should we implement a hybrid approach?

@CollinM