One of the folks working at Google Cloud here (and was on the BigQuery team until very recently).
While the blog post is a great POV from someone who's very familiar with Redshift, I'd like to offer some details around BigQuery that are easy to overlook at [0]. It IS a fundamentally different technology, so many assumptions must be revisited.
For example, in my opinion any and all benchmarks for analytic workloads should include considerations for encryption, truly ad-hoc analytics (not optimized for sort keys), and concurrency. Likewise, any and all cost estimates should include a real measure of volatility, as analytics workloads tend to be incredibly volatile.
And ICYMI: BigQuery received a major upgrade 3 weeks ago [1]
While the blog post is a great POV from someone who's very familiar with Redshift, I'd like to offer some details around BigQuery that are easy to overlook at [0]. It IS a fundamentally different technology, so many assumptions must be revisited.
For example, in my opinion any and all benchmarks for analytic workloads should include considerations for encryption, truly ad-hoc analytics (not optimized for sort keys), and concurrency. Likewise, any and all cost estimates should include a real measure of volatility, as analytics workloads tend to be incredibly volatile.
And ICYMI: BigQuery received a major upgrade 3 weeks ago [1]
[0] https://medium.com/@thetinot/15-awesome-things-you-probably-...
[1] https://cloud.google.com/blog/big-data/2016/09/bigquery-intr...