Dehydration:

When an orchestration has been idle for a while, the orchestration engine will save the state information of the instance and free up memory resources.

Rehydration:

When a message is received, or else when a timeout has expired, the orchestration engine can be automatically triggered to rehydrate the instance – it is at this point that the orchestration engine loads the saved instance of the orchestration into memory, restores the state, and runs its from the point it left off.