CAFault: Enhance Fault Injection Technique in Practical Distributed Systems via...
summary
A framework that generates and prunes FDModel configurations by observing coverage changes and then drives focused testing with fault-handling guided fuzzing on HDFS, ZooKeeper, MySQL‑Cluster and IPFS. It extracts exception‑handling code, mutates configuration groups with binary‑search minimization to find fault‑dependent items, and steers injections to exercise fault‑tolerance code. Compared to CrashFuzz, Mallory, and Chronos it covers 31.5%, 29.3%, and 81.5% more fault‑tolerance logic and uncovers 16 previously unknown serious bugs.
AI-written summary · about
chapters (11)
- 0:00CAFault introduction
- 2:08Existing fault injection challenges
- 3:33Fault dependent testing need
- 6:00CAFault’s two processes
- 9:51Fault tolerance extraction & fuzzing
- 11:37CAFault core components
- 12:42Real-world bug detection
- 13:47Fault handling coverage comparison
- 15:12Fault model performance impact
- 16:18Guided fuzzing improvements
- 17:43Future work
related talks
39mMaking Fuzzing Part of Your Software Development Lifecycle
14mModel-guided Fuzzing of Distributed Systems
04mOrchestrating chaos: Applying database research in the wild
18mEfficient Exposure of Partial Failure Bugs in Distributed Systems with Inferred...
17mOne-Size-Fits-None: Understanding and Enhancing Slow-Fault Tolerance in Modern...
53mFuzz Testing for Fun and Profit
01h 00mClusterFuzz: Fuzzing at Google Scale
31mDistributed system fault injection testing with Docker