Many customers are curious about TiDB X and ask me - TiDB is based on object storage, like S3, but S3's latency is poor. How do you handle this chanllenge for #OLTP? At the #CNCF meeting, jinpeng.zhang, the maintainer of #TiKV, revealed the secret. youtu.be/jDlAdMEMZgs?si…

Congratulations to the #1 ranked CMU Database Group PhD student Wan Shen Lim (Wan Shen Lim (@wslim.bsky.social)) for successfully passing his doctoral defense. Wan has been working on hard AF database research with me for the last *nine* years (undergrad+grad) at CMU. He also hates chickens.

Performance Hints Over the years, my colleague Sanjay Ghemawat and I have done a fair bit of diving into performance tuning of various pieces of code. We wrote an internal Performance Hints document a couple of years ago as a way of identifying some general principles and we've

i am convinced that software devs have a speed problem they think the #1 issues is writing code faster... its not. its fixing the code that is already there to stop being utter garbage (as a garbage code connoisseur) quality is really lacking these days, yet quantity has never

The next generation of databases is built for change at scale. TiDB Cloud Premium: Adding UNIQUE KEY on 14.1B rows (5.8 TiB) in production 42m31s · 5.53M rows/s p99: 11ms → 13ms This is TiDB X: scale without operational fear.

Postgres dialect for Turso… 👀

In 2026, VLDB is returning to the Boston area 5 decades after it was born here (first VLDB was in Framingham). A good opportunity to get your company's name on the program (and earn the everlasting gratitiude of the organizing committee) vldb.org/2026/sponsorsh…

great talk .. and the most important takeaway ..

Looks like Glauber Costa has been cooking. In-process Postgres on top of Turso. 👀

"Can you imagine selling 1 million databases to a single customer?" I've pitched this to dozens of partners. They all laugh. Then I explain. One of our customers builds AI agents. Each agent gets its own database — not a row in a shared table, its own isolated instance. Because

My book Latency has been now out for six months. It's gotten a lot of good feedback, so I decided to read it myself for the first time. Here is my summary and reflection of Chapter 1!

Riffing on the recent Jepsen report, I took a look at MariaDB's open-source multi-master clustering database, Galera. I was impressed both by the work MariaDB has done to make REPEATABLE READ more in line with the literature and also with Galera itself.

"Give me your tired, your poor, your huddled masses yearning to breathe free" 🗽 As Charles Barkley observed, we are falling short of our own ideals.

Ten years ago on April 1, we open sourced TiKV. Today, on another April 1, TiDB Cloud Lake enters Private Preview. A new cloud-native analytics experience inside TiDB Cloud: - elastic warehouses - usage-based pricing - auto-suspend for idle compute - simpler analytics

Nice deep dive to testing with Antithesis!

Congratulations to Matei Zaharia on winning this year's ACM Prize in Computing! This is very well-deserved! 🎉

Over the last couple weeks, I've been experimenting with a new way to teach folks about stability and metastable failures in distributed systems.

The founders of FloeDB (Mark Cusack + Kurt Westerfeld) gave an interesting talk with CMU Database Group about their new Apache Iceberg-compatible query engine. Two key takeaways from their talk: 1⃣ Floe is a hard fork of Yellowbrick Data | SQL Data Platform. 2⃣ Floe is building a "catalog-of-catalogs"

Read the lost thesis of Dennis Ritchie, creator of the C programming language & co-creator of Unix: bit.ly/RitchiePhD Ritchie never got his PhD b/c he didn't want to pay Harvard the thesis binding fee. (v/IEEE Spectrum)

Dario is wrong. He knows absolutely nothing about the effects of technological revolutions on the labor market. Don't listen to him, Sam, Yoshua, Geoff, or me on this topic. Listen to economists who have spent their career studying this, like Aghion Philippe , Erik Brynjolfsson ,