This is a really well-written and informative blog post, and it's nice to see a new storage engine option for MySQL.
Very happy to see continuing innovation on MySQL and contribution to open source from Facebook. I wonder if there is potential to get more than 2x less space usage than InnoDB + compression. It seems like a no brainer to build a new storage engine for 10x but it's a lot of work to deploy and "harden" this for 2x. EDIT: I'm guessing the value is more about solving the write amplification problem, where it looks like it's about 10-12x reduced over Inno+compression. Nice work! Replacing petabytes of SSDs isn't cheap.
As someone developing with rocksdb currently, this is very exciting! Can't wait to give it a shot.