Orchestrating chaos: Applying database research in the wild

on YouTube ↗04m

" summary

Claims that lineage-driven fault injection uses database provenance to cut the fault space from astronomically many (2^100 ~ 10^30) down to a tiny set worth testing by inferring which component failures actually affect observable outputs. Describes an implemented LDFI toolchain that drives targeted fault injection from provenance traces, reports early field results showing far fewer experiments needed than random chaos or manual Jepsen style hunts, and outlines near term extensions and scalability limits.

AI-written summary · about

" chapters (4)

1. 0:00Intro
2. 1:00What is the shift in database research
3. 1:40Too many faults
4. 2:40Two giants

" related talks

• 19mCAFault: Enhance Fault Injection Technique in Practical Distributed Systems via...Yu Jiang, Fuchen Ma, Yuanliang Chen, Yuanhang Zhou, Zhen Yan · USENIX ATC '25
• 18mEfficient Exposure of Partial Failure Bugs in Distributed Systems with Inferred...Peng Huang, Jia Pan, Haoze Wu · 21st USENIX Symposium on Networked Systems Design and Implementation (NSDI '24)
• 57mChaos Engineering PrimerSergiu Bodiu · Voxxed Days Singapore 2017
• 33mChaos engineering for the rest of usMarouane Gazanayi · BlaBlaConf 2022
• 32mBuilding Reliable Infrastructure in RustTyler Neely · RustFest Paris 2018
• 17mOne-Size-Fits-None: Understanding and Enhancing Slow-Fault Tolerance in Modern...Yuxuan Jiang, Peng Huang, Guangtao Xue, Ruiming Lu, Yunchi Lu · NSDI '25
• 47mApproximate Lifted Inference with Probabilistic DatabasesWolfgang Gatterbauer · Uncertainty in Computation 2016
• 26mDebugging Large-Scale Datalog: A Scalable Provenance Evaluation StrategyDavid Zhao, Bernhard Scholz, Pavle Subotic · POPL 2021