Data-flow Availability: Achieving Timing Assurance in Autonomous Systems
Talkat 18th USENIX Symposium on Operating Systems Design and Implementation (OSDI '24) · 2024-07-10..07-12·
summary
A survey of 189 timing bugs across 7 open-source CPS projects finds only 9% use real-time theory. The paper defines Data-flow Availability (DFA) to encode freshness, consistency, stability as timing constraints on data flows and implements Kairos: API-based annotations, run-time program instrumentation, profile-based testing using control-state deviation as an oracle, and cross-layer scheduler coordination. Kairos fixes ~60–70% of timing bugs, cuts control deviation by 74% and collisions by 62% while adding 2.8% runtime overhead.
AI-written summary · about
chapters (11)
- 0:00Cyber-physical systems timing
- 1:31Predictability over speed
- 2:58Testing vs scheduling theory
- 4:23Timing bugs in open-source
- 7:24Key timing properties
- 9:12Chyros timing framework
- 11:44Chyros workflow
- 12:48Scheduling policies for timing
- 14:13Chyros on real platforms
- 15:40Results: 74% fewer deviations, 62% fewer collisions
- 16:25Conclusion and challenges
related talks
01h 01mFeedback Control Theory: Architectures and Tools for Real-Time Decision Making I
13mChallenges in High Performance Robotics Systems
27mdora-rs: Modern Dataflow for Robotics
20m"Rust in Motion: Building Reliable and Performant Robotics Systems"
27mDeterministic fleet management for autonomous mobile robots using Rust
01h 04mMeshing Robots and Humans with Distributed Apps
14mReliability Engineering for Hybrid Robot-Cloud Systems
01h 17mFeedback Control Theory: Architectures and Tools for Real-Time Decision Making II