HDD Defect Mapping Using Partial Scans and Embedded ML Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current storage devices face challenges in improving processing power and complexity due to limited internal resources and hardware-based solutions that require significant capital investment and are not scalable for varying operational needs.
Innovation Solution
Implementing machine learning methods within the System on a Chip (SoC) of storage devices to dynamically generate and adjust models for processing tasks, allowing for efficient defect management without additional hardware changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If additional processors or specialized components are added to handle processing tasks, then processing power is improved, but hardware complexity and manufacturing cost increase significantly
Solution Approach 1:
The patent replaces hardware-based processing solutions with software-based machine learning algorithms executed on the existing SoC processor. Instead of adding physical processors or specialized components to handle defect detection and management tasks, the system uses machine learning models (such as neural networks) that can be implemented through software, thereby avoiding the complexity and cost of hardware modifications while still achieving improved processing capabilities.
Solution Approach 2:
The patent changes the approach from hardware parameter changes (adding more processors or specialized components) to software parameter changes (implementing machine learning algorithms). By transitioning to a software-based solution with machine learning models, the system achieves improved processing power without physically modifying the hardware architecture, thus avoiding increased device complexity.
2Power
If hardware-based solutions are implemented to improve processing, then processing capability is enhanced, but capital investment cost increases by millions of dollars
Solution Approach 1:
The patent substitutes expensive hardware-based processing solutions with cost-effective software-based machine learning algorithms. By implementing defect detection and management through machine learning models executed on existing hardware, the system achieves enhanced processing capability without requiring the millions of dollars in capital investment that would be needed for hardware modifications.
Solution Approach 2:
The patent uses software-based machine learning models that can be updated and replaced without physical hardware changes. This approach allows the system to achieve improved processing capabilities through software updates rather than expensive hardware replacements, significantly reducing manufacturing and deployment costs.
3Productivity
If external processing is offloaded to host or remote systems, then processing load is reduced, but system complexity and communication dependency increase
Solution Approach 1:
The patent implements machine learning models that execute directly within the storage device's SoC, enabling the system to perform defect detection and management tasks autonomously without requiring external processing resources. This self-service approach improves processing efficiency by eliminating communication delays while avoiding the complexity of external system integration and communication protocols.
Data Source
AI summary
Methods are provided for managing defects in Hard Disk Drive (HDD) storage devices. In particular, only a portion of the cylinders of an HDD is tested. A bag of machine learning models is used to reconstruct the data for the untested cylinders. A defect file for the HDD is generated, a classifier model may be applied to the defect file, and one or more neural network models may be applied. If the defects are unsuitable for use by the models, then a scan of the entire HDD is run instead. An HDD comprises a rotating disk and a read/write head actuated above the disk surface. The disk may be formatted into concentric data tracks, with each track being divided into sectors. The tracks may be organized into zones (groups of tracks called cylinders), and the axially parallel sectors in each cylinder may be organized into wedges.