US20180203797A1 - Compression and Decompression of Data at High Speed in Solid State Storage - Google Patents
Compression and Decompression of Data at High Speed in Solid State Storage Download PDFInfo
- Publication number
- US20180203797A1 US20180203797A1 US15/912,352 US201815912352A US2018203797A1 US 20180203797 A1 US20180203797 A1 US 20180203797A1 US 201815912352 A US201815912352 A US 201815912352A US 2018203797 A1 US2018203797 A1 US 2018203797A1
- Authority
- US
- United States
- Prior art keywords
- data
- decompression
- blocks
- flash
- decoders
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
- 238000007906 compression Methods 0.000 title claims abstract description 54
- 230000006835 compression Effects 0.000 title claims abstract description 54
- 230000006837 decompression Effects 0.000 title claims abstract description 45
- 239000007787 solid Substances 0.000 title abstract description 16
- 238000000034 method Methods 0.000 claims description 35
- 230000015654 memory Effects 0.000 claims description 26
- 238000013459 approach Methods 0.000 claims description 5
- 238000012937 correction Methods 0.000 claims description 5
- 239000000872 buffer Substances 0.000 claims 6
- 230000008569 process Effects 0.000 description 12
- 238000004891 communication Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 3
- 238000004590 computer program Methods 0.000 description 3
- 238000013500 data storage Methods 0.000 description 3
- 230000000991 decompressive effect Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 2
- RYGMFSIKBFXOCR-UHFFFAOYSA-N Copper Chemical compound [Cu] RYGMFSIKBFXOCR-UHFFFAOYSA-N 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000010979 ruby Substances 0.000 description 1
- 229910001750 ruby Inorganic materials 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000004557 technical material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F12/00—Accessing, addressing or allocating within memory systems or architectures
- G06F12/02—Addressing or allocation; Relocation
- G06F12/0223—User address space allocation, e.g. contiguous or non contiguous base addressing
- G06F12/023—Free address space management
- G06F12/0238—Memory management in non-volatile memory, e.g. resistive RAM or ferroelectric memory
- G06F12/0246—Memory management in non-volatile memory, e.g. resistive RAM or ferroelectric memory in block erasable memory, e.g. flash memory
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0602—Interfaces specially adapted for storage systems specifically adapted to achieve a particular effect
- G06F3/0608—Saving storage space on storage systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0628—Interfaces specially adapted for storage systems making use of a particular technique
- G06F3/0638—Organizing or formatting or addressing of data
- G06F3/064—Management of blocks
- G06F3/0641—De-duplication techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0668—Interfaces specially adapted for storage systems adopting a particular infrastructure
- G06F3/0671—In-line storage system
- G06F3/0673—Single storage device
- G06F3/0679—Non-volatile semiconductor memory device, e.g. flash memory, one time programmable memory [OTP]
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C7/00—Arrangements for writing information into, or reading information out from, a digital store
- G11C7/10—Input/output [I/O] data interface arrangements, e.g. I/O data control circuits, I/O data buffers
- G11C7/1006—Data managing, e.g. manipulating data before writing or reading out, data bus switches or control circuits therefor
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
- H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
- H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
- H03M7/3084—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction using adaptive string matching, e.g. the Lempel-Ziv method
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2212/00—Indexing scheme relating to accessing, addressing or allocation within memory systems or architectures
- G06F2212/40—Specific encoding of data in memory or cache
- G06F2212/401—Compressed data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2212/00—Indexing scheme relating to accessing, addressing or allocation within memory systems or architectures
- G06F2212/72—Details relating to flash memory management
- G06F2212/7202—Allocation control and policies
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2213/00—Indexing scheme relating to interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
- G06F2213/0038—System on Chip
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C16/00—Erasable programmable read-only memories
- G11C16/02—Erasable programmable read-only memories electrically programmable
- G11C16/06—Auxiliary circuits, e.g. for writing into memory
- G11C16/10—Programming or data input circuits
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C2207/00—Indexing scheme relating to arrangements for writing information into, or reading information out from, a digital store
- G11C2207/10—Aspects relating to interfaces of memory device to external buses
- G11C2207/102—Compression or decompression of data before storage
Definitions
- the present invention relates generally to computer software, computer program architecture, computer memory, and data storage and compression. More specifically, techniques for compression and decompression of data at high speed in solid state storage are described.
- Hard disk based storage traditionally has been used to store computer data. Hard disk capacity has increased dramatically over the years; however, hard disk performance in terms of Input/Output Operations per Second (hereafter “IOPS”) has scaled more slowly and has reached a saturation point.
- IOPS Input/Output Operations per Second
- the terms “hard disk drive” and “hard disk” are used interchangeably herein to refer to computer-based non-volatile magnetic data storage.
- Flash memory is typically implemented as solid state storage, such as a solid state drive (SSD), which is packaged similar to a hard disk and has a similar interface and programming model.
- SSD solid state drive
- Flash On a cost per gigabyte (GB) basis, Flash is more expensive than hard disks. Therefore, it is desirable to save costs by compressing data before storing it in Flash, such that the data may be saved using less Flash memory. This also has the advantage of reducing the amount of data that needs to be written to Flash, which improves performance.
- GB gigabyte
- a conventional architecture for performing such lossless compression can include a unit configured to implement Lempel-Ziv compression followed by another unit configured to implement variable length coding.
- Some Lempel-Ziv compression algorithm solutions include a hash module, a module for resolving collision chains, and a module for performing longest match searches.
- Some Lempel-Ziv compression algorithm solutions also include various units of RAM memory for storing hash data or collision chains data.
- Variable length coding following Lempel-Ziv compression is used conventionally to maintain superior compression ratios; however conventional compression and decompression solutions involving variable length coding cannot be scaled to achieve 4 GB/s.
- FIG. 1 illustrates an exemplary Flash memory system comprising a plurality of Flash chips
- FIG. 2 illustrates an exemplary block diagram showing compression and storage of data in Flash chips
- FIG. 3 illustrates an exemplary decompressor system for decompression of data stored in Flash chips to reconstruct original data
- FIG. 4 illustrates an exemplary process for compression and storage of data in Flash chips to enable two-stage decompression
- FIG. 5 illustrates an exemplary computer system suitable for implementing compression and decompression of data at high speed in solid state storage.
- hard disk drive and “hard disk” are used interchangeably herein to refer computer-based non-volatile magnetic data storage.
- hard disk-based storage may be implemented in a variety of forms, including, but not limited to Direct Attached Storage (DAS), Network Attached Storage (NAS) and Storage Area Networks (SAN).
- DAS Direct Attached Storage
- NAS Network Attached Storage
- SAN Storage Area Networks
- a single apparatus i.e., device, machine, system, or the like
- SSD solid state drives
- HDD hard disk drives
- a single monolithic device having both Flash and hard disk drives may be used.
- multiple, disparate (i.e., separate) storage facilities in different apparatuses may be used.
- the techniques described herein may be used with any type of digital memory without limitation or restriction.
- the described techniques may be implemented, in whole or in part, as a computer program or application (“application”) or as a plug-in, module, or sub-component of another application.
- the described techniques may be implemented using various types of programming, development, scripting, or formatting languages, frameworks, syntax, applications, protocols, objects, or techniques, including JavaTM, JavascriptTM, Ruby, Rails, C, Objective C, C++, C#, Adobe® Integrated RuntimeTM (Adobe® AIRTM), ActionScriptTM, FlexTM, LingoTM, Ajax, Perl, COBOL, Fortran, ADA, XML, MXML, HTML, DHTML, XHTML, HTTP, XMPP, and others. Design, publishing, and other types of applications such as Dreamweaver®, Shockwave®, Adobe Flash®, and Fireworks® may also be used to implement the described techniques.
- the described techniques may be varied and are not limited to the examples or descriptions provided.
- the overall system performance of solid state storage may be improved by optimizing decompression rates, using the compression and decompression techniques described herein.
- Most applications are very sensitive to read latency (i.e., the time it takes to complete a read), but not very sensitive to write latency. Since decompression is in the read path, it is desirable to maintain decompression performance at rates matching the aggregate bandwidth of all Flash chips in the system. On the other hand, since compression is in the write path, it is less important for compression to be performed at such high speeds. Described herein are techniques for compression and decompression of data in solid state storage to achieve decompression performance at rates approximating the aggregate bandwidth of a memory system implementing multiple Flash chips.
- FIG. 1 illustrates an exemplary Flash memory system comprising a plurality of Flash chips.
- system 100 includes Flash controller 102 and Flash chips 104 - 108 .
- Flash chips 104 - 108 may be implemented in system 100 , the actual number based on implementation considerations such as cost, power dissipation and available physical space.
- an additional thirteen (13) additional Flash chips may be implemented between Flash chips 106 and 108 , for a total of sixteen (16) Flash chips.
- each Flash chip may store data in blocks where each block is equal to the size of an Error Correction Code block (hereafter “ECC block”), a block of data over which an error code is computed. Flash memory has fairly high error rates, thus it is essential to have some error correction to reduce the error rate experienced by the end user.
- the error correction may be block-based, meaning that the error code is computed over an entire block of data and the correction also is done on the entire block.
- the size of an ECC block hereafter “N ECC bytes”
- N ECC bytes may be 512 bytes, which matches the industry standard sector size for disks.
- blocks of data may be arranged differently according to different purposes, and also in different sizes.
- each of the Flash chips may be compliant with the Open NAND Flash Interface (ONFI) 2.0 specification, and thus may supply data at rates of approximately 166 MB/s. In other examples, the Flash chips may comply with other specifications.
- ECC may be done separately for each of Flash chips 104 - 108 , in which case they each may comprise an entire ECC block of data.
- each of Flash chips 104 - 108 may comprise N ECC bytes of data.
- the numbers inside each of the Flash chips are exemplary byte addresses.
- the ECC block of data in Flash chip 104 may comprise bytes 0, 16, 32, and so on through byte 8176 for a total of 512 bytes of data in Flash chip 104 .
- the ECC block of data in Flash chip 106 may comprise bytes 1, 17, 33, and so on through byte 8177 for a total of 512 bytes of data in Flash chip 106 .
- the ECC block of data in Flash chip 108 may comprise bytes 15, 31, 47, and so on through byte 8191 for a total of 512 bytes of data in Flash chip 108 . Since all of the Flash chips may be accessed simultaneously, in an implementation with 16 Flash chips up to 16 ECC blocks of data may be retrieved in a single read operation. In some examples, the data may arrive on a 16-byte wide bus, wherein the first byte from all of the Flash chips arrives in a first clock cycle, the second byte from all of the Flash chips arrives in a second clock cycle, and so on. Thus, in the example shown in FIG.
- system 100 may be implemented with more or fewer Flash chips.
- FIG. 2 illustrates an exemplary block diagram showing compression and storage of data in Flash chips.
- diagram 200 includes uncompressed data 202 , compressed data 204 , and blocks 206 - 212 of padded compressed data.
- Compression algorithms are broadly classified into two categories: lossless and lossy. Lossless compression generally is designed to reproduce original data more accurately upon decompression. Lossy compression, while less accurate, instead provides higher compression ratios. Since solid state storage is often used for a wide variety of data types, lossless compression is desired for implementing compression of data in solid state storage.
- lossless compression may be performed using Lempel-Ziv (hereafter “LZ” or “Lempel-Ziv”) compression followed by variable length coding.
- LZ compression may be implemented by finding repeated patterns in data and replacing all but the first occurrence of a pattern with a pointer to the original pattern. As used herein, a byte that is not replaced with a pointer (i.e., the first occurrence of the pattern) is called “a literal,” and the pointer information for a string of bytes that have been replaced is called “a match.” In some examples, further compression then may be achieved using variable length coding.
- variable length coding assigns shorter codes to symbols that occur more frequently, and longer codes to symbols that occur less frequently. Encoding data in this manner often leads to an average code length that is less than the original, thus making the data smaller. As described in more detail below, these processes may be reversed during decompression to reconstruct the original data.
- a two-stage compression may be implemented using other compression algorithms, wherein the first stage of decompression is slower relative to the second stage of decompression (e.g., the second stage of compression may be implemented using arithmetic coding instead of variable length coding).
- uncompressed data 202 may comprise 4096 bytes of data.
- uncompressed data 202 may be compressed to the size of compressed data 204 , which may comprise 1890 bytes.
- Compressed data 204 then may be padded up to blocks of N ECC bytes to become four blocks of padded compressed data (e.g., blocks 206 - 212 ), each comprising N ECC bytes (e.g., 512 bytes) of data.
- blocks 206 - 212 combine for a total of 2048 bytes.
- padding may be implemented by including a particular code to indicate that the rest of the output is padded up to the next ECC block boundary (i.e., the boundary of a block of a predetermined size).
- block 206 may comprise bytes 0 to N ECC ⁇ 1
- block 208 may comprise bytes N ECC to 2N ECC ⁇ 1
- block 210 may comprise bytes 2N ECC to 3N ECC ⁇ 1
- block 212 may comprise bytes 3N ECC to 4N ECC ⁇ 1.
- block 206 then may be written to a first Flash chip (e.g., Flash 0 as shown), block 208 may be written to a second Flash chip (e.g., Flash 1 as shown), block 210 may be written to a third Flash chip (e.g., Flash 2 as shown), and block 212 may be written to a fourth Flash chip (e.g., Flash 3 as shown).
- a first Flash chip e.g., Flash 0 as shown
- block 208 may be written to a second Flash chip (e.g., Flash 1 as shown)
- block 210 may be written to a third Flash chip (e.g., Flash 2 as shown)
- block 212 may be written to a fourth Flash chip (e.g., Flash 3 as shown).
- the compression technique displayed in diagram 200 serves to arrange compressed output data in a fashion that makes it possible to decompress at a faster rate.
- FIG. 3 illustrates an exemplary decompressor system for decompression of data stored in Flash chips to reconstruct original data.
- system 300 includes Variable Length Decoders (VLDs) 302 - 304 , First-In-First-Out circuits (FIFOs) 308 - 312 , multiplexer (MUX) 314 , and reverse LZ unit (hereafter “UNLZ”) 316 .
- VLDs Variable Length Decoders
- FIFOs First-In-First-Out circuits
- MUX multiplexer
- UNLZ reverse LZ unit
- all of the ECC blocks (e.g., comprising N ECC bytes) that were a part of that compressed block may be accessed simultaneously and each VLD may decode the ECC block retrieved from the Flash chip with which it is associated. Since it is desirable that no variable length code be split across ECC block boundaries, the technique of padding described above in FIG. 2 may be implemented so that each VLDs has the full information required to decode a variable length code (i.e., that no variable length code is split such that part of it is output to one VLD and the rest to another VLD). In some examples, there is a VLD for each Flash chip.
- each VLD may decode compressed data from an associated Flash chip.
- a variable length code from Flash 0 may be decoded by VLD 302
- another variable length code from Flash 1 may be decoded by VLD 304
- VLD 306 which receives and decodes a variable length code from Flash 15.
- the output from VLDs 302 - 306 may comprise literals and matches.
- the VLD may function to pad the match information with additional bytes, where necessary, such that the total space occupied by the match and the pad bytes equals the length of the original string.
- VLD 302 may pad the string with 4 bytes of arbitrary data so that the code is 7 bytes, to match the length of the replaced string.
- the length of the matched string were 8 bytes, also with a 2 byte backwards distance and a 1 byte length, the arbitrary data portion would be 5 bytes.
- VLD 302 also may generate a 1 bit flag for each byte to assist UNLZ 316 with identifying whether a particular byte is a literal or part of a match. For example, one value of a flag (e.g., 0) may indicate a literal and another value (e.g., 1) may indicate a match.
- VLD 304 and VLD 306 may function similarly to VLD 302 with respect to data from Flash 1 and Flash 15, respectively.
- each of VLD 302 - 306 may be provided to an associated FIFO.
- FIFO 308 may be associated with VLD 302 and receive VLD 302 's output
- FIFO 310 may be associated with VLD 304 and receive VLD 304 's output and so on through to FIFO 312 , which receives VLD 306 's output.
- FIFO 308 - 312 may be configured so that its input is written one byte at a time, and its output is read 16 bytes at a time.
- system 300 may be implemented such that different numbers of bytes may be written to, or read from, FIFO 308 - 312 , based upon implementation considerations. Reading multiple bytes in parallel from FIFO 308 - 312 allows the second stage of decompression (e.g., reverse LZ compression, described below) to run faster.
- second stage of decompression e.g., reverse LZ compression, described below
- UNLZ 316 may read data from FIFOs 308 - 312 , and all FIFOs in between, which are associated with Flash 0-15, respectively, each of which may comprise N ECC bytes of compressed data. In some examples, UNLZ 316 may receive this data via MUX 314 , and it may read 16 bytes in a single clock cycle. For example, MUX 314 may process data output from FIFO 308 (e.g., 16 bytes at time), then process data output from FIFO 310 , and so on through FIFO 312 , and then output data (e.g., in order) to UNLZ 316 .
- FIFOs 308 - 312 and all FIFOs in between, which are associated with Flash 0-15, respectively, each of which may comprise N ECC bytes of compressed data.
- UNLZ 316 may receive this data via MUX 314 , and it may read 16 bytes in a single clock cycle.
- MUX 314 may process data output
- UNLZ 316 may decompress data from FIFO 308 first, then decompress data from FIFO 310 , and so on until an entire block of data is decompressed. In some examples, such a block of data may not use all 16 FIFOs (e.g., does not use VLDs and Flash chips). In other examples, system 300 may be implemented with more or fewer Flash chips, VLD units and FIFO units.
- UNLZ 316 may examine the flag for each byte to determine if it is a literal or a match. For example, if the byte is a literal, UNLZ 316 may simply pass the data through as original data, or part of original data. If, in another example, the byte is the start of a match, then UNLZ 316 may reconstruct the original data associated with that string by reading the length and the backward distance of the match, retrieving the previous occurrence of the matching string suing the backward distance, and replacing the appropriate number of bytes according to data from the matching string.
- UNLZ 316 may operate on 16 bytes in a single clock cycle.
- a span of 16 bytes may include multiple matches.
- UNLZ 316 may operate to replace all of the matches in the 16-byte span in a single clock cycle.
- a minimum length of matches may be four (4) bytes.
- a 16-byte span may have up to 4 matches, if the span begins with the start of a match. Since a match may start on any byte, in another example, a 16-byte span may comprise three (3) complete matches and portions of two (2) additional matches.
- a memory used to store previous bytes must be operable to conduct at least five (5) lookups in a single clock cycle, for example if each match has a different backward distance.
- the memory may be implemented as a multiported memory with 5 read ports and 1 write port.
- multiple memories e.g., 5) may be implemented, each comprising 1 of 5 copies of the same data, each with 1 write port and 1 read port.
- the maximum number of matches may depend on a given implementation and compression process, and that other implementations may require more or fewer ports, for example, depending on the compression hardware.
- UNLZ 316 may operate at a clock frequency of 250 MHz, thus yielding a peak performance of 4 GB/s, when processing 16 bytes per clock cycle. In other examples, UNLZ 316 may operate faster or slower, depending on the processing technology used, yielding higher or lower performance. In still other examples, system 300 may be implemented with more or fewer components.
- FIG. 4 illustrates an exemplary process for compression and storage of data in Flash chips to enable two-stage decompression.
- process 400 may begin with compressing data until a limit of a block is detected by a compressor ( 402 ).
- the limit of the block may be associated with a predetermined block size (e.g., N ECC bytes).
- the compressed data may be padded ( 404 ) up to the predetermined block size (e.g., padding up to N ECC bytes, as described above re FIG. 2 ).
- a determination may be made whether there is more uncompressed data to process (i.e., a remaining uncompressed portion of data) ( 406 ).
- a loop variable e.g., i
- FIG. 5 illustrates an exemplary computer system suitable for implementing compression and decompression of data at high speed in solid state storage.
- computer system 500 may implement computer programs, applications, methods, processes, or other software or hardware to perform the above-described techniques.
- Computer system 500 includes a bus 502 or other communication mechanism for communicating information, which interconnects subsystems and devices, such as processor 504 , system memory 506 (e.g., RAM), storage device 508 (e.g., Flash, ROM, etc.), disk drive 510 (e.g., magnetic or optical), communication interface 512 (e.g., modem or Ethernet card), display 514 (e.g., CRT or LCD), input device 516 (e.g., keyboard), and cursor control 518 (e.g., mouse or trackball).
- processor 504 system memory 506 (e.g., RAM), storage device 508 (e.g., Flash, ROM, etc.), disk drive 510 (e.g., magnetic or optical), communication interface 512 (e.
- computer system 500 performs specific operations by processor 504 executing one or more sequences of one or more instructions stored in system memory 506 . Such instructions may be read into system memory 506 from another computer readable medium, such as static storage device 508 or disk drive 510 . In some examples, hard-wired circuitry may be used in place of or in combination with software instructions for implementation.
- Non-volatile media includes, for example, optical or magnetic disks, such as disk drive 510 .
- Volatile media includes dynamic memory, such as system memory 506 .
- Computer readable media includes, for example, floppy disk, flexible disk, hard disk, magnetic tape, any other magnetic medium, CD-ROM, any other optical medium, punch cards, paper tape, any other physical medium with patterns of holes, RAM, PROM, EPROM, FLASH-EPROM, any other memory chip or cartridge, or any other medium from which a computer can read.
- Transmission medium may include any tangible or intangible medium that is capable of storing, encoding or carrying instructions for execution by the machine, and includes digital or analog communications signals or other intangible medium to facilitate communication of such instructions.
- Transmission media includes coaxial cables, copper wire, and fiber optics, including wires that comprise bus 502 for transmitting a computer data signal.
- execution of the sequences of instructions may be performed by a single computer system 500 .
- two or more computer systems 500 coupled by communication link 520 may perform the sequence of instructions in coordination with one another.
- Computer system 500 may transmit and receive messages, data, and instructions, including program, i.e., application code, through communication link 520 and communication interface 512 .
- Received program code may be executed by processor 504 as it is received, and/or stored in disk drive 510 , or other non-volatile storage for later execution.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Human Computer Interaction (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
Description
- This patent application is a continuation of U.S. patent application having Ser. No. 13/880,728, filed on Apr. 20, 2013 in the name of the same inventors and having a title of “Compression and Decompressive of Data at High Speed in Solid State Storage,” which further claims the priority benefit of the PCT application having a serial No. PCT/1N2011/000725 filed Oct. 20, 2011 entitled “Compression and Decompressive of Data at High Speed in Solid State Storage,” which claims the priority benefit of the Indian provisional patent number 2942/MUM/2010, and entitled “Compression and Decompressive of Data at High Speed in Solid State Storage”, filed on Oct. 22, 2010, all of which are incorporated herein by reference in their entirety.
- The present invention relates generally to computer software, computer program architecture, computer memory, and data storage and compression. More specifically, techniques for compression and decompression of data at high speed in solid state storage are described.
- Hard disk based storage traditionally has been used to store computer data. Hard disk capacity has increased dramatically over the years; however, hard disk performance in terms of Input/Output Operations per Second (hereafter “IOPS”) has scaled more slowly and has reached a saturation point. The terms “hard disk drive” and “hard disk” are used interchangeably herein to refer to computer-based non-volatile magnetic data storage.
- Digital memories based on electronic circuitry perform faster than conventional hard disks. Thus, storing data in digital memory is an effective way to increase and improve upon the performance of hard disks. However, there are problems associated with conventional digital memory solutions. Traditionally, Dynamic Random Access Memory (DRAM) has been used to provide fast access to data. More recently, NAND Flash™ memory (hereafter “Flash” or “Flash chip”) has surpassed DRAM in terms of cost and capacity and is becoming a popular choice for high performance storage. Flash memory is typically implemented as solid state storage, such as a solid state drive (SSD), which is packaged similar to a hard disk and has a similar interface and programming model.
- On a cost per gigabyte (GB) basis, Flash is more expensive than hard disks. Therefore, it is desirable to save costs by compressing data before storing it in Flash, such that the data may be saved using less Flash memory. This also has the advantage of reducing the amount of data that needs to be written to Flash, which improves performance.
- However, the challenge in introducing compression in solid state storage is that compression and decompression engines must process data at very high rates to achieve performance rates similar to those of Flash-based memories. This poses a challenge because compression and decompression algorithms are compute-intensive (i.e., require a large number of computing operations). Conventional software-based approaches, running compression and decompression algorithms in a host processor, are unlikely to be able to keep up with the performance requirements. Also, there are performance limitations with conventional dedicated hardware for compression and decompression. For example, an SSD implementation can require many Flash chips in parallel to read and write data at very high rates. Some Flash chips, such as those compliant with the Open NAND Flash Interface (ONFI) 2.0 specification, can supply data at rates of approximately 166 megabytes (MB) per second. Further, it is not uncommon to operate 16 to 24 such Flash chips in parallel, which results in data rates as high as 4 GB/s, which is difficult to approximate using conventional compression and decompression solutions. It is desirable to achieve decompression at rates approaching 4 GB/s, which conventional software and hardware compression and decompression tools cannot achieve without incurring significant expense and complex architectures.
- One conventional solution uses a lossless compression algorithm, such as Lempel-Ziv compression algorithms and variable length coding. A conventional architecture for performing such lossless compression can include a unit configured to implement Lempel-Ziv compression followed by another unit configured to implement variable length coding. Some Lempel-Ziv compression algorithm solutions include a hash module, a module for resolving collision chains, and a module for performing longest match searches. Some Lempel-Ziv compression algorithm solutions also include various units of RAM memory for storing hash data or collision chains data. Variable length coding following Lempel-Ziv compression is used conventionally to maintain superior compression ratios; however conventional compression and decompression solutions involving variable length coding cannot be scaled to achieve 4 GB/s.
- Thus, what is needed is a solution for decompression of data at high speeds without the limitations of conventional techniques.
- Various embodiments (“examples”) of the invention are disclosed in the following detailed description and the accompanying drawings:
-
FIG. 1 illustrates an exemplary Flash memory system comprising a plurality of Flash chips; -
FIG. 2 illustrates an exemplary block diagram showing compression and storage of data in Flash chips; -
FIG. 3 illustrates an exemplary decompressor system for decompression of data stored in Flash chips to reconstruct original data; -
FIG. 4 illustrates an exemplary process for compression and storage of data in Flash chips to enable two-stage decompression; and -
FIG. 5 illustrates an exemplary computer system suitable for implementing compression and decompression of data at high speed in solid state storage. - Various embodiments or examples may be implemented in numerous ways, including as a system, a process, an apparatus, or a series of program instructions on a computer readable medium such as a computer readable storage medium or a computer network where the program instructions are sent over optical, electronic, or wireless communication links. In general, operations of disclosed processes may be performed in an arbitrary order, unless otherwise provided in the claims.
- A detailed description of one or more examples is provided below along with accompanying figures. The detailed description is provided in connection with such examples, but is not limited to any particular example. The scope is limited only by the claims and numerous alternatives, modifications, and equivalents are encompassed. Numerous specific details are set forth in the following description in order to provide a thorough understanding. These details are provided for the purpose of example and the described techniques may be practiced according to the claims without some or all of these specific details. For clarity, technical material that is known in the technical fields related to the examples has not been described in detail to avoid unnecessarily obscuring the description.
- The described techniques may be implemented as hardware, firmware, circuitry, software or a combination thereof for purposes of providing computational and processing capabilities. The terms “hard disk drive” and “hard disk” are used interchangeably herein to refer computer-based non-volatile magnetic data storage. As used herein, hard disk-based storage may be implemented in a variety of forms, including, but not limited to Direct Attached Storage (DAS), Network Attached Storage (NAS) and Storage Area Networks (SAN). In some examples, a single apparatus (i.e., device, machine, system, or the like) may include both Flash and hard disk-based storage facilities (e.g., solid state drives (SSD), hard disk drives (HDD), or others). In other words, a single monolithic device having both Flash and hard disk drives may be used. In other examples, multiple, disparate (i.e., separate) storage facilities in different apparatuses may be used. Further, the techniques described herein may be used with any type of digital memory without limitation or restriction. In some examples, the described techniques may be implemented, in whole or in part, as a computer program or application (“application”) or as a plug-in, module, or sub-component of another application. If implemented as software, the described techniques may be implemented using various types of programming, development, scripting, or formatting languages, frameworks, syntax, applications, protocols, objects, or techniques, including Java™, Javascript™, Ruby, Rails, C, Objective C, C++, C#, Adobe® Integrated Runtime™ (Adobe® AIR™), ActionScript™, Flex™, Lingo™, Ajax, Perl, COBOL, Fortran, ADA, XML, MXML, HTML, DHTML, XHTML, HTTP, XMPP, and others. Design, publishing, and other types of applications such as Dreamweaver®, Shockwave®, Adobe Flash®, and Fireworks® may also be used to implement the described techniques. The described techniques may be varied and are not limited to the examples or descriptions provided.
- The overall system performance of solid state storage may be improved by optimizing decompression rates, using the compression and decompression techniques described herein. Most applications are very sensitive to read latency (i.e., the time it takes to complete a read), but not very sensitive to write latency. Since decompression is in the read path, it is desirable to maintain decompression performance at rates matching the aggregate bandwidth of all Flash chips in the system. On the other hand, since compression is in the write path, it is less important for compression to be performed at such high speeds. Described herein are techniques for compression and decompression of data in solid state storage to achieve decompression performance at rates approximating the aggregate bandwidth of a memory system implementing multiple Flash chips.
-
FIG. 1 illustrates an exemplary Flash memory system comprising a plurality of Flash chips. Here,system 100 includesFlash controller 102 and Flash chips 104-108. In some examples, more or fewer Flash chips may be implemented insystem 100, the actual number based on implementation considerations such as cost, power dissipation and available physical space. In an example, an additional thirteen (13) additional Flash chips (not shown) may be implemented betweenFlash chips - In some examples, each Flash chip may store data in blocks where each block is equal to the size of an Error Correction Code block (hereafter “ECC block”), a block of data over which an error code is computed. Flash memory has fairly high error rates, thus it is essential to have some error correction to reduce the error rate experienced by the end user. The error correction may be block-based, meaning that the error code is computed over an entire block of data and the correction also is done on the entire block. In some examples, the size of an ECC block (hereafter “NECC bytes”), may be 512 bytes, which matches the industry standard sector size for disks. In other examples, blocks of data may be arranged differently according to different purposes, and also in different sizes. Also, in some examples, each of the Flash chips may be compliant with the Open NAND Flash Interface (ONFI) 2.0 specification, and thus may supply data at rates of approximately 166 MB/s. In other examples, the Flash chips may comply with other specifications.
- In some examples, ECC may be done separately for each of Flash chips 104-108, in which case they each may comprise an entire ECC block of data. Thus, each of Flash chips 104-108 may comprise NECC bytes of data. As shown, the numbers inside each of the Flash chips are exemplary byte addresses. For example, the ECC block of data in
Flash chip 104 may comprisebytes byte 8176 for a total of 512 bytes of data inFlash chip 104. In this example, the ECC block of data inFlash chip 106 may comprisebytes byte 8177 for a total of 512 bytes of data inFlash chip 106. In this example, the ECC block of data in Flash chip 108 (e.g., the 16 Flash chip) may comprisebytes byte 8191 for a total of 512 bytes of data inFlash chip 108. Since all of the Flash chips may be accessed simultaneously, in an implementation with 16 Flash chips up to 16 ECC blocks of data may be retrieved in a single read operation. In some examples, the data may arrive on a 16-byte wide bus, wherein the first byte from all of the Flash chips arrives in a first clock cycle, the second byte from all of the Flash chips arrives in a second clock cycle, and so on. Thus, in the example shown inFIG. 1 , when all 16 Flash chips are read,bytes 0 through 15 arrive in one clock cycle (e.g., each from a separate Flash chip), then bytes 16 through 31 arrive in the next clock cycle, and so on. In other examples,system 100 may be implemented with more or fewer Flash chips. -
FIG. 2 illustrates an exemplary block diagram showing compression and storage of data in Flash chips. Here, diagram 200 includesuncompressed data 202,compressed data 204, and blocks 206-212 of padded compressed data. Compression algorithms are broadly classified into two categories: lossless and lossy. Lossless compression generally is designed to reproduce original data more accurately upon decompression. Lossy compression, while less accurate, instead provides higher compression ratios. Since solid state storage is often used for a wide variety of data types, lossless compression is desired for implementing compression of data in solid state storage. - In some examples, lossless compression may be performed using Lempel-Ziv (hereafter “LZ” or “Lempel-Ziv”) compression followed by variable length coding. LZ compression may be implemented by finding repeated patterns in data and replacing all but the first occurrence of a pattern with a pointer to the original pattern. As used herein, a byte that is not replaced with a pointer (i.e., the first occurrence of the pattern) is called “a literal,” and the pointer information for a string of bytes that have been replaced is called “a match.” In some examples, further compression then may be achieved using variable length coding. While each symbol of computer data is described with a fixed length code, which is almost always 8 bits, variable length coding assigns shorter codes to symbols that occur more frequently, and longer codes to symbols that occur less frequently. Encoding data in this manner often leads to an average code length that is less than the original, thus making the data smaller. As described in more detail below, these processes may be reversed during decompression to reconstruct the original data. In other examples, a two-stage compression may be implemented using other compression algorithms, wherein the first stage of decompression is slower relative to the second stage of decompression (e.g., the second stage of compression may be implemented using arithmetic coding instead of variable length coding).
- In an example,
uncompressed data 202 may comprise 4096 bytes of data. In this example,uncompressed data 202 may be compressed to the size ofcompressed data 204, which may comprise 1890 bytes.Compressed data 204 then may be padded up to blocks of NECC bytes to become four blocks of padded compressed data (e.g., blocks 206-212), each comprising NECC bytes (e.g., 512 bytes) of data. In this example, blocks 206-212 combine for a total of 2048 bytes. In some examples, padding may be implemented by including a particular code to indicate that the rest of the output is padded up to the next ECC block boundary (i.e., the boundary of a block of a predetermined size). The compressor outputs this code when it sees that the output stream is approaching an ECC block boundary and then fills any remaining bits or bytes of the current ECC block with arbitrary data, thereby adding a portion of arbitrary data to the ECC block. In some examples, a compressor (not shown) may be implemented to detect when it is approaching the boundary of an ECC block and will apply padding to complete the ECC block. The compressor may then start the next block with a new variable length code. Thus, in some examples, block 206 may comprisebytes 0 to NECC−1, block 208 may comprise bytes NECC to 2NECC−1, block 210 may comprise bytes 2NECC to 3NECC−1, and block 212 may comprise bytes 3NECC to 4NECC−1. In other examples, other techniques for padding data encoded by variable length coding may be implemented. In some examples, block 206 then may be written to a first Flash chip (e.g.,Flash 0 as shown), block 208 may be written to a second Flash chip (e.g.,Flash 1 as shown), block 210 may be written to a third Flash chip (e.g.,Flash 2 as shown), and block 212 may be written to a fourth Flash chip (e.g.,Flash 3 as shown). A person of ordinary skill in the art will recognize that the sizes described herein are for purposes of illustration only, and that the actual sizes of uncompressed and compressed data may vary, and thus also the number of Flash chips used, depending on the nature of the data and the implementation of the compressor. As shown, the compression technique displayed in diagram 200 serves to arrange compressed output data in a fashion that makes it possible to decompress at a faster rate. -
FIG. 3 illustrates an exemplary decompressor system for decompression of data stored in Flash chips to reconstruct original data. Here,system 300 includes Variable Length Decoders (VLDs) 302-304, First-In-First-Out circuits (FIFOs) 308-312, multiplexer (MUX) 314, and reverse LZ unit (hereafter “UNLZ”) 316. As shown here, for each Flash chip (e.g.,Flash 0,Flash 1 throughFlash 15, etc.), there may be a separate VLD. In some examples, when a particular block of compressed data needs to be accessed, all of the ECC blocks (e.g., comprising NECC bytes) that were a part of that compressed block may be accessed simultaneously and each VLD may decode the ECC block retrieved from the Flash chip with which it is associated. Since it is desirable that no variable length code be split across ECC block boundaries, the technique of padding described above inFIG. 2 may be implemented so that each VLDs has the full information required to decode a variable length code (i.e., that no variable length code is split such that part of it is output to one VLD and the rest to another VLD). In some examples, there is a VLD for each Flash chip. In this first stage of decompression, each VLD may decode compressed data from an associated Flash chip. For example, as shown, a variable length code fromFlash 0 may be decoded byVLD 302, another variable length code fromFlash 1 may be decoded byVLD 304, and so on throughVLD 306, which receives and decodes a variable length code fromFlash 15. In some examples, the output from VLDs 302-306 may comprise literals and matches. In the case of a match, the VLD may function to pad the match information with additional bytes, where necessary, such that the total space occupied by the match and the pad bytes equals the length of the original string. For example, if a string of 7 bytes fromFlash 0 was found during compression to have a match with an earlier occurrence of 7 bytes, and if that match information comprised a 2 byte backwards distance identifying the previous occurrence with a 1 byte length, the match information thus comprising 3 bytes. In some examples,VLD 302 may pad the string with 4 bytes of arbitrary data so that the code is 7 bytes, to match the length of the replaced string. In another example, if the length of the matched string were 8 bytes, also with a 2 byte backwards distance and a 1 byte length, the arbitrary data portion would be 5 bytes. In some examples,VLD 302 also may generate a 1 bit flag for each byte to assistUNLZ 316 with identifying whether a particular byte is a literal or part of a match. For example, one value of a flag (e.g., 0) may indicate a literal and another value (e.g., 1) may indicate a match.VLD 304 andVLD 306 may function similarly toVLD 302 with respect to data fromFlash 1 andFlash 15, respectively. - In some examples, the output of each of VLD 302-306 may be provided to an associated FIFO. For example,
FIFO 308 may be associated withVLD 302 and receiveVLD 302's output,FIFO 310 may be associated withVLD 304 and receiveVLD 304's output and so on through toFIFO 312, which receivesVLD 306's output. In some examples, FIFO 308-312 may be configured so that its input is written one byte at a time, and its output is read 16 bytes at a time. A person of ordinary skill in the art will recognize thatsystem 300 may be implemented such that different numbers of bytes may be written to, or read from, FIFO 308-312, based upon implementation considerations. Reading multiple bytes in parallel from FIFO 308-312 allows the second stage of decompression (e.g., reverse LZ compression, described below) to run faster. - In some examples,
UNLZ 316 may read data from FIFOs 308-312, and all FIFOs in between, which are associated with Flash 0-15, respectively, each of which may comprise NECC bytes of compressed data. In some examples,UNLZ 316 may receive this data viaMUX 314, and it may read 16 bytes in a single clock cycle. For example,MUX 314 may process data output from FIFO 308 (e.g., 16 bytes at time), then process data output fromFIFO 310, and so on throughFIFO 312, and then output data (e.g., in order) to UNLZ 316. In some examples,UNLZ 316 may decompress data fromFIFO 308 first, then decompress data fromFIFO 310, and so on until an entire block of data is decompressed. In some examples, such a block of data may not use all 16 FIFOs (e.g., does not use VLDs and Flash chips). In other examples,system 300 may be implemented with more or fewer Flash chips, VLD units and FIFO units. - In some examples,
UNLZ 316 may examine the flag for each byte to determine if it is a literal or a match. For example, if the byte is a literal,UNLZ 316 may simply pass the data through as original data, or part of original data. If, in another example, the byte is the start of a match, then UNLZ 316 may reconstruct the original data associated with that string by reading the length and the backward distance of the match, retrieving the previous occurrence of the matching string suing the backward distance, and replacing the appropriate number of bytes according to data from the matching string. - As mentioned above,
UNLZ 316 may operate on 16 bytes in a single clock cycle. In some examples, a span of 16 bytes may include multiple matches. In such examples,UNLZ 316 may operate to replace all of the matches in the 16-byte span in a single clock cycle. In some examples, a minimum length of matches may be four (4) bytes. In such an example, a 16-byte span may have up to 4 matches, if the span begins with the start of a match. Since a match may start on any byte, in another example, a 16-byte span may comprise three (3) complete matches and portions of two (2) additional matches. In some examples, to enable this flexibility, a memory used to store previous bytes must be operable to conduct at least five (5) lookups in a single clock cycle, for example if each match has a different backward distance. For example, the memory may be implemented as a multiported memory with 5 read ports and 1 write port. Alternatively, multiple memories (e.g., 5) may be implemented, each comprising 1 of 5 copies of the same data, each with 1 write port and 1 read port. A person of ordinary skill in the art will understand that the maximum number of matches may depend on a given implementation and compression process, and that other implementations may require more or fewer ports, for example, depending on the compression hardware. - In some examples,
UNLZ 316 may operate at a clock frequency of 250 MHz, thus yielding a peak performance of 4 GB/s, when processing 16 bytes per clock cycle. In other examples,UNLZ 316 may operate faster or slower, depending on the processing technology used, yielding higher or lower performance. In still other examples,system 300 may be implemented with more or fewer components. -
FIG. 4 illustrates an exemplary process for compression and storage of data in Flash chips to enable two-stage decompression. Here,process 400 may begin with compressing data until a limit of a block is detected by a compressor (402). The limit of the block may be associated with a predetermined block size (e.g., NECC bytes). Then, in some examples, the compressed data may be padded (404) up to the predetermined block size (e.g., padding up to NECC bytes, as described above reFIG. 2 ). Once a block of compressed data is padded, a determination may be made whether there is more uncompressed data to process (i.e., a remaining uncompressed portion of data) (406). If yes, then the compressor continues to compress and pad the uncompressed data into blocks of the predetermined block size. If no, then a loop variable (e.g., i) is initialized (e.g., i=0) (408). As shown, an ith (e.g., 0th) block of data, of the predetermined block size, then may be stored to a Flash chip designated Flash i (e.g., Flash 0) (410). After an ith block of data is stored in a Flash chip, the loop variable may be advanced (e.g., i=i+1) (412), and then a determination made whether the system is done storing data (e.g., whether there is more compressed data to store) (414). If the system is not done storing data, then the next ih block of data (e.g., i=1) may be stored to the next Flash i (e.g., Flash 1), and so on. If the system is done storing data (e.g., there is no more compressed data to store), then process 400 ends. -
FIG. 5 illustrates an exemplary computer system suitable for implementing compression and decompression of data at high speed in solid state storage. In some examples,computer system 500 may implement computer programs, applications, methods, processes, or other software or hardware to perform the above-described techniques.Computer system 500 includes abus 502 or other communication mechanism for communicating information, which interconnects subsystems and devices, such asprocessor 504, system memory 506 (e.g., RAM), storage device 508 (e.g., Flash, ROM, etc.), disk drive 510 (e.g., magnetic or optical), communication interface 512 (e.g., modem or Ethernet card), display 514 (e.g., CRT or LCD), input device 516 (e.g., keyboard), and cursor control 518 (e.g., mouse or trackball). - According to some examples,
computer system 500 performs specific operations byprocessor 504 executing one or more sequences of one or more instructions stored insystem memory 506. Such instructions may be read intosystem memory 506 from another computer readable medium, such asstatic storage device 508 ordisk drive 510. In some examples, hard-wired circuitry may be used in place of or in combination with software instructions for implementation. - The term “computer readable medium” refers to any tangible medium that participates in providing instructions to
processor 504 for execution. Such a medium may take many forms, including but not limited to, non-volatile media and volatile media. Non-volatile media includes, for example, optical or magnetic disks, such asdisk drive 510. Volatile media includes dynamic memory, such assystem memory 506. - Common forms of computer readable media includes, for example, floppy disk, flexible disk, hard disk, magnetic tape, any other magnetic medium, CD-ROM, any other optical medium, punch cards, paper tape, any other physical medium with patterns of holes, RAM, PROM, EPROM, FLASH-EPROM, any other memory chip or cartridge, or any other medium from which a computer can read.
- Instructions may further be transmitted or received using a transmission medium. The term “transmission medium” may include any tangible or intangible medium that is capable of storing, encoding or carrying instructions for execution by the machine, and includes digital or analog communications signals or other intangible medium to facilitate communication of such instructions. Transmission media includes coaxial cables, copper wire, and fiber optics, including wires that comprise
bus 502 for transmitting a computer data signal. - In some examples, execution of the sequences of instructions may be performed by a
single computer system 500. According to some examples, two ormore computer systems 500 coupled by communication link 520 (e.g., LAN, PSTN, or wireless network) may perform the sequence of instructions in coordination with one another.Computer system 500 may transmit and receive messages, data, and instructions, including program, i.e., application code, throughcommunication link 520 andcommunication interface 512. Received program code may be executed byprocessor 504 as it is received, and/or stored indisk drive 510, or other non-volatile storage for later execution. - Although the foregoing examples have been described in some detail for purposes of clarity of understanding, the invention is not limited to the details provided. There are many alternative ways of implementing the invention. The disclosed examples are illustrative and not restrictive.
Claims (20)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US15/912,352 US20180203797A1 (en) | 2010-10-22 | 2018-03-05 | Compression and Decompression of Data at High Speed in Solid State Storage |
Applications Claiming Priority (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
IN2942/MUM/2011 | 2010-10-22 | ||
IN2942MU2010 | 2010-10-22 | ||
PCT/IN2011/000725 WO2012053015A2 (en) | 2010-10-22 | 2011-10-20 | Compression and decompression of data at high speed in solid state storage |
US201313880728A | 2013-04-20 | 2013-04-20 | |
US15/912,352 US20180203797A1 (en) | 2010-10-22 | 2018-03-05 | Compression and Decompression of Data at High Speed in Solid State Storage |
Related Parent Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/IN2011/000725 Continuation WO2012053015A2 (en) | 2010-10-22 | 2011-10-20 | Compression and decompression of data at high speed in solid state storage |
US13/880,728 Continuation US9940230B2 (en) | 2010-10-22 | 2011-10-20 | Compression and decompression of data at high speed in solid state storage |
Publications (1)
Publication Number | Publication Date |
---|---|
US20180203797A1 true US20180203797A1 (en) | 2018-07-19 |
Family
ID=45464059
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US13/880,728 Active - Reinstated 2034-11-29 US9940230B2 (en) | 2010-10-22 | 2011-10-20 | Compression and decompression of data at high speed in solid state storage |
US15/912,352 Abandoned US20180203797A1 (en) | 2010-10-22 | 2018-03-05 | Compression and Decompression of Data at High Speed in Solid State Storage |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US13/880,728 Active - Reinstated 2034-11-29 US9940230B2 (en) | 2010-10-22 | 2011-10-20 | Compression and decompression of data at high speed in solid state storage |
Country Status (2)
Country | Link |
---|---|
US (2) | US9940230B2 (en) |
WO (1) | WO2012053015A2 (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2020139418A1 (en) * | 2018-12-26 | 2020-07-02 | Advanced Micro Devices, Inc. | Compressing data for storage in cache memories in a hierarchy of cache memories |
Families Citing this family (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2013048530A1 (en) * | 2011-10-01 | 2013-04-04 | Intel Corporation | Method and apparatus for high bandwidth dictionary compression technique using set update dictionary update policy |
US9571698B1 (en) * | 2012-03-30 | 2017-02-14 | EMC IP Holding Company LLC | Method and system for dynamic compression module selection |
US9843802B1 (en) | 2012-03-30 | 2017-12-12 | EMC IP Holding Company LLC | Method and system for dynamic compression module selection |
US9843702B1 (en) | 2012-03-30 | 2017-12-12 | EMC IP Holding Company LLC | Method and system for dynamic compression module selection |
US20150026509A1 (en) * | 2013-07-22 | 2015-01-22 | Kabushiki Kaisha Toshiba | Storage device having a data stream converter |
US20150067436A1 (en) * | 2013-09-03 | 2015-03-05 | Sandisk Technologies Inc. | Nonvolatile Memory System Compression |
KR102017807B1 (en) * | 2013-12-31 | 2019-09-03 | 에스케이하이닉스 주식회사 | Apparatus for processing data and method for processing data |
US9337862B2 (en) * | 2014-06-09 | 2016-05-10 | Tidal Systems, Inc. | VLSI efficient Huffman encoding apparatus and method |
TWI545576B (en) * | 2014-10-27 | 2016-08-11 | 群聯電子股份有限公司 | Data writing method, memory control circuit unit and memory storage apparatus |
TWI639159B (en) | 2015-05-28 | 2018-10-21 | 東芝記憶體股份有限公司 | Semiconductor device |
DE102015220205A1 (en) * | 2015-10-16 | 2017-04-20 | Continental Automotive Gmbh | Device and method for data transmission between a digital memory module and a microcontroller for a motor vehicle |
US10437667B2 (en) * | 2016-03-29 | 2019-10-08 | International Business Machines Corporation | Raid system performance enhancement using compressed data |
US10390114B2 (en) * | 2016-07-22 | 2019-08-20 | Intel Corporation | Memory sharing for physical accelerator resources in a data center |
US10027984B2 (en) | 2016-11-30 | 2018-07-17 | Hewlett Packard Enterprise Development Lp | Methods and systems for efficiently reading a data block from a data seglet with compressed data blocks |
US10404836B2 (en) * | 2016-12-26 | 2019-09-03 | Intel Corporation | Managing state data in a compression accelerator |
US10592170B2 (en) * | 2017-01-25 | 2020-03-17 | International Business Machines Corporation | Optimizing dual-layered compression in storage systems |
CN110784225A (en) * | 2018-07-31 | 2020-02-11 | 华为技术有限公司 | Data compression method, data decompression method, related device, electronic equipment and system |
US10831497B2 (en) | 2019-01-31 | 2020-11-10 | International Business Machines Corporation | Compression/decompression instruction specifying a history buffer to be used in the compression/decompression of data |
US10630312B1 (en) * | 2019-01-31 | 2020-04-21 | International Business Machines Corporation | General-purpose processor instruction to perform compression/decompression operations |
US11221778B1 (en) * | 2019-04-02 | 2022-01-11 | Pure Storage, Inc. | Preparing data for deduplication |
US11048413B2 (en) * | 2019-06-12 | 2021-06-29 | Samsung Electronics Co., Ltd. | Method for reducing read ports and accelerating decompression in memory systems |
US12009842B2 (en) * | 2019-12-20 | 2024-06-11 | Intel Corporation | Error correcting code circuitry compatible with multi-width interfaces |
US11243890B2 (en) * | 2020-01-14 | 2022-02-08 | EMC IP Holding Company LLC | Compressed data verification |
KR102267477B1 (en) | 2020-02-20 | 2021-06-22 | 삼성전자주식회사 | Storage device and operating method of the same |
US20220414014A1 (en) * | 2021-06-24 | 2022-12-29 | Intel Corporation | Technology for early abort of compression acceleration |
CN117472285A (en) * | 2023-12-26 | 2024-01-30 | 深圳市领德创科技有限公司 | Intelligent operation acceleration method for solid state disk, computer equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6310563B1 (en) * | 2000-05-12 | 2001-10-30 | International Business Machines Corporation | Method and apparatus for enhanced decompressor parsing |
US20100088464A1 (en) * | 2008-10-06 | 2010-04-08 | Xueshi Yang | Compression Based Wear Leveling for Non-Volatile Memory |
US20110107052A1 (en) * | 2009-10-30 | 2011-05-05 | Senthilkumar Narayanasamy | Virtual Disk Mapping |
US20120260009A1 (en) * | 2009-07-23 | 2012-10-11 | Stec, Inc. | Data storage system with compression/decompression |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5603001A (en) * | 1994-05-09 | 1997-02-11 | Kabushiki Kaisha Toshiba | Semiconductor disk system having a plurality of flash memories |
US5689255A (en) | 1995-08-22 | 1997-11-18 | Hewlett-Packard Company | Method and apparatus for compressing and decompressing image data |
US6604158B1 (en) * | 1999-03-11 | 2003-08-05 | Realtime Data, Llc | System and methods for accelerated data storage and retrieval |
US6903668B1 (en) * | 2003-11-18 | 2005-06-07 | M-Systems Flash Disk Pioneers Ltd. | Decompression accelerator for flash memory |
US8019959B2 (en) * | 2007-02-09 | 2011-09-13 | Marvell World Trade Ltd. | Nonvolatile memory system |
KR101513028B1 (en) * | 2007-07-02 | 2015-04-17 | 엘지전자 주식회사 | broadcasting receiver and method of processing broadcast signal |
WO2009061814A2 (en) * | 2007-11-05 | 2009-05-14 | University Of Florida Research Foundation, Inc. | Lossless data compression and real-time decompression |
US7728742B2 (en) * | 2008-09-30 | 2010-06-01 | Apple Inc. | Method and apparatus for compressing and decompressing data |
-
2011
- 2011-10-20 WO PCT/IN2011/000725 patent/WO2012053015A2/en active Application Filing
- 2011-10-20 US US13/880,728 patent/US9940230B2/en active Active - Reinstated
-
2018
- 2018-03-05 US US15/912,352 patent/US20180203797A1/en not_active Abandoned
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6310563B1 (en) * | 2000-05-12 | 2001-10-30 | International Business Machines Corporation | Method and apparatus for enhanced decompressor parsing |
US20100088464A1 (en) * | 2008-10-06 | 2010-04-08 | Xueshi Yang | Compression Based Wear Leveling for Non-Volatile Memory |
US20120260009A1 (en) * | 2009-07-23 | 2012-10-11 | Stec, Inc. | Data storage system with compression/decompression |
US20110107052A1 (en) * | 2009-10-30 | 2011-05-05 | Senthilkumar Narayanasamy | Virtual Disk Mapping |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2020139418A1 (en) * | 2018-12-26 | 2020-07-02 | Advanced Micro Devices, Inc. | Compressing data for storage in cache memories in a hierarchy of cache memories |
Also Published As
Publication number | Publication date |
---|---|
WO2012053015A3 (en) | 2012-10-04 |
WO2012053015A2 (en) | 2012-04-26 |
US9940230B2 (en) | 2018-04-10 |
US20130290615A1 (en) | 2013-10-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20180203797A1 (en) | Compression and Decompression of Data at High Speed in Solid State Storage | |
US11334429B2 (en) | Non-volatile memory apparatus and reading method thereof | |
KR100602394B1 (en) | Dual mode data compression for operating code | |
US10224959B2 (en) | Techniques for data compression verification | |
US8988257B2 (en) | Data compression utilizing variable and limited length codes | |
US8407378B2 (en) | High-speed inline data compression inline with an eight byte data path | |
KR101737451B1 (en) | Evaluating alternative encoding solutions during data compression | |
US8892809B2 (en) | Data compression and encoding in a memory system | |
US20130297575A1 (en) | Data management systems and methods using compression | |
JP6009676B2 (en) | Data compression device and data decompression device | |
US10048867B2 (en) | Method to shorten hash chains in lempel-ziv compression of data with repetitive symbols | |
KR20150020137A (en) | Variable-size flash translation layer | |
US8495464B2 (en) | Reliability support in memory systems without error correcting code support | |
US10303402B2 (en) | Data compression using partial statistics | |
CN114222973A (en) | Decompression engine for decompressing compressed input data comprising multiple data streams | |
US9998142B1 (en) | Techniques for invariant-reference compression | |
CN107402725B (en) | Nonvolatile memory device and data deduplication method thereof | |
JP7282066B2 (en) | Data compression device and data compression method | |
US11461008B2 (en) | Memory system for improving compression performance of a dictionary coder circuit | |
US11397546B2 (en) | Memory system | |
US20190109603A1 (en) | Accelerated Processing for Maximum Distance Separable Codes Using Composite Field Extensions | |
US11561738B2 (en) | Memory system | |
US20240022260A1 (en) | Low complexity optimal parallel huffman encoder and decoder | |
JP2023132713A (en) | Data expansion device, memory system, and data expansion method | |
CN117394865A (en) | Memory device, compression method of symbol stream and generation method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: INDRA NETWORKS INC., DELAWARE Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:SHAH, MONISH;REEL/FRAME:045234/0616 Effective date: 20180315 Owner name: CNEX LABS, INC., CALIFORNIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:INDRA NETWORKS INC.;REEL/FRAME:045234/0783 Effective date: 20180315 |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: NON FINAL ACTION MAILED |
|
STCB | Information on status: application discontinuation |
Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION |
|
AS | Assignment |
Owner name: POINT FINANCIAL, INC., ARIZONA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:CNEX LABS, INC.;REEL/FRAME:058951/0738 Effective date: 20220128 |