How do to do research with Systems Researchers
Talkat Algorithmic Foundations for Emerging Computing Technologies Boot Camp 2025 · 2025-09-08..09-12·
summary
Covers concrete claims and artifacts: replacing B-trees with write‑optimized structures (a B-epsilon tree) yields up to a 200x speedup on random inserts, closes OLTP/OLAP gaps, and leads to a commercial MySQL backend sold as Tokutek to customers like Kayak and Walmart; explains implementations (B^ε trees, LSM variants), BetrFS file‑system work with multiple FAST awards, why patents and university licensing matter, and practical VC/negotiation lessons.
AI-written summary · about
chapters (22)
- 0:00Cross-disciplinary research intro
- 1:23Good, bad, and ugly of cross-disciplinary
- 3:32Experimental research types
- 6:04Systems skepticism on algorithms
- 8:36Research ownership in collaborations
- 13:42Successful collaboration examples
- 18:05Curse of knowledge in talks
- 19:32BetrFS project and problem-solving
- 23:07Everyday versus speculative research
- 27:51Mosaic project from TLB to theory
- 28:13Pre-Tokutek database history
- 34:00B-tree problems in databases
- 37:52Inspiration for Tokutek and B-epsilon
- 41:52Founding Tokutek
- 42:13Tokutek product and customers
- 48:26Post-Tokutek database impact
- 51:18Steps to start a company
- 54:32Patent licensing with universities
- 1:00:13Raising venture capital
- 1:05:36Systems vs theory culture
- 1:09:34Final thoughts on cross-disciplinary
- 1:13:36Q&A
related talks
45mThe Database: A Sophisticated and Special Data Structure
52mHow shit works: Databases
01h 16mOn-Disk Data Structures: B-Trees, LSM-Trees and Fractal Trees
01h 16mDeep Dive into On-Disk Data Structures
13mClosing the B+-tree vs. LSM-tree Write Amplification Gap on Modern Storage Hardware...
35mQuestioning Database Claims: Design Patterns of Storage Engine
39mAn Introduction to B-trees and LSM-trees for DEVs and DevOps
17mTheory matters: how to build a successful company by trusting theoreticians