# US 2025/0147671 A1 · Solid-state disks wear leveling with predictive migration of data devices in a raid container system > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2025/0147671 A1 - Document ID: US20250147671A1 - Status: Published application - Kind: A1 - Application number: 18/502,147 - Filed: 2023-11-06 - Published: 2025-05-08 - Earliest claimed: 2023-11-06 - Assignees: Dell Products L.P. - Applicants: Dell Products L.P. - Inventors: Kuolin Hua; Kunxiu Gao; Malak Alshawabkeh - Web page: https://patentgrep.com/patent/US20250147671A1 ## CPC - G06F 3/0616: {in relation to life time, e.g. increasing Mean Time Between Failures [MTBF]} - G06F 3/0644: {Management of space entities, e.g. partitions, extents, pools} - G06F 3/0647: {Migration mechanisms} - G06F 3/0688: {Non-volatile semiconductor memory arrays} - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract A container system facilitates selective migrations of data from high-wear inducing data devices to spare data devices to promote SSD wear levelling. Storage capacity is configured into same-size cells with each container having the same number of cells as RAID width W. The cells are distributed over W SSDs. The containers are subdivided into equal-size data devices, each distributed over the same set of W SSDs as its associated container. Candidate migration source-target pairs are selected using heuristics guided by a prediction model that correlates disk wear rate with total writes of existing workloads. Wear rate predictions are computed using a weighted sum of write rates of all data devices on a disk. Wear rate predictions for candidate migration pairs are computed inclusive of additional wear caused by data migration to select a migration plan. ## Claims ### Claim 1 A method comprising: configuring storage capacity of an array of solid-state drives (SSDs) by: creating equal-size redundant array of independent disk (RAID) containers that each have W same-size cells corresponding to a RAID width W; distributing the W cells per container over W SSDs; subdividing each of the containers into equal-size data devices that are logical disk drives that exhibit non-linear wear rates that increase and decrease, with a non-zero number of empty data devices reserved as spare storage capacity for RAID failure recovery; and distributing each data device over W SSDs; and computing and implementing SSD wear leveling by: computing predicted future wear rates of each of the data devices using time-series forecasting based on Support Vector Regression (SVR); and migrating data from high-wear rate data devices on high-wear level SSDs to empty data devices on SSDs that are not high-wear level. ### Claim 2 (depends on 1) The method of claim 1 further comprising computing a predicted wear rate of each of the SSDs, where wear rates increase and decrease and indicate change in wear level over time. ### Claim 3 (depends on 2) The method of claim 2 further comprising computing the predicted wear rate of each of the SSDs as a weighted sum of all write rates of all data devices on the SSD. ### Claim 4 (depends on 3) The method of claim 3 further comprising performing computations using non-linear kernel SVR. ### Claim 5 (depends on 3) The method of claim 3 further comprising computing heuristic candidate migration plans for pairings of high-wear rate data devices on high-wear level SSDs and empty data devices on SSDs that are not high-wear level. ### Claim 6 (depends on 5) The method of claim 5 further comprising computing predicted wear rate of each SSD resulting from each plan, inclusive of the data movement required to implement the plan, and computing and implementing the plan characterized by minimal maximum SSD wear rate inclusive of the data movement required to implement the plan. ### Claim 7 (depends on 5) The method of claim 5 further comprising excluding migrated data device data from being migrated again for a predetermined time period. ### Claim 8 An apparatus comprising: a data storage node in which storage capacity of an array of solid-state drives (SSDs) is characterized by equal-size redundant array of independent disk (RAID) containers that each have W same-size cells corresponding to a RAID width W, the W cells per container distributed over W SSDs, each of the containers subdivided into equal-size data devices that are logical disk drives that exhibit non-linear wear rates that increase and decrease, with a non-zero number of empty data devices reserved as spare storage capacity for RAID failure recovery, and each data device distributed over W SSDs; and an SSD wear control program configured to compute and implement SSD wear leveling by: computing predicted future wear rates of each of the data devices using time-series forecasting based on Support Vector Regression (SVR); and causing data to migrate from high-wear rate data devices on high-wear level SSDs to empty data devices on SSDs that are not high-wear level. ### Claim 9 (depends on 8) The apparatus of claim 8 further comprising the wear control program configured to compute a predicted wear rate of each of the SSDs, where wear rates increase and decrease and indicate change in wear level over time. ### Claim 10 (depends on 9) The apparatus of claim 9 further comprising the wear control program configured to compute the predicted wear rate of each of the SSDs as a weighted sum of all write rates of all data devices on the SSD. ### Claim 11 (depends on 10) The apparatus of claim 10 further comprising the wear control program configured to compute predicted wear rates using non-linear kernel SVR. ### Claim 12 (depends on 10) The apparatus of claim 10 further comprising the wear control program configured to compute heuristic candidate migration plans for pairings of high-wear rate data devices on high-wear level SSDs and empty data devices on SSDs that are not high-wear level. ### Claim 13 (depends on 12) The apparatus of claim 12 further comprising the wear control program configured to compute predicted wear rate of each SSD resulting from each plan, inclusive of the data movement required to implement the plan, and compute and implement the plan characterized by minimal maximum SSD wear rate inclusive of the data movement required to implement the plan. ### Claim 14 (depends on 13) The apparatus of claim 13 further comprising the wear control program configured to exclude migrated data device data from being migrated again for a predetermined time period. ### Claim 15 A non-transitory computer-readable storage medium stores instructions that when executed by one or more computers cause the computers to perform a method comprising: configuring storage capacity of an array of solid-state drives (SSDs) by: creating equal-size redundant array of independent disk (RAID) containers that each have W same-size cells corresponding to a RAID width W; distributing the W cells per container over W SSDs; subdividing each of the containers into equal-size data devices that are logical disk drives that exhibit non-linear wear rates that increase and decrease, with a non-zero number of empty data devices reserved as spare storage capacity for RAID failure recovery; and distributing each data device over W SSDs; and computing and implementing SSD wear leveling by: computing predicted future wear rates of each of the data devices using time-series forecasting based on Support Vector Regression (SVR); and migrating data from high-wear rate data devices on high-wear level SSDs to empty data devices on SSDs that are not high-wear level. ### Claim 16 (depends on 15) The non-transitory computer-readable storage medium of claim 15 in which the method further comprises computing a predicted wear rate of each of the SSDs, where wear rates increase and decrease and indicate change in wear level over time. ### Claim 17 (depends on 16) The non-transitory computer-readable storage medium of claim 16 in which the method further comprises computing the predicted wear rate of each of the SSDs as a weighted sum of all write rates of all data devices on the SSD. ### Claim 18 (depends on 17) The non-transitory computer-readable storage medium of claim 17 in which the method further comprises performing computations using non-linear kernel SVR. ### Claim 19 (depends on 17) The non-transitory computer-readable storage medium of claim 17 in which the method further comprises computing heuristic candidate migration plans for pairings of high-wear rate data devices on high-wear level SSDs and empty data devices on SSDs that are not high-wear level. ### Claim 20 (depends on 19) The non-transitory computer-readable storage medium of claim 19 in which the method further comprises computing predicted wear rate of each SSD resulting from each plan, inclusive of the data movement required to implement the plan, and computing and implementing the plan characterized by minimal maximum SSD wear rate inclusive of the data movement required to implement the plan.