InnoDB or the other default MySQL storage engine MyISAM provide reasonable performance on analytical query workloads up to 100,000 rows or tables under a million rows. Performance is harder to tune and maintain beyond that, Thompson wrote in a blog post.
ColumnStore is suitable for reporting or analysis on much larger data sets. Mobile applications of customer Pinger, for example, process millions of text messages and phone calls with ColumnStore, in addition to more than 1.5 billion rows of logs per day.
Looks interesting. We’ve had performance issues with our large db tables for years that we’ve managed to work around, but it’s likely time to take a look at a redesign around a newer technology.