Machine Learning Supplemented Storage Device Inspections

    公开(公告)号:US20230367667A1

    公开(公告)日:2023-11-16

    申请号:US17741217

    申请日:2022-05-10

    IPC分类号: G06F11/07 G06N3/08

    摘要: Methods are provided for utilizing machine learning operations configured for use in processing missing pieces of visual data in image data to predict potential location of defects and/or damage in storage device disks. These predictions can allow for a sufficient ability to categorize disks during storage device quality inspections. This can allow for quality inspections to conclude before all areas of the disk surface are scanned. Because less surface area of the disks within the storage device require scanning, the time required for quality inspection scanning prior to deployment can be greatly reduced. Additionally, the partial scans occurring prior to deployment may be supplemented or updated after deployment through the performance of a dense scan. These secondary scans can be configured to scan all previously unscanned areas during storage device downtimes or in response to an environmental trigger such that the storage device will not exhibit any loss in performance.

    DATA STORAGE DEVICE DYNAMICALLY RELOCATING DATA SECTORS BASED ON MAP-OUT VALUE

    公开(公告)号:US20220027226A1

    公开(公告)日:2022-01-27

    申请号:US16935763

    申请日:2020-07-22

    摘要: A data storage device is disclosed comprising a non-volatile storage medium (NVSM) having a plurality of data sectors and a plurality of reserve sectors. A map-out value is generated for each of a first plurality of the data sectors based on a read latency of each of the first plurality of data sectors, and when the map-out value of a first data sector in the first plurality of data sectors exceeds a threshold, a first logical block address (LBA) is mapped from the first data sector to a first reserve sector. When the map-out value of a second data sector in the first plurality of data sectors exceeds the map-out value of the first data sector, the first LBA is mapped from the first reserve sector back to the first data sector, and a second LBA is mapped from the second data sector to the first reserve sector.

    Machine learning supplemented storage device inspections

    公开(公告)号:US12032434B2

    公开(公告)日:2024-07-09

    申请号:US17741217

    申请日:2022-05-10

    IPC分类号: G06F11/00 G06F11/07 G06N3/08

    摘要: Methods are provided for utilizing machine learning operations configured for use in processing missing pieces of visual data in image data to predict potential location of defects and/or damage in storage device disks. These predictions can allow for a sufficient ability to categorize disks during storage device quality inspections. This can allow for quality inspections to conclude before all areas of the disk surface are scanned. Because less surface area of the disks within the storage device require scanning, the time required for quality inspection scanning prior to deployment can be greatly reduced. Additionally, the partial scans occurring prior to deployment may be supplemented or updated after deployment through the performance of a dense scan. These secondary scans can be configured to scan all previously unscanned areas during storage device downtimes or in response to an environmental trigger such that the storage device will not exhibit any loss in performance.