Instead of storing the person object in the Execution table and having the entire object cascade, we should just reference the person ID (similar to workflowId). If we wanted to resolve that later, we could then do personDao.find(id).