Autonomous Data Protection via ML Anomaly Quarantine
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data replication techniques for protecting data in storage area networks are inefficient due to high operational overhead and risks, and they do not effectively address anomalies that could indicate data corruption or loss.
Innovation Solution
Implementing continuous data protection techniques using machine learning engines for anomaly detection, which take rolling snapshots of data storage elements, model data storage change rate trends, and quarantine potential anomalies, while dynamically allocating memory and applying reinforcement learning to improve anomaly detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data replication techniques are used to protect data in storage area networks, then data protection is achieved, but operational overhead increases and efficiency decreases
Solution Approach 1:
The patent extracts the data protection function from traditional replication mechanisms and implements it through anomaly detection and quarantine. Instead of replicating all data continuously, the system identifies and isolates only the anomalous data blocks that indicate corruption or compromise, thereby maintaining protection while reducing operational overhead.
Solution Approach 2:
The system employs autonomous anomaly detection that operates independently without requiring manual intervention. The machine learning models continuously monitor storage devices, automatically identify anomalies, and enforce quarantine actions, enabling the system to protect itself without increasing operational burden on users or administrators.
2Reliability
If traditional data protection techniques are used, then data safeguarding is provided, but they fail to effectively address anomalies indicating data corruption or loss
Solution Approach 1:
The system implements continuous feedback loops where machine learning models analyze storage device behavior patterns in real-time. Anomaly detection models receive feedback from monitored I/O operations, change rates, and data access patterns, enabling them to progressively improve their detection accuracy and respond to evolving threat patterns.
Solution Approach 2:
The patent replaces traditional mechanical replication mechanisms with intelligent anomaly detection systems powered by machine learning. Instead of relying on rigid replication protocols that cannot detect subtle corruption patterns, the system uses AI-driven analysis to identify anomalous behavior indicating data corruption or compromise.
3Measurement precision
If continuous monitoring and snapshotting of storage devices is implemented, then anomaly detection accuracy is improved, but storage resource consumption increases
Solution Approach 1:
The system applies partial monitoring by focusing computational resources on identifying and analyzing only the anomalous data blocks rather than continuously processing all storage operations. The machine learning models prioritize detection based on risk assessment, applying intensive monitoring only where anomalies are detected, thereby maintaining high detection accuracy while reducing overall storage resource consumption.
Data Source
AI summary
Embodiments of the present disclosure relate to autonomous data protection. One or more input/output (I/O) streams can be received by one or more storage devices. One or more snapshots of each storage device can be obtained. One or more anomalies can be identified based on a change rate corresponding to each storage device's memory allocated to store user data included in the I/O streams.