Abstract: Methods, non-transitory machine readable media, and computing devices that manage storage operations directed to dual-port solid state disks (SSDs) coupled to multiple hosts are disclosed. With this technology, context metadata comprising a checksum is retrieved based on a first physical address mapped, in a cached zoned namespace (ZNS) mapping table, to a logical address. The logical address is extracted from a request to read a portion of a file. A determination is made when the checksum is valid based on a comparison to identification information extracted from the request and associated with the file portion. At least the first physical address is replaced in the cached ZNS mapping table with a second physical address retrieved from an on-disk ZNS mapping table, when the determination indicates the checksum is invalid. The file portion retrieved from a dual-port SSD using the second physical address is returned to service the request.

Abstract: Methods, non-transitory machine-readable media, and computing devices that manage resources between multiple hosts coupled to dual-port solid-state disks (SSDs) are disclosed. With this technology, in-core conventional namespace (CNS) and zoned namespace (ZNS) mapping tables are synchronized by a host flash translation layer with on-disk CNS and ZNS mapping tables, respectively. An entry in one of the in-core CNS or ZNS mapping tables is identified based on whether a received storage operation is directed to a CNS or a ZNS of the dual-port SSD. The entry is further identified based on a logical address extracted from the storage operation. The storage operation is serviced using a translation in the identified entry for the logical address, when the storage operation is directed to the CNS, or a zone identifier in the identified entry for a zone of the ZNS, when the storage operation is directed to the ZNS.

Abstract: Methods, non-transitory machine readable media, and computing devices that manage resources between multiple hosts coupled to dual-port solid-state disks (SSDs) are disclosed. With this technology, in-core conventional namespace (CNS) and zoned namespace (ZNS) mapping tables are synchronized by a host flash translation layer with on-disk CNS and ZNS mapping tables, respectively. An entry in one of the in-core CNS or ZNS mapping tables is identified based on whether a received storage operation is directed to a CNS or a ZNS of the dual-port SSD. The entry is further identified based on a logical address extracted from the storage operation. The storage operation is serviced using a translation in the identified entry for the logical address, when the storage operation is directed to the CNS, or a zone identifier in the identified entry for a zone of the ZNS, when the storage operation is directed to the ZNS.

Abstract: Methods, non-transitory machine readable media, and computing devices that manage storage operations directed to dual-port solid state disks (SSDs) coupled to multiple hosts are disclosed. With this technology, context metadata comprising a checksum is retrieved based on a first physical address mapped, in a cached zoned namespace (ZNS) mapping table, to a logical address. The logical address is extracted from a request to read a portion of a file. A determination is made when the checksum is valid based on a comparison to identification information extracted from the request and associated with the file portion. At least the first physical address is replaced in the cached ZNS mapping table with a second physical address retrieved from an on-disk ZNS mapping table, when the determination indicates the checksum is invalid. The file portion retrieved from a dual-port SSD using the second physical address is returned to service the request.

Abstract: Methods, non-transitory machine readable media, and computing devices that manage storage operations directed to dual-port solid state disks (SSDs) coupled to multiple hosts are disclosed. With this technology, context metadata comprising a checksum is retrieved based on a first physical address mapped, in a cached zoned namespace (ZNS) mapping table, to a logical address. The logical address is extracted from a request to read a portion of a file. A determination is made when the checksum is valid based on a comparison to identification information extracted from the request and associated with the file portion. At least the first physical address is replaced in the cached ZNS mapping table with a second physical address retrieved from an on-disk ZNS mapping table, when the determination indicates the checksum is invalid. The file portion retrieved from a dual-port SSD using the second physical address is returned to service the request.

Abstract: Methods, non-transitory machine readable media, and computing devices that manage resources between multiple hosts coupled to dual-port solid-state disks (SSDs) are disclosed. With this technology, in-core conventional namespace (CNS) and zoned namespace (ZNS) mapping tables are synchronized by a host flash translation layer with on-disk CNS and ZNS mapping tables, respectively. An entry in one of the in-core CNS or ZNS mapping tables is identified based on whether a received storage operation is directed to a CNS or a ZNS of the dual-port SSD. The entry is further identified based on a logical address extracted from the storage operation. The storage operation is serviced using a translation in the identified entry for the logical address, when the storage operation is directed to the CNS, or a zone identifier in the identified entry for a zone of the ZNS, when the storage operation is directed to the ZNS.

Abstract: Methods, non-transitory machine readable media, and computing devices that manage resources between multiple hosts coupled to dual-port solid-state disks (SSDs) are disclosed. With this technology, in-core conventional namespace (CNS) and zoned namespace (ZNS) mapping tables are synchronized by a host flash translation layer with on-disk CNS and ZNS mapping tables, respectively. An entry in one of the in-core CNS or ZNS mapping tables is identified based on whether a received storage operation is directed to a CNS or a ZNS of the dual-port SSD. The entry is further identified based on a logical address extracted from the storage operation. The storage operation is serviced using a translation in the identified entry for the logical address, when the storage operation is directed to the CNS, or a zone identifier in the identified entry for a zone of the ZNS, when the storage operation is directed to the ZNS.

Abstract: Methods, non-transitory machine readable media, and computing devices that manage storage operations directed to dual-port solid state disks (SSDs) coupled to multiple hosts are disclosed. With this technology, context metadata comprising a checksum is retrieved based on a first physical address mapped, in a cached zoned namespace (ZNS) mapping table, to a logical address. The logical address is extracted from a request to read a portion of a file. A determination is made when the checksum is valid based on a comparison to identification information extracted from the request and associated with the file portion. At least the first physical address is replaced in the cached ZNS mapping table with a second physical address retrieved from an on-disk ZNS mapping table, when the determination indicates the checksum is invalid. The file portion retrieved from a dual-port SSD using the second physical address is returned to service the request.

Abstract: Methods, non-transitory machine readable media, and computing devices that manage storage operations directed to dual-port solid state disks (SSDs) coupled to multiple hosts are disclosed. With this technology, context metadata comprising a checksum is retrieved based on a first physical address mapped, in a cached zoned namespace (ZNS) mapping table, to a logical address. The logical address is extracted from a request to read a portion of a file. A determination is made when the checksum is valid based on a comparison to identification information extracted from the request and associated with the file portion. At least the first physical address is replaced in the cached ZNS mapping table with a second physical address retrieved from an on-disk ZNS mapping table, when the determination indicates the checksum is invalid. The file portion retrieved from a dual-port SSD using the second physical address is returned to service the request.

Abstract: Methods, non-transitory machine readable media, and computing devices that manage resources between multiple hosts coupled to dual-port solid-state disks (SSDs) are disclosed. With this technology, in-core conventional namespace (CNS) and zoned namespace (ZNS) mapping tables are synchronized by a host flash translation layer with on-disk CNS and ZNS mapping tables, respectively. An entry in one of the in-core CNS or ZNS mapping tables is identified based on whether a received storage operation is directed to a CNS or a ZNS of the dual-port SSD. The entry is further identified based on a logical address extracted from the storage operation. The storage operation is serviced using a translation in the identified entry for the logical address, when the storage operation is directed to the CNS, or a zone identifier in the identified entry for a zone of the ZNS, when the storage operation is directed to the ZNS.

Abstract: In a data storage system, the available space of a magnetic storage device is divided into multiple sequential write regions for storing sequentially written data, where the regions are each separated from adjacent sequential write regions by a guard space. Object data, such as key-value pairs, are written sequentially to a particular sequential write region, in blocks of data that correspond to the contents of a nonvolatile buffer being flushed to the magnetic storage device. When a key-value pair stored in the magnetic storage device is subsequently updated, the original key-value pair is not overwritten. Instead, the new version of the key-value pair is included in the next block of data to be written to the magnetic storage device, and a mapping table tracks the location of the newest version of each key-value pair stored in the magnetic storage device.

Abstract: Methods, non-transitory machine readable media, and computing devices that facilitate efficient storage operations using host-managed solid-state disks (SSDs) are disclosed. With this technology, a direct memory access (DMA) transfer is initiated of a data block from a location indicated in an application write request to a write buffer in a device memory of an SSD. A determination is made when write rule(s) are satisfied based on content of the write buffer including at least the data block and other data block(s) previously transferred to the write buffer. A copy request is issued to transfer a portion of the content to flash media of the SSD, when the write rule(s) are satisfied. This technology does not require host memory for write buffering or processor cycles for copying data from application data buffers to a write buffer in host memory, and thereby significantly improves resource utilization of host devices managing SSDs.

Abstract: Methods, non-transitory machine readable media, and computing devices that facilitate efficient storage operations using host-managed solid-state disks (SSDs) are disclosed. With this technology, a direct memory access (DMA) transfer is initiated of a data block from a location indicated in an application write request to a write buffer in a device memory of an SSD. A determination is made when write rule(s) are satisfied based on content of the write buffer including at least the data block and other data block(s) previously transferred to the write buffer. A copy request is issued to transfer a portion of the content to flash media of the SSD, when the write rule(s) are satisfied. This technology does not require host memory for write buffering or processor cycles for copying data from application data buffers to a write buffer in host memory, and thereby significantly improves resource utilization of host devices managing SSDs.

Abstract: A solid-state drive includes a flash memory device, a power loss protection circuit, a dynamic random access memory (RAM) coupled to the power loss protection circuit, and a controller configured to direct I/O requests to either the flash memory drive or the RAM. Because the controller can direct I/O request to the RAM, the RAM is revealed as a separate mass storage device to a host. Consequently, the RAM provides additional and significantly higher performance storage capacity to the solid-state drive.

Abstract: Object data, such as a key-value pair, are stored in a disk drive in conjunction with metadata associated with the object data. A key-value pair and metadata associated therewith are written in different locations in the disk drive, but as part of a single sequential write operation, such as when contents of a key-value buffer containing one or more key-value pairs are flushed to the disk drive. The key-value pair may be written during a first phase of the sequential write operation and the metadata may be written during a second phase of the sequential write operation that does not immediately precede or follow the first phase.

Abstract: In a data storage system, the available space of a magnetic storage device is divided into multiple sequential write regions for storing sequentially written data, where the regions are each separated from adjacent sequential write regions by a guard space. Object data, such as key-value pairs, are written sequentially to a particular sequential write region, in blocks of data that correspond to the contents of a nonvolatile buffer being flushed to the magnetic storage device. When a key-value pair stored in the magnetic storage device is subsequently updated, the original key-value pair is not overwritten. Instead, the new version of the key-value pair is included in the next block of data to be written to the magnetic storage device, and a mapping table tracks the location of the newest version of each key-value pair stored in the magnetic storage device.

Abstract: In a data storage system, the available space of a magnetic storage device is divided into multiple sequential write regions for storing sequentially written data, where the regions are each separated from adjacent sequential write regions by a guard space. Object data, such as key-value pairs, are written sequentially to a particular sequential write region, in blocks of data that correspond to the contents of a nonvolatile buffer being flushed to the magnetic storage device. When a key-value pair stored in the magnetic storage device is subsequently updated, the original key-value pair is not overwritten. Instead, the new version of the key-value pair is included in the next block of data to be written to the magnetic storage device, and a mapping table tracks the location of the newest version of each key-value pair stored in the magnetic storage device.

Abstract: A solid-state drive includes a flash memory device, a power loss protection circuit, a dynamic random access memory (RAM) coupled to the power loss protection circuit, and a controller configured to direct I/O requests to either the flash memory drive or the RAM. Because the controller can direct I/O request to the RAM, the RAM is revealed as a separate mass storage device to a host. Consequently, the RAM provides additional and significantly higher performance storage capacity to the solid-state drive.

Abstract: In a data storage system, the available space of a magnetic storage device is divided into multiple sequential write regions for storing sequentially written data, where the regions are each separated from adjacent sequential write regions by a guard space. Object data, such as key-value pairs, are written sequentially to a particular sequential write region, in blocks of data that correspond to the contents of a nonvolatile buffer being flushed to the magnetic storage device. When a key-value pair stored in the magnetic storage device is subsequently updated, the original key-value pair is not overwritten. Instead, the new version of the key-value pair is included in the next block of data to be written to the magnetic storage device, and a mapping table tracks the location of the newest version of each key-value pair stored in the magnetic storage device.

Abstract: Object data, such as a key-value pair, are stored in a disk drive in conjunction with metadata associated with the object data. A key-value pair and metadata associated therewith are written in different locations in the disk drive, but as part of a single sequential write operation, such as when contents of a key-value buffer containing one or more key-value pairs are flushed to the disk drive. The key-value pair may be written during a first phase of the sequential write operation and the metadata may be written during a second phase of the sequential write operation that does not immediately precede or follow the first phase.