I really like where you are going with this, but I'm having trouble visualizing what this looks like or maybe what an optimum scenario would be from a source perspective. For example, I'm building a new system on Google Cloud with Dgraph and I'm going to want to do some sophisticated data analysis from a lot of angles. How would something like that work? How is it mapped, and ingested? What about the latency on data movement if you have a very large volume of data? One of the amazing things with Sisense was how fast it was with such a disparate choice of data sources.