How can confidence be instilled in results whose computation involves applying a series of processing steps in a Big Data environment?