A few tips inspired by the exam guide outline. When looking at data processing, consider the frequency of results needed. For example, is the business purpose to produce a report once a month? If so you might have a very different design than if the report had to be generated daily or even on-demand. Another tip, what was the output side? What about the input side? Is the data usable as is or does it need to be transformed to be useful? If it doesn't need to be transformed, for example some CSV files, maybe you would leave them in Cloud Storage. On the other hand, if transformation is required, you can start to consider which service can and should do that. How frequently and how significantly will resource demands change and how is this going to be handled? For example, is it an annual report that determines whether the Bigtable cluster needs in other instance? Is this a manual process or is there some automated method of handling growth? Such as using a serverless service. Monitoring is covered in another module.