Patent Grant
9671962
The present invention relates generally to a storage control system and more particularly to a system for data management of parity.
Data storage, often called storage or memory, refers to computer components and recording media that retain digital data. Data storage is a core function and fundamental component of consumer and industrial electronics, especially devices such as computers, televisions, cellular phones, mobile devices, and digital video cameras.
Recently, forms of long-term storage other than electromechanical hard disks have become feasible for use in computers. NOT-AND (NAND) flash is one form of non-volatile memory used in solid-state storage devices. The memory cells are arranged in typical row and column fashion with circuitry for accessing individual cells. The memory transistors of those cells are placed to store an analog value that can be interpreted to hold two logical states in the case of Single Level Cell (SLC) or more than two logical states in the case of Multi Level Cell (MLC).
A flash memory cell is light in weight, occupies very little space, and consumes less power than electromechanical disk drives. Construction of a storage system with this type of memory allows for much higher bandwidths and input/output operations per second (IOPS) than typical electromechanical disk drives. More importantly, it is especially rugged and can operate at a much high temperature range. It will withstand without adverse effects repeated drops, each of which would destroy a typical electromechanical hard disk drive. A problem exhibited by flash memory is that it tends to have a limited life in use.
Thus, a need still remains for better data management devices. In view of the increasing demand for data management devices, it is increasingly critical that answers be found to these problems. In view of the ever-increasing commercial competitive pressures, along with growing consumer expectations and the diminishing opportunities for meaningful product differentiation in the marketplace, it is critical that answers be found for these problems. Additionally, the need to reduce costs, improve efficiencies and performance, and meet competitive pressures adds an even greater urgency to the critical necessity for finding answers to these problems.
Solutions to these problems have been long sought but prior developments have not taught or suggested any solutions and, thus, solutions to these problems have long eluded those skilled in the art.
The present invention provides a method of operation of a storage control system, including: partitioning memory channels with memory devices; selecting a super device with one of the memory devices from one of the memory channels; selecting a super block associated with the super device; and determining a location of a parity within the super block when the super block is formed.
The present invention provides a storage control system, including: a channel module for partitioning memory channels with memory devices; a device module, coupled to the channel module, for selecting a super device with one of the memory devices from one of the memory channels; a block module, coupled to the device module, for selecting a super block associated with the super device; and a parity module, coupled to the block module, for determining a location of a parity within the super block when the super block is formed.
Certain embodiments of the invention have other steps or elements in addition to or in place of those mentioned above. The steps or elements will become apparent to those skilled in the art from a reading of the following detailed description when taken with reference to the accompanying drawings.
The following embodiments are described in sufficient detail to enable those skilled in the art to make and use the invention. It is to be understood that other embodiments would be evident based on the present disclosure, and that system, process, or mechanical changes may be made without departing from the scope of the present invention.
In the following description, numerous specific details are given to provide a thorough understanding of the invention. However, it will be apparent that the invention may be practiced without these specific details. In order to avoid obscuring the present invention, some well-known circuits, system configurations, and process steps are not disclosed in detail.
The drawings showing embodiments of the system are semi-diagrammatic and not to scale and, particularly, some of the dimensions are for the clarity of presentation and are shown exaggerated in the drawing FIGs.
Where multiple embodiments are disclosed and described having some features in common, for clarity and ease of illustration, description, and comprehension thereof, similar and like features one to another will ordinarily be described with similar reference numerals. The embodiments have been numbered first embodiment, second embodiment, etc. as a matter of descriptive convenience and are not intended to have any other significance or provide limitations for the present invention.
The term “module” referred to herein can include software, hardware, or a combination thereof in the present invention in accordance with the context in which the term is used. For example, the software can be machine code, firmware, embedded code, and application software. Also for example, the hardware can be circuitry, processor, computer, integrated circuit, integrated circuit cores, a microelectromechanical system (MEMS), passive devices, environmental sensors including temperature sensors, or a combination thereof.
Referring now to
The memory controller 104 provides data control and management of the memory array 106. The memory controller 104 interfaces with the host system 108 and controls the memory array 106 to transfer data between the host system 108 and the memory array 106.
The memory array 106 includes an array of memory devices 110 including flash memory devices or non-volatile memory devices. The memory array 106 can include pages of data or information. The host system 108 can request the memory controller 104 for reading, writing, and deleting data from or to a logical address space of a storage device or the memory sub-system 102 that includes the memory array 106.
The memory devices 110 can include chip selects 112, which are defined as control inputs, for enabling the memory devices 110. Each of the chip selects 112 can be used to control the operation of one of the memory devices 110. Each of the chip selects 112 can be used to control sub addresses or logical units (LUN) within each of the memory devices 110. When the chip selects 112 are enabled, the memory devices 110 are in active state for operation including reading, writing, or erasing.
Referring now to
The control unit 202 can be implemented in a number of different manners. For example, the control unit 202 can be a processor, an embedded processor, a microprocessor, a hardware control logic, a hardware finite state machine (FSM), a digital signal processor (DSP), or a combination thereof.
The control interface 210 can be used for communication between the control unit 202 and other functional units in the memory controller 104. The control interface 210 can also be used for communication that is external to the memory controller 104.
The control interface 210 can receive information from the other functional units or from external sources, or can transmit information to the other functional units or to external destinations. The external sources and the external destinations refer to sources and destinations external to the memory controller 104.
The control interface 210 can be implemented in different ways and can include different implementations depending on which functional units or external units are being interfaced with the control interface 210. For example, the control interface 210 can be implemented with a dedicated hardware including an application-specific integrated circuit (ASIC), a configurable hardware including a field-programmable gate array (FPGA), a discrete electronic hardware, or a combination thereof.
The storage unit 204 can include both hardware and the software 212. For example, the software 212 can include control firmware. The storage unit 204 can include a volatile memory, a nonvolatile memory, an internal memory, an external memory, or a combination thereof. For example, the storage unit 204 can be a nonvolatile storage such as non-volatile random access memory (NVRAM), Flash memory, disk storage, or a volatile storage such as static random access memory (SRAM).
The storage unit 204 can include a storage interface 214. The storage interface 214 can also be used for communication that is external to the memory controller 104. The storage interface 214 can receive information from the other functional units or from external sources, or can transmit information to the other functional units or to external destinations. The external sources and the external destinations refer to sources and destinations external to the memory controller 104.
The storage interface 214 can include different implementations depending on which functional units or external units are being interfaced with the storage unit 204. The storage interface 214 can be implemented with technologies and techniques similar to the implementation of the control interface 210.
The memory interface unit 206 can enable external communication to and from the memory controller 104. For example, the memory interface unit 206 can permit the memory controller 104 to communicate with the memory array 106 of
The memory interface unit 206 can include a memory interface 216. The memory interface 216 can be used for communication between the memory interface unit 206 and other functional units in the memory controller 104. The memory interface 216 can receive information from the other functional units or can transmit information to the other functional units.
The memory interface 216 can include different implementations depending on which functional units are being interfaced with the memory interface unit 206. The memory interface 216 can be implemented with technologies and techniques similar to the implementation of the control interface 210.
The host interface unit 208 allows the host system 108 of
The control unit 202 can operate the host interface unit 208 to send control or status information generated by the memory controller 104 to the host system 108. The control unit 202 can also execute the software 212 for the other functions of the memory controller 104. The control unit 202 can further execute the software 212 for interaction with the memory array 106 via the memory interface unit 206.
The functional units in the memory controller 104 can work individually and independently of the other functional units. For illustrative purposes, the memory controller 104 is described by operation of the memory controller 104 with the host system 108 and the memory array 106. It is understood that the memory controller 104, the host system 108, and the memory array 106 can operate any of the modules and functions of the memory controller 104.
Referring now to
Each of the memory groups 302 can be identified as being associated with a number of the memory devices 110. Although each of the memory groups 302 is shown associated with eight of the memory devices 110 for illustration purposes, it is understood that each of the memory groups 302 can be associated with any number of the memory devices 110.
Each of the squares in
Each of memory channels 304 can be associated with a number of the memory devices 110. Each of the memory channels 304 is defined as a physical bank or a physical group of the memory devices 110 that share a common data bus and a common set of associated control signals.
The memory channels 304 can represent an independent bus or communication path. Each of the memory channels 304 can represent an independent execution unit. Each of the memory channels 304 can include an ability to perform independent and parallel operations with respect to other instances of the memory channels 304.
Actions, commands, and status information transmitted within one of the memory channels 304 can be limited to the memory devices 110 that are connected to the one of the memory channels 304 and can be independent from other instances of the memory channels 304 within a storage device. Data bus and control signals of one of the memory channels 304 cannot be shared with the other instances of the memory channels 304.
Each of the memory channels 304 can include a number of bits of data based on a type of the memory devices 110. For example, each of the memory channels 304 can include 8 bits, 16 bits, or any number of bits of data bus depending on the type of the memory devices 110 used and the construction of the memory controller 104 of
For example, one of the memory channels 304 can control or have ownership of a set of the memory devices 110 that share a common data bus and control signals. Also for example, the memory devices 110 can represent flash devices or flash components that independently operate from each other.
The super devices 306, which are defined as groups of the memory devices 110, can be identified or selected. The super devices 306 can include portions or all of the memory channels 304. The super devices 306 can be selected with one of the memory devices 110 from one of the memory channels 304. The term “super” in the super devices 306 refers to a group of multiple hardware instances of the memory devices 110.
The memory devices 110 from the memory channels 304 can be grouped to form the super devices 306. For example, if there are 16 of the memory channels 304 and 8 of the memory devices 110 per each of the memory channels 304, there can be 8 of the memory devices 110, one from each of the memory channels 304, grouped together for each of the super devices 306 and 16 of the super devices 306 as shown in
The super devices 306 can also include multiple of the memory devices 110 from each of the memory channels 304. At least one of the super devices 306 can be formed by taking multiple passes through the memory channels 304. In addition, at least one of the super devices 306 can span the memory channels 304 as much as possible but it is not strictly bounded by the memory channels 304.
Each of the super devices 306 can be associated with one of the memory devices 110 from each of the memory channels 304. Although each of the super devices 306 is shown having 8 of the memory devices 110 for illustration purposes, it is understood that each of the super devices 306 can include any number of the memory devices 110. For example, each of the super devices 306 can include up to a number of the memory channels 304. Also for example, all of the super devices 306 can have the same structure and size. Each of the super devices 306 can include any number of the memory devices 110 per each of the memory channels 304. In this case, there can be multiple passes made or accessed through the memory channels 304. For example, each of the super devices 306 can include a number of the memory channels 304 with 2 of the memory devices 110 per each of the memory channels 304.
Each of the super devices 306 can include one of the memory devices 110 from each of the memory channels 304. Each of the super devices 306 can include a number of the super blocks 308, each of which can include a number of erase blocks 310 (EB). The super blocks 308 are defined as logical groupings of the erase blocks 310 in the memory devices 110. Although there are 16 of the super devices 306 shown for illustration purposes, there can be any number of the super devices 306 identified. The number of the super devices 306 can be based on a number of independently accessible instances of the memory devices 110 on the memory channels 304 selected by Chip Enables (CEs), logical units (LUN), or a combination thereof.
The erase blocks 310 are units of a non-volatile memory device or other types of block oriented non-volatile memory devices. For example, the erase blocks 310 can be one of atomic units of a NAND flash memory device or other types of block oriented non-volatile memory devices, which no finer control over erasing memory elements is capable. Also for example, the erase blocks 310 can form a part of a logically addressable memory space within a NAND flash or other type of block oriented non-volatile memory devices.
The erase blocks 310 can represent the smallest unit of memory that can be erased. The memory devices 110 can represent NAND packages that can include multiple dies or logical units (LUN).
As an example, each of the super devices 306 can access one of the super blocks 308 at any given instance. Each of the super blocks 308 can include a number of the erase blocks 310 that are accessible by one of the super devices 306 at any given time. There can be one of the erase blocks 310 active for each of the memory channels 304 within one of the super devices 306. The erase blocks 310 can be assigned across the memory channels 304 so that they can be accessed in a parallel approach. One of the erase blocks 310 can be assigned per each of the memory channels 304 to form one of the super blocks 308 in the super devices 306.
In an example of one of the super devices 306 having 8 of the memory devices 110, there can be 8 active instances of the erase blocks 310 associated with the one of the super devices 306. The 8 active instances of the erase blocks 310 can belong to a specific instance of the super blocks 308. Also as an example, there can be multiple active instances of the erase blocks 310 per each of the memory devices 110.
The present invention can include architecture and use of the memory devices 110 including NAND flash or other block oriented non-volatile memory organized in efficient groups to reduce the overhead of and runtime management in a solid state storage device. The present invention can include the super blocks 308 as an organization of memory blocks for parallel use by multiple memory controllers and the logical and physical linkage between the memory blocks. The present invention can provide additional organization for data protection via RAID operations with reduced overhead for both runtime recycling and drive initialization speed. The present invention can provide an additional structure for managing bad block memory elements without detriment or loss of system performance or efficiency.
The super blocks 308 provide the following effects. One of the effects is amount of metadata that is stored in each of the erase blocks 310 is minimized or reduced. The metadata can now be stored at a super-block level or in one of the super blocks 308. Another of the effects is fewer number of the erase blocks 310 are managed as well as reduction in time to initialize these data structures.
For example, the metadata can include non-host related information including logical-to-physical mapping information of the host data within the memory devices 110. Also for example, the metadata can include non-host related information including logical block address (LBA) information, super block identification (ID) values, program/erase (PE) cycles, aging data, and other NAND statistics that is collected as a consequence of normal operations.
Another of the effects is the host data is written to the super blocks 308 in a deterministic order. Another of the effects is recycling can traverse or access the super blocks 308 in the same order by which the host system 108 of
It has been discovered that the super devices 306 provide improved performance by grouping of the memory devices 110 that are controlled in parallel across a number of the memory channels 304.
Referring now to
Each of the super blocks 308 is a group of the erase blocks 310 that span across multiple of the memory channels 304 within one of the super devices 306. As an example, each of the memory devices 110 in the super blocks 308 can contribute a pair of the erase blocks 310 that can be accessed together. Each of the super blocks 308 can be selected to be associated with the super devices 306. Each of the super blocks 308 can be accessed via and associated with only one of the super devices 306. Since the super devices 306 span across multiple of the memory channels 304, the super blocks 308 can also span across multiple of the memory channels 304. The term “spanning across” used above for the super blocks 308 refers to portions of different instances of the memory channels 304 are included in each of the super blocks 308.
There can be a number of the super devices 306 included in a solid-state drive or a storage system dictated by a number of addressable elements within the memory channels 304 that can be controlled in a parallel method of operation. There can be a number of the super blocks 308 per each of the super devices 306. Any of the super blocks 308 can be associated with one and only one of the super devices 306.
All the chip selects 112 of
One of the super chip selects 404 is selected to be associated with one of the super blocks 308 and a group of the chip selects 112 of the memory devices 110 that are associated with the one of the super blocks 308. The one of the super chip selects 404 can be used to control the group of the chip selects 112. The super chip selects 404 are associated with the super blocks 308. The super chip selects 404 are provided to control operation of the memory devices 110.
The super chip selects 404 can be used to assert or de-assert the chip selects 112. The chip selects 112 can be asserted or de-asserted to enable or disable, respectively, the memory devices 110. For example, the chip selects 112 can be asserted or de-asserted by assigning the super chip selects 404 to an active or inactive state such that the chip selects 112 are driven to logic ‘1’ or ‘0’, respectively, to enable or disable the memory devices 110.
Each of the super blocks 308 can span or include only one of the super chip selects 404. By having multiple of the super chip selects 404 for multiple of the super blocks 308, one of the super chip selects 404 can be available for writing host data 405 from the host system 108 of
The partition of the chip selects 112 and the provision of the super chip selects 404 eliminate conflict between host write and block erase operations because these operations are performed with difference instances of the super chip selects 404. Furthermore, conflict between host read and host write (or recycle) operations can also be eliminated by mutual exclusion between operations on the super chip selects 404.
If there are several pending host reads, there is a good chance that one of them can be located on one of the super chip selects 404 that are idle. If data requested for a host read operation was written by a large-block host write operation, there is a good chance that all the data can be read with one of the super chip selects 404 that was used for the large-block host write operation.
The last instance of the erase blocks 310 at the end of one of the super blocks 308 can be written. The last instance of the erase blocks 310 can be written with a parity 406 for RAID-5 architectures. As such, the size of each of the super blocks 308 can determine the overhead in the RAID-5 architectures. The last instance of the erase blocks 310 is the last block to be accessed in the order by which the erase blocks 310 in the super blocks 308 are written by the host system 108. RAID-5 refers to an implementation for redundant array of independent disks (RAID) systems that uses block-level striping with parity data distributed across all member disks.
The parity 406 is defined as redundancy data that is used to rebuild or reconstruct a memory array to recover lost or corrupt data. The parity 406 is used as a form of error detecting code. The parity 406 can be calculated using exclusive-OR (XOR) binary operation performed on the data to be stored.
An allocation of the erase blocks 310 within one of the super blocks 308 can provide additional area for storage of the parity 406 including RAID parity data. The parity 406 can only be complete after a full stripe across all the erase blocks 310 within the super blocks 308 are written. The parity 406 can be stored within one of the super blocks 308.
Other effects include, for RAID-5 operations, the last of the erase blocks 310 in the super blocks 308 can be reserved for parity bits. As a result, parity pages are recycled along with data pages. Having the parity pages in a dedicated instance of the erase blocks 310, possibility of read disturb on these pages is significantly reduced.
The parity pages can be read only if there is an error correction code (ECC) error in one of the data pages in the super blocks 308. The parity pages can only be used in a case of an uncorrectable error and thus can have an extremely low read count, making them less susceptible to read disturb issues. Data retention in parity blocks can be better than that of data blocks in the super blocks 308 that are read a large number of times.
A read disturb in the memory devices 110 is a change of values of memory cells by an act of reading a memory cell or other memory cells within the memory devices 110. For example, in a NAND flash, read disturbs can be limited to cells within one of the erase blocks 310. An example of this is reading from one page of memory within one of the erase blocks 310 in the memory devices 110 can corrupt and change memory cells in other pages within the one of the erase blocks 310. The read disturb affect refers to the process of reading memory will create errors in the memory. Read disturbs are not always permanent and can be corrected when the memory is rewritten.
Each time a page is read in one of the erase blocks 310, a small amount of charge change is experienced by other pages within the one of the erase blocks 310. Therefore, by reading pages in one of the erase blocks 310, the bit error rate (BER) of the other pages can increase. By locating the RAID-5 parity data in the erase blocks 310 other than the erase blocks 310 being used for user data, the parity data does not suffer the same read disturbs as the normal data.
Read-write and read-erase conflicts can be minimized. The conflicts can be minimized by partitioning the super blocks 308 using the super chip selects 404. By rebalancing the super blocks 308 when recycling, a number of valid pages can be rebalanced between the chip selects 112.
There are positive factors of having the super blocks 308. The positive factors include fewer blocks to manage, fewer pages of metadata 407, infrastructure support for a RAID-5 implementation, and minimization of writes to a logical-to-physical (L2P) log during recycling. The metadata 407 is defined as non-host related information that is in addition to and stored along with host data and that aids or supports retrieval, organization, health of, and current condition or state of the memory devices 110. The metadata 407 is the metadata described in
The metadata 407 can be written to include logical page indexes. Pages of the metadata 407 can be grouped and written once per each of the super blocks 308 as opposed to keeping the metadata 407 in each of the erase blocks 310. This grouping of the metadata 407 can also make traversing it during drive initialization more efficient.
As an example, there can be transient metadata held or stored in the storage unit 204 of
The memory devices 110 that form the super devices 306 can be fixed through the life of a storage drive. It is understood that this assignment of the memory devices 110 to the super devices 306 can be determined in a number of ways because there can be unpopulated instances of the memory devices 110 in the storage drive.
Although there are three of the super blocks 308 shown in
A number of concurrent writes can be per logical unit (LUN) or dies. Each of the pages can be specifically on separate instances of the memory channels 304, or LUNs, and not just within one of the erase blocks 310.
Multiple of the erase blocks 310 grouped into each of the super blocks 308 provide several advantages. One of the advantages is that there are fewer of the super blocks 308 than a number of the erase blocks 310 to manage and initialize. Instead of managing lists of the erase blocks 310, lists of the super blocks 308 can be managed. As a result, data management resources including management data structures for managing the super blocks 308 are smaller or less.
By creating the super blocks 308 that span multiple of the chip selects 112 of the memory devices 110, additional advantages are achieved. One advantage is that the super blocks 308 provide an infrastructure that supports an internal redundant array of independent disks (RAID) architecture including RAID-5. The super blocks 308 can be created in a way that facilitates efficient host performance, recycling, RAID, and load balancing.
The term “internal” used above for RAID refers to RAID within and specific to one of the super blocks 308. RAID parity information can be part of and limited to contents of the one of the super blocks 308. As such, operations to the one of the super blocks 308 self-contained with respect to providing RAID error protection to the host data 405 and any stored metadata information.
A number of the erase blocks 310 in each of the super blocks 308 can be determined by the super devices 306. For example, each of the super blocks 308 can include at least 16 of the erase blocks 310. Another advantage of the super blocks 308 is that the super blocks 308 can include a number of the erase blocks 310 that are bad or remapped. If one of the super blocks 308 includes a number of the erase blocks 310 that are bad blocks, the entirety of the one of the super blocks 308 would not be rendered as useless. The erase blocks 310 that are bad or unusable can be remapped.
Within one of the super blocks 308, each of the erase blocks 310 does not have to be associated with the same erase block number (or address) in each of the elements. As an example, a portion of the 23rd instance of the super blocks 308 can include a number of the erase blocks 310 as follows.
Channel 0, CE0, LUN0, erase block 23
Channel 1, CE0, LUN0, erase block 23
Channel 2, CE0, LUN0, erase block 25
Channel 3, CE0, LUN0, erase block 23
Channel 3, CE0, LUN0, erase block 27
In this example, “Channel” is a corresponding channel number of the memory channels 304, “CE0” is chip enable 0, “LUN0” is logical unit 0, and “erase block” is an nth instance of the erase blocks 310.
Holes in contiguous addressing of erase blocks will exist from both factory marked bad erase blocks and erase blocks that have been determined to be bad during normal operation of the storage device. Remapped specifically refers to the substitution of one physical erase block for another.
Implementation of RAID-5 presents significant problems for recycling because parity pages need to be recycled or regenerated along with data pages. Parity data does not get relocated with the valid data during recycling. Instead, the parity data can be regenerated when recycle writes are performed. By implementing RAID within the super blocks 308, data and parity pages are recycling together.
The super blocks 308 also provide a way for a recycling operation to traverse data in the order by which that data is written. This operation minimizes the number of writes to the logical-to-physical log, especially when the host system 108 writes larger data blocks.
Other advantages include fewer statistics, which can be included in metadata pages 408, need to be managed. These statistics can include a program-erase count 410 and a valid-page count 412. There can be fewer of the metadata pages 408. The metadata pages 408 are defined as memory components within each of the erase blocks 310 that are programmed as individual units for the metadata 407.
The program-erase count 410 is defined as a numerical value indicating a number of times or cycles the memory devices 110 associated with the corresponding instances of the super blocks 308 are programmed or erased. The valid-page count 412 is defined as a numerical value indicating a number of pages that are valid in each of the super blocks 308.
Since the host data 405 can be written in the order it arrives from the host and because all data pages within the memory devices 110 can be written sequentially and contiguously, one or more new version of the same host logical sector can exist within the super blocks 308. In this case, only one of the host sectors can be valid, making other copies in the super blocks 308 invalid. The valid state of the pages within one of the super blocks 308 can be held in both the metadata pages 408 stored within the super blocks 308 and in the logical-to-physical (L2P) tabling system within the runtime volatile memory in the storage device.
Each of the super blocks 308 can include one of the metadata pages 408. Each of the metadata pages 408 can span multiple physical pages and is not limited to a single physical page depending on a number of elements in each of the super blocks 308 and the amount of metadata required for each element. Each of the super blocks 308 can include only one of the metadata pages 408. The metadata pages 408 can include logical page indexes of the super blocks 308. If the size of the super blocks 308 is properly determined, each of the metadata pages 408 can have a size that matches a size of a physical page in the memory devices 110.
The metadata pages 408 can include any additional information stored along with the host data 405 that can aid in retrieval, organization, health of, and current condition or state of the memory devices 110. The metadata pages 408 can include but is not limited to a logical to physical mapping information of the host data 405 within the memory devices 110. The metadata pages 408 that are held within one of the super blocks 308 can represent metadata specific to the one of the super blocks 308.
During portions of drive operations such as but not limited to drive initialization, the metadata pages 408 can be used to restore a host logical to physical mapping of data and a valid state of data stored within the one of the super blocks 308. The metadata pages 408 can also be used for storing the health and use status of the super blocks 308. The metadata pages 408 can be used to record the current recycling or runtime state of the super blocks 308.
Recycling operations can traverse or access data by following the same order by which host-write operations follow. As such, the original striping can be preserved. A number of operations for writing to the L2P log can be minimized.
The super-block organization is conducive to implementing RAID 4 within the storage drive. The last instance of the erase blocks 310 in each of the super blocks 308 can include parity pages. For example, parity can be generated using exclusive-OR operations (XOR). The memory controller 104 of
Factory and run-time defects can be handled. If one or more of the erase blocks 310 within one of the super blocks 308 is bad, the one or more of the erase blocks 310 can be removed from the one of the super blocks 308. In addition, bad instances of the erase blocks 310 in the one of the super blocks 308 can be skipped or ignored.
Statistics for the erase blocks 310 can be used to estimate remaining lifetime of the storage drive or to extend the life of the storage drive if the super blocks 308 are provided with erase-block granularity. Depending on how the memory devices 110 are populated in the storage drive with reduced capacity, the chip selects 112 of the memory devices 110 that are not used can be processed.
Each of the super blocks 308 can include the program-erase count 410. The program-erase count 410 and the valid-page count 412 can be managed at a super-block level. The program-erase count 410 can be calculated each time the memory devices 110 are accessed for a program operation or an erase operation. The program-erase count 410 can be used to determine if wear begins to deteriorate the integrity of the memory devices 110.
The super blocks 308 can include any number of unusable blocks 414, which are defined as blocks that are no longer available for access. If one of the erase blocks 310 is one of the unusable blocks 414, writing and recycling operations on the corresponding instance of the super blocks 308 can skip over the one of the erase blocks 310.
The super blocks 308 can include remapped blocks for the unusable blocks 414. The erase blocks 310 can represent spare blocks 416, which are blocks that are available for use. If the memory devices 110 include the spare blocks 416, one of the erase blocks 310 that is one of the unusable blocks 414 can be remapped to one of the spare blocks 416. Remapping is preferably not spanned over different instances of the memory devices 110.
There can be multiple levels of performance improvements having to do with both the super blocks 308 and the use of remapping of the erase blocks 310. The performance improvements from the use of the super blocks 308 can be from but not limited to the centralized storage of the metadata pages 408 in a single location instead of in each of the erase blocks 310 in the super blocks 308. Each access to one of the erase blocks 310 can incur a latency overhead of reading and writing. By only having to access one of the erase blocks 310 in the super blocks 308, there can be an effective elimination of multiple erase block accesses to retrieve the same amount of information. For reading and writing of the host data 405, the remapping function can allow each of the super blocks 308 to have a full complement of the erase blocks 310. All of the erase blocks 310 within one of the super devices 306 can be accessed in a parallel approach with the one of the super devices 306 active.
It has been discovered that the super blocks 308 spanning multiple of the memory channels 304 provide improved performance since multiple pages within each of the super blocks 308 are capable of being written concurrently.
It has also been discovered that multiple of the erase blocks 310 grouped into each of the super blocks 308 provide improved performance with reduced data management resources since there are fewer of the super blocks 308 than a number of the erase blocks 310 to manage and initialize.
It has further been discovered that the super blocks 308 spanning multiple of the chip selects 112 of the memory devices 110 provide improved reliability with an infrastructure that supports an internal redundant array of independent disks (RAID) architecture including RAID-5.
It has further been discovered that the super blocks 308 including a number of the erase blocks 310 that are bad or remapped provides improved memory resource utilization since the super blocks 308 with the unusable blocks 414 do not have to be thrown away but instead the unusable blocks 414 are able to be remapped with the spare blocks 416 thereby eliminating the shortcoming of traditional super blocks where a bad block renders the entire super block useless.
It has further been discovered that the super blocks 308 provide improved performance since there are fewer statistics, including the program-erase count 410 and the valid-page count 412, included in the metadata pages 408 that need to be managed.
It has further been discovered that remapping of the unusable blocks 414 provide improved performance since the super blocks 308 that are incomplete cause write performance degradation.
Referring now to
The structure of the super blocks 308 is provided to allow dynamic location of redundant array of independent disks (RAID) parity with selective assignments of the erase blocks 310 based on drive life and wear. The data structure of the super blocks 308 can be saved or stored across a power cycle where a solid-state drive (SSD) employs the use of MLC NAND flash.
Each of the memory channels 304 can be associated with a number of the memory devices 110 of
For example, for NAND flash memories, the memory devices 110 can represent dies or logical units (LUN). An important point is that the memory devices 110 can execute commands independently from each other allowing parallel operations. This independent operation is used to gain greater bandwidths and high I/O operations (IOPS).
Each of the memory devices 110 can include multiple planes 504, each of which can include a number of the erase blocks 310. Each of the erase blocks 310 can include a number of pages 506. The pages 506 are defined as memory components within each of the erase blocks 310 that are programmed as individual units. The pages 506 are smallest groups of data bytes that are read from or written to in each of the erase blocks 310. Each of the pages 506 can include a number of bytes.
The pages 506 can be grouped into each of the erase blocks 310, which are the smallest area that can be erased. Each of the pages 506 is the smallest region that is programmed. The erase blocks 310 can be organized in each of the planes 504, which can be grouped into a logical unit (LUN).
For example, a die can include approximately 4096 of the erase blocks 310 divided into two of the planes 504 with 2048 of the erase blocks 310 per each of the planes 504. Also for example, each of the erase blocks 310 can include 256 pages, each of which can include 8K bytes, where each 1K bytes includes 1024 bytes.
The first example diagram depicts one of the super blocks 308 having the host data 405, the metadata 407, and the parity 406, each of which can include any number of the pages 506. The parity 406 can be at any location within each of the super blocks 308. For illustrative purposes, the parity 406 is shown in one of the planes 504 of one of the memory channels 304, although it is understood that the parity 406 can be at any location within each of the super blocks 308.
The pages 506 of the metadata 407 can include a parity map field 508. The parity map field 508 is defined as information that is associated with the parity 406. The parity map field 508 identifies the location 502 of the parity 406 within each of the super blocks 308. The parity map field 508 can represent a parity location map or parity mapping.
The parity 406 can include any number of the pages 506. The parity 406 can be located in any of the planes 504 per each of the memory channels 304 or spanning across any number of the planes 504 within any number of the memory channels 304. The parity 406 can be located in any of the memory channels 304 or spanning across any number of the memory channels 304. The parity 406 can include a number of the pages 506 that are contiguous, the pages 506 that are separate from and not abut each other, or a combination thereof.
The parity map field 508 can optionally include flags 510, which are defined as information that is associated with the parity 406. The flags 510 can include information associated with test mode parameters that can be used to change the endurance and retention qualities of the parity 406. The test mode parameters can be used when a recovery needs to take place. The recovery does not happen very often and the time it takes to re-write the test mode parameters to use the parity 406 can be tolerated. The test mode parameters can be used to manipulate or change the erase blocks 310 to have more retention as opposed to more endurance, as an example.
A number of methods can be implemented to build the parity map field 508. For example, the host data 405 can be stored in the erase blocks 310 at the beginning of the super blocks 308 and the parity 406 in the last instances of the erase blocks 310. Also for example, the erase blocks 310 in the super blocks 308 do not need to be monotonically increasing in regards to the numbers of the memory channels 304. As a specific example, 16 of the memory channels 304 used in the erase blocks 310 can be ordered as 5, 15, 12, 11, 0, 2, 7, 9, 10, 1, 3, 13, 14, 4, 8, 6. In this case, the erase blocks 310 on channel 6 would contain the parity 406.
A means for identifying the pages 506 of the metadata 407 versus the pages 506 of the parity 406 can be determine by a signature value 512 in the pages 506 of the metadata 407. The signature value 512 is defined as information that is used to identify the pages 506 of the metadata 407. The signature value 512 is a sequence of data bytes that are unique such that they are different from the XOR parity data of the parity 406 in the same location within the erase blocks 310. The parity map field 508 can be redundant or the same in all of the pages 506 of the metadata 407 so it can be clearly or distinctly identified from the parity 406.
The super blocks 308 can be traversed by taking any number of passes through the memory channels 304. As an example, the super blocks 308 can be traversed by taking 2 passes through the memory channels 304.
Referring now to
The metadata 407 stored across multiple of the memory channels 304 or different instances of the planes 504 can include redundant parity mapping structures so that the parity map field 508 does not have to be stored in a given location. Any location within the super blocks 308 can be selected for the location of the parity 406. As an example, the parity 406 can be located in one of the planes 504 in one of the memory channels 304. As a specific example,
Referring now to
Each of the super blocks 308 can represent the set. The members of the set of each the super blocks 308 can include the host data 405, which can include a number of the pages 506 of
The members of the set of each the super blocks 308 can also include information for the metadata 407, which can include extra information in the pages 506 of the host data 405 and extra information in the pages 506 that are stand-alone and not associated with the host data 405. The members of the set of each the super blocks 308 can also include the parity 406, including RAID parity information, which can provide protection for the pages 506 of both the host data 405 and the metadata 407.
Since there can be multiple members in each type of information, subsets can be formed within each of the super blocks 308 as the main set. The subsets are portions of the set. There can be intersections 702 between different memory types within each of the super blocks 308. For example, a number of the pages 506 of the metadata 407 can be held in or associated with the host data 405.
In each of the super blocks 308, membership is dynamic and can be “reassigned” in different physical instances of the erase blocks 310. This is to tolerate loss of functions due to the erase blocks 310 that are bad or unusable in one of the elements of the set of each of the super blocks 308.
Referring now to
Protection of data can be implemented with FRAME using the parity 406. FRAME is a stripe of the host data 405 of
The memory controller 104 can include a channel module 802 for performing operations associated with the memory channels 304 of
The memory controller 104 can include a device module 804 for performing operations associated with the super devices 306 of
The device module 804 can group the memory devices 110 from the memory channels 304 to form the super devices 306. The super devices 306 can include a number of the memory devices 110 that are controlled in parallel across a number of the memory channels 304. The super devices 306 can be constructed with or associated with any number of the memory channels 304 that are executed in parallel. The number of the super devices 306 can be based on a number of independently accessible instances of the memory devices 110 on the memory channels 304 selected by Chip Enables (CEs), logical units (LUN), or a combination thereof.
One or more of the super devices 306 can be formed by taking multiple passes through the memory channels 304. For example, if there are 16 of the memory channels 304, one or more of the super devices 306 can be formed across 8 of the memory channels 304 or 16 of the memory channels 304. Also for example, any number of passes or partial passes can be taken through the memory channels 304 to traverse the super devices 306. As a specific example, 1.5 or 2 passes can be taken through the memory channels 304. If one of the memory channels 304 is used twice, two different dies can be used on the one of the memory channels 304 so that the parity 406 provides protection from a plane failure.
The memory controller 104 can include a block module 806 for performing operations associated with the super blocks 308. Each of the super blocks 308 can be identified and selected by the block module 806. Each of the super blocks 308 is a group of the erase blocks 310 of
As an example, each of the memory devices 110 in the super blocks 308 can contribute a pair of the erase blocks 310 that can be accessed together. Each of the super blocks 308 can be selected to be associated with the super devices 306. Each of the super blocks 308 can be accessed via and associated with only one of the super devices 306.
Since the super devices 306 span across multiple of the memory channels 304, the super blocks 308 can also span across multiple of the memory channels 304. The term “spanning across” used above for the super blocks 308 refers to portions of different instances of the memory channels 304 are included in each of the super blocks 308.
All the chip selects 112 of
The super chip selects 404 can be used to assert or de-assert the chip selects 112. The chip selects 112 can be asserted or de-asserted to enable or disable, respectively, the memory devices 110. For example, the chip selects 112 can be asserted or de-asserted by the block module 806 assigning the super chip selects 404 to an active or inactive state such that the chip selects 112 are driven to logic ‘1’ or ‘0’, respectively, to enable or disable the memory devices 110.
Read-write and read-erase conflicts can be minimized by the block module 806. The conflicts can be minimized by partitioning the super blocks 308 using the super chip selects 404. By rebalancing the super blocks 308 when recycling, a number of valid pages can be rebalanced between the chip selects 112.
The pages 506 of
Multiple of the erase blocks 310 grouped into each of the super blocks 308 by the block module 806 provide several advantages. One of the advantages is that there are fewer of the super blocks 308 than a number of the erase blocks 310 to manage and initialize by the block module 806. Instead of managing lists of the erase blocks 310, lists of the super blocks 308 can be managed by the block module 806. As a result, data management resources including management data structures for managing the super blocks 308 are smaller or less.
A number of the erase blocks 310 in each of the super blocks 308 can be determined by the super devices 306 of
The memory controller 104 can include a parity module 808 for determining the parity 406 and the location 502 of
During the normal operation of a solid-state drive (SSD), the use of NAND wears the NAND to an unreliable state. This wear is not always even in a pool of the erase blocks 310 and some NAND wears much faster than others do. One way to protect against this uneven wear is to provide a RAID type parity protection. One problem that exists is when the logical unit (LUN) or die is selected, for use as the parity information, is either unreliable or worse is completely unusable.
To solve the problem above, the location 502 of the parity 406 is determined based on a dynamic mapping of the parity 406 that is tracked at the super block level. The term “dynamic” in the dynamic mapping refers to determining the location 502 of the parity 406 within one of the super blocks 308 whenever the one of the super blocks 308 is formed. As such, the dynamic mapping allows the location 502 of the parity 406 to be different whenever the one of the super blocks 308 is formed as opposed to the location 502 of the parity 406 being fixed in the super blocks 308. As the erase blocks 310 in the super blocks 308 become unreliable or unusable, the super blocks 308 are formed or regenerated during normal operation of the SSD using the erase blocks 310 that are reliable or usable.
The location 502 is determined based on the erase blocks 310 that are reliable or usable. For example, the erase blocks 310 are determined as reliable or usable when the erase blocks 310 are written with predetermined data and subsequently read with the same data that was written. Also for example, the erase blocks 310 are determined as reliable or usable when a number of the program/erase cycles of the erase blocks 310 is below a predetermined threshold.
The location 502 of the parity 406 of each of the super blocks 308 can be provided in a header 810 of each of the super blocks 308. The header 810 is defined as a portion of each of the super blocks 308 that is used to store information associated with the super blocks 308. For example, the header 810 can represent a top portion of each of the super blocks 308.
The header 810 can include a table that describes participants or members of the super blocks 308. For example, the participants or the members can include the host data 405, the metadata 407, the parity 406, and any other types of information stored in the super blocks 308.
The location 502 of the parity 406 of each of the super blocks 308 can be provided by the parity map field 508 of
There can be multiple of the pages 506 of the metadata 407 for storing multiple redundant instances of the parity map field 508. The multiple redundant instances of the parity map field 508 provide additional protection for the location 502 of the parity 406. When one or more of the pages 506 of the metadata 407 that includes one of the multiple redundant instances of the parity map field 508 is unreadable, other instances of the pages 506 of the metadata 407 that include the multiple redundant instances can be read. When two or more of the multiple redundant instances are compared to be the same, the parity map field 508 is determined to be valid.
The parity module 808 can determine a parity size 814 of the parity 406. The parity size 814 is defined as information that indicates an amount of the parity 406. The parity size 814 can be calculated based on a number of bytes, the pages 506, the erase blocks 310, or any other units that indicate amount of information.
The parity size 814 can be determined based on a number of the erase blocks 310 that are reliable or usable in the super blocks 308. The parity 406 can provide sizable parity protection mapped in each individual instance of the super blocks 308. For example, the parity size 814 of the parity 406 in one of the super blocks 308 can be greater than that of the parity 406 in another of the super blocks 308 with fewer of the erase blocks 310 than the one of the super blocks 308. In other words, the parity 406 covers fewer amount of information if there are fewer of the erase blocks 310 in the super blocks 308.
The parity module 808 can determine a parity type 816 for the parity 406. The parity type 816 is defined as an identification of information stored within one of the super blocks 308 that the parity 406 protects. The parity type 816 can be determined for each of the participants or the members in each of the super blocks 308. Different types of data stored in the super blocks 308 can be allowed to have different amounts of parity protection using multiple of the parity 406 within the super blocks 308.
The parity type 816 can be determined based on the kinds of data stored within the super blocks 308. The data stored within the super blocks 308 that the parity 406 protects can include hot host data 818, cool host data 820, different types of multiple of the metadata 407, drive-level logging functions 822, and SCRAM or drive-related emergency shutdown information. The drive-level logging functions 822 are information that is associated with control or status of a data storage device. For example, the drive-level logging functions 822 can be associated with informational messages, error events, mapping transactions, and unbound metadata.
The hot host data 818 is defined as the host data 405 stored in a portion of one of the super blocks 308 that gets more number of data write accesses compared to the host data 405 stored in other portions of the one of the super blocks 308. The cool host data 820 is defined as the host data 405 stored in a portion of one of the super blocks 308 that gets less number of data write accesses compared to the host data 405 stored in other portions of the one of the super blocks 308.
The hot host data 818 can get more parity protection than the cool host data 820 does due to different retention or endurance rates between the hot host data 818 and the cool host data 820. The parity size 814 of the parity 406 for the hot host data 818 can be greater than that of the parity 406 for the cool host data 820.
There can be a number of the location 502 for a number of the parity 406, each of which provides protection for different data based on the parity type 816. For example, the number of the parity 406 can provide protection for the hot host data 818, the cool host data 820, the metadata 407, the drive-level logging functions 822, and the SCRAM or the drive-related emergency shutdown information.
For example, the parity map field 508 can identify multiple of the location 502 of multiple of the parity 406. Also for example, there can be multiple of the parity map field 508 in one of the super blocks 308, each of which identifies the location 502 of each of multiple of the parity 406, providing the one of the super blocks 308 having more than one parity layout. Each of the multiple of the parity map field 508 can be located at different sections of the one of the super blocks 308.
The parity 406 can cover the host data 405 based on plane failures associated with the planes 504 of
For example, when one of the planes 504, such as an odd or an even plane, in one of the memory channels 304 fails or is unusable for reliable operation, the parity 406 can be allocated to another of the planes 504 in the one of the memory channels 304 or one of the planes 504 in another of the memory channels 304 in the same instance of the super blocks 308. As a specific example, when one of the planes 504 of one of the memory channels 304, shown as plane 1 in channel 15 in
The parity module 808 can determine the parity 406 based on an age 824 of each the super blocks 308, for which the parity 406 protects. The age 824 associated with one of the super blocks 308 is defined as duration of how long the one of the super blocks 308 is used since it is formed or regenerated. The super blocks 308 having different values of the age 824 can dynamically have different parity strengths, which are mapped on a per super block basis. For example, one of the super blocks 308 can have the parity 406 with the parity size 814 greater than the parity size 814 of the parity 406 of another of the super blocks 308 when the one of the super blocks 308 has the age 824 greater than the age 824 of the another of the super blocks 308.
Having one of the erase blocks 310 with the parity 406 per each of the super blocks 308 provides plane-level protection. Having one or more of the erase blocks 310 with two of the parity 406 per each of the super blocks 308 provides die-level protection provided that there are two physical planes in a die.
Functions or operations of the memory controller 104 as described above can be implemented with the control unit 202 of
As an example, the block module 806 can be implemented with the control unit 202 to select the super blocks 308 associated with an instance of the super devices 306 or the super chip selects 404. As another example, the parity module 808 can be implemented with the control unit 202 to determine the location 502 of the parity 406 within the super blocks 308 when the super blocks 308 are formed.
The device module 804 can be coupled to the channel module 802. The block module 806 can be coupled to the device module 804. The parity module 808 can be coupled to the block module 806.
It has been discovered that the location 502 of the parity 406 within the super blocks 308 provides improved reliability by determining the location 502 based on the erase blocks 310 that are reliable or usable whenever the super blocks 308 are formed. During the normal operation of the SSD, the use of NAND wears the NAND to an unreliable state. This wear is not always even in the pool of erase blocks and some NANDs wear much faster than others do. One way to protect against this uneven wear is to provide a RAID type parity protection. One problem that exists is when the logical unit (LUN) or die selected, for use as the parity information, is either unreliable or worse is completely unusable. The location 502 of the parity 406 dynamically determined based on a dynamic mapping of the parity 406 based on the erase blocks 310 that are reliable or usable whenever the super blocks 308 are formed during the operation of the SSD solves this problem thereby providing the improved reliability.
It has also been discovered that the parity size 814 provides improved reliability since the parity size 814 is determined based on a number of the erase blocks 310 that are reliable or usable in the super blocks 308.
It has further been discovered that the parity type 816 provides improved reliability since the parity type 816 identifies types of data stored in the super blocks 308 that a number of the parity 406 protect.
It has further been discovered that the age 824 provides improved reliability by the age 824 indicating duration of how long the super blocks 308 have been formed for the parity 406 to be generated to reliably protect the super blocks 308 since the super blocks 308 having different values of the age 824 need different parity strengths.
The storage control system 100 of
Furthermore, data generated in one module can be used by another module without being directly coupled to each other. For example, the parity module 808 can receive information associated with the memory channels 304 partitioned by the channel module 802 to determine the location 502 of the parity 406 within the erase blocks 310 in one of the memory channels 304. The channel module 802, the device module 804, the block module 806, and the parity module 808 can be implemented as hardware accelerators (not shown) within the control unit 202 or can be implemented as hardware accelerators (not shown) in the memory controller 104 or outside of the memory controller 104.
Referring now to
Thus, it has been discovered that the storage control system of the present invention furnishes important and heretofore unknown and unavailable solutions, capabilities, and functional aspects for a storage control system with data management mechanism. The resulting method, process, apparatus, device, product, and/or system is straightforward, cost-effective, uncomplicated, highly versatile, accurate, sensitive, and effective, and can be implemented by adapting known components for ready, efficient, and economical manufacturing, application, and utilization.
Another important aspect of the present invention is that it valuably supports and services the historical trend of reducing costs, simplifying systems, and increasing performance.
These and other valuable aspects of the present invention consequently further the state of the technology to at least the next level.
While the invention has been described in conjunction with a specific best mode, it is to be understood that many alternatives, modifications, and variations will be apparent to those skilled in the art in light of the aforegoing description. Accordingly, it is intended to embrace all such alternatives, modifications, and variations that fall within the scope of the included claims. All matters hithertofore set forth herein or shown in the accompanying drawings are to be interpreted in an illustrative and non-limiting sense.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4048481 | Bailey, Jr. et al. | Sep 1977 | A |
| 4839587 | Flatley et al. | Jun 1989 | A |
| 5034744 | Obinata | Jul 1991 | A |
| 5210854 | Beaverton et al. | May 1993 | A |
| 5311395 | McGaha et al. | May 1994 | A |
| 5450354 | Sawada et al. | Sep 1995 | A |
| 5479638 | Assar et al. | Dec 1995 | A |
| 5784174 | Fujino et al. | Jul 1998 | A |
| 5790828 | Jost | Aug 1998 | A |
| 5930504 | Gabel | Jul 1999 | A |
| 5949785 | Beasley | Sep 1999 | A |
| 5963983 | Sakakura et al. | Oct 1999 | A |
| 6034897 | Estakhri et al. | Mar 2000 | A |
| 6069827 | Sinclair | May 2000 | A |
| 6091652 | Haehn et al. | Jul 2000 | A |
| 6275436 | Tobita et al. | Aug 2001 | B1 |
| 6345367 | Sinclair | Feb 2002 | B1 |
| 6356447 | Scafidi | Mar 2002 | B2 |
| 6381176 | Kim et al. | Apr 2002 | B1 |
| 6381670 | Lee et al. | Apr 2002 | B1 |
| 6412080 | Fleming et al. | Jun 2002 | B1 |
| 6529997 | Debiez et al. | Mar 2003 | B1 |
| 6552581 | Gabara | Apr 2003 | B1 |
| 6587915 | Kim | Jul 2003 | B1 |
| 6618249 | Fairchild | Sep 2003 | B2 |
| 6661503 | Yamaguchi et al. | Dec 2003 | B1 |
| 6728913 | Parker | Apr 2004 | B1 |
| 6763424 | Conley | Jul 2004 | B2 |
| 6775792 | Ulrich et al. | Aug 2004 | B2 |
| 6778387 | Fairchild | Aug 2004 | B2 |
| 6850443 | Lofgren et al. | Feb 2005 | B2 |
| 6854070 | Johnson et al. | Feb 2005 | B2 |
| 6871304 | Hadjihassan et al. | Mar 2005 | B2 |
| 6903972 | Lasser et al. | Jun 2005 | B2 |
| 6906961 | Eggleston et al. | Jun 2005 | B2 |
| 6975028 | Wayburn et al. | Dec 2005 | B1 |
| 7082495 | DeWhitt et al. | Jul 2006 | B2 |
| 7107389 | Inagaki et al. | Sep 2006 | B2 |
| 7139864 | Bennett et al. | Nov 2006 | B2 |
| 7233497 | Simon et al. | Jun 2007 | B2 |
| 7243186 | Liang et al. | Jul 2007 | B2 |
| 7298888 | Hamar | Nov 2007 | B2 |
| 7330927 | Reeve et al. | Feb 2008 | B1 |
| 7333364 | Yu et al. | Feb 2008 | B2 |
| 7350101 | Nguyen et al. | Mar 2008 | B1 |
| 7355896 | Li et al. | Apr 2008 | B2 |
| 7434122 | Jo | Oct 2008 | B2 |
| 7441067 | Gorobets et al. | Oct 2008 | B2 |
| 7516267 | Coulson et al. | Apr 2009 | B2 |
| 7613871 | Tanaka et al. | Nov 2009 | B2 |
| 7620710 | Kottomtharayil et al. | Nov 2009 | B2 |
| 7620769 | Lee et al. | Nov 2009 | B2 |
| 7639532 | Roohparvar et al. | Dec 2009 | B2 |
| 7661054 | Huffman et al. | Feb 2010 | B2 |
| 7679948 | Park et al. | Mar 2010 | B2 |
| 7693422 | Alicherry et al. | Apr 2010 | B2 |
| 7738502 | Chang et al. | Jun 2010 | B2 |
| 7743216 | Lubbers et al. | Jun 2010 | B2 |
| 7818525 | Frost et al. | Oct 2010 | B1 |
| 7827348 | Lee et al. | Nov 2010 | B2 |
| 7830164 | Earle et al. | Nov 2010 | B2 |
| 7853749 | Kolokowsky | Dec 2010 | B2 |
| 7979614 | Yang | Jul 2011 | B1 |
| 8000161 | Stan et al. | Aug 2011 | B2 |
| 8001135 | Perlmutter et al. | Aug 2011 | B2 |
| 8010738 | Chilton et al. | Aug 2011 | B1 |
| 8028123 | Kiizer et al. | Sep 2011 | B2 |
| 8046645 | Hsu et al. | Oct 2011 | B2 |
| 8051241 | Feldman et al. | Nov 2011 | B2 |
| 8072805 | Chou et al. | Dec 2011 | B2 |
| 8095724 | Ji et al. | Jan 2012 | B2 |
| 8095765 | Asnaashari et al. | Jan 2012 | B2 |
| 8117396 | Fair et al. | Feb 2012 | B1 |
| 8127202 | Cornwell et al. | Feb 2012 | B2 |
| 8145984 | Sommer et al. | Mar 2012 | B2 |
| 8154921 | Mokhlesi et al. | Apr 2012 | B2 |
| 8169825 | Shalvi et al. | May 2012 | B1 |
| 8205028 | Sakarda | Jun 2012 | B1 |
| 8209677 | Shintani et al. | Jun 2012 | B2 |
| 8219724 | Caruso et al. | Jul 2012 | B1 |
| 8219776 | Forhan et al. | Jul 2012 | B2 |
| 8228701 | Sokolov et al. | Jul 2012 | B2 |
| 8245101 | Olbrich et al. | Aug 2012 | B2 |
| 8250621 | Cha | Aug 2012 | B2 |
| 8254172 | Kan | Aug 2012 | B1 |
| 8259506 | Sommer et al. | Sep 2012 | B1 |
| 8289801 | Smith et al. | Oct 2012 | B2 |
| 8296534 | Gupta et al. | Oct 2012 | B1 |
| 8332578 | Frickey, III et al. | Dec 2012 | B2 |
| 8363413 | Paquette et al. | Jan 2013 | B2 |
| 8369141 | Sommer et al. | Feb 2013 | B2 |
| 8386700 | Olbrich et al. | Feb 2013 | B2 |
| 8386860 | Tseng et al. | Feb 2013 | B2 |
| 8397101 | Goss et al. | Mar 2013 | B2 |
| 8407409 | Kawaguchi | Mar 2013 | B2 |
| 8464106 | Filor et al. | Jun 2013 | B2 |
| 8503238 | Wu et al. | Aug 2013 | B1 |
| 8521981 | Strauss et al. | Aug 2013 | B2 |
| 8560770 | Haines et al. | Oct 2013 | B2 |
| 8601203 | Holbrook et al. | Dec 2013 | B2 |
| 8612669 | Syu et al. | Dec 2013 | B1 |
| 8612804 | Kang et al. | Dec 2013 | B1 |
| 8661184 | Wood et al. | Feb 2014 | B2 |
| 8694811 | Raju et al. | Apr 2014 | B2 |
| 8750052 | Aoki et al. | Jun 2014 | B2 |
| 8793556 | Northcott et al. | Jul 2014 | B1 |
| 8799747 | Goss et al. | Aug 2014 | B2 |
| 8832506 | Griffin et al. | Sep 2014 | B2 |
| 8880838 | Kaiser et al. | Nov 2014 | B2 |
| 8984216 | Fillingim | Mar 2015 | B2 |
| 9043668 | Goss et al. | May 2015 | B2 |
| 9063844 | Higgins et al. | Jun 2015 | B2 |
| 9069468 | Mehra et al. | Jun 2015 | B2 |
| 9116401 | Kim et al. | Aug 2015 | B2 |
| 9201728 | Patapoutian et al. | Dec 2015 | B2 |
| 20020056025 | Qiu et al. | May 2002 | A1 |
| 20020156891 | Ulrich et al. | Oct 2002 | A1 |
| 20020159285 | Morley et al. | Oct 2002 | A1 |
| 20030033308 | Patel et al. | Feb 2003 | A1 |
| 20030046603 | Harari et al. | Mar 2003 | A1 |
| 20030074592 | Hasegawa | Apr 2003 | A1 |
| 20030163633 | Aasheim et al. | Aug 2003 | A1 |
| 20040080985 | Chang et al. | Apr 2004 | A1 |
| 20040088511 | Bacon et al. | May 2004 | A1 |
| 20040252670 | Rong et al. | Dec 2004 | A1 |
| 20050021904 | Iaculo et al. | Jan 2005 | A1 |
| 20050038792 | Johnson | Feb 2005 | A1 |
| 20050073884 | Gonzalez et al. | Apr 2005 | A1 |
| 20050076102 | Chen et al. | Apr 2005 | A1 |
| 20050144516 | Gonzalez et al. | Jun 2005 | A1 |
| 20060015683 | Ashmore et al. | Jan 2006 | A1 |
| 20060020745 | Conley et al. | Jan 2006 | A1 |
| 20060022054 | Elhamias et al. | Feb 2006 | A1 |
| 20060080505 | Arai et al. | Apr 2006 | A1 |
| 20060136682 | Haridas et al. | Jun 2006 | A1 |
| 20060143365 | Kikuchi | Jun 2006 | A1 |
| 20060143475 | Herbert et al. | Jun 2006 | A1 |
| 20060253641 | Gatzemeier et al. | Nov 2006 | A1 |
| 20060256624 | Eggleston et al. | Nov 2006 | A1 |
| 20060282644 | Wong | Dec 2006 | A1 |
| 20060294574 | Cha | Dec 2006 | A1 |
| 20070050536 | Kolokowsky | Mar 2007 | A1 |
| 20070061511 | Faber | Mar 2007 | A1 |
| 20070067598 | Fujimoto | Mar 2007 | A1 |
| 20070079152 | Winick et al. | Apr 2007 | A1 |
| 20070083779 | Misaka et al. | Apr 2007 | A1 |
| 20070226592 | Radke | Sep 2007 | A1 |
| 20070234004 | Oshima et al. | Oct 2007 | A1 |
| 20070260811 | Merry, Jr. et al. | Nov 2007 | A1 |
| 20070263444 | Gorobets et al. | Nov 2007 | A1 |
| 20070276973 | Tan et al. | Nov 2007 | A1 |
| 20080028246 | Witham | Jan 2008 | A1 |
| 20080046630 | Lasser | Feb 2008 | A1 |
| 20080052446 | Lasser et al. | Feb 2008 | A1 |
| 20080082736 | Chow et al. | Apr 2008 | A1 |
| 20080126720 | Danilak | May 2008 | A1 |
| 20080183918 | Dhokia et al. | Jul 2008 | A1 |
| 20080189588 | Tanaka et al. | Aug 2008 | A1 |
| 20080263289 | Hosoya et al. | Oct 2008 | A1 |
| 20080313505 | Lee et al. | Dec 2008 | A1 |
| 20090006900 | Lastras-Montano et al. | Jan 2009 | A1 |
| 20090019321 | Radke | Jan 2009 | A1 |
| 20090070651 | Diggs et al. | Mar 2009 | A1 |
| 20090083587 | Ng et al. | Mar 2009 | A1 |
| 20090089485 | Yeh | Apr 2009 | A1 |
| 20090091990 | Park et al. | Apr 2009 | A1 |
| 20090109786 | Ye et al. | Apr 2009 | A1 |
| 20090125670 | Keays | May 2009 | A1 |
| 20090138654 | Sutardja | May 2009 | A1 |
| 20090146721 | Kurooka et al. | Jun 2009 | A1 |
| 20090157948 | Trichina et al. | Jun 2009 | A1 |
| 20090164702 | Kern | Jun 2009 | A1 |
| 20090164710 | Choi et al. | Jun 2009 | A1 |
| 20090172248 | You | Jul 2009 | A1 |
| 20090172262 | Olbrich et al. | Jul 2009 | A1 |
| 20090179707 | Higashino | Jul 2009 | A1 |
| 20090228634 | Nakamura et al. | Sep 2009 | A1 |
| 20090228761 | Perlmutter et al. | Sep 2009 | A1 |
| 20090259819 | Chen et al. | Oct 2009 | A1 |
| 20090259896 | Hsu et al. | Oct 2009 | A1 |
| 20090271562 | Sinclair | Oct 2009 | A1 |
| 20090287975 | Kim et al. | Nov 2009 | A1 |
| 20090300238 | Panabaker et al. | Dec 2009 | A1 |
| 20090323419 | Lee et al. | Dec 2009 | A1 |
| 20090327581 | Coulson | Dec 2009 | A1 |
| 20090327591 | Moshayedi | Dec 2009 | A1 |
| 20100017650 | Chin et al. | Jan 2010 | A1 |
| 20100023674 | Aviles | Jan 2010 | A1 |
| 20100050053 | Wilson et al. | Feb 2010 | A1 |
| 20100082890 | Heo et al. | Apr 2010 | A1 |
| 20100122019 | Flynn et al. | May 2010 | A1 |
| 20100128537 | Suhail et al. | May 2010 | A1 |
| 20100138592 | Cheon | Jun 2010 | A1 |
| 20100165689 | Rotbard et al. | Jul 2010 | A1 |
| 20100169541 | Freikorn | Jul 2010 | A1 |
| 20100172179 | Gorobets et al. | Jul 2010 | A1 |
| 20100174845 | Gorobets et al. | Jul 2010 | A1 |
| 20100217898 | Priborsky et al. | Aug 2010 | A1 |
| 20100217915 | O'Connor et al. | Aug 2010 | A1 |
| 20100223531 | Fukutomi et al. | Sep 2010 | A1 |
| 20100228928 | Asnaashari et al. | Sep 2010 | A1 |
| 20100262792 | Hetzler et al. | Oct 2010 | A1 |
| 20100262795 | Hetzler et al. | Oct 2010 | A1 |
| 20100262875 | Hetzler et al. | Oct 2010 | A1 |
| 20100287328 | Feldman et al. | Nov 2010 | A1 |
| 20100293367 | Berke et al. | Nov 2010 | A1 |
| 20100312954 | Jeon et al. | Dec 2010 | A1 |
| 20100318719 | Keays et al. | Dec 2010 | A1 |
| 20100332726 | Wang | Dec 2010 | A1 |
| 20110002224 | Tamura | Jan 2011 | A1 |
| 20110016239 | Stenfort | Jan 2011 | A1 |
| 20110055455 | Post et al. | Mar 2011 | A1 |
| 20110055468 | Gonzalez et al. | Mar 2011 | A1 |
| 20110066788 | Eleftheriou et al. | Mar 2011 | A1 |
| 20110072423 | Fukata | Mar 2011 | A1 |
| 20110078393 | Lin | Mar 2011 | A1 |
| 20110099342 | Ozdemir | Apr 2011 | A1 |
| 20110107144 | Ohara | May 2011 | A1 |
| 20110131365 | Zhang et al. | Jun 2011 | A1 |
| 20110131447 | Prakash et al. | Jun 2011 | A1 |
| 20110132000 | Deane et al. | Jun 2011 | A1 |
| 20110138100 | Sinclair | Jun 2011 | A1 |
| 20110145473 | Maheshwari | Jun 2011 | A1 |
| 20110161775 | Weingarten | Jun 2011 | A1 |
| 20110173395 | Bhattacharjee et al. | Jul 2011 | A1 |
| 20110190963 | Glassl et al. | Aug 2011 | A1 |
| 20110191522 | Condict et al. | Aug 2011 | A1 |
| 20110191649 | Lim et al. | Aug 2011 | A1 |
| 20110209032 | Choi et al. | Aug 2011 | A1 |
| 20110238892 | Tsai et al. | Sep 2011 | A1 |
| 20110239088 | Post | Sep 2011 | A1 |
| 20110258496 | Tseng et al. | Oct 2011 | A1 |
| 20110314219 | Ulrich et al. | Dec 2011 | A1 |
| 20110320687 | Belluomini et al. | Dec 2011 | A1 |
| 20120008401 | Katz et al. | Jan 2012 | A1 |
| 20120011336 | Saika | Jan 2012 | A1 |
| 20120047318 | Yoon et al. | Feb 2012 | A1 |
| 20120047320 | Yoo et al. | Feb 2012 | A1 |
| 20120047409 | Post et al. | Feb 2012 | A1 |
| 20120066450 | Yochai et al. | Mar 2012 | A1 |
| 20120079348 | Naeimi | Mar 2012 | A1 |
| 20120079355 | Patapoutian et al. | Mar 2012 | A1 |
| 20120096217 | Son et al. | Apr 2012 | A1 |
| 20120124046 | Provenzano | May 2012 | A1 |
| 20120124273 | Goss et al. | May 2012 | A1 |
| 20120151260 | Zimmermann et al. | Jun 2012 | A1 |
| 20120170365 | Kang et al. | Jul 2012 | A1 |
| 20120185706 | Sistla et al. | Jul 2012 | A1 |
| 20120203951 | Wood et al. | Aug 2012 | A1 |
| 20120213004 | Yun et al. | Aug 2012 | A1 |
| 20120216085 | Weingarten et al. | Aug 2012 | A1 |
| 20120236656 | Cometti | Sep 2012 | A1 |
| 20120239858 | Melik-Martirosian | Sep 2012 | A1 |
| 20120254686 | Esumi et al. | Oct 2012 | A1 |
| 20120266011 | Storer et al. | Oct 2012 | A1 |
| 20120266048 | Chung et al. | Oct 2012 | A1 |
| 20120278530 | Ebsen | Nov 2012 | A1 |
| 20120278531 | Horn | Nov 2012 | A1 |
| 20120284587 | Yu et al. | Nov 2012 | A1 |
| 20120297113 | Belluomini et al. | Nov 2012 | A1 |
| 20120311402 | Tseng et al. | Dec 2012 | A1 |
| 20120317334 | Suzuki et al. | Dec 2012 | A1 |
| 20120324191 | Strange et al. | Dec 2012 | A1 |
| 20120331207 | Lassa et al. | Dec 2012 | A1 |
| 20130007380 | Seekins et al. | Jan 2013 | A1 |
| 20130007543 | Goss et al. | Jan 2013 | A1 |
| 20130054881 | Ellis et al. | Feb 2013 | A1 |
| 20130060994 | Higgins et al. | Mar 2013 | A1 |
| 20130061019 | Fitzpatrick et al. | Mar 2013 | A1 |
| 20130073788 | Post et al. | Mar 2013 | A1 |
| 20130080691 | Weingarten et al. | Mar 2013 | A1 |
| 20130094289 | Sridharan et al. | Apr 2013 | A1 |
| 20130100600 | Yin et al. | Apr 2013 | A1 |
| 20130104005 | Weingarten et al. | Apr 2013 | A1 |
| 20130124792 | Melik-Martirosian et al. | May 2013 | A1 |
| 20130151753 | Jeon et al. | Jun 2013 | A1 |
| 20130198436 | Bandic et al. | Aug 2013 | A1 |
| 20130205102 | Jones et al. | Aug 2013 | A1 |
| 20130232290 | Ish et al. | Sep 2013 | A1 |
| 20130238833 | Vogan et al. | Sep 2013 | A1 |
| 20130265825 | Lassa | Oct 2013 | A1 |
| 20130297986 | Cohen | Nov 2013 | A1 |
| 20130297988 | Wu et al. | Nov 2013 | A1 |
| 20130332791 | Chu | Dec 2013 | A1 |
| 20140036589 | Parthasarathy et al. | Feb 2014 | A1 |
| 20140059359 | Bahirat | Feb 2014 | A1 |
| 20140108891 | Strasser et al. | Apr 2014 | A1 |
| 20140129874 | Zaltsman et al. | May 2014 | A1 |
| 20140158525 | Greene | Jun 2014 | A1 |
| 20140181370 | Cohen et al. | Jun 2014 | A1 |
| 20140208174 | Ellis et al. | Jul 2014 | A1 |
| 20140372777 | Reller et al. | Dec 2014 | A1 |
| 20150177999 | Gakhal et al. | Jun 2015 | A1 |
| Number | Date | Country |
|---|---|---|
| 1 956 489 | Aug 2008 | EP |
| 1 990 921 | Nov 2008 | EP |
| 2 498 259 | Sep 2012 | EP |
| 2012129859 | Jul 2012 | JP |
| WO 2009042298 | Apr 2009 | WO |
| WO 2011156466 | Dec 2011 | WO |
| Entry |
|---|
| Online Merriam Webster Dictionary, defintion of “Distinct” from Jun. 12, 2011, https://web.archive.org/web/20110612181129/http://www2.merriam-webster.com/cgi-bin/mwdictadu?book=Dictionary&va=distinct. |
| Cooke, “Introduction to Flash Memory (T1A),” Flash Memory Summit, Aug. 22, 2008, Micron Technology, Inc., 102 pages. |
| Gal at al., “Algorithms and Data Structures for Flash Memories,” ACM Computing Surveys, Jun. 2005, vol. 37, No. 2, 30 pages. |
| IBM Corporation, “Systems Management, Work Management,” Version 5, Release 4, 9th Edition, Feb. 2006, pp. 1-21. |
| O'Brien, “SMART Storage Systems Optimus SAS Enterprise SSD Review” SMART Storage Systems, Oct. 9, 2012, 44 pages. |
| Spanjer, “Flash Management—Why and How?” Smart Modular Technologies, Nov. 2009, http://www.scantec.de/fileadmin/pdf/Smart—Modular/Flash-Management.pdf, 14 pages. |
| Texas Instruments, “Power Management IC for Digital Set Top Boxes,” SLVSA10A, Sep. 2009, pp. 1-22. |
| International Search Report and Written Opinion dated Dec. 20, 2013, received in PCT/US2013/045282, which corresponds to U.S. Appl. No. 13/493,949, 7 pages (Ellis). |
| International Search Report and Written Opinion dated Nov. 7, 2014, received in International Patent Application No. PCT/US2014/049732, which corresponds to U.S. Appl. No. 14/334,350, 13 pages (Fitzpatrick). |
| International Search Report and Written Opinion dated Oct. 17, 2014, received in International Patent Application No. PCT/US2014/049734, which corresponds to U.S. Appl. No. 14/337,259, 8 pages (Higgins). |
| International Search Report and Written Opinion dated Oct. 23, 2014, received in International Patent Application No. PCT/US2014/049736, which corresponds to U.S. Appl. No. 14/446,249 8 pages (Fitzpatrick). |
| International Search Report and Written Opinion dated Nov. 5, 2014, received in International Patent Application No. PCT/US2014/049282, which corresponds to U.S. Appl. No. 13/957,407, 12 pages (Fitzpatrick). |
| Ulinktech, “ATA Command Table (in Alphabetic Order).” Feb. 6, 2011, https://web.archive.org/web/20110206060820/http://www.ulinktech.com/downloads/AT, 6 pages. |
| International Search Report and Written Opinion dated May 14, 2014, received in International Patent Application No. PCT/US2014/017168, which corresponds to U.S. Appl. No. 14/076,115, 6 pages (Fitzpatrick). |
| International Search Report and Written Opinion dated May 14, 2014, received in International Patent Application No. PCT/US2014/017169, which corresponds to U.S. Appl. No. 14/076,148, 6 pages (Fitzpatrick). |
| International Search Report dated Mar. 25, 2014, received in International Patent Application No. PCT/US2013/072400, which corresponds to U.S. Appl. No. 13/690,337, 3 pages (Ellis). |
| International Search Report and Written Opinion dated Jun. 12, 2014, received in PCT/US2014/018972, which corresponds to U.S. Appl. No. 13/779,352, 12 pages (Schmier). |
| Invitation to Pay Additional Fees dated Jul. 25, 2014, received in International Patent Application No. PCT/US2014/021290, which corresponds to U.S. Appl. No. 13/791,797, 9 pages (Dean). |
| International Search Report and Written Opinion dated Jul. 31, 2014, received in International Patent Application No. PCT/US2014/031465, which corresponds to U.S. Appl. No. 13/851,928, 13 pages (Ellis). |
| International Search Report and Written Opinion dated Jul. 31, 2014, received in International Patent Application No. PCT/US2014/033876, which corresponds to U.S. Appl. No. 13/861,326, 9 pages (Fitzpatrick). |
| Narayanan et al., “Migrating Server Storage to SSDs: Analysis of Tradeoffs,” Computer Systems, Apr. 2009, 12 pages. |
| Shiraz et al., “Block Aging Prevention Technique (BAP) for Flash Based Solid State Disks,” 7th International Conference on Emerging Technologies (ICET), Sep. 5, 2011, 6 pages. |
| Tai et al, “Prolongation of Lifetime and the Evaluation Method of Dependable SSD,” 25 International Symposium on Defect and Fault Tolerance in VLSI Systems, 2010, NJ, USA, 8 pages. |
| Yimo et al., “WeLe-RAID: A SSD-Based RAID for System Endurance and Performance,” Jan. 2011, Network and Parallel Computing, Springer, 14 pages. |
| International Search Report and Written Opinion dated Jan. 9, 2015, received in International Patent Application No. PCT/US2014/049731, whcih corresponds to U.S. Appl. No. 14/334,324, 9 pages (Fitzpatrick). |
| International Search Report and Written Opinion dated Feb. 18, 2015, received in International Application No. PCT/US2014/065401, which corresponds to U.S. Appl. No. 14/082,031, 9 pages (Higgins). |
| International Search Report dated Apr. 15, 2014, received in International Patent Application No. PCT/US2013/078340, which corresponds to U.S. Appl. No. 13/746,542, 11 pages (Ellis). |
| Tseng et al., “Understanding the Impact of Power Loss on Flash Memory,” DAC'11, Jun. 5-10, 2011, San Diego, California, 6 pages. |
| International Search Report and Written Opinion dated Aug. 22, 2014, received in International Patent Application No. PCT/US2014/032978, which corresponds to U.S. Appl. No. 14/081,992, 10 pages (Ellis). |
| Number | Date | Country | |
|---|---|---|---|
| 20140156966 A1 | Jun 2014 | US |
