SYSTEMS, METHODS, AND DEVICES FOR DATA PROPAGATION IN GRAPH PROCESSING

    公开(公告)号:US20220107844A1

    公开(公告)日:2022-04-07

    申请号:US17170881

    申请日:2021-02-08

    Abstract: A method of partitioning a graph for processing may include sorting two or more vertices of the graph based on incoming edges and outgoing edges, placing a first one of the vertices with fewer incoming edges in a first partition, and placing a second one of the vertices with fewer outgoing edges in a second partition. The first one of the vertices may have a lowest number of incoming edges, and the first one of the vertices may be placed in a first available partition. The second one of the vertices may have a lowest number of outgoing edges, and the second one of the vertices may be placed in a second available partition. A method for updating vertices of a graph may include storing a first update in a first buffer, storing a second update in a second buffer, and transferring the first and second updates to a memory using different threads.

    MULTI-NON-VOLATILE MEMORY SOLID STATE DRIVE BLOCK-LEVEL FAILURE PREDICTION WITH SEPARATE LOG PER NON-VOLATILE MEMORY

    公开(公告)号:US20230037270A1

    公开(公告)日:2023-02-02

    申请号:US17964013

    申请日:2022-10-11

    Abstract: A storage device is disclosed. A first storage media may store data. The first storage media may be of a first storage type and may be organized into at least two blocks. A second storage media may also store data. The second storage media may be of a second storage type different from the first type, and may also be organized into at least two blocks. A controller may manage reading data from and writing data to the first storage media and the second storage media. Metadata storage may store device-based log data for errors in the storage device. The drive-based log data may include a first log data for the first storage media and a second log data for the second storage media. An identification circuit may identify a suspect block in the at least two blocks in the first storage media and the second storage media, responsive to the device-based log data.

    SYSTEMS AND METHODS FOR PREDICTING STORAGE DEVICE FAILURE USING MACHINE LEARNING

    公开(公告)号:US20210264294A1

    公开(公告)日:2021-08-26

    申请号:US15931573

    申请日:2020-05-13

    Abstract: A method for predicting a time-to-failure of a target storage device may include training a machine learning scheme with a time-series dataset, and applying the telemetry data from the target storage device to the machine learning scheme which may output a time-window based time-to-failure prediction. A method for training a machine learning scheme for predicting a time-to-failure of a storage device may include applying a data quality improvement framework to a time-series dataset of operational and failure data from multiple storage devices, and training the scheme with the pre-processed dataset. A method for training a machine learning scheme for predicting a time-to-failure of a storage device may include training the scheme with a first portion of a time-series dataset of operational and failure data from multiple storage devices, testing the machine learning scheme with a second portion of the time-series dataset, and evaluating the machine learning scheme.

    MULTI-NON-VOLATILE MEMORY SOLID STATE DRIVE BLOCK-LEVEL FAILURE PREDICTION WITH UNIFIED DEVICE LOG

    公开(公告)号:US20220066897A1

    公开(公告)日:2022-03-03

    申请号:US17093626

    申请日:2020-11-09

    Abstract: A storage device is disclosed. A first storage media may store data. The first storage media may be of a first storage type, and may be organized into at least two blocks. A second storage media may also store data. The second storage media may be of a second storage type different from the first type, and may also be organized into at least two blocks. A controller may manage reading data from and writing data to the first storage media and the second storage media. Metadata storage may store device-based log data for errors in the storage device. The drive-based log data for errors may include a unified log data for the first storage media and the second storage media. An identification circuit may identify a suspect block in the at least two blocks in the first storage media and the second storage media, responsive to the unified log data. Parameters for the first storage media and the second storage media may be derived from a unified parameter in the unified log data in proportion to the number of write operations and the number of read operations to each storage media, relative to the number of write operations and the number of read operations of the storage device.

    MULTI-NON-VOLATILE MEMORY SOLID STATE DRIVE BLOCK-LEVEL FAILURE PREDICTION WITH SEPARATE LOG PER NON-VOLATILE MEMORY

    公开(公告)号:US20220066683A1

    公开(公告)日:2022-03-03

    申请号:US17093620

    申请日:2020-11-09

    Abstract: A storage device is disclosed. A first storage media may store data. The first storage media may be of a first storage type and may be organized into at least two blocks. A second storage media may also store data. The second storage media may be of a second storage type different from the first type, and may also be organized into at least two blocks. A controller may manage reading data from and writing data to the first storage media and the second storage media. Metadata storage may store device-based log data for errors in the storage device. The drive-based log data may include a first log data for the first storage media and a second log data for the second storage media. An identification circuit may identify a suspect block in the at least two blocks in the first storage media and the second storage media, responsive to the device-based log data.

    FIRMWARE-BASED SSD BLOCK FAILURE PREDICTION AND AVOIDANCE SCHEME

    公开(公告)号:US20210124502A1

    公开(公告)日:2021-04-29

    申请号:US16701133

    申请日:2019-12-02

    Abstract: A Solid State Drive (SSD) is disclosed. The SSD may comprise flash storage for data, the flash storage organized into a plurality of blocks. A controller may manage reading data from and writing data to the flash storage. Metadata storage may store device-based log data for errors in the SSD. Identification firmware may identify a block responsive to the device-based log data. In some embodiments of the inventive concept, verification firmware may determine whether the suspect block is predicted to fail responsive to both precise block-based data and the device-based log data.

    PARTITIONING GRAPH DATA FOR LARGE SCALE GRAPH PROCESSING

    公开(公告)号:US20200183604A1

    公开(公告)日:2020-06-11

    申请号:US16255815

    申请日:2019-01-23

    Abstract: Inventive aspects include a method, apparatus, and system for partitioning and processing graph data for large-scale graphs. Such may include, in a partitioning pre-processing step, assigning a plurality of destination vertices to a plurality of partitions such that each destination vertex of the plurality of destination vertices is uniquely assigned to only one partition from among the plurality of partitions. Such may also include, in a main execution of external graph processing step, (i) loading a given partition of destination vertices from among the plurality of partitions from a solid state drive (SSD) into a main memory of a computing machine, (ii) streaming one or more chunks of source vertex data from the SSD into the main memory of the computing machine, and (iii) performing graph processing based at least on the loaded given partition of destination vertices and the streamed one or more chunks of source vertex data.

    MULTI-NON-VOLATILE MEMORY SOLID STATE DRIVE BLOCK-LEVEL FAILURE PREDICTION WITH SEPARATE LOG PER NON-VOLATILE MEMORY

    公开(公告)号:US20250130912A1

    公开(公告)日:2025-04-24

    申请号:US19000666

    申请日:2024-12-23

    Abstract: A storage device is disclosed. A first storage media may store data. The first storage media may be of a first storage type and may be organized into at least two blocks. A second storage media may also store data. The second storage media may be of a second storage type different from the first type, and may also be organized into at least two blocks. A controller may manage reading data from and writing data to the first storage media and the second storage media. Metadata storage may store device-based log data for errors in the storage device. The drive-based log data may include a first log data for the first storage media and a second log data for the second storage media. An identification circuit may identify a suspect block in the at least two blocks in the first storage media and the second storage media, responsive to the device-based log data.

    SYSTEMS AND METHODS FOR PREDICTING STORAGE DEVICE FAILURE USING MACHINE LEARNING

    公开(公告)号:US20230281489A1

    公开(公告)日:2023-09-07

    申请号:US18197717

    申请日:2023-05-15

    CPC classification number: G06N5/04 G06N20/00 G06F11/16

    Abstract: A method for predicting a time-to-failure of a target storage device may include training a machine learning scheme with a time-series dataset, and applying the telemetry data from the target storage device to the machine learning scheme which may output a time-window based time-to-failure prediction. A method for training a machine learning scheme for predicting a time-to-failure of a storage device may include applying a data quality improvement framework to a time-series dataset of operational and failure data from multiple storage devices, and training the scheme with the pre-processed dataset. A method for training a machine learning scheme for predicting a time-to-failure of a storage device may include training the scheme with a first portion of a time-series dataset of operational and failure data from multiple storage devices, testing the machine learning scheme with a second portion of the time-series dataset, and evaluating the machine learning scheme.

Patent Agency Ranking