Abstract: A method for securing a data storage device (DSD) against rogue behaviour by a host, the method executed by a controller of the DSD and comprising: determining a host type of the host; detecting one or more access activities performed by the host on the DSD; processing the one or more access activities to determine a security threat level of the host, wherein the security threat level is determined by weighting one or more corresponding access activity parameters by one or more impact weights; and in response to determining that the security threat level of the host is greater than or equal to a rogue host threat level, controlling the access activities performable by the host on the DSD to safeguard the DSD against the host, wherein the one or more impact weights are dynamically determined based on the host type.

Abstract: Aspects of a storage device including a memory and a controller are provided. The controller may collect, by an association rule mining (ARM) model, file system data from a host file system, the file system data defining at least one attribute of a file. The controller may receive, from the host, a memory command associated with the file. The controller can associate, by the ARM model, the at least one attribute with the file. The controller may perform the memory command based on the association of the at least one attribute with the file.

Abstract: A data storage device includes a non-volatile memory device including one or more memory dies and a controller. The controller is configured to receive a pseudocode file and a search key from one or more external devices and perform an index search based on the received pseudocode and search key. The controller may further determine a data file associated with the performed index search and output the determined data file to the one or more external device.

Abstract: Methods and apparatus are provided for arbitrating access to, and usage of, various device resources of a data storage device (DSD) configured for Machine Learning with Low-Power. The data storage device may include a TinyML controller with an artificial intelligence (AI) accelerator integrated with a data storage controller on a system-on-a-chip (SoC). The device resources may be, e.g., storage resources such as random access memory (RAM) devices, non-volatile memory (NVM) arrays, and latches formed on NVM dies of the NVM arrays. The resource arbitration may be based, for example, on parameters pertaining to ML operations performed by an ML controller that includes the AI accelerator, such as a turnaround time of an ML epoch or a stage-wise execution time. The resource arbitration is configured to provide for the efficient interleaving of the ML/AI operations performed by the ML controller and data storage operations performed by the data storage controller.

Abstract: Aspects of a storage device including a memory and a controller are provided. The controller may collect, by an association rule mining (ARM) model, file system data from a host file system, the file system data defining at least one attribute of a file. The controller may receive, from the host, a memory command associated with the file. The controller can associate, by the ARM model, the at least one attribute with the file. The controller may perform the memory command based on the association of the at least one attribute with the file.

Abstract: A method for securing a data storage device (DSD) against rogue behaviour by a host, the method executed by a controller of the DSD and comprising: determining a host type of the host; detecting one or more access activities performed by the host on the DSD; processing the one or more access activities to determine a security threat level of the host, wherein the security threat level is determined by weighting one or more corresponding access activity parameters by one or more impact weights; and in response to determining that the security threat level of the host is greater than or equal to a rogue host threat level, controlling the access activities performable by the host on the DSD to safeguard the DSD against the host, wherein the one or more impact weights are dynamically determined based on the host type.

Abstract: Storage devices store not only host data, but also control data related to operations of the storage device associated with the data. Control data is stored within blocks of memory called control blocks. As storage devices are configured to communicate with numerous hosts and namespaces, proper management of the available control blocks can avoid numerous problems such as mixed control block usage that punish certain hosts based on actions of other, non-related hosts. Methods and devices that efficiently manage control blocks include a control block management logic configured to process a request to generate a new namespace and parse the namespace request to determine the number of control blocks required. These separate control blocks for the requested namespace can be partitioned. Finally, control block activity for the requested namespace is directed only to the partitioned control blocks. Thus, the various control block activity between hosts is separate and not mixed.

Abstract: Aspects of a storage device including a memory and a controller are provided. The controller may receive a prefetch request to retrieve data for a host having a promoted stream. The controller may access a frozen time table indicating hosts for which data has been prefetched and frozen times associated with the host and other hosts. The controller can determine whether the host has a higher priority over other hosts included in the frozen time table based on corresponding frozen times and data access parameters associated with the host. The controller may determine to prefetch the data for the host in response to the prefetch request when the host has a higher priority than the other hosts. The controller can receive a host read command associated with the promoted stream from the host and provide the prefetched data to the host in response to the host read command.

Abstract: Methods and apparatus are provided for arbitrating access to, and usage of, various device resources of a data storage device (DSD) configured for Machine Learning with Low-Power. The data storage device may include a TinyML controller with an artificial intelligence (AI) accelerator integrated with a data storage controller on a system-on-a-chip (SoC). The device resources may be, e.g., storage resources such as random access memory (RAM) devices, non-volatile memory (NVM) arrays, and latches formed on NVM dies of the NVM arrays. The resource arbitration may be based, for example, on parameters pertaining to ML operations performed by an ML controller that includes the AI accelerator, such as a turnaround time of an ML epoch or a stage-wise execution time. The resource arbitration is configured to provide for the efficient interleaving of the ML/AI operations performed by the ML controller and data storage operations performed by the data storage controller.

Abstract: A data storage device includes a non-volatile memory device including one or more memory dies and a controller. The controller is configured to receive a pseudocode file and a search key from one or more external devices and perform an index search based on the received pseudocode and search key. The controller may further determine a data file associated with the performed index search and output the determined data file to the one or more external device.

Abstract: Aspects of a storage device including a memory and a controller are provided. The controller may receive a prefetch request to retrieve data for a host having a promoted stream. The controller may access a frozen time table indicating hosts for which data has been prefetched and frozen times associated with the host and other hosts. The controller can determine whether the host has a higher priority over other hosts included in the frozen time table based on corresponding frozen times and data access parameters associated with the host. The controller may determine to prefetch the data for the host in response to the prefetch request when the host has a higher priority than the other hosts. The controller can receive a host read command associated with the promoted stream from the host and provide the prefetched data to the host in response to the host read command.

Abstract: Systems and methods are disclosed for providing parallel data transfer. In certain embodiments, a data storage device includes a non-volatile memory and a controller configured to: receive a command from a host to obtain a file stored in the non-volatile memory; determine a plurality of channels available between the host and the data storage device; dynamically divide the file into a plurality of chunks based at least in part on the plurality of channels; perform load balancing to determine a first chunk of the plurality of chunks to be sent over a first channel of the plurality of channels and at least a second chunk of the plurality of chunks to be sent over a second channel of the plurality of channels; and simultaneously transmit the first chunk over the first channel and the second chunk over the second channel.

Abstract: Storage devices store not only host data, but also control data related to operations of the storage device associated with the data. Control data is stored within blocks of memory called control blocks. As storage devices are configured to communicate with numerous hosts and namespaces, proper management of the available control blocks can avoid numerous problems such as mixed control block usage that punish certain hosts based on actions of other, non-related hosts. Methods and devices that efficiently manage control blocks include a control block management logic configured to process a request to generate a new namespace and parse the namespace request to determine the number of control blocks required. These separate control blocks for the requested namespace can be partitioned. Finally, control block activity for the requested namespace is directed only to the partitioned control blocks. Thus, the various control block activity between hosts is separate and not mixed.

Abstract: Aspects of a storage device including a memory and a controller are provided which reduces or eliminates garbage collection in zoned namespace (ZNS) architectures by mapping zones to sub-blocks of blocks of the memory. Each zone includes a plurality of logical addresses. The controller determines a number of open zones, and maps the open zones to the sub-blocks in response to the number of open zones meeting a threshold. Thus, larger numbers of open blocks typically present in ZNS may be reduced, and increased block sizes due to scaling may be accommodated in ZNS. In some aspects, the controller receives a request from a host device to write data associated with the zones in sub-blocks, and maps each of the zones to at least one of the sub-blocks in response to the request. The request may indicate zones are partially unused. Thus, out of zone conditions may also be avoided.

Abstract: Systems and methods are disclosed for providing parallel data transfer. In certain embodiments, a data storage device includes a non-volatile memory and a controller configured to: receive a command from a host to obtain a file stored in the non-volatile memory; determine a plurality of channels available between the host and the data storage device; dynamically divide the file into a plurality of chunks based at least in part on the plurality of channels; perform load balancing to determine a first chunk of the plurality of chunks to be sent over a first channel of the plurality of channels and at least a second chunk of the plurality of chunks to be sent over a second channel of the plurality of channels; and simultaneously transmit the first chunk over the first channel and the second chunk over the second channel.

Abstract: Aspects of a storage device including a memory and a controller are provided which reduces or eliminates garbage collection in zoned namespace (ZNS) architectures by mapping zones to sub-blocks of blocks of the memory. Each zone includes a plurality of logical addresses. The controller determines a number of open zones, and maps the open zones to the sub-blocks in response to the number of open zones meeting a threshold. Thus, larger numbers of open blocks typically present in ZNS may be reduced, and increased block sizes due to scaling may be accommodated in ZNS. In some aspects, the controller receives a request from a host device to write data associated with the zones in sub-blocks, and maps each of the zones to at least one of the sub-blocks in response to the request. The request may indicate zones are partially unused. Thus, out of zone conditions may also be avoided.