Tokutek’s Chief Scientist Discusses TokuDB v5.0, Part 2

Keeping from Getting Buried in Big Data

Boy, what a way to go into spring. We are expecting up to 1/2″ of rain (or perhaps sleet) today. This follows the 3″ of snow we had on Friday (Shaq — watch out we long ago got Nate – you are next).

Here is the picture from outside our offices on Friday:

Anyway, as I mentioned in my post last Monday, our Chief Scientist and co-founder Martín Farach-Colton had the privilege of sitting down with Sheeri Cabral and Sarah Novotny for their weekly MySQL Database Community Podcast. It turned into a two-part series, and the second one recently was released (see part 1 and now part 2).

In it, Martín speaks about Tokutek and TokuDB v5.0, which launched last month (see press release, blog, quotes). In Part 2, you can hear Martín talk further on TokuDB v5.0 as well as general database issues and even some good beach reading.  This week’s installment includes:

• ACID on Tokutek
• Checkpoints and fast recovery times
• Automated crash testing
• Current methodology and future  plans for backup
• Making enterprise features available in MySQL
• Avoiding partitioning while still having very quick deletions
• “Wonking-out” with cache-oblivious algorithmics
• Unlearning the “hacks” that people have gotten used to
• Improvements in performance with clustering and memory constrained indexes
• Graphic novels on the Foundations of Mathematics
• Making peace between analysts and DBAs

Enjoy.

And, as I mentioned in my last piece I wasn’t going to quite hang up the boots yet with “spring” officially here. So glad they are still out…