# US 2025/0147682 A1 · Random Storage Access and Data Erasure for Improved Performance and Reduced Write Amplification > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2025/0147682 A1 - Document ID: US20250147682A1 - Status: Published application - Kind: A1 - Application number: 19/014,020 - Filed: 2025-01-08 - Published: 2025-05-08 - Earliest claimed: 2022-07-20 - Assignees: Micron Technology, Inc. - Applicants: Micron Technology, Inc. - Inventors: Luca Bert - Web page: https://patentgrep.com/patent/US20250147682A1 ## CPC - G06F 3/064: {Management of blocks} - G06F 3/061: {Improving I/O performance} - G06F 3/0656: {Data buffering arrangements} - G06F 3/0659: {Command handling arrangements, e.g. command buffers, queues, command scheduling} - G06F 3/0604: {Improving or facilitating administration, e.g. storage management} - G06F 3/0679: {Non-volatile semiconductor memory device, e.g. flash memory, one time programmable memory [OTP]} - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract A host system to query, during booting up of the host system, a superblock size in a connected memory sub-system. The host system can place write requests into separate streams and send the streams to the memory sub-system to store data of the write requests into separate sets of superblocks for the streams respectively. The host system can allocate, a plurality of log buffers for the streams respectively and record, into the log buffers, sequences of logical addresses as in the streams respectively. The host system can trim a stream, among the plurality of streams, by issuing commands to the memory sub-system to erase, according to the superblock size, an amount of data from a portion of a sequence of logical addresses recorded in a log buffer for the stream, causing the memory sub-system to free at least one superblock. ## Claims ### Claim 1 A device, comprising: memory cells configured as a plurality of superblocks; and a controller configured to receive requests to write data to a storage space provided by the memory cells, the requests configured as a plurality of streams; and wherein the controller is configured to aggregate placement of data of different streams into different superblocks and configured not to reclaim a superblock that is to be partially erased according to commands received in the device. ### Claim 2 (depends on 1) The device of claim 1, wherein the controller is further configured to: receive, during booting up of a system containing the device, a query; and providing, in response to the query, a superblock size of the plurality of superblocks. ### Claim 3 (depends on 2) The device of claim 2, wherein the controller is further configured to: maintain a plurality of stream cursors for the plurality of different streams respectively. ### Claim 4 (depends on 3) The device of claim 3, wherein each of the plurality of stream cursors is configured to use a different superblock among the plurality of superblocks. ### Claim 5 (depends on 4) The device of claim 4, wherein the controller is further configured to: allocate storage spaces for writing by the plurality of streams according to the plurality of stream cursors. ### Claim 6 (depends on 5) The device of claim 5, wherein each respective stream cursor, among the plurality of stream cursors is configured to identify a next free space in a respective superblock to host a logical block address to be written. ### Claim 7 (depends on 6) The device of claim 6, wherein the controller is further configured to: receive a request to trim a stream, among the plurality of streams; identify at least one first superblock that is full and that stores data for the stream; and erase the at least one first superblock; wherein the stream contains data in a second superblock that is not full; and wherein the controller is configured to exclude the second superblock from being reclaimed in responding to the request. ### Claim 8 A method, comprising: configuring memory cells in a device as a plurality of superblocks; receiving, in the device, requests to write data to a storage space provided by the memory cells, the requests configured as a plurality of streams; and aggregating, by the device, placement of data of different streams into different superblocks, wherein the device is configured not to reclaim a superblock that is to be partially erased according to commands received in the device. ### Claim 9 (depends on 8) The method of claim 8, further comprising: receive, during booting up of a system containing the device, a query; and providing, in response to the query, a superblock size of the plurality of superblocks. ### Claim 10 (depends on 8) The method of claim 8, further comprising: maintaining, in the device, a plurality of stream cursors for the plurality of different streams respectively. ### Claim 11 (depends on 10) The method of claim 10, wherein each of the plurality of stream cursors is configured to identify a different superblock among the plurality of superblocks. ### Claim 12 (depends on 11) The method of claim 11, further comprising: allocating storage spaces for writing by the plurality of streams according to the plurality of stream cursors. ### Claim 13 (depends on 12) The method of claim 12, wherein each respective stream cursor, among the plurality of stream cursors is configured to identify a next free space in a respective superblock to host a logical block address to be written. ### Claim 14 (depends on 13) The method of claim 13, further comprising: receiving, in the device, a request to trim a stream, among the plurality of streams, wherein the stream contains data stored in at least one first superblock that is full, and wherein the stream further contains data stored in a second superblock that is not full; identifying at least one first superblock; erasing the at least one first superblock; and excluding the second superblock from being reclaimed during responding to the request. ### Claim 15 A non-transitory computer storage medium storing instructions which, when executed by a device, cause the device to perform a method, comprising: configuring memory cells in the device as a plurality of superblocks; receiving, in the device, requests to write data to a storage space provided by the memory cells, the requests configured as a plurality of streams; and aggregating, by the device, placement of data of different streams into different superblocks, wherein the device is configured not to reclaim a superblock that is to be partially erased according to commands received in the device. ### Claim 16 (depends on 15) The non-transitory computer storage medium of claim 15, wherein the method further comprises: receive, during booting up of a system containing the device, a query; and providing, in response to the query, a superblock size of the plurality of superblocks. ### Claim 17 (depends on 15) The non-transitory computer storage medium of claim 15, wherein the method further comprises: maintaining, in the device, a plurality of stream cursors for the plurality of different streams respectively; wherein each of the plurality of stream cursors is configured to identify a different superblock among the plurality of superblocks. ### Claim 18 (depends on 17) The non-transitory computer storage medium of claim 17, wherein the method further comprises: allocating storage spaces for writing by the plurality of streams according to the plurality of stream cursors. ### Claim 19 (depends on 18) The non-transitory computer storage medium of claim 18, wherein each respective stream cursor, among the plurality of stream cursors is configured to identify a next free space in a respective superblock to host a logical block address to be written. ### Claim 20 (depends on 19) The non-transitory computer storage medium of claim 19, wherein the method further comprises: receiving, in the device, a request to trim a stream, among the plurality of streams, wherein the stream contains data stored in at least one first superblock that is full, and wherein the stream further contains data stored in a second superblock that is not full; identifying at least one first superblock; erasing the at least one first superblock; and excluding the second superblock from being reclaimed during responding to the request.