STORAGE DEVICE READ-DISTURB-BASED READ TEMPERATURE MAP UTILIZATION SYSTEM

Abstract

A storage device read-disturb-based read temperature map utilization system includes a storage device chassis housing a storage subsystem. A local read temperature utilization subsystem in the storage device chassis determines read disturb information for a plurality of blocks in the storage subsystem, uses it to identify a subset of rows in block(s) in the storage subsystem that have a relatively higher read temperature and, based on those read temperature identifications, generates a local logical storage element read temperature map that identifies a subset of logical storage elements associated with the storage subsystem that have a relatively higher read temperature. The local read temperature utilization subsystem then moves data from first block(s) in the storage subsystem to second block(s) in the storage subsystem based on relative read temperatures identified in the local logical storage element read temperature map.

Description

BACKGROUND

The present disclosure relates generally to information handling systems, and more particularly to a storage device in an information handling system utilizing a logical storage location read-disturb-based read temperature map generated by that storage device.

As the value and use of information continues to increase, individuals and businesses seek additional ways to process and store information. One option available to users is information handling systems. An information handling system generally processes, compiles, stores, and/or communicates information or data for business, personal, or other purposes thereby allowing users to take advantage of the value of the information. Because technology and information handling needs and requirements vary between different users or applications, information handling systems may also vary regarding what information is handled, how the information is handled, how much information is processed, stored, or communicated, and how quickly and efficiently the information may be processed, stored, or communicated. The variations in information handling systems allow for information handling systems to be general or configured for a specific user or specific use such as financial transaction processing, airline reservations, enterprise data storage, or global communications. In addition, information handling systems may include a variety of hardware and software components that may be configured to process, store, and communicate information and may include one or more computer systems, data storage systems, and networking systems.

Information handling systems such as, for example, server devices and/or storage systems, and/or other computing devices known in the art, includes storage systems having one or more storage devices (e.g., Solid State Drive (SSD) storage devices) for storing data generated by the computing device. In some situations, it may be desirable to identify how often logical storage locations associated with any storage device are read. For example, different storage devices with different capabilities are associated with different costs (e.g., storage devices with relatively “higher” capabilities are more expensive than storage devices with relative “lower” capabilities), while different data stored in a storage system may have different characteristics, with some data being read relatively often (also referred to as data having a relatively “hot” read temperature) and other data being read relatively less often (also referred to as data having a relatively “cold” read temperature). As will be appreciated by one of skill in the art in possession of the present disclosure, financial margins of storage providers (e.g., entities that provide storage for customers) may be improved by offering particular storage Service Level Agreements (SLAs) while using relatively cheaper storage devices, and value can be passed on to customers by providing improved storage SLAs for data with relatively “hot” read temperatures without incurring higher costs for all storage devices in the storage system (e.g., by storing data with relatively “hot” read temperatures on relatively higher capability/cost storage devices, and storing data with relatively “cold” read temperatures on relatively lower capability/cost storage devices).

Conventional read temperature identification systems typically utilize a host processor (or a storage processor) and a host memory in a server device and/or storage system to identify read temperatures of logical storage locations in SSD storage device(s) included in, connected to, and/or otherwise coupled to that server device and/or storage system. For example, a Logical Block Address (LBA) range may be divided into smaller address ranges or logical “chunks” (e.g., 128KB chunks). A counter (e.g., a Dynamic Random Access Memory (DRAM) counter) in the host memory may then be assigned to track read access to each logical chunk, and when the host processor performs read operations to read data from each of the SSD storage device(s) in the server device and/or storage system, the host processor will map the LBA range of that read operation to the corresponding logical chunk(s) being read, and increment the counter(s) for those physical storage element chunk(s) in the host memory. However, such conventional read temperature identification systems suffer from a number of issues.

For example, the conventional read temperature identification systems discussed above require dedicated host memory (e.g., for a 16TB SSD storage device with 128KB logical chunks, 32MB of dedicated host memory is required if 8 bit counters are utilized), and the read temperature information identified will not be power-fail safe without a persistent power implementation (e.g., a battery backup, the use of Storage Class Memory (SCM) devices, etc.), each of which increases costs. In another example, the conventional read temperature identification systems discussed above increase complexity, as for a High Availability (HA) system each of multiple host processors included in a server device and/or storage system must generate its own read temperature map that tracks read temperatures of its storage devices in that server device and/or storage system, and then those host processors must synchronize their respective read temperature maps. Further complexity may be introduced when more Input/Output (I/O) initiators are utilized (e.g., when additional host processors are utilized in Non-Volatile Memory express over Fabrics (NVMe-oF) Just a Bunch Of Drives (JBOD) systems, disaggregated storage systems, and/or other systems that would be apparent to one of skill in the art in possession of the present disclosure).

In yet another example, the conventional read temperature identification systems discussed above may be inaccurate in some situations, as read temperature identification operations may be performed “in the background” with a “best effort” approach, and when host processors in a server device and/or storage system are busy performing other operations, those read temperature identification operations may not be performed in order to prevent I/O latency and/or other performance issues. While the host processors in a server device and/or storage system may sometimes only delay the read temperature identification operations in those situations, in some cases the read temperature identification operations may simply not be performed. In yet another example, the conventional read temperature identification systems discussed above can introduce a performance impact to data path(s) in a server device and/or storage system due to the use of the host processor and the host memory bus in performing the read temperature identification (e.g., via Read Modify Write (RMW) operations to provide these relatively small read temperature data writes via 64 byte cache line host memory entries, resulting in increased cache thrashing operations).

One conventional read temperature identification solution to the issues discussed above is to assume or characterize (a priori) the read temperatures of a storage device based on the type of data being read (e.g., metadata vs customer data), the type of application instructing the read operation (e.g., Relational Database Management System (RDBMS) applications vs. social media post applications (e.g., applications provided “tweets” via the TWITTERⓇ social networking service available from TWITTERⓇ of San Francisco, California, United States) vs. video streaming applications), the type of workload being performed (e.g., 4K vs. 8K video streaming workloads, sequential access vs. random access workloads, etc.). However, such conventional read temperature identification solutions suffer from a number of issues as well.

For example, the conventional read temperature identification solutions discussed above require pre-qualification or classification of data attributes, and cannot provide granularity beyond the particular classification that is used. In another example, conventional read temperature identification solutions do not allow for sub-classifications of data (e.g., a video type of the read data) that may be useful, will not allow data (e.g., video data such as that providing a movie) that is read often to be provided on a faster storage device or replicated at additional storage locations, and present issues with tracking effective “hits” per storage device and load balancing (as conventional read temperature identification solutions are typically limited to tracking data requests (e.g., video data requests) at the application level). In yet another example, conventional read temperature identification solutions require modification of software when new types of data, applications, and/or workloads are introduced and, as such, are less resilient with regard to optimizing read performance for use cases that emerge over time, and present additional costs associated with research and development to qualify new workloads or applications, develop software, test that software, perform software patch updates on server devices and/or storage systems that will use that software, and/or introduce other added cost factors that would be apparent to one of skill in the art in possession of the present disclosure.

Accordingly, it would be desirable to provide read temperature identification system that addressees the issues discussed above.

SUMMARY

According to one embodiment, a storage device includes a processing system; and a memory system that is coupled to the processing system and that includes instructions that, when executed by the processing system, cause the processing system to provide a local read temperature utilization engine that is configured to: determine, for a plurality of blocks in a storage subsystem, read disturb information; identify, using the read disturb information, a subset of rows in at least one block in the storage subsystem that have a higher read temperature than the other rows in the at least one block in the storage subsystem; generate, based on the identification of the subset of rows in the at least one block in the storage subsystem that have the higher read temperature than the other rows in the at least one block in the storage subsystem, a local logical storage element read temperature map that identifies a subset of logical storage elements associated with the storage subsystem that have a higher read temperature than the other logical storage elements associated with the storage subsystem; and move data from at least one first block in the storage subsystem to at least one second block in the storage subsystem based on relative read temperatures identified in the local logical storage element read temperature map.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic view illustrating an embodiment of an Information Handling System (IHS).

FIG. 2 is a schematic view illustrating an embodiment of a computing device that may include the read-disturb-based read temperature identification system of the present disclosure.

FIG. 3A is a schematic view illustrating an embodiment of a storage device that may be included in the computing device of FIG. 2 and that may provide the read-disturb-based read temperature identification system of the present disclosure.

FIG. 3B is a schematic view illustrating an embodiment of a storage subsystem that may be included in the storage device of FIG. 3A.

FIG. 3C is a schematic view illustrating an embodiment of NAND die that may be included in the storage subsystem of FIG. 3B.

FIG. 3D is a graph view illustrating an embodiment of voltages/values available in a cell of a NAND wordline in a NAND block that is included in the NAND die of FIG. 3C.

FIG. 3E is a schematic view illustrating an embodiment of a NAND block that is abstracted with “NAND rows” for purposes of describing different embodiments the read-disturb-based read temperature identification system of the present disclosure.

FIG. 4 is a flow chart illustrating an embodiment of a method for providing read-disturb-based read temperature identification.

FIG. 5A is a schematic view illustrating an embodiment of the storage device of FIG. 3A operating during the method of FIG. 4.

FIG. 5B is a schematic view illustrating an embodiment of a read disturb signature identified for the NAND rows in the NAND block of FIG. 3E during the method of FIG. 4.

FIG. 5C is a graph view illustrating an embodiment of voltage skewing in a cell that provides a portion of the read disturb signature of FIG. 5B during the method of FIG. 4.

FIG. 5D is a schematic view illustrating an embodiment of the storage device of FIG. 3A operating during the method of FIG. 4.

FIG. 6 is a schematic view illustrating an embodiment of the storage device of FIG. 3A operating during the method of FIG. 4.

FIG. 7A is a schematic view illustrating an embodiment of local logical storage element read temperature map generated by a first storage device in the computing device of FIG. 2 during the method of FIG. 4.

FIG. 7B is a schematic view illustrating an embodiment of local logical storage element read temperature map generated by a second storage device in the computing device of FIG. 2 during the method of FIG. 4.

FIG. 8A is a schematic view illustrating an embodiment of the computing device of FIG. 2 operating during the method of FIG. 4.

FIG. 8B is a schematic view illustrating an embodiment of global logical storage element read temperature map generated by the computing device of FIG. 2 during the method of FIG. 4.

FIG. 9 is a schematic view illustrating an embodiment of a storage device that may be included in the computing device of FIG. 2 and that may provide the storage device read-disturb-based read temperature map utilization system of the present disclosure.

FIG. 10 is a flow chart illustrating an embodiment of a method for storage device utilization of a read-disturb-based read temperature map.

FIG. 11 is a schematic view illustrating an embodiment of the storage device of FIG. 9 operating during the method of FIG. 10.

FIG. 12 is a schematic view illustrating an embodiment of data movement by the storage device of FIG. 9 during the method of FIG. 10 based on a read-disturb-based read temperature map.

FIG. 13 is a schematic view illustrating an embodiment of data movement by the storage device of FIG. 9 during the method of FIG. 10 based on a read-disturb-based read temperature map.

DETAILED DESCRIPTION

For purposes of this disclosure, an information handling system may include any instrumentality or aggregate of instrumentalities operable to compute, calculate, determine, classify, process, transmit, receive, retrieve, originate, switch, store, display, communicate, manifest, detect, record, reproduce, handle, or utilize any form of information, intelligence, or data for business, scientific, control, or other purposes. For example, an information handling system may be a personal computer (e.g., desktop or laptop), tablet computer, mobile device (e.g., personal digital assistant (PDA) or smart phone), server (e.g., blade server or rack server), a network storage device, or any other suitable device and may vary in size, shape, performance, functionality, and price. The information handling system may include random access memory (RAM), one or more processing resources such as a central processing unit (CPU) or hardware or software control logic, ROM, and/or other types of nonvolatile memory. Additional components of the information handling system may include one or more disk drives, one or more solid state drives, one or more network ports for communicating with external devices as well as various input and output (I/O) devices, such as a keyboard, a mouse, touchscreen and/or a video display. The information handling system may also include one or more buses operable to transmit communications between the various hardware components.

In one embodiment, IHS 100, FIG. 1, includes a processor 102, which is connected to a bus 104. Bus 104 serves as a connection between processor 102 and other components of IHS 100. An input device 106 is coupled to processor 102 to provide input to processor 102. Examples of input devices may include keyboards, touchscreens, pointing devices such as mouses, trackballs, and trackpads, and/or a variety of other input devices known in the art. Programs and data are stored on a mass storage device 108, which is coupled to processor 102. Examples of mass storage devices may include hard discs, optical disks, magneto-optical discs, solid-state storage devices, and/or a variety of other mass storage devices known in the art. IHS 100 further includes a display 110, which is coupled to processor 102 by a video controller 112. A system memory 114 is coupled to processor 102 to provide the processor with fast storage to facilitate execution of computer programs by processor 102. Examples of system memory may include random access memory (RAM) devices such as dynamic RAM (DRAM), synchronous DRAM (SDRAM), solid state memory devices, and/or a variety of other memory devices known in the art. In an embodiment, a chassis 116 houses some or all of the components of IHS 100. It should be understood that other buses and intermediate circuits can be deployed between the components described above and processor 102 to facilitate interconnection between the components and the processor 102.

Referring now to FIG. 2, an embodiment of a computing device 200 is illustrated that may include the read-disturb-based read temperature identification system of the present disclosure. In an embodiment, the computing device 200 may be provided by the IHS 100 discussed above with reference to FIG. 1 and/or may include some or all of the components of the IHS 100, and in the specific examples below is illustrated and described as being provided by a server device and/or a storage system. However, while illustrated and discussed as being provided by particular computing devices, one of skill in the art in possession of the present disclosure will recognize that the functionality of the computing device 200 discussed below may be provided by other devices that are configured to operate similarly as the computing device 200 discussed below. In the illustrated embodiment, the computing device 200 includes a chassis 202 that houses the components of the computing device 200, only some of which are illustrated and discussed below. For example, the chassis 202 may house a processing system (not illustrated, but which may include the processor 102 discussed above with reference to FIG. 1 that may be provided by a Central Processing Unit (CPU) and/or other processing systems that one of skill in the art in possession of the present disclosure would recognize as providing a computing device host processor) and a memory system (not illustrated, but which may include the memory 114 discussed above with reference to FIG. 1) that is coupled to the processing system and that includes instructions that, when executed by the processing system, cause the processing system to provide a global read temperature identification engine 204 that is configured to perform the functionality of the global read temperature identification engines and/or computing devices discussed below.

The memory system housed in the chassis 202 may also include a global read temperature identification database 206 that is configured to store any of the information utilized by the global read temperature identification engine 204 discussed below. The chassis 202 may also house a storage system 208 that, in the illustrated embodiment, includes a plurality of storage devices 210a, 210b, and up to 210c. In the specific examples below, each of the storage devices 210a-210c in the storage system 208 are described as being provided by particular Solid State Drive (SSD) storage devices, but one of skill in the art in possession of the present disclosure will appreciate how the teachings of the present disclosure may benefit other storage device technologies, and thus storage devices utilizing those other types of storage device technologies are envisioned as falling within the scope of the present disclosure as well. However, while a specific computing device 200 has been illustrated and described, one of skill in the art in possession of the present disclosure will recognize that computing devices (or other devices operating according to the teachings of the present disclosure in a manner similar to that described below for the computing device 200) may include a variety of components and/or component configurations for providing conventional computing device functionality, as well as the functionality discussed below, while remaining within the scope of the present disclosure as well.

Referring now to FIG. 3A, an embodiment of a storage device 300 is illustrated that may provide any or each of the storage devices 210a-210c in the storage system 208 discussed above with reference to FIG. 2. In an embodiment, the storage device 300 may be provided by the IHS 100 discussed above with reference to FIG. 1 and/or may include some or all of the components of the IHS 100, and in the specific examples below is illustrated and described as being provide by an SSD storage device (e.g., a Non-Volatile Memory express (NVMe) SSD storage device). However, while illustrated and discussed as being provided by a particular storage device, one of skill in the art in possession of the present disclosure will appreciate that the teachings of the present disclosure may be implemented in other storage devices that are configured to operate similarly as the storage device 200 discussed below. In the illustrated embodiment, the storage device 200 includes a chassis 302 that houses the components of the storage device 300, only some of which are illustrated and discussed below.

For example, the chassis 302 may house a storage device processing system 304 (which may include the processor 102 discussed above with reference to FIG. 1 such as a Central Processing Unit (CPU), storage device controller, and/or other processing systems that one of skill in the art in possession of the present disclosure would recognize as being provided in an SSD storage device) that is coupled via a first memory interface 306 (e.g., a Dual Data Rate (DDR) interface) to a first memory system 308 (which may include the memory 114 discussed above with reference to FIG. 1 such as Dynamic Random Access Memory (DRAM) devices and/or other memory systems that would be apparent to one of skill in the art in possession of the present disclosure). As illustrated in the specific examples provided herein, the first memory system 308 may include instructions that, when executed by the storage processing system 304, cause the storage device processing system 304 to provide a local read temperature identification engine 310 that is configured to perform the functionality of the local read temperature identification engines and/or storage devices discussed below.

As also illustrated in the specific examples provided herein, the first memory system 308 may include a local read temperature identification database 312 that is configured to store any of the information utilized by the local read temperature identification engine 310 discussed below. However, one of skill in the art in possession of the present disclosure will recognize that other embodiments of the present disclosure may provide the local read temperature identification database 312 in other locations while remaining within the scope of the present disclosure as well. For example, as illustrated, the storage device processing system 304 may also be coupled via a second memory interface 314 (e.g., a Storage Class Memory (SCM) interface) to a second memory system 316 (which may include the memory 114 discussed above with reference to FIG. 1 such as SCM devices and/or other memory systems that would be apparent to one of skill in the art in possession of the present disclosure), and the local read temperature identification database 312 may be provided by the second memory system 316 while remaining within the scope of the present disclosure as well.

The storage device processing system 304 may also be coupled via a storage interface 318 to a storage subsystem 320. With reference to FIG. 3B, in some embodiments, the storage subsystem 320 may include a storage subsystem chassis 320a that supports a plurality of NAND die 322. With reference to FIG. 3C, each NAND die 322 may include a chassis 322a that supports a plurality of NAND blocks 324, with each NAND block 324 including a chassis 324a that supports a plurality of NAND wordlines 326. Furthermore, each NAND wordline 326 may include a plurality of cells that provide a plurality of data portions 326a, and a respective error check portion 326b (e.g., a Cyclic Redundancy Check (CRC)portion and/or other error check data known in the art) may be associated with each of those data portions 326a. However, one of skill in the art in possession of the present disclosure will appreciate how in some embodiments the data written to a NAND block 324 may include “padding” data or other data which conventionally does require the writing of associated error check portions.

To provide a specific example, the storage subsystem 320 may include 128, 256, or 512 NAND die, with each NAND die including approximately 2000 NAND blocks, and with each NAND block including NAND wordlines grouped into 100-200 NAND layers (although forecasts predict that NAND wordlines will be grouped into up to 800 layers by the year 2030). As will be appreciated by one of skill in the art in possession of the present disclosure, conventional Triple Level Cell (TLC) technology typically allows on the order of tens to hundreds of K of data (e.g., 96KiB on a NAND wordline, 48KiB on a NAND wordline with two NAND wordlines activated at any particular time, up to hundreds of KiB when more planes are utilized, etc.) to be stored per NAND wordline (i.e., in the data portions of those NAND wordlines), resulting in NAND wordlines with ~250 K cells.

With reference to FIG. 3D, a simplified representation of how data may be stored in a cell 328 is provided, and one of skill in the art in possession of the present disclosure will appreciate how data may be stored in any of the plurality of cells in any of the plurality of NAND wordlines discussed above in the manner described below. The data storage representation of the cell 328 in FIG. 3D includes a graph 330 with voltage 330a on the X-axis, and illustrates how different voltages of the cell 228 may be associated with different values for that cell 328, which in specific example illustrated in FIG. 3E includes values “A”, “B”, “C”, “D”, “E”, “F”, “G”, and “H”. Furthermore, the data storage representation of the cell 328 also illustrated how reference voltages may be defined to distinguish whether a voltage in the cell provide a particular value, with a B reference (“B REF”) distinguishing between a value “A” or a value “B” for the cell 328, a C reference (“C REF”) distinguishing between a value “B” or a value “C” for the cell 328, a D reference (“D REF”) distinguishing between a value “C” or a value “D” for the cell 328, an E reference (“E REF”) distinguishing between a value “D” or a value “E” for the cell 328, an F reference (“F REF”) distinguishing between a value “E” or a value “F” for the cell 328, a G reference (“G REF”) distinguishing between a value “F” or a value “G” for the cell 328, an H reference (“H REF”) distinguishing between a value “G” or a value “H” for the cell 328.

As such, when the cell 328 includes a voltage below “B REF” it will provide a value “A”, when the cell 328 includes a voltage between “B REF” and “C REF” it will provide a value “B”, when the cell 328 includes a voltage between “C REF” and “D REF” it will provide a value “C”, when the cell 328 includes a voltage between “D REF” and “E REF” it will provide a value “D”, when the cell 328 includes a voltage between “E REF” and “F REF” it will provide a value “E”, when the cell 328 includes a voltage between “F REF” and “G REF” it will provide a value “F”, when the cell 328 includes a voltage between “G REF” and “H REF” it will provide a value “G”, when the cell 328 includes a voltage over “H REF” it will provide a value “H”. While not illustrated or described in detail herein, one of skill in the art in possession of the present disclosure will appreciate that each value A-H illustrated in FIG. 3D may be configured to store more than one bit depending on the amount of voltage that is provided to indicate that value (e.g., a first voltage level between “B REF” and “C REF” will provide a first set of bits for the value “B”, a second voltage level between “B REF” and “C REF” will provide a second set of bits for the value “B”, and so on).

As will be appreciated by one of skill in the art in possession of the present disclosure, different storage device manufacturers/providers may configure the NAND wordlines/NAND layers in NAND blocks differently, with some storage devices including NAND blocks with separate NAND wordlines, some storage devices including NAND blocks with NAND layers that each include a plurality of NAND wordlines, and some storage devices including NAND blocks with groups of NAND layers that each include a plurality of NAND wordlines. As such, with reference to FIG. 3E, the present disclosure abstracts the physical implementation of NAND wordlines and NAND layers into “NAND rows”, with each NAND block 324 discussed in the examples below including a plurality of NAND rows 332. In other words, any one of the NAND rows 332 may include NAND wordline(s), NAND layer(s) each including a plurality of NAND wordlines, or group(s) of NAND layers that each include a plurality of NAND wordlines. As will be appreciated by one of skill in the art in possession of the present disclosure, the read disturb signatures discussed below may vary based on the design of the storage subsystem/storage device, as it may effect a NAND wordline or group of NAND wordlines, and thus the abstraction of the physical implementation of NAND wordlines into NAND rows is provided to simplify the discussion below while encompassing such different storage subsystem/storage device designs.

However, while the specific examples discussed above describes the storage device 300 as including the storage interface 318 that may be provided by a flash device interface and the storage subsystem 320 that is described as being provided by NAND devices (e.g., NAND flash devices), one of skill in the art in possession of the present disclosure will appreciate how the teachings of the present disclosure may benefit other storage technologies, and thus storage devices utilizing those other types of storage technologies are envisioned as falling within the scope of the present disclosure as well. Furthermore, while a specific example of cells that may store 8 values (“A”-“H” in the examples above) are provided, one of skill in the art in possession of the present disclosure will appreciate how the cells may store 2 values (e.g., “A”/“0” and “B”/“1”), 4 values (e.g., “A”/“00”, “B”/“01”, “C”/“10”, and “D”/“11”), or more than 8 values while remaining within the scope of the present disclosure as well). Furthermore, one of skill in the art in possession of the present disclosure will appreciate how different NAND rows 332 in any particular NAND block 324 of the storage subsystem 302 may use different value encoding techniques (e.g., “A” and “B”, “A”-“D”, “A”-“H” in the examples above), and such mixed encoding NAND rows 332 will fall within the scope of the present disclosure.

In the illustrated embodiment, a power loss prevention system 334 is housed in the chassis 302 and coupled to the storage subsystem 320, and in specific examples may be provided by a Power Loss Prevention (PLP) capacitor and/or other power storage/provisioning subsystems that would be apparent to one of skill in the art in possession of the present disclosure. In the illustrated embodiments, the storage device processing system 304 is also coupled to an expansion bus 336 such as, for example, a Peripheral Component Interconnect express (PCIe) expansion bus that may provide the connection to the global read temperature identification engine 204, as well as to one or more storage device processing system peripherals 338. Furthermore, the expansion bus 336 may provide one or more connections for performing operations associated with the storage device 300 (e.g., connection(s) for reading/writing, connections for managing any of the data/information discussed below, etc.), and may also provide out-of-band interface(s), side channel interface(s), and/or other interfaces that provide access to the storage device processing system 304 for other systems. However, while a specific storage device 300 has been illustrated and described, one of skill in the art in possession of the present disclosure will recognize that storage devices (or other devices operating according to the teachings of the present disclosure in a manner similar to that described below for the storage device 300) may include a variety of components and/or component configurations for providing conventional storage device functionality, as well as the functionality discussed below, while remaining within the scope of the present disclosure as well.

Referring now to FIG. 4, an embodiment of a method 400 for providing read-disturb-based read temperature identification is illustrated. As discussed below, the systems and methods of the present disclosure utilize the read disturb effect that produces noise in adjacent NAND rows when any particular NAND row is read in order to identify NAND rows that are read more often than other NAND rows and thus have higher “read temperatures” than those other NAND rows. For example, the read-disturb-based read temperature identification system of the present disclosure may include storage device(s) that each determine read disturb information for each block in that storage device, use that read disturb information to identify a subset of rows in at least one block in that storage device that have a higher read temperature than the other rows in the at least one block in that storage device and, based on that identification, generate and store a local logical storage element read temperature map that identifies a subset of logical storage elements associated with that storage device that have a higher read temperature than the other logical storage elements associated with that storage device. A global read temperature identification subsystem coupled to the storage device(s) may then retrieve the local logical storage element read temperature map generated by each of the storage device(s) and use them to generate a global logical storage element read temperature map.

As such, the read disturb effect, which occurs automatically in response to conventional read operations and persists across power cycles, may be leveraged to generate read temperature maps for storage devices and storage systems, thus addressing many of the issues with conventional read temperature identification systems discussed above. As will be appreciated by one of skill in the art in possession of the present disclosure, the local logical storage element read temperature maps of the present disclosure are generated by the storage devices themselves (rather than the host processor/storage processor in the server device and/or storage system in which they are located), limiting read temperature identification operations performed by that host processor/storage processor to the utilization of those local logical storage element read temperature maps to generate a global logical storage element read temperature map in embodiments of the present disclosure. Furthermore, the local logical storage element read temperature maps of the present disclosure may be generated without any knowledge of the type of data being read, the application performing the read operation, or the workload being performed that resulted in the read operation.

The method 400 begins at block 402 where storage device(s) determine read disturb information for each of a plurality of blocks in that storage device. During or prior to the method 400, the computing device 200 may be utilized to write data to the storage devices 210a, 210b, and up to 210c in the storage system 208, and then read that data from those storage device(s). As will be appreciated by one of skill in the art in possession of the present disclosure, a simplified example of the writing of data to a NAND block in a storage device using the specific examples provided above includes a processing system in the computing device 200 (e.g., the processing system that provides the global read temperature identification engine 204) erasing all of the NAND rows in that NAND block to set each of their cells to the “A” value, and then selectively applying voltages across “vertical” bitlines in the NAND block and one or more “horizontal” NAND wordline(s) in NAND row(s) in that NAND block in order to cause the cells in those one or more NAND wordlines(s) to switch from the “A” value to a value indicated by a higher voltage in that cell (e.g., one of the values “B”, “C”, “D”, “E”, “F”, “G”, and “H” in the example above), resulting in each of those cells in the one or more NAND row(s) in that NAND block storing some number of electrons to provide one of the values “A” “B”, “C”, “D”, “E”, “F”, “G”, or “H” discussed above. As will be appreciated by one of skill in the art in possession of the present disclosure, the selective application of voltages discussed above may include no application of a voltage for a cell that is desired to have an “A” value.

Subsequently, data may be read from a NAND block by determining what values the cells in its NAND rows store. As will be appreciated by one of skill in the art in possession of the present disclosure, a simplified example of the reading of data from a first NAND row in a NAND block in a storage device includes a processing system in the computing device 200 (e.g., the processing system that provides the global read temperature identification engine 204) “selecting” the first NAND row by providing a voltage across the “vertical” bitlines in the NAND block, with the electrons stored in the cells in the first NAND row (i.e., to provide the values discussed above) operating to reduce the current that is sensed at the bottom of the “vertical” bitlines in the NAND block and that is produced in response to the applied voltage (with the sensing of that reduced current operating to identify particular values in the cells in the first NAND row). However, in order to prevent other second NAND rows in that NAND block that are not being read from effecting the current resulting from the voltage provided across the “vertical” bitlines in that NAND block (i.e., in order to ensure the effect on that current by the electrons stored in the cells of the first NAND row may be sensed as discussed above), those second NAND rows are “deselected” by providing a “bypass” voltage across each of those “horizontal” second NAND rows that forces its cell(s) to conduct current on the bitline.

As will be appreciated by one of skill in the art in possession of the present disclosure, the provisioning of that “bypass” voltage across each of the “horizontal” second NAND rows results in the “micro-programming” of those second NAND rows caused by electrons accumulating in those second NAND rows each time the first NAND row is read (i.e., due to the “bypass” voltage used to “deselect” them attracting electrons out of the bitline), which is referred to as the “read disturb effect” herein. Furthermore, one of skill in the art in possession of the present disclosure will appreciate how the read disturb effect/microprogramming/electron accumulation discussed above is higher in second NAND rows that are closer to the first NAND row being read, and the amount of read disturb effect/microprogramming/electron accumulation in NAND rows will depend on the type of SSD technology used in the SSD storage device.

As will be appreciated by one of skill in the art in possession of the present disclosure, while each NAND row includes many cells that may each identify multiple different values, any read of a NAND row operates to read all the cells in that NAND row and, as such, it is the read temperature of the NAND row that is of concern. Furthermore, while it is possible to read a portion of a NAND row (some subset of the NAND wordlines in that NAND row), that read operation will still apply a voltage to that entire NAND row in order to accomplish the read, thus introducing the same read disturb effect in that NAND row that would have occurred if the entire NAND row had been read.

Conventionally, the read disturb effect discussed above is considered a problem that must be corrected, as the microprogramming/electron accumulation in second NAND row(s) adjacent a first NAND row that is read often can cause a desired value in one or more of the cells in those second NAND row(s) to be mistakenly read as a different value, which one of skill in the art in possession of the present disclosure will appreciate results in a number of incorrect or “flipped” bits (i.e., bits that do not match their originally written value) that provide a “fail bit count” and must be corrected (e.g., using the error check portion 326b associated with the data portion 326a provided by the cell in the NAND wordline) to reverse the change in the value read for that cell in that NAND row. However, as discussed in further detail above, the inventors of the present disclosure have discovered that the read disturb effect may be leveraged in order to identify the read temperature of NAND rows in NAND blocks in a manner that eliminates many of the issues present in conventional read temperature identification systems.

As such, subsequent to the writing of data to the storage device(s) 210a-210c and the reading of that data from those storage device(s) 210a-210c, any or all of those storage device(s) 210a-210c/300 may operate at block 402 to determine read disturb information for each of the NAND blocks 324 included in the NAND die 322 in the storage subsystem 320 in that storage device. As will be appreciated by one of skill in the art in possession of the present disclosure, the discussion below of the determination of read disturb information by the storage device 300 may apply to any of the storage devices 210a-210c, and may be performed upon startup, reset, or other initialization of the storage device 300, periodically by the storage device during runtime, at the request of a user of the computing device 200, and/or on any other schedule or at any other time that would be apparent to one of skill in the art in possession of the present disclosure.

With reference to FIG. 5A, in an embodiment of block 402, the storage device processing system 304 in the storage device 300 may perform read disturb information retrieval operations 506 that may include the storage device processing system 304 retrieving, via the storage interface 318, read disturb information associated with each of the NAND blocks 324 included in the NAND die 322 in the storage subsystem 320. The inventors of the present disclosure have developed techniques for retrieving read disturb information that are described in U.S. Pat. Application Serial No. ______, attorney docket number 126901.01, filed ______, the disclosure of which is incorporated by reference herein in its entirety. However, while the retrieval of read disturb information from each of the NAND rows 332 in each of the NAND blocks 324 in the storage subsystem 320 is described, one of skill in the art in possession of the present disclosure will appreciate how the retrieval of read disturb information for a subset of NAND rows 332 in a subset of NAND blocks 324 in the storage subsystem 320 will fall within the scope of the present disclosure as well (e.g., when a previously “hot” subset of NAND rows in NAND block(s) are being checked to determine whether they are still “hot”).

For example, the read disturb information retrieval operations 506 performed at block 402 may include the storage device processing system 304 in the storage device 300 accessing each NAND block 324 to identify a read disturb “signature” for each of a plurality of NAND rows 332a-332i in that NAND block 324 that may be provided by fail bit counts in one or more adjacent NAND rows. With reference to FIG. 5B, a specific example of the read disturb information for the NAND row 332e in a NAND block 324 is provided, and illustrates a read disturb signature 502 provided by fail bit counts for some of the NAND rows 332a-332d and 332f-332i.

In particular, the read disturb signature 502 illustrated in FIG. 5B includes a fail bit count portion 502a associated with the NAND rows 332a-332e on a “first side” of the NAND row 332e, and a fail bit count portion 502b associated with the NAND rows 332f-332i on a “second side” of the NAND row 332e. As will be appreciated by one of skill in the art in possession of the present disclosure, the distribution of the read disturb signature 502 provides a histogram across all the cells in NAND wordlines of the NAND rows (e.g., with some cells in the NAND wordline(s) in the NAND row 332f relatively more effected by the read disturb effect than other cells in that NAND wordlines in that NAND row 332f), with all of the NAND wordlines in the NAND rows impacted by the read disturb effect to some extent (i.e., due to electron accumulation prior to the attenuation effects discussed above). However, while a particular example is provided, one of skill in the art in possession of the present disclosure will appreciate that other storage subsystem technologies (e.g., SCM storage devices, Dual Data Rate (DDR) storage devices, etc.) provide similar effects (e.g., DDR storage devices experience a “row hammer” effect) that will fall within the scope of the present disclosure as well. As will be appreciated by one of skill in the art in possession of the present disclosure, the fail bit count portion 502a illustrates how the NAND row 332d experiences a higher fail bit count than the NAND row 332c, which experiences a higher fail bit count than the NAND row 332b, and so on due to their relative proximity to the NAND row 332e. Similarly, the fail bit count portion 502b illustrates how the NAND row 332f experiences a higher fail bit count than the NAND row 332g, which experiences a higher fail bit count than the NAND row 332h, and so on due to their relative proximity to the NAND row 332e.

Furthermore, the fail bit count portions 502a and 502b also illustrate how the NAND row 332d experiences a higher fail bit count than the NAND row 332f due to the NAND row 332d seeing the “full” current resulting from the voltage applied to the “vertical” bitlines in the NAND block when performing a read of the NAND row 332e relative to the current that is attenuated by the charge in the cell of the NAND row 332e(e.g., if the charge in the cell(s) of a NAND row provides a relatively low voltage value (e.g., the “A” value discussed below), the “downstream” NAND rows will see relatively more electrons than the “upstream” NAND rows, while if the charge in the cell(s) of a NAND row provides a relatively high voltage value (e.g., the “H” value discussed below), the “downstream” NAND rows will see relatively less electrons than the “upstream” NAND rows).

With reference to FIG. 5C, the simplified representation from FIG. 3D of how data may be stored in a cell 328 is reproduced, but with the data storage representation of the cell 328 in FIG. 5C including the graph 330 with voltage 330a on the X-axis and probability 504 on the Y-axis. The graph 330 in FIG. 5C also includes voltage/value probabilities 504a (provided in solid lines) for each value “A”-“H” available in the cell 328 that, as can be seen, is highest in the middle of the voltage range for each value “A”-“H”, and reduces to near-zero near the bounds of the voltage range for each value “A”-“H” (e.g., the voltage/value probability for the value “A” is highest midway between “0” and “B REF” and reduces to near-zero at both “0” and “B REF”, the voltage/value probability for the value “B” is highest midway between “B REF” and “C REF” and reduces to near-zero at both “B REF” and “C REF”, and so on).

As will be appreciated by one of skill in the art in possession of the present disclosure, the cell 328 associated with the graph 330 in FIG. 5C has experienced the read disturb effect (e.g., it is a cell in one of the NAND rows 332b-d, 332f, or 332g in FIG. 5B), and the graph 330 illustrates a read disturb effect skew 504b (provided in dashed lines) that illustrates how the read disturb effect skews the voltage/value probabilities 504a for each value “A”-“H” available in the cell 328. As discussed above, after a plurality of reads to an adjacent NAND row (e.g., the NAND row 332e) that causes the accumulation of charge in a particular NAND row (e.g., the NAND row 332d), a desired value in some cells may be mistakenly read as a different value due to the voltage in those cells crossing the reference voltage that defines that different value.

For example, FIG. 5C illustrates how the accumulation of charge in the NAND row 332d may introduce the read disturb effect skew 504b for one or more of the values “A”-“” that can cause at least a portion of the voltage/value probabilities 504a for those values to shift across the reference voltage for an adjacent value. As can be seen in FIG. 5C, the read disturb effect skew 504b to the voltage/value probability 504a for the value “A” causes that voltage/value probability 504a to skew past the B REF, and thus some reads of voltages in the cell 328 that are desired to provide the value “A” will instead mistakenly provide the value “B” (i.e., due to the actual voltage read being between the B REF and the C REF because it was “pushed” in that “direction” due to the read disturb effect). Furthermore, while a single example is provided, one of skill in the art in possession of the present disclosure will appreciate how the read disturb effect skew 504b to the voltage/value probability 504a for any of the values “B”-“H” can result in the identification of a mistaken value in a similarly manner.

One of skill in the art in possession of the present disclosure will recognize how conventional systems (e.g., SSD storage device firmware and controllers) may utilize software to shift the reference voltages for one or more values in a cell to compensate for this read disturb effect. However, at block 402, the storage device processing system 304 may instead identify this read disturb information for each NAND row in each of a plurality of NAND blocks 324 in its storage device 300, and one of skill in the art in possession of the present disclosure will appreciate how the read disturb signature for each of those NAND rows will differ depending on whether that NAND row has been read a relatively higher number of times (in which case its read disturb signature will include relatively high fail bit counts for its adjacent NAND rows), whether that NAND row has been read a relatively lower number of times (in which case its read disturb signature will include relatively lower fail bit counts for its adjacent NAND rows), whether that NAND row has been read a relatively intermediate number of times (in which case its read disturb signature will include relatively intermediate fail bit counts for its adjacent NAND rows), etc.

Techniques for using fail bit counts that provide read disturb signatures in order to determine read disturb information are described by the inventors of the present disclosure in more detail in U.S. Pat. Application Serial No. ______, attorney docket number 126899.01, filed ______; and U.S. Patent Application Serial No. ______, attorney docket number 126900.01, filed ______; the disclosures of which are incorporated by reference herein in their entirety. However, one of skill in the art in possession of the present disclosure will appreciate that other techniques may be utilized to determine read disturb information while remaining within the scope of the present disclosure. For example, the inventors of the present disclosure have developed techniques for determining read disturb information without the need to explicitly identify failed bit counts, which are described in U.S. Patent Application Serial No. ______, attorney docket number 126898.01, filed ______, the disclosure of which is incorporated by reference herein in its entirety.

With reference to FIG. 5D, the storage device processing system 304 may then perform read disturb information storage operations 506 that include accessing the first memory system 308 vis the first memory interface 306 and storing the read disturb information in the local read temperature identification database 312. As such, following block 402, each of the storage devices 210a-210c/300 in the computing device 200 may have determined and stored read disturb information for each NAND row 332a-332i included in each of its NAND blocks 324 provided by each of its NAND dies 322 in its storage subsystem 320.

In some embodiments, at block 402 and prior to or subsequent to storing the read disturb information in the local read temperature identification database 312, the storage device processing system 304 in the storage device 300 may perform read disturb information isolation operations in order to isolate data in the read disturb information determined for each NAND row 332a-332i included in each of its NAND blocks 324 provided by each of its NAND dies 322 in its storage subsystem 320. The inventors of the present disclosure have developed several techniques for isolating read disturb information that are described in U.S. Pat. Application Serial No. ______, attorney docket number 126901.01, filed ______, the disclosure of which is incorporated by reference herein in its entirety. As described in those patent documents, the read disturb information determined at block 402 by the storage device processing system 304 for each NAND row 332a-332i included in each of its NAND blocks 324 provided by each of its NAND dies 322 in its storage subsystem 320 may include “noise” and/or other information artifacts that are not indicative of the read disturb effect, and thus different isolation techniques may be performed on the read disturb information in order to allow the storage device processing system 304 to more accurately characterized the read disturb effect for each NAND row 332a-332i included in each of its NAND blocks 324 provided by each of its NAND dies 322 in its storage subsystem 320. As such, in some embodiments, the read disturb information stored in the local read temperature identification database 312 in each storage device 210a-210c/300 may be isolated read disturb information.

The method 400 then proceeds to block 404 where the storage device(s) use the read disturb information to identify a subset of rows in block(s) in that storage device that have higher read temperatures than other rows in the block(s) in that storage device. With reference to FIG. 6, in an embodiment of block 404, the storage device processing system 304 may perform read temperature identification operations 600 that may include accessing the read disturb information stored in the local read temperature identification database 312 (e.g., via the first memory interface 306 and the first memory system 308), and identifying relative read temperatures of each NAND row 332a-332i included in each of its NAND blocks 324 provided by each of its NAND dies 322 in its storage subsystem 320. However, while the read temperature information is described as being stored in the local read temperature identification database 312 prior to being accessed and used to identify relative read temperatures of NAND rows, one of skill in the art in possession of the present disclosure will recognize that the read temperature information may be used to identify relative read temperatures of NAND rows upon its collection and without storing it in the local read temperature identification database 312 (e.g., relative read temperatures of NAND rows may be identified “on the fly” as read disturb information is collected at block 402) while remaining within the scope of the present disclosure as well.

As discussed above, the read disturb signature determined for each NAND row will differ depending on whether that NAND row has been read a relatively higher number of times, whether that NAND row has been read a relatively lower number of times, whether that NAND row has been read a relatively intermediate number of times, etc. Furthermore, one of skill in the art in possession of the present disclosure will appreciate how NAND rows that have been read a relatively higher number of times may be identified as having a relatively high read temperature, NAND rows that have been read a relatively lower number of times have be identified as having a relatively low read temperature, NAND rows that have been read a relatively intermediate number of times may be identified as having a relatively intermediate temperature, and so on.

As such, in some embodiments of block 404, the storage device processing system 304 may analyze each read disturb signature determined for each NAND row 328a-328i included in each of its NAND blocks 324 provided by each of its NAND dies 322 in its storage subsystem 320, and determine whether that read disturb signature identifies a relatively high read temperature (e.g., because the NAND rows adjacent the NAND row for which that read disturb signature was generated experienced relatively high fail bit counts), whether that read disturb signature identifies a relatively low read temperature (e.g., because the NAND rows adjacent the NAND row for which that read disturb signature was generated experienced relatively low fail bit counts), whether that read disturb signature identifies a relatively intermediate read temperature (e.g., because the NAND rows adjacent the NAND row for which that read disturb signature was generated experiences relatively intermediate fail bit counts), and/or whether that read disturb signature identifies other read temperature granularities that are distinguishable from the relatively high read temperature, the relatively low read temperature, and the relatively intermediate read temperature discussed above.

Furthermore, in some embodiments of block 404, the storage device 300 may operate to process current and previously determined read disturb information for a NAND row in order to generate a read temperature for that NAND row, and the inventors of the present disclosure describe techniques for storage devices to generate read temperature in U.S. Patent Application Serial No. ______, attorney docket no. 126892.01, filed ______; U.S. Patent Application Serial No. ______, attorney docket no. 126893.01, filed ______; and U.S. Pat. Application Serial No. ______, attorney docket no. 126894.01, filed ______; the disclosures of which are incorporated by reference herein in their entirety.

As will be appreciated by one of skill in the art in possession of the present disclosure, any relative read temperature metrics may be assigned to the relatively high read temperature, relatively low read temperature, relatively intermediate read temperature, and so on, in order to indicate the different read temperatures for each of the NAND rows. For example, the NAND row read temperatures identified as discussed above using NAND row read disturb signatures may not provide exact numerical read temperatures (e.g., as may be provided in conventional read temperature identification systems may operate to record the exact number of reads of a NAND row), but the inventors of the present disclosure have found that the use of the NAND row read disturb signatures to provide NAND row read temperatures as discussed above provide for the accurate identification of relative temperatures of the different NAND rows in a storage subsystem. Furthermore, as read disturb signature analysis becomes more accurate in the future, the inventors of the present disclosure expect that any particular read disturb signature may then be associated with a number of reads of a NAND row, and thus envision doing so while remaining within the scope of the present disclosure as well.

While the embodiments discussed below describe the use of the read temperature determined as described above to generate a local logical storage element read temperature map, the inventors of the present disclosure have developed techniques for storage devices to identify and use read temperatures based on the read disturb effect to move data without the need to generate a local logical storage element read temperature map that are described in U.S. Pat. Application Serial No. ______, attorney docket number 126887.01, filed ______, the disclosure of which is incorporated by reference herein in its entirety.

The method 400 then proceeds to block 406 where each of the storage device(s) generate a local logical storage element read temperature map identifying a subset of logical storage elements that are associated with that storage device and that have higher read temperatures than other logical storage elements associated with that storage device, as well as to block 408 where the storage device(s) store the local storage element read temperature map generated by that storage device. The inventors of the present disclosure have developed several techniques for generating and/or maintaining a local logical storage element read temperature map that are described in U.S. Pat. Application Serial No. ______, attorney docket number 126897.01, filed ______; U.S. Pat Application Serial No. ______, attorney docket number 126896.01, filed ______; U.S. Pat Application Serial No. ______, attorney docket number 126895.01, filed ______; U.S. Pat Application Serial No. ______, attorney docket number 126891.01, filed ; the disclosures of which are incorporated by reference herein in their entirety. As described in those patent documents, local logical storage element read temperature maps may be generated by mapping read temperatures identified for physical storage to a logical-to-physical mapping (e.g., via a “reverse lookup”), and may be maintained by identifying any historical read temperature(s) associated with a logical storage element when its data is moved to a new storage element (i.e., by writing that data to a physical storage location mapped to that new storage element), and then mapping those historical read temperatures to that new storage element in the logical-to-physical mapping as well (and in some cases, persistently storing those historical read temperatures in the NAND block that stores the corresponding data, in a metadata NAND block in the storage subsystem 320, and/or in other storage locations that would be apparent to one of skill in the art in possession of the present disclosure).

In an embodiment, at block 406, the storage device processing system 304 in the storage device 300 may generate a local logical storage element read temperature map using the relative read temperatures identified for each NAND row 332a-332i included in each of its NAND blocks 324 provided by each of its NAND dies 322 in its storage subsystem 320 at block 404. In an embodiment, the storage device 300 may utilize Logical Block Addressing (LBA), which one of skill in the art in possession of the present disclosure will recognize logically divides the storage subsystem 320 into logical storage elements (e.g., 512 byte to 4096 byte sectors), with the first logical storage element identified as logical block 0, the second logical storage element identified as logical block 1, and so on, and with each logical storage element mapped to a respective NAND row 332a-332i included in the NAND blocks 324 provided by the NAND dies 322 in the storage subsystem 320.

As such, block 406 may include the storage device processing system 304 mapping, in a local logical storage element read temperature map, the read temperature determined for each NAND row at block 404 to the logical storage element that is mapped to that NAND row. Thus, continuing with the example provided above, a read temperature identified for a first NAND row may be mapped, in the local logical storage element read temperature map, to an LBA block 0 that is mapped to the first NAND row; a read temperature identified for a second NAND row may be mapped, in the local logical storage element read temperature map, to an LBA block 1 that is mapped to the second NAND row; a read temperature identified for a third NAND row may be mapped, in the local logical storage element read temperature map, to an LBA block 2 that is mapped to the third NAND row; and so on until a read temperature is mapped to each of the LBA blocks.

In a specific example, any NAND row may be mapped to one or more logical storage elements, and in the event a NAND row has a particular read temperature, each logical storage element mapped to that NAND row will have that particular read temperature. As such, the present disclosure may provide read temperature granularity at the NAND row level. Furthermore, while most storage device implementations today map logical blocks to NAND rows such that each logical block is fully contained within that NAND row, one of skill in the art in possession of the present disclosure will appreciate that a logical block may “straddle” multiple NAND rows, and in such cases read temperatures of that logical block may be computed by combining the read temperatures determined for those multiple NAND rows using any of variety of techniques that would be apparent to one of skill in the art in possession of the present disclosure. Thus, following block 406, each of the storage devices 210a-210c/300 may have generated a respective local logical storage element read temperature map, and at block 408 each of the storage devices 210a-210c/300 may have stored that local logical storage element read temperature map in its local read temperature identification database 312.

With reference to FIG. 7A, an example of a local logical storage element read temperature map 700 is illustrated that may have been generated and stored by a first of the storage devices 210a-210c/300 in its local read temperature identification database 312. In the illustrated example, the local logical storage element read temperature map 700 includes relatively high read temperature logical storage elements 700a (illustrated as black boxes in the local logical storage element read temperature map 700), relatively low read temperature logical storage elements 700b (illustrated as white boxes in the local logical storage element read temperature map 700), relatively high-intermediate read temperature logical storage elements 700c (illustrated as dark grey boxes in the local logical storage element read temperature map 700), and relatively low-intermediate read temperature logical storage elements 700d (illustrated as light grey boxes in the local logical storage element read temperature map 700). However, one of skill in the art in possession of the present disclosure will recognize that the inclusion of other levels of read temperature granularity in local logical storage element read temperature maps will fall within the scope of the present disclosure as well. As will be appreciated by one of skill in the art in possession of the present disclosure, the dashed line in FIG. 7A is provided to indicate that the local logical storage element read temperature map 700 is one of a plurality of local logical storage element read temperature maps that provide the total logical storage space for the computing device 200.

With reference to FIG. 7B, an example of a local logical storage element read temperature map 702 is illustrated that may have been generated and stored by a second of the storage devices 210a-210c/300 in its local read temperature identification database 312. In the illustrated example, the local logical storage element read temperature map 702 includes relatively high read temperature logical storage elements 702a (illustrated as black boxes in the local logical storage element read temperature map 702), relatively low read temperature logical storage elements 702b (illustrated as white boxes in the local logical storage element read temperature map 702), relatively high-intermediate read temperature logical storage elements 702c (illustrated as dark grey boxes in the local logical storage element read temperature map 70), and relatively low-intermediate read temperature logical storage elements 702d (illustrated as light grey boxes in the local logical storage element read temperature map 702). However, one of skill in the art in possession of the present disclosure will recognize that the inclusion of other levels of read temperature granularity in local logical storage element read temperature maps will fall within the scope of the present disclosure as well. As will be appreciated by one of skill in the art in possession of the present disclosure, the dashed line in FIG. 7B is provided to indicate that the local logical storage element read temperature map 702 is one of a plurality of local logical storage element read temperature maps that provide the total logical storage space for the computing device 200. Furthermore, while examples of only two local logical storage element read temperature maps 700 and 702 generated by two storage devices are provided, one of skill in the art in possession of the present disclosure will appreciate that storage systems may include many more storage devices, and each of those storage devices may generate a local logical storage element read temperature map while remaining within the scope of the present disclosure as well.

Some of the embodiments discussed below describe the use of local logical storage element read temperature maps from different storage devices by the global read temperature identification engine 204 in the computing device 200 to generate a global logical storage element read temperature map. However, as also described in further detail below, storage devices may use the local logical storage element read temperature maps that they generate to move data independently of the global read temperature identification engine 204 in the computing device 200. As described below, such storage-device-independent data movement leverages read-disturb-based read temperatures to provide for more efficient data placement within the storage devices without the need for intervention from other subsystems in the computing device 200.

The method 400 then proceeds to block 410 where a global read temperature identification subsystem retrieves the local logical storage element read temperature map(s) generated by the storage device(s). With reference to FIG. 8A, in an embodiment of block 410, the global read temperature identification engine 204 in the computing device 200 may perform local logical storage element read temperature map retrieval operations 800 in order to retrieve the local logical storage element read temperature maps generated and stored by the storage devices 210a, 210b, and up to 210c at blocks 406 and 408. The inventors of the present disclosure have developed several techniques for accessing and utilizing local logical storage element read temperature maps, information provided therein, and/or associated information, which are described in U.S. Pat Application Serial No. ______, attorney docket number 126890.01, filed ______; and U.S. Pat Application Serial No. ______, attorney docket number 126889.01, filed ______; the disclosures of which are incorporated by reference herein in their entirety.

As will be appreciated by one of skill in the art in possession of the present disclosure, in some examples the global read temperature identification engine 204 in the computing device 200 may access the local read temperature identification databases 312 in the first memory system 308 in each of the storage devices 210a-210c/300 in order to retrieve the local logical storage element read temperature maps stored therein, while in other embodiments the global read temperature identification engine 204 in the computing device 200 may provide requests for those local logical storage element read temperature maps such that each of the storage devices 210a-210c/300 transmit them to the global read temperature identification engine 204.

The method 400 then proceeds to block 412 where the global read temperature identification subsystem uses the local logical storage element read temperature map(s) to generate a global logical storage element read temperature map, as well as to optional block 414 where the global read temperature identification subsystem stores the global logical storage element read temperature map. In some embodiment, at block 412, the global read temperature identification engine 204 in the computing device 200 may concatenate the local logical storage element read temperature maps retrieved from the storage devices 210a-210c/300 to generate a global logical storage element read temperature map that it then stores in the global read temperature identification database 206. For example, FIG. 8B illustrates an example of a global logical storage element read temperature map 802 that may have been generated by the global read temperature identification engine 204 via concatenation of the local logical storage element read temperature maps 700 and 702 discussed above with reference to FIGS. 7A and 7B, and then stored by the global read temperature identification engine 204 in its global read temperature identification database 312.

However, one of skill in the art in possession of the present disclosure will appreciate how the generation of a global logical storage element read temperature map via concatenation of local logical storage element read temperature maps provides a simplified example of the use of local logical storage element read temperature maps to generate a global logical storage element read temperature map, and that the local logical storage element read temperature maps discussed above may be utilized to generate the global logical storage element read temperature map in other manners that will fall within the scope of the present disclosure as well. For example, the generation of the global logical storage element read temperature map using the local logical storage element read temperature maps may depend on how the global read temperature identification engine 204 is configured to track read temperatures, how the global read temperature identification engine 204 is configured to organize data (e.g., data may be “striped” across the storage devices 210a-210c), and/or based on other factors that would be apparent to one of skill in the art in possession of the present disclosure.

For example, in some embodiments the computing device 200 may store data in the storage devices 210a-210b independently of each other (e.g., as the logical blocks discussed above), in which case the generation of the global logical storage element read temperature map via concatenation of local logical storage element read temperature maps may be appropriate. However, in other embodiments, the computing device 200 may utilize more complex software that organizes the storage of the data in the storage devices 210a-210c in “groups” of logical blocks. For example, for performance considerations a data group of data A, B, and C may be provided by three respective logical blocks, and may be written to each of three respective storage devices, but the computing device 200 may view that data group as “atomic” such that the read temperature that matters is the read temperature of that data group. In such an embodiment, a “higher level” “group global logical storage element read temperature map” may be generated in order to allow the tracking of data group read temperatures, and the inventors of the present disclosure are developing techniques for doing so. One example of such an embodiment is a Redundant Array of Independent Drives (RAID) storage system, but one of skill in the art in possession of the present disclosure will appreciate that other storage systems may introduce similar considerations as well.

Furthermore, the inventors of the present disclosure have developed techniques for generating global logical storage element read temperature maps using local logical storage element read temperature maps which are described in U.S. Pat Application Serial No. ______, attorney docket number 126889.01, filed ______, the disclosure of which is incorporated by reference herein in its entirety. As described in that patent document, data in local logical storage element read temperature maps generated by different storage devices may be scaled relative to each other so that the read temperatures of storage locations in different storage devices with different utilizations (e.g., a first storage device with one million reads and a second storage device with ten million reads) may be accurately compared relative to each other (i.e., “hot” storage locations in the first storage device with one million reads may not be “hot” relative to “hot” storage locations in the second storage device with ten million reads).

In some embodiments, a subset of the local logical storage element read temperature maps generated by the storage devices 210a-210c may be retrieved and used by the global read temperature identification engine 204 at block 412. For instance, if the memory system in the computing device 300 does not have sufficient space to store the global logical storage element read temperature map (or for other reasons that memory space is allocated for the storage of other data), a global logical storage element read temperature map may be generated that only identifies logical storage elements with relatively “hot” red temperatures. As such, in some embodiments, the local logical storage element read temperature maps retrieved from the storage devices 210a-210c may only identify logical storage elements having a particular read temperature (e.g., those with relatively “hot” read temperatures), allowing for the generation of the global logical storage element read temperature map identifying logical storage elements with that particular temperature as well. However, in other embodiments, the global read temperature identification engine 204 may be configured to retrieve the local logical storage element read temperature map(s) from the storage devices 210a-210c, and then utilize a filter to generate a filtered global logical storage element read temperature map that identifies particular read temperatures from the local logical storage element read temperature map(s).

In an embodiment, following block 412, the computing device 200 (e.g., the processing system in the computing device 200) may utilize the global logical storage element read temperature map (which may have been stored in the global read temperature identification database 206) in order to provide for the storage of data, movement of data, and/or other data operations that would be apparent to one of skill in the art in possession of the present disclosure. For example, data stored in NAND rows with relatively high read temperatures may be moved to relatively high capability/cost storage devices, data stored in NAND rows with relatively low read temperatures may be moved to relatively low capability/cost storage devices, data stored in NAND rows with relatively intermediate read temperatures may be moved to relatively intermediate capability/cost storage devices, etc. In another example, the computing device 200 (e.g., the processing system in the computing device 200) may utilize the global logical storage element read temperature map in order to perform load balancing (e.g., when the storage devices 210a-210c are the same type of capability/cost storage device, load balancing reads to those storage devices can result in a higher performing storage system (relative to the performance of that storage system without the load balancing). However, while a few specific examples of operations based on identified read temperatures have been described, one of skill in the art in possession of the present disclosure will appreciate how the read temperatures identified in the global logical storage element read temperature map may be utilized to perform any of a variety of read-temperature-based operations while remaining within the scope of the present disclosure as well.

As will be appreciated by one of skill in the art in possession of the present disclosure, global read temperature identification engine 204 that created the global logical storage element read temperature map 802 may perform any of the read-temperature-based operations discussed above. However, one of skill in the art in possession of the present disclosure will also recognize that other subsystems in the computing device 200 (i.e., other than the global read temperature identification engine 204) may perform the read-temperature-based operations discussed above while remaining within the scope of the present disclosure. Furthermore, subsystems outside the computing device 200 may perform the read-temperature-based operations discussed above while remaining within the scope of the present disclosure as well. As such, access to the global logical storage element read temperature map 802 and/or the local logical storage element read temperature maps 700 and 702 may be provided to subsystems other than the storage devices and global read temperature identification engine 204 discussed above, allowing the information stored therein (as well as information used to generated those maps) to be utilized by those other subsystems in any of a variety of manners that will be apparent to one of skill in the art in possession of the present disclosure.

Thus, systems and methods have been described that utilize the read disturb effect that produces noise in adjacent NAND rows when any particular NAND row is read to identify NAND rows that are read more often than other NAND rows and thus have higher “read temperatures” than those other NAND rows. For example, the read-disturb-based read temperature identification system of the present disclosure may include storage device(s) that each determine read disturb information for each block in that storage device, use that read disturb information to identify a subset of rows in at least one block in that storage device that have a higher read temperature than the other rows in the at least one block in that storage device and, based on that identification, generate and store a local logical storage element read temperature map that identifies a subset of logical storage elements associated with that storage device that have a higher read temperature than the other logical storage elements associated with that storage device. A global read temperature identification subsystem coupled to the storage device(s) may then retrieve the local logical storage element read temperature map generated by each of the storage device(s) and use them to generate a global logical storage element read temperature map.

As such, the read disturb effect that happens automatically in response to conventional read operations and that persists across power cycles may be leveraged to generate read temperature maps for storage devices and storage systems, thus addressing many of the issues with conventional read temperature identification systems discussed above. As will be appreciated by one of skill in the art in possession of the present disclosure, systems and methods of the present disclosure allow a determination of the relative read temperatures of data within storage devices by the storage device themselves (i.e., without requiring processing cycles of a host processor in the server device and/or storage system in which they are located), and with the advent of Storage Class Memory (SCM) devices and low-cost NAND devices that is causing the storage landscape to fracture further than it already has, the opportunity and value associated with placing particular data in the most efficient storage media has increased, and may be realized with the novel read-disturb-based read temperature identification techniques described herein.

Referring now to FIG. 9, an embodiment of the storage device 300 discussed above with reference to FIGS. 3A-3E is illustrated. In the embodiments illustrated and discussed below, the first memory system 308 in the storage device 300 that is coupled to the storage device processing system 304 may include instructions that, when executed by the storage device processing system 304, cause the storage device processing system 304 to provide a local read temperature utilization engine 900 that is configured to perform the functionality of the local read temperature utilization engines and/or storage devices discussed below. As discussed in further detail below, the local read temperature utilization engine 900 may be configured to move data within the storage subsystem 320 based on the local logical storage element read temperature map generated by the storage device as discussed above with reference to the method 400. In some examples, the local read temperature identification engine 204 discussed above and the local read temperature utilization engine 900 described below may be integrated as part of the same engine, although systems with separate local read temperature identification engines and local read temperature utilization engines are envisioned as falling within the scope of the present disclosure as well.

As also illustrated in the specific examples provided herein, the first memory system 308 in the storage device 300 may also include a local read temperature utilization database 902 that is configured to store any of the information utilized by the local read temperature utilization engine 900 discussed below. In some examples, the local read temperature identification database 312 discussed above and the local read temperature utilization database 902 described below may be integrated as part of the same database, although systems with separate local read temperature identification databases and local read temperature utilization databases are envisioned as falling within the scope of the present disclosure as well. However, while a specific storage device 300 has been illustrated and described, one of skill in the art in possession of the present disclosure will recognize that storage devices (or other devices operating according to the teachings of the present disclosure in a manner similar to that described below for the storage device 300) may include a variety of components and/or component configurations for providing conventional storage device functionality, as well as the functionality discussed below, while remaining within the scope of the present disclosure as well.

Referring now to FIG. 10, an embodiment of a method 1000 for storage device utilization of read-disturb-based read temperature maps is illustrated. As discussed below, the systems and methods of the present disclosure provide for the utilization of a local logical storage element read temperature map by the storage device that generated that local logical storage element read temperature map in order to place or move data within that storage device. For example, the storage device read-disturb-based read temperature map utilization system of the present disclosure may include a storage device chassis housing a storage subsystem. A local read temperature utilization subsystem in the storage device chassis determines read disturb information for a plurality of blocks in the storage subsystem, uses it to identify a subset of rows in block(s) in the storage subsystem that have a relatively higher read temperature and, based on those read temperature identifications, generates a local logical storage element read temperature map that identifies a subset of logical storage elements associated with the storage subsystem that have a relatively higher read temperature. The local read temperature utilization subsystem then moves data from first block(s) in the storage subsystem to second block(s) in the storage subsystem based on relative read temperatures identified in the local logical storage element read temperature map. As such, a storage device may move data internally (i.e., between storage locations in its storage subsystem) based on read temperatures for that data that are identified in the local logical storage element read temperature map that it generated according to the teachings of the present disclosure, which allows for more efficient placement of that data within the storage device during garbage collection operations, storage endurance optimization operations, read performance optimization operations, and/or other storage device operations that would be apparent to one of skill in the art in possession of the present disclosure.

The method 1000 begins at block 1002 where a storage device determines read disturb information for blocks in the storage device. In an embodiment, at block 1002, the local read temperature utilization engine 900 in the storage device 300 may operate substantially as described above with reference to block 402 of the method 400 in order to determine read disturb information for a plurality of NAND blocks 324 (and in some embodiment, all of the NAND blocks 324) in the storage subsystem 320. As will be appreciated by one of skill in the art in possession of the present disclosure, the method 1000 may be performed by any of the storage devices 210a-210c in the storage system 208 of the computing device 200 prior to (or in response to) determining data should be moved within the storage device 300. As such, block 1002 may be performed periodically so that the local logical storage element read temperature map has already been updated when it is needed to move data within the storage device, in response to determining that data should be moved as part of garbage collection operations, storage endurance optimization operations, read performance optimization operations, and/or other storage device operations, and/or in other scenarios that would be apparent to one of skill in the art in possession of the present disclosure. Thus, following block 1002, the storage device 300 may have determined read disturb information for a plurality of the NAND rows 332 in a plurality of the NAND blocks 324 (and in some cases, all of the NAND rows 332 in all of the NAND blocks 324) in its storage subsystem 320.

The method 1000 then proceeds to block 1004 where the storage device uses the read disturb information to identify rows in block(s) in the storage device that have higher read temperatures than other rows in block(s) in the storage device. In an embodiment, at block 1004, the local read temperature utilization engine 900 in the storage device 300 may operate substantially as described above with reference to block 404 of the method 400 in order to identify a subset of NAND rows 332 in NAND block(s) 324 in the storage subsystem 320 of the storage device 300 that have higher read temperatures than other NAND rows 332 in NAND block(s) 324 in the storage subsystem 320 of the storage device 300. As discussed in detail above, the read disturb information determined at block 1002 allows the local read temperature utilization engine 900 in the storage device 300 to determine the relative read temperatures of the NAND rows 332 in NAND blocks 324 in the storage subsystem 320 of the storage device 300, which in the examples herein include NAND rows 332 with relatively “hot” read temperatures, NAND rows 332 with relatively “high intermediate” read temperatures, NAND rows 332 with relatively “low intermediate” read temperatures, and NAND rows 332 with relatively “low” read temperatures. However, while four particular relative read temperatures are described in the examples herein, different read temperature granularities (e.g., two read temperatures: “hot” and “cold”, three read temperatures, more than four read temperatures, etc.) are envisioned as falling within the scope of the present disclosure as well. As such, following block 1004, a plurality of NAND rows 332 in a plurality of NAND blocks 324 (and in some cases, all of the NAND rows 332 in all of the NAND blocks 324) in the storage subsystem 320 of the storage device 300 may be associated with a read temperature.

The method 1000 then proceeds to block 1006 where the storage device generates a local logical storage element read temperature map that identifies logical storage elements associated with the storage device with higher read temperatures than other logical storage elements associated with the storage device. In an embodiment, at block 1006, the local read temperature utilization engine 900 in the storage device 300 may operate substantially as described above with reference to block 406 of the method 400 in order to generate a local logical storage element read temperature map that identifies logical storage elements associated with the storage device 300 with higher read temperatures than other logical storage elements associated with the storage device 300. One of skill in the art in possession of the present disclosure will appreciate how, in the specific examples provided below, the storage device 300 generates the local logical storage element read temperature map 700 described above with reference to FIG. 7A. However, local logical storage element read temperature maps generated by different storage devices may differ depending on the read activities performed by (or via) that storage device 300. As such, following block 1006, a local logical storage element read temperature map describing the read temperatures of data stored in the storage subsystem 320 of the storage device 300 may be stored in the local read temperature utilization database 902.

The method 1000 then proceeds to block 1008 where the storage device moves data from first block(s) in the storage device to second block(s) in the storage device based on relative read temperatures identified in the local logical storage element read temperature map. With reference to FIG. 11, in an embodiment of block 1008, the local read temperature utilization engine 900 in the storage device 300 may perform local logical storage element read temperature map access operations 1100 via the first memory interface 306 the first memory system 308 to access the local logical storage element read temperature map generated and stored in its local read temperature utilization database 902 at block 1006, and utilize that local logical storage element read temperature map to perform data movement operations 1102 that include moving data from a first NAND block 324 in its storage subsystem 320 to at least one second NAND block 324 in its storage subsystem 320 via the storage interface 318. As discussed below, how data is moved from the first NAND block 324 to the at least one second NAND block 324 in the storage subsystem 320 based on the local logical storage element read temperature map may depend on the particular data movement operations being performed, and while specific examples of garbage collection operations, storage endurance optimization operations, and read performance optimization operations are described below, one of skill in the art in possession of the present disclosure will appreciate how other data movement operations may be performed at block 1008 utilizing the local logical storage element read temperature map while remaining within the scope of the present disclosure as well.

In one example of block 1008, the data movement operations 1102 may be performed as part of garbage collection operations. As will be appreciated by one of skill in the art in possession of the present disclosure, garbage collection operations for storage devices such as the Solid State Drive (SSD) storage device discussed above operate to optimize space, improve write performance for the storage device by pro-actively eliminating the need for whole NAND block erasure prior to every write operation, and improve efficiency of the storage device. For example, as “valid” data stored in a first NAND block is updated (e.g., a newer version of that data is written to a second NAND block) or deleted, that “valid” data becomes “obsolete” data that remains in the first NAND block, and garbage collection operations may operate to identify the first NAND block when it has a threshold amount of “obsolete” data, move any “valid” data remaining in the first NAND block to a second NAND block, and then erase the first NAND block so that data may later be relatively quickly written to that first NAND block (without the need to perform the erasure of the NAND block at the time of the write operation).

As such, garbage collection operations may involve moving any “valid” data in a NAND block 324 upon which garbage collection operations are being performed. In an embodiment, the local read temperature utilization engine 900 in the storage device 300 may perform garbage collection operations as a background process in the storage device 300, and the data movement operations 1102 performed as part of those garbage collection operations may utilize the local logical storage element read temperature map generated at block 1006 that identifies the read temperatures of the data that is moved in order to provide for more efficient placement of that data within the storage subsystem 320 than is realized with conventional garbage collection operations, as discussed in further detail below.

In another example of block 1008, the data movement operations 1102 may be performed as part of storage endurance optimization operations. As will be appreciated by one of skill in the art in possession of the present disclosure, storage endurance optimization operations for storage devices such as the SSD storage device discussed above operate to place data within the storage subsystem 320 of the storage device 300 in a manner that increases the “endurance” or life of the storage device 300 by, for example, reducing write amplification in the storage device 300. For example, each time data is provided by a host or other subsystem for writing to the storage device 300, the storage device 300 may have to perform multiple write operations to move “valid” data, erase “obsolete” data, and/or perform other data writes that allow the data provided by the host to be written to the storage device 300. As will be appreciated by one of skill in the art in possession of the present disclosure, the number of times data may be written to and erased from the storage subsystem 320 in the storage device 300 is limited, and by reducing the write amplification discussed above the “endurance” or life of the storage device 300 may be extended.

As such, the data movement operations 1102 may utilize the local logical storage element read temperature map to move data in order to reduce the write amplification discussed above by placing data in the storage subsystem 320 in a manner that reduces the number of write operations performed by the storage device 300 in response to any single write request from a host (e.g., moving first data in a NAND block with a relatively “high” read temperature into a first NAND block and second data in a NAND block with a relatively “low” read temperature into a second NAND block so that subsequent moves of the first data do not require corresponding moves of the second data, grouping data in a NAND block with a relatively “high” read temperature in adjacent NAND rows so that read disturb effects associated with any of that data only effects other relatively “high” read temperature data, grouping data in a NAND block in manner that reduces the number of data moves required to erase that NAND block during garbage collection operations, etc.).

Furthermore, as the number of reads from a NAND block 324 increase, the ability to perform error correction for that data reduces (e.g., as reads increase, errors increase, and because Error Correction Codes (ECCs) provide a limit to the number of data errors that can be corrected, the capabilities of ECC to correct errors may be exceeded after a threshold number of reads). As such, storage endurance may be optimized by ensuring that any movement/placement of data within the storage subsystem 320 will provide that data in a location from which it will not be moved for as long as possible (e.g., due to read disturb issues, error correction limits being reached/exceeded, etc.) As will be appreciated by one of skill in the art in possession of the present disclosure, each time data is read from a NAND block 324, the error correction information associated with that data may indicate how reliable that data is and may require movement of that data in the event its reliability drops below a reliability threshold.

As such, the local read temperature utilization engine 900 may operate to determine that the data in a first NAND block in the storage subsystem 320 has reached a reliability threshold and, in response, may move the data from the first NAND block in the storage subsystem 320 to at least one second NAND block in the storage subsystem 320 as part of the data movement operations 1102. In an embodiment, the local read temperature utilization engine 900 in the storage device 300 may perform storage endurance optimization operations as a background process in the storage device 300, and the data movement operations 1102 performed as part of those storage endurance optimization operations may utilize the local logical storage element read temperature map generated at block 1006 that identifies the read temperatures of the data that is moved in order to provide for more efficient placement of that data within the storage subsystem 320 than is realized with conventional storage endurance optimization operations, as discussed in further detail below.

In some embodiments, data may be moved into different NAND subsystem groups within the storage subsystem 320 as part of the storage endurance optimization operations described above, and one of skill in the art in possession of the present disclosure will appreciate that by separating the relatively “high” read temperature data and the relatively “low” read temperature data in different NAND subsystems will improve the “endurance” or lifespan of the storage device 300 by, for example, placing relatively “high” read temperature data in different NAND blocks than relatively “low” read temperature data (e.g., with later movement of the relatively “high” read temperature data to a different NAND block not requiring movement of the relatively “low” read temperature data that is stored on different NAND block(s), thus reducing the number of writes utilized with the relatively “low” read temperature data).

In another example of block 1008, the data movement operations 1102 may be performed as part of read performance optimization operations. As will be appreciated by one of skill in the art in possession of the present disclosure, read performance optimization operations for storage devices such as the SSD storage device discussed above operate to place data within the storage subsystem 320 of the storage device 300 in order to reduce data retrieval delays due to storage subsystem bandwidth issues. For example, the storage subsystem 320 may introduce limits to concurrent accesses of NAND blocks 324 due to the bandwidth available via the physical channel provided to the NAND die 322 that includes those NAND blocks 324. As such, if too many read requests are received for data stored in NAND blocks 324 in the same NAND die 322, a queue may be created so that reads of any particular NAND row 332 in those NAND blocks 324 may be made sequentially via the physical channel provided to the NAND die 322, thus delaying read operations placed further down in the queue. In an embodiment, the local read temperature utilization engine 900 in the storage device 300 may perform read performance optimization operations as a background process in the storage device 300, and the data movement operations 1102 performed as part of those read performance optimization operations may utilize the local logical storage element read temperature map generated at block 1006 that identifies the read temperatures of the data that is moved in order to provide for more efficient placement of that data within the storage subsystem 320 than is realized with conventional read performance optimization operations, as discussed in further detail below.

In some embodiments, data may be moved into different NAND subsystem groups within the storage subsystem 320 as part of the read performance optimization operations described above, and one of skill in the art in possession of the present disclosure will appreciate that by storing the relatively “high” read temperature data across different NAND subsystems will improve the read performance of the storage device 300 by, for example, placing relatively “high” read temperature data on different NAND die 322 (e.g., splitting up the relatively “high” read temperature data in the storage subsystem 320 across different NAND die 322 such that the bandwidth of the physical channels to those NAND die 322 will not be exceeded when concurrent accesses of that relatively “high” read temperature data are performed), or placing relatively “high” read temperature data on different NAND blocks 324 (e.g., splitting up the relatively “high” read temperature data in the storage subsystem 320 across different NAND blocks 324 to prevent “read contention″/degraded Quality of Service (QoS) that would otherwise result if that relatively “high” read temperature data were all stored on the same NAND block 324).

As such, the local read temperature utilization engine 900 may determine that the read performance of the storage subsystem 320 is below a read performance threshold and, in response, may move data from first NAND blocks in the storage subsystem 320 to second NAND blocks in the storage subsystem 320 such that the read performance of the storage subsystem 320 is improved. As will be appreciated by one of skill in the art in possession of the present disclosure, such data movement operations 1102 performed as part of read performance optimization operations may be configured to distribute the data across the different NAND die to optimize the average read response time, QoS, and/or other performance characteristics of the storage device 300 that would be apparent to one of skill in the art in possession of the present disclosure. In some embodiments, different data that is accessed often (e.g., relatively “high” read temperature data) may be correlated and then placed in NAND blocks in different NAND die 322 as part of the read performance optimization operations. Furthermore, while the discussion above focuses on the optimization of read response, one of skill in the art in possession of the present disclosure will appreciate how other storage device characteristics (e.g., wear) may be optimized (e.g., to reduce wear on the storage device) in a similar manner while remaining within the scope of the present disclosure as well.

With reference to FIG. 12, a specific embodiment of the data movement operations 1102 discussed above is illustrated using the example of the local logical storage element read temperature map 700 of FIG. 7A that is stored in the local read temperature utilization database 902 and that identifies read temperatures of logical storage elements that, as discussed above, were mapped to physical storage locations in the storage subsystem 320. FIG. 12 also illustrates how the storage subsystem 320 may have had its NAND subsystems (NAND dies, NAND blocks, NAND rows, NAND pages, etc.) grouped into relative performance NAND subsystems. As will be appreciated by one of skill in the art in possession of the present disclosure, different NAND subsystems (NAND dies, NAND blocks, NAND rows, NAND pages, etc.) in the storage device 300 may have different read speeds or other storage device characteristics known in the art, and thus those with higher performance characteristics (e.g., relatively faster read speeds) may be grouped separately from those with lower performance characteristics (e.g., relatively slower read speeds). As such, the illustrated example provides the grouping of relatively “high” performance NAND subsystem(s) 1200a, relatively “high intermediate” performance NAND subsystem(s) 1200b, relatively “low intermediate” performance NAND subsystem(s) 1200c, and relatively “low” performance NAND subsystem(s) 1200d within the storage subsystem 320.

The specific example in FIG. 12 illustrates how the data movement operations 1102 performed at block 1008 may include relatively “high” read temperature data moves 1202a, 1202b, 1202c, 1202d, and 1202e that move data associated with the logical storage elements in the local logical storage element read temperature map 700 that have relatively “high” read temperatures (e.g., the black logical storage elements in FIG. 12) from their current NAND subsystem to the high performance NAND subsystem(s) 1200a in the storage subsystem 320. Furthermore, while not explicitly illustrated, one of skill in the art in possession of the present disclosure will appreciate how the data movement operations 1102 performed at block 1008 may include relatively “high-intermediate” read temperature data moves that move data associated with the logical storage elements in the local logical storage element read temperature map 700 that have relatively “high-intermediate” read temperatures (e.g., the dark grey logical storage elements in FIG. 12) from their current NAND subsystem to the high-intermediate performance NAND subsystem(s) 1200b in the storage subsystem 320, relatively “low-intermediate” read temperature data moves that move data associated with the logical storage elements in the local logical storage element read temperature map 700 that have relatively “low-intermediate” read temperatures (e.g., the light grey logical storage elements in FIG. 12) from their current NAND subsystem to the low-intermediate performance NAND subsystem(s) 1200c in the storage subsystem 320, and relatively “low” read temperature data moves that move data associated with the logical storage elements in the local logical storage element read temperature map 700 that have relatively “low-intermediate” read temperatures (e.g., the white logical storage elements in FIG. 12) from their current NAND subsystem to the low performance NAND subsystem(s) 1200d in the storage subsystem 320.

In some embodiments, the movement of data into different performance NAND subsystem groups may be performed as part of the garbage collection operations described above, and one of skill in the art in possession of the present disclosure will appreciate that by placing the relatively “high” read temperature data in the high performance NAND subsystems 1200a will improve the read performance of the storage device 300 (e.g., reads of more frequently read data will be quicker because that data is now stored in NAND subsystems that provide for faster reads). As such, data movements performed as part of garbage collection operations may leverage the local logical read temperature maps of the present disclosure to place data within the storage device 300 in a manner that improves the efficiency, performance, life, and/or other characteristics of the storage device 300.

With reference to FIG. 13, a specific embodiment of the data movement operations 1102 discussed above is illustrated using the example of the local logical storage element read temperature map 700 of FIG. 7A that is stored in the local read temperature utilization database 902 and that identifies read temperatures of logical storage elements that, as discussed above, were mapped to physical storage locations in the storage subsystem 320. FIG. 13 also illustrates how the storage subsystem 320 may have a NAND block 1300 with NAND row groups including a NAND row group 1302 having adjacent NAND rows 1302a, 1302b, and up to 1302c; a NAND row group 1304 having adjacent NAND rows 1304a, 1304b, and up to 1304c; a NAND row group 1306 having adjacent NAND rows 1306a, 1306b, and up to 1306c; and a NAND row group 1308 having adjacent NAND rows 1308a, 1308b, and up to 1308c.

The specific example in FIG. 13 illustrates how the data movement operations 1102 performed at block 1008 may include relatively “high” read temperature data moves 1310a, 1310b, 1310c, 1310d, and 1310e that move data associated with the logical storage elements in the local logical storage element read temperature map 700 that have relatively “high” read temperatures (e.g., the black logical storage elements in FIG. 13) from their current NAND rows to the NAND rows 1302a-1302a in the NAND row group 1302 in the storage subsystem 320. Furthermore, while not explicitly illustrated, one of skill in the art in possession of the present disclosure will appreciate how the data movement operations 1100 performed at block 1008 may include relatively “high-intermediate” read temperature data moves that move data associated with the logical storage elements in the local logical storage element read temperature map 700 that have relatively “high-intermediate” read temperatures (e.g., the dark grey logical storage elements in FIG. 12) from their current NAND rows to the NAND rows 1304a-1304a in the NAND row group 1304 in the storage subsystem 320, relatively “low-intermediate” read temperature data moves that move data associated with the logical storage elements in the local logical storage element read temperature map 700 that have relatively “low-intermediate” read temperatures (e.g., the light grey logical storage elements in FIG. 12) from their current NAND rows to the NAND rows 1306a-1306a in the NAND row group 1306 in the storage subsystem 320, and relatively “low” read temperature data moves that move data associated with the logical storage elements in the local logical storage element read temperature map 700 that have relatively “low-intermediate” read temperatures (e.g., the white logical storage elements in FIG. 12) from their current NAND rows to the NAND rows 1308a-1308a in the NAND row group 1308 in the storage subsystem 320.

In some embodiments, the movement of data into different NAND row groups in a NAND block may be performed as part of the storage endurance optimization operations described above, and one of skill in the art in possession of the present disclosure will appreciate that by placing the relatively “high” read temperature data in the adjacent NAND rows 1302a-1302c in the NAND group 1302, the relatively “high-intermediate” read temperature data in the adjacent NAND rows 1304a-1304c in the NAND group 1304, the relatively “low-intermediate” read temperature data in the adjacent NAND rows 1306a-1306c in the NAND group 1306, and the relatively “low” read temperature data in the adjacent NAND rows 1308a-1308c in the NAND group 1308, will allow the movement of the relatively “high” read temperature data (e.g., when associated error correction limits are reached) without the need to move some or all of the relatively lower read temperature data stored in the NAND block 1300.

Thus, systems and methods have been described that provide for the efficient placement or movement of data in a storage device using a local logical storage element read temperature map that is generated by that storage device based on read disturb effects in its physical storage media. For example, the read-disturb-based read temperature map utilization system of the present disclosure may include a storage device chassis housing a storage subsystem. A local read temperature utilization subsystem in the storage device chassis determines read disturb information for a plurality of blocks in the storage subsystem, uses it to identify a subset of rows in block(s) in the storage subsystem that have a relatively higher read temperature and, based on those read temperature identifications, generates a local logical storage element read temperature map that identifies a subset of logical storage elements associated with the storage subsystem that have a relatively higher read temperature. The local read temperature utilization subsystem then moves data from first block(s) in the storage subsystem to second block(s) in the storage subsystem based on relative read temperatures identified in the local logical storage element read temperature map. As such, storage devices may internally move data based on read temperatures for that data that are identified in the local logical storage element read temperature maps of the present disclosure, which allows for more efficient placement of that data within the storage device during garbage collection operations, storage endurance optimization operations, read performance optimization operations, and/or other storage device operations that would be apparent to one of skill in the art in possession of the present disclosure.

Although illustrative embodiments have been shown and described, a wide range of modification, change and substitution is contemplated in the foregoing disclosure and in some instances, some features of the embodiments may be employed without a corresponding use of other features. Accordingly, it is appropriate that the appended claims be construed broadly and in a manner consistent with the scope of the embodiments disclosed herein.

Claims

1. A storage device read-disturb-based read temperature map utilization system, comprising: a storage device chassis;a storage subsystem that is included in the storage device chassis;a local read temperature utilization subsystem that is included in the storage device chassis and that is configured to: determine, for a plurality of blocks in the storage subsystem, read disturb information;identify, using the read disturb information, a subset of rows in at least one block in the storage subsystem that have a higher read temperature than the other rows in the at least one block in the storage subsystem;generate, based on the identification of the subset of rows in the at least one block in the storage subsystem that have the higher read temperature than the other rows in the at least one block in the storage subsystem, a local logical storage element read temperature map that identifies a subset of logical storage elements associated with the storage subsystem that have a higher read temperature than the other logical storage elements associated with the storage subsystem; andmove data from at least one first block in the storage subsystem to at least one second block in the storage subsystem based on relative read temperatures identified in the local logical storage element read temperature map.

2. The system of claim 1, wherein the moving the data from the at least one first block in the storage subsystem to the at least one second block in the storage subsystem includes: moving data associated with the subset of logical storage elements that have a higher read temperature than the other logical storage elements to the at least one second block in the storage subsystem that has higher performance data access capabilities than the at least one first block.

3. The system of claim 1, wherein the moving the data from the at least one first block in the storage subsystem to the at least one second block in the storage subsystem includes: grouping data associated with the subset of logical storage elements that have a higher read temperature than the other logical storage elements in adjacent rows in the at least one second block in the storage subsystem.

4. The system of claim 1, wherein the moving the data from the at least one first block in the storage subsystem to the at least one second block in the storage subsystem includes: moving subsets of data associated with the subset of logical storage elements that have a higher read temperature than the other logical storage elements to the at least one second block in the storage subsystem such that each subset of the data is moved to a different die in the storage subsystem.

5. The system of claim 1, wherein the local read temperature utilization subsystem is configured to: determine that garbage collection operations should be performed on the at least one first block in the storage subsystem and, in response, move the data from the at least one first block in the storage subsystem to the at least one second block in the storage subsystem.

6. The system of claim 1, wherein the local read temperature utilization subsystem is configured to: determine that the data in at least one first block in the storage subsystem has reached a reliability threshold and, in response, move the data from the at least one first block in the storage subsystem to the at least one second block in the storage subsystem.

7. The system of claim 1, wherein the local read temperature utilization subsystem is configured to: determine that the read performance of the storage subsystem is below a read performance threshold and, in response, move the data from the at least one first block in the storage subsystem to the at least one second block in the storage subsystem.

8. A storage device, comprising: a processing system; anda memory system that is coupled to the processing system and that includes instructions that, when executed by the processing system, cause the processing system to provide a local read temperature utilization engine that is configured to: determine, for a plurality of blocks in a storage subsystem, read disturb information;identify, using the read disturb information, a subset of rows in at least one block in the storage subsystem that have a higher read temperature than the other rows in the at least one block in the storage subsystem;generate, based on the identification of the subset of rows in the at least one block in the storage subsystem that have the higher read temperature than the other rows in the at least one block in the storage subsystem, a local logical storage element read temperature map that identifies a subset of logical storage elements associated with the storage subsystem that have a higher read temperature than the other logical storage elements associated with the storage subsystem; andmove data from at least one first block in the storage subsystem to at least one second block in the storage subsystem based on relative read temperatures identified in the local logical storage element read temperature map.

9. The storage device of claim 8, wherein the moving the data from the at least one first block in the storage subsystem to the at least one second block in the storage subsystem includes: moving data associated with the subset of logical storage elements that have a higher read temperature than the other logical storage elements to the at least one second block in the storage subsystem that has higher performance data access capabilities than the at least one first block.

10. The storage device of claim 8, wherein the moving the data from the at least one first block in the storage subsystem to the at least one second block in the storage subsystem includes: grouping data associated with the subset of logical storage elements that have a higher read temperature than the other logical storage elements in adjacent rows in the at least one second block in the storage subsystem.

11. The storage device of claim 8, wherein the moving the data from the at least one first block in the storage subsystem to the at least one second block in the storage subsystem includes: moving subsets of data associated with the subset of logical storage elements that have a higher read temperature than the other logical storage elements to the at least one second block in the storage subsystem such that each subset of the data is moved to a different die in the storage subsystem.

12. The storage device of claim 8, wherein the local read temperature utilization engine is configured to: determine that garbage collection operations should be performed on the at least one first block in the storage subsystem and, in response, move the data from the at least one first block in the storage subsystem to the at least one second block in the storage subsystem.

13. The storage device of claim 8, wherein the local read temperature utilization engine is configured to: determine that the data in at least one first block in the storage subsystem has reached a reliability threshold and, in response, move the data from the at least one first block in the storage subsystem to the at least one second block in the storage subsystem.

14. A method for storage device utilization of a read-disturb-based read temperature map, comprising: determining, by a storage device for a plurality of blocks in the storage device, read disturb information;identifying, using the read disturb information, a subset of rows in at least one block in the storage device that have a higher read temperature than the other rows in the at least one block in the storage device;generating, based on the identification of the subset of rows in the at least one block in the storage device that have the higher read temperature than the other rows in the at least one block in the storage device, a local logical storage element read temperature map that identifies a subset of logical storage elements associated with the storage device that have a higher read temperature than the other logical storage elements associated with the storage device; andmoving data from at least one first block in the storage device to at least one second block in the storage device based on relative read temperatures identified in the local logical storage element read temperature map.

15. The method of claim 14, wherein the moving the data from the at least one first block in the storage device to the at least one second block in the storage device includes: moving data associated with the subset of logical storage elements that have a higher read temperature than the other logical storage elements to the at least one second block in the storage device that has higher performance data access capabilities than the at least one first block.

16. The method of claim 14, wherein the moving the data from the at least one first block in the storage device to the at least one second block in the storage device includes: grouping data associated with the subset of logical storage elements that have a higher read temperature than the other logical storage elements in adjacent rows in the at least one second block in the storage device.

17. The method of claim 14, wherein the moving the data from the at least one first block in the storage device to the at least one second block in the storage device includes: moving subsets of data associated with the subset of logical storage elements that have a higher read temperature than the other logical storage elements to the at least one second block in the storage device such that each subset of the data is moved to a different die in the storage device.

18. The method of claim 14, further comprising: determining, by the storage device, that garbage collection operations should be performed on the at least one first block in the storage device and, in response, move the data from the at least one first block in the storage device to the at least one second block in the storage device.

19. The method of claim 14, further comprising: determining, by the storage device, that the data in at least one first block in the storage device has reached a reliability threshold and, in response, move the data from the at least one first block in the storage device to the at least one second block in the storage device.

20. The method of claim 14, further comprising: determining, by the storage device, that the read performance of the storage device is below a read performance threshold and, in response, move the data from the at least one first block in the storage device to the at least one second block in the storage device.