SSD Anomaly Detection Using Multi-Source Data Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current anomaly detection methods for solid-state drives (SSDs) rely on limited information from Self-Monitoring, Analysis, and Reporting Technology (S.M.A.R.T.) and lack automation and comprehensive analysis, leading to inefficiencies and incomplete understanding of anomaly causes.
Innovation Solution
A method and device for anomaly detection in SSDs that collect and analyze multiple types of data, including S.M.A.R.T., NAND flash cell threshold voltage distribution, and bit error rate eye diagram data, using trained models for feature extraction and anomaly cause analysis to improve automation and comprehensiveness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple types of test data are collected and analyzed using trained models, then the accuracy and comprehensiveness of anomaly detection is improved, but the device complexity and processing requirements increase
Solution Approach 1:
The patent segments the anomaly detection system into multiple specialized components: S.M.A.R.T. data analysis module, NAND flash cell threshold voltage distribution analysis module, bit error rate eye diagram data analysis module, and anomaly cause analysis module. Each module processes specific data types using dedicated trained models, allowing complex multi-data-type analysis to be divided into manageable, specialized segments that can be executed independently and efficiently
Solution Approach 2:
The patent introduces trained machine learning models as intermediaries between the raw test data and the anomaly detection process. These models (including random forest, support vector machine, and neural network models) serve as mediators that automatically extract features and identify patterns from multiple data types, reducing the complexity of direct data analysis while improving detection accuracy
2Productivity
If automated anomaly detection and cause analysis are implemented using trained models, then the productivity and efficiency are improved, but the device complexity increases
Solution Approach 1:
The patent implements self-service automation where the system automatically collects test data from multiple sources, processes the data through trained models, identifies anomalies, and determines anomaly causes without manual intervention. The anomaly detection processing device autonomously performs the entire workflow from data collection to cause analysis, significantly improving productivity while the modular architecture keeps complexity manageable
Solution Approach 2:
The patent applies preliminary action by pre-training multiple specialized models (random forest, support vector machine, neural network models) before deployment. These models are trained offline on historical data and then used for rapid automated detection. The feature extraction and model preparation are performed in advance, enabling efficient real-time anomaly detection without complex runtime processing
3Loss of information
If deep analysis of anomaly causes is performed using trained models, then the information completeness is improved, but the loss of time for manual operations is reduced
Solution Approach 1:
The patent implements feedback mechanisms where the anomaly cause analysis model continuously refines its predictions based on the test data and detected anomalies. The system provides feedback loops that allow the trained models to learn from previous detections and improve their accuracy over time, ensuring complete information extraction while maintaining efficient processing speeds
Solution Approach 2:
The patent replaces manual analysis operations with automated machine learning models. Instead of human testers manually examining logs and determining reliability, the system uses trained random forest, support vector machine, and neural network models to automatically perform deep analysis of anomaly causes. This substitution eliminates manual operations and reduces analysis time while providing comprehensive information
Data Source
AI summary
An anomaly detection processing method and device for solid-state drive (SSD) are provided. The anomaly detection processing method for SSD including collecting test data of an SSD, the test data including at least one of self-monitoring, analysis and reporting technology S.M.A.R.T. data, NAND flash cell threshold voltage distribution data, and bit error rate eye diagram data, determining whether the SSD has an anomaly based on the test data, and determining an anomaly cause of the SSD based on a subset of test data, the subset including specific test data based on which the SSD has been determined to have an anomaly may be provided.