I was lucky enough to attend SQL Saturday Austin 2025 a little over a week ago in conjunction with some work at the Redgate office. The opening keynote at the event was given by Conor Cunningham, who is an architect at Microsoft and runs the engineering team in Austin for the Data Platform. His talk was very interesting and engaging. This was about half the room below.
I’ll describe the talk, which was great, but I’m probably mis-stating something. Keep in mind I’m writing this a few days later from memory.
One of the interesting things Conor talked about was the engineering process in Redmond. The main thrust was that they work in 6 month planning cycles. Work gets submitted, voted on, works through groups, and as a result, most of the things approved are related to more revenue in some way.
Good for Microsoft, not something many of us love.
In Austin, they tackled things differently. You can see the outline in the image below. Conor hired a few engineers here, usually out of college, and they work on new features. Things that are too small to make the list in Redmond, but this also helps grow/train engineers on how to write production quality code.
What do they work on? He had a few slides, but IS [NOT[ DISTICNT was one. He described this a bit, but the cool thing was a bunch of the SQL 2022 features were from Austin. String_Split with the ordinal, the bit functions, and a bunch more.
Basically all the cool features I appreciated, including DATETRUNC, came from this small team. I am very glad they exist.
He also delved into the challenges of doing remote work. Building SQL Server is a non trivial procss, and they’ve been working to try and make this easier. They refactor code, they try to break things out so engineers can get quicker feedback, write more tests, etc. to make their software engineering easier.
He also talked about their work with hardware manufacturers and some of the optimizations he’s done with CPU and the ATX instructions to make SQL Server a little faster. For some types of queries, they’ve great improved the speed of internal SQL Server processing.
It was a great talk and quite entertaining. Hopefully some of you will see it elsewhere, or he’ll do it again in Austin and you can come.