Seagate Technology Incorporated

Harry Tiotantra

Wenxiang Xie

Kai C Chen

Andrew Michael Kowles

Feng Shen

Ajay Narayan Kulkarni

Thein Than Zaw

CheeHou Peng

PohGuat Bay

Yan Naing

Mark A Gaertner

Jun Cai

Wei Zhang

Jian Qiang

Utt Heng Kan

AikChuan Lim

Jeetandra Kella

KayHee Tang

Shen Feng

WeiQing Zhou

Chee Hou Peng

Xu Huang

David Andrew Rice

ChoonWei Ng

Chen Kai

Xie WenXiang

TzePin Chin

Ricardo Lim Soon Lian

Ronny Fnu

Li Xin

Described are systems, devices and methods for regulating speculative reading in a hard disk drive that has a read reliability budget. Regulating speculative reading under a read reliability budget involves selecting an amount of read look ahead (RLA) in response to determining whether or not a read rate exceeds the read reliability budget. When the read rate is lower than the read reliability budget, a full RLA amount can be used, thus ensuring high performance. When the read rate exceeds the read reliability budget, the RLA amount can be adjusted down while maintaining performance at acceptable levels. Adjustments to RLA may be made at periodic intervals, and in response to whether the workload is more sequential or more random.

Speculative read regulation in hard disk drives having read reliability budgets

A method includes creating a table with a plurality of cells indicative of intersections of tracks and wedges on a data storage disc surface. The method also includes storing disc pending command information for different disc pending commands in corresponding different cells of the plurality of cells according to locations for the different disc pending commands on the data storage disc surface. The method further includes searching the table for one or more next disc pending commands of the different disc pending commands to schedule for execution.

Efficient scheduling of data storage disc input/output

A data storage device can have one or more rotating data media with data tracks that are radially disposed from a central spindle. The data tracks may be logically divided into multiple regions while a write strategy is generated with a region module to set a sequence of different regions for future data writes. Receipt of a data write request to the data storage media from a host can prompt the region module to classify the data write request as a sequential or random write in order to intelligently select a region to satisfy the data write request based on the write strategy to maximize data writing consistency for data associated with the data write request.

Intelligent region utilization in a data storage device

A method includes receiving at a storage device a command from a host. When learning is active on the storage device, an initial parameter value of a plurality of parameter values is used for performing a first action of a plurality of actions for the command. The first action is performed using the initial parameter value of the plurality of parameter values for the command The first parameter value is incremented to a next parameter value of the plurality of parameter values for the command for use in reperforming the first action.

Onboard machine learning for storage device

A method includes writing first data to a first track of a magnetic recording medium of a storage device. First parity sectors corresponding to the first data are written. The first parity sectors have a first size. Second parity sectors corresponding to the first data are written. The second parity sectors have a second size. Second data is written to a second track of the magnetic recording medium. The second track is adjacent to the first track. It is determined whether an unrecoverable data error has occurred on the second track. After writing to the second track and determining that no unrecoverable data error has occurred, the first and second parity sectors corresponding to the first data are released.

Mode hop mitigation solutions in a heat-assisted magnetic recording head

Apparatus and method for managing data in a hybrid data storage device. In some embodiments, a hybrid data storage device operates in a normal mode in which data transfer commands received from a host device are serviced by a top level controller circuit which directs a first subset of the received data transfer commands to a non-volatile rotatable recording memory and a second subset of the received data transfer commands are directed to a non-volatile solid-state memory. The controller circuit transitions to a performance mode responsive to a detected first parameter so that subsequently received data transfer commands are directed exclusively to the solid-state memory. The controller circuit terminates the performance mode and resumes the normal mode responsive to a detected second parameter.

Hybrid data storage device with performance mode data path

A method includes generating a map for a storage tier. The map maps a plurality of extents to physical locations on the storage tier and is implemented as a tree with each extent to physical location mapping being one of a plurality of leaf extent nodes of the tree. Individual ones of the plurality of leaf extent nodes are compressed into small extent nodes. Different groups of the small extent nodes are associated with carrier nodes. A command that involves an extent of the plurality of extents, which is represented by a small extent node in one of the carrier nodes, is received from a host. A decompression operation is performed on the small extent node in the carrier node to provide a leaf extent node for the extent associated with the host command. An extent node operation is performed on the leaf extent node for the extent.

Mapping system for data storage devices

Implementations disclosed herein include a method comprising receiving a TRIM request for a media cache in a storage media from a host, determining whether a TRIM range in the TRIM request overlaps with an active operation, invalidating the TRIM range responsive to determining that the TRIM range does not overlap with an active operation, and inserting at least one null node in the media cache. The method may comprise receiving a host read request, performing a media cache search for a read LBA range, and determining where data is located. If all the data is located in the media cache, the system reads from the media cache. If the data is located in a null node, patterning for the null node occurs. If the data is located partially in the main store, the media cache, or a null node, the data may be read, combined, and returned to the host.

Shingled magnetic recording trim operation

Apparatus and method for managing metadata in a data storage device. In some embodiments, a metadata object has entries that describe data sets stored in a non-volatile write cache. During an archival (persistence) operation, the metadata object is divided into portions, and the portions are copied in turn to a non-volatile memory at a rate that maintains a measured latency within a predetermined threshold. A journal is formed of time-ordered entries that describe changes to the metadata object after the copying of the associated portions to the non-volatile memory. The journal is subsequently stored to the non-volatile memory, and may be subsequently combined with the previously stored portions to recreate the metadata object in a local memory. The measured performance latency may be related to a specified customer command completion time (CCT) for host commands.

Latency sensitive metadata object persistence operation for storage device

An exemplary write method disclosed herein includes receiving a request to write data to a consecutive sequence of logical block addresses (LBAs); identifying a first non-contiguous sequence of data tracks mapped to a first portion of the consecutive sequence of LBAs; and identifying a second non-contiguous sequence of data tracks mapped to a second portion of the consecutive sequence of LBAs, where the second portion sequentially follows the first portion. The method further includes writing the data of the second portion of the consecutive sequence of LBAs to the first non-contiguous sequence of data tracks during a first pass of a transducer head through the radial zone and writing the data of the first portion of the consecutive sequence of LBAs to the second non-contiguous sequence of data tracks during a second, subsequent pass of the transducer head through the radial zone.

Two-pass logical write for interlaced magnetic recording

An exemplary write method disclosed herein includes receiving a request to write data to a consecutive sequence of logical block addresses (LBAs) that is the mapped to a non-contiguous sequence of data tracks on a storage medium, and writing the data of the consecutive sequence of LBAs to a non-contiguous sequence of data tracks on the storage medium and according to a consecutive track order.

Growing community of inventors

Singapore, Singapore

Harry Tiotantra