International Business Machines Corporation

Lincoln T Simmons

Timothy J Fisher

Aaron Daniel Fry

Charles John Camp

Roman Alexander Pletka

Nikolas Ioannou

Sasa Tomic

Adalberto Guillermo Yanes

Jason Szecheong Ma

Andrew Dale Walls

Ioannis Koltsidas

Gary A Tressler

Kevin E Sallese

Franck Excoffier

David A Pierce

Adam C Chunn

Razik S Ahmed

Matthew R Orr

Ashwitha Krishna Kumar

Samuel K Ingram

A technique for garbage collection in a storage system includes generating regrouping metadata for one or more pages of at least two logical erase blocks (LEB). The regrouping metadata indicates an associated stream for each of the pages. Multiple of the LEBs that include valid pages associated with a first stream are selected, based on the regrouping metadata, for regrouping. The valid pages associated with the first stream from the selected LEBs are regrouped into a new LEB.

Regrouping data during relocation to facilitate write amplification reduction

A controller-implemented method, according to one embodiment, includes: examining, by the controller, each of a plurality of journal entries from at least one journal beginning with a most recent one of the journal entries in a most recent one of the at least one journal and working towards an oldest one of the journal entries in an oldest one of the at least one journal, the journal entries corresponding to one or more updates made to one or more logical to physical table (LPT) entries of a LPT; determining, by the controller, whether a current LPT entry, which corresponds to a currently examined journal entry, has already been updated; and discarding, by the controller, the currently examined journal entry in response to determining that the current LPT entry has already been updated.

Logical to physical table restoration from stored journal entries

In at least one embodiment, a controller of a non-volatile memory array iteratively performs a merged background management process independently of any host system's demand requests targeting the memory array. During an iteration of the merged background management process, the controller performs a read sweep by reading data from each of a plurality of page groups within the memory array and recording page group error statistics regarding errors detected by the reading for each page group, where each page group is formed of a respective set of one or more physical pages of storage in the memory array. During the iteration of the merged background management process, the controller employs the page group error statistics recorded during the read sweep in another background management function.

Storage array management employing a merged background management process

According to one embodiment, a method includes issuing a read request to read one or more units of data from at least one non-volatile random access memory (NVRAM) device. The read request includes one or more read voltage thresholds. The method also includes receiving the one or more data units and read command parameters used to read the one or more data units from the at least one NVRAM device. Moreover, the method includes storing error-free data units, the read command parameters used to read the error-free data units from the at least one NVRAM device, and a read completion status to one of a plurality of read buffers. The read completion status indicates a completed read when a data unit is error-free and indicates an incomplete read when a data unit is errored.

Read buffer architecture supporting integrated XOR-reconstructed and read-retry for non-volatile random access memory (NVRAM) systems

In at least one embodiment, a controller of a non-volatile memory array including a plurality of subdivisions stores write data within the non-volatile memory array utilizing a plurality of block stripes of differing numbers of blocks, where all of the blocks within each block stripe are drawn from different ones of the plurality of subdivisions. The controller builds new block stripes for storing write data from blocks selected based on estimated remaining endurances of blocks in each of the plurality of subdivisions.

Wear leveling of a memory array

A method, according to one embodiment, includes: receiving a recirculation command; performing a coarse page lookup to determine valid ones of logical pages to be recirculated; requesting performance of a fine page lookup on source physical addresses containing the valid logical pages to verify the valid logical pages; and sending write commands corresponding to verified valid logical pages from the fine page lookup. Other systems, methods, and computer program products are described in additional embodiments.

Accelerated non-volatile memory recirculation processing

A flash memory codeword architecture is provided. A non-integer count of logical pages is packed into a codeword payload data container. A codeword payload header is generated. The codeword payload header includes an offset to a first logical page that is packed, at least in part, into the codeword payload data container. The codeword payload data container and the codeword payload header are concatenated to generate a codeword payload. Error-correcting code data is generated based, at least in part, on the codeword payload using a systematic error-correcting code. The codeword payload and error-correcting code data is concatenated to generate a codeword. A physical page is programmed with the codeword.

Validation bits and offsets to represent logical pages split between data containers

A controller-implemented method, according to one embodiment, includes: restoring a valid snapshot of a LPT from the non-volatile random access memory, examining each journal entry from at least one journal beginning with a most recent one of the journal entries in a most recent one of the at least one journal and working towards an oldest one of the journal entries in an oldest one of the at least one journal, the journal entries corresponding to updates made to one or more entries of the LPT, determining whether a current LPT entry which corresponds to a currently examined journal entry has already been updated, using the currently examined journal entry to update the current LPT entry in response to determining that the current LPT entry has not already been updated, and discarding the currently examined journal entry in response to determining that the current LPT entry has already been updated.

A technique for garbage collection in a data storage system includes determining a dirty physical byte count for each of a plurality of candidate garbage collection units. The dirty physical byte count provides a total amount of dirty bytes. At least one of a dirty physical codeword container count and a dirty physical page count is determined for each of the candidate garbage collection units. The dirty physical codeword container count provides an amount of physical codeword containers that are completely dirty and the dirty physical page count provides an amount of physical pages that are completely dirty. A garbage collection unit, included in the candidate garbage collection units, is selected for garbage collection based on the dirty physical byte count and at least one of the dirty physical codeword container count and the dirty physical page count.

Garbage collection techniques for a data storage system

A method, according to one embodiment, includes: receiving a recirculation command, performing a coarse page lookup to determine valid ones of logical pages to be recirculated, issuing write commands for the valid logical pages, requesting performance of a fine page lookup on source physical addresses containing the valid logical pages to verify the valid logical pages, receiving verified valid logical pages resulting from the fine page lookup, and sending the write commands corresponding to the verified valid logical pages. Other systems, methods, and computer program products are described in additional embodiments.

A computer processor determines an over-provisioning ratio and a host write pattern. The computer processor determines a write amplification target based on the host write pattern and the over-provisioning ratio. The computer processor determines a staleness threshold, wherein the staleness threshold corresponds to a ratio of valid pages of a block to total pages of the block. The computer processor erases a first block having a staleness which exceeds the staleness threshold.

Garbage collection scaling

A data retention methodology for use in electrically rewritable nonvolatile storage systems is disclosed. The methodology collects characteristic data of the storage system (e.g., time stamps, program/erase cycles, sensed temperature over time, and others) and uses at least a portion of that data to associate various data retention schemes with the collected characteristic data. At power-on, the methodology determines a duration of a time during which power was not supplied to the storage system. The methodology also uses the power-on time to trigger the selection of an appropriate data protection scheme (e.g., enhanced ECC and/or read/scrub frequency). Dynamic selection and adjustment of the applied protection scheme may be based on the predetermined and/or calculated association between various protection schemes and the collected characteristic data of the storage system.

Growing community of inventors

Houston, TX, United States of America

Lincoln T Simmons