MSFRD: Mutation Similarity based SSD Failure Rating and Diagnosis for Complex...
summary
Proposes MSFRD, which uses dynamic mutation extraction from telemetry (51M+ T logs, 85 custom attributes) and a weighted mutation-similarity metric to score each SSD against historical failures and healthy devices. It averages the top-three similarity neighbors, computes failure/health/outlier tendencies, and assigns a four-level failure rating with progressive, automated diagnosis for level escalation. Evaluated on Samsung and Alibaba datasets, it boosts precision by 23.8% and recall by 38.9% for failure prediction versus prior schemes.
AI-written summary · about
chapters (10)
- 0:00MSFRD research and SSD failures
- 2:36MSFRD scheme and failure rating
- 4:49Dynamic mutation extraction
- 7:18Classification and anomaly detection
- 9:24Four-level SSD failure rating
- 11:35MSFRD process and similarity measurement
- 14:51Progressive diagnosis strategy
- 15:56Effectiveness evaluation results
- 18:07Key findings and implications
- 19:30Conclusion and Q&A
related talks
17mMulti-view Feature-based SSD Failure Prediction: What, When, and Why
14mFailureMiner: A Joint Key Decision Mining Scheme for Practical SSD Failure Prediction...
05mGenerating Realistic Wear Distributions for SSDs (Work-in-Progress Report)
27mMaking Disk Failure Predictions SMARTer!
40mDemystifying File Similarity for Malware Detection
25mAI Reliability in Security: A New Data Science Paradigm
26mHow to Apply the Machine Learning Appropriate Way for Your Security Operation
49mShifting Safety Techniques To a Statistical World