Maximilian Kuschewski

What do io_uring, a racing car, and the fastest financial transactions database have in common? Maybe the coolest story in recent tech lore!

Vortex is BOTH 38% smaller and has 10–25x faster scans than Parquet+ZStd for TPC-H SF10 We implemented BtrBlocks-style cascading compression in Vortex that recursively tries codecs like ALP, FSST, and bit-packing, letting the data pick the best encoding spiraldb.com/post/cascading…

@andreaskipf Link to paper: cs.cit.tum.de/fileadmin/w00c…

SELECT * FROM aws; DROP mic; Presenting cloudspecs.fyi at CIDR, the coolest database conference out there 🤌 Thanks to @andreaskipf for the photo!

Our next South Bay Systems event will be on January 21st, featuring speakers from CIDR! Come to Databricks to hear about: * xNVMe by @pinartozun of ITU * Spilling in QP by Maximilian Kuschewski of TUM * NPUs in DBs by Alexander Baumstark of TU-Ilmenau luma.com/8a54z94d

New year brain food? "Our Future in the Cloud" landed on btrblocks.com, our take on where hyperscalers are heading 📊and why it matters. Happy New Year and happy reading! #CloudComputing #TechTrends

Interested in io_uring? Check out our new paper, joint work with my former colleagues from TU Darmstadt (@cbinnig and @m4ttze) and from TU Munich (@melhindiCS and @viktor_leis). arxiv.org/abs/2512.04859…

@iavins Gotta plug this: you'll really enjoy my colleague Marcus's new paper: B-Trees Are Back: Engineering Fast and Pageable Node Layouts cs.cit.tum.de/fileadmin/w00c…

The B Tree data structure fascinates me. Databases use B Trees to store data on disk, organizing everything into pages that typically range from 4kb to 8kb. All I/O operations happen in units of these pages. The page looks like this... (1/9)

@TigerBeetleDB is now 20% faster end-to-end thanks to our new radix sort github.com/tigerbeetle/ti…. Curious about the optimizations + methodology? github.com/tigerbeetle/ti… introduces the radix sort (only ~100 LOC)

𝗕𝘁𝗿𝗕𝗹𝗼𝗰𝗸𝘀 𝗻𝗲𝘄𝘀! After securing a 𝗘𝗥𝗖 𝗣𝗿𝗼𝗼𝗳 𝗼𝗳 𝗖𝗼𝗻𝗰𝗲𝗽𝘁 grant in July (press release: #c155602" target="_blank" rel="nofollow noopener">tum.de/en/news-and-ev…), 𝘁𝗼𝗱𝗮𝘆 𝗺𝗮𝗿𝗸𝘀 𝘁𝗵𝗲 𝗼𝗳𝗳𝗶𝗰𝗶𝗮𝗹 𝘀𝘁𝗮𝗿𝘁 of our first PoC project in the context of #BtrBlocks.

The production quality on the #SD25 conference videos is really fantastic. The @TigerBeetleDB team puts so much love and effort into everything 🤩

The #SD25 lightning talk session, hosted by @natalievais, is premiering now! Our speakers are: @asubiotto, @TylerFCloutier, Joris Portegies Zwart, @maxikuschewski & @denis_bazhenov. Join us in the live chat now: youtube.com/watch?v=k58vJl…

@andrewlamb1111 Link: db.in.tum.de/~schmidt/paper…

@andrewlamb1111 One improvement regarding benchmaxxing is having thousands of diverse benchmark queries instead of dozens. Plugging the new SQLStorm paper below ;)

Benchmaxxing (verb): to add specific optimizations that only impact benchmark results and are not widely applicable to real world use cases.

Sweet VLDB paper from TUM (Mateusz Gienieczko / github.com/v0ldek) proposing extending @ApacheParquet using user defined encodings (via WASM). Favorite image shows the ease of integrating into @ApacheDataFusio gienieczko.com/anyblox-paper

Big thanks to @natalievais for the ⚡ Talks & @TigerBeetleDB for a fantastic conference! Definitely the sickest-looking stage I've ever been on 😄

@wangbin579 Please share!

I have one of the most challenging papers in database theory—so difficult that it was rejected for being too advanced. If anyone dares to test themselves, you can ask me for a copy.

It is neat to see @ApacheDataFusio mentioned in multiple talks at @cidrdb this year. I think there a large untapped opportunity for academic/industrial collaboration using DataFusion for academic research

@nabla3110 @andrewlamb1111 @ApacheDataFusio @cidrdb Yes

@andrewlamb1111 @ApacheDataFusio @cidrdb Is the person in the first photo Viktor Leis?

@SpiralDB Cool stuff!

This Halloween, we wanted to take a moment to reflect on another month of building Vortex in the open blog.spiraldb.com/trick-or-treat…