I read this as we have so many query / data platforms so we're going to build another query standard to standardize all the query languages. Just another flavor of the standards problem.

I see F1 as: ``what is the largest SQL subset we can support without compromising scalability.''

30000 queries per day, and 550000 queries per second seem like two radically different things.

This is for offline data analysis, more like Google's bigquery / Dremel. They are different things.

What's relation with BigQuery?

Bq is based on Dremel, which is for data warehouse, not facing user traffic. You can find the original paper for Dremel easily.