JP2014225297A - Flash memory module and storage device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve data capacity efficiency by overlapping or eliminating data to which different codes are allocated for the respective data.SOLUTION: A storage device includes a flash memory controller in which one or a plurality of flash memory modules is incorporated. Each of the flash memory modules comprises: at least one flash memory chip providing a storage region; and a controller controlling write/read of data including user data and a guarantee code annexed to the user data to/from the storage region of the flash memory chip. The controller divides each of a plurality of pieces of the data to which the user data is common into the user data and the guarantee code, stores one of the user data in a predetermined unit region of the storage region, and stores the respective guarantee codes annexed to a plurality of the user data in a predetermined unit region of the storage region by coupling the guarantee codes.

Description

  The present invention relates to a flash memory module and a storage apparatus, and is particularly suitable for application to a flash memory module, a storage apparatus, and a data control method that use a device whose rewrite is restricted as a storage medium.

  Conventionally, in a storage device, for example, a non-volatile storage medium capable of random access such as a magnetic disk or an optical disk is used as a data storage medium. In particular, recently, storage devices having a large number of small disk drives have become mainstream.

  Along with recent advancement of semiconductor technology, a batch erasable nonvolatile semiconductor memory has been developed. An example of such a nonvolatile semiconductor memory is a flash memory. A storage apparatus using a flash memory as a storage medium is considered to save power and have a high access time and the like compared to a storage apparatus having a large number of small disk drives.

  The above-described flash memory has a feature that data cannot be directly rewritten to a previously recorded data area. For this reason, when rewriting the recording data, it is necessary to erase the recording area after reading the recording data and write the update data to the erased unwritten area. However, the flash memory erase time is longer than the write time. Generally, when writing data, the old data is read once and then integrated with the write data to generate update data. A method is adopted in which the update data is written in another unwritten area and the original recording area is invalidated (a process for making the reference from the host device impossible). When the unwritten area is exhausted, the invalidated area is erased and a new unwritten area is generated.

  However, in the flash memory, there is a limit to the number of times data can be erased, and the data area where the number of times of erasure has increased due to concentrated data rewriting cannot be erased and becomes unusable. For this reason, it is necessary to prevent the data erasure process from being concentrated in a specific data area. In addition, the flash memory has a read error rate that increases with time even for a page that has been written once, and therefore, an operation called “refresh” is required to write a page that has passed for a fixed time to another page.

  In a flash memory having such characteristics, a deduplication technique is used in order to shorten the data erasure time or reduce the number of erasures (for example, Patent Document 1). In the deduplication technology, when the logical address provided to the host device and the physical address of the actual storage area are managed in association with each other, the data stores a plurality of logical address spaces in which the same data is stored. This is a technique for associating with one physical address space. By using the de-duplication technology, the amount of data written can be reduced, and in a storage device using the flash memory as a storage medium, the life of the flash memory can be extended and the performance stability can be improved.

JP 2009-87021 A

By the way, in the storage device, a guarantee code is assigned to the write data in order to prevent the data stored in the data area from being garbled due to hardware failure or the like, or writing to or reading from an illegal address due to an internal control error. The In general, this guarantee code has a combination of a part calculated from data contents for the purpose of providing redundancy to data and a part whose value can be freely set for each vendor. When deduplication is performed on data with such a guarantee code, the guarantee code part is different for each data, so even if the data content is the same, there is a difference in the guarantee code, so deduplication is performed. There was a problem that could not.
The present invention has been made in view of the above points, and proposes a flash memory module and a storage device capable of improving data capacity efficiency by deduplicating data with different guarantee codes for each data. It is something to try.

  In order to solve such a problem, the present invention includes a plurality of flash memory chips and a memory controller, and each flash memory chip of the plurality of flash memory chips includes a plurality of blocks, and each block includes a plurality of blocks. Each physical page is a data write / read unit, and the memory controller includes a first data set, a first guarantee code set, a second data set, and a second guarantee code set. The first data set includes a plurality of first user data, the first guarantee code set includes a plurality of first guarantee codes, and each first of the plurality of first guarantee codes. Is associated with one of the plurality of first user data, and the second data set is a plurality of second user data. The second warranty code set includes a plurality of second warranty codes, and each second warranty code of the plurality of second warranty codes is associated with one of the plurality of second user data. The memory controller stores the same data set in a first physical page when the first data set and the second data set are the same data set; and A flash memory module is provided, wherein the second guarantee code set is stored in a second physical page.

  According to such a configuration, among a plurality of data written to the flash memory chip, for a plurality of data having common user data, the user data and the guarantee code are separated, one user data is stored in the physical page, and a plurality of data is stored. A guarantee code corresponding to user data is concatenated and stored in a physical page. As a result, it is possible to extend the life of the flash memory by deduplicating the data to which the guarantee code is assigned and reducing the amount of data written to the storage medium. In addition, in storage media such as flash memory that employs the write-once type writing method, the unused area that is not used by the user is increased, and the unused area is used as a data write-on area. Can be improved.

  According to the present invention, it is possible to improve data capacity efficiency by deduplicating data to which a different guarantee code is assigned for each data.

It is a block diagram which shows the hardware constitutions of the computer system which concerns on the 1st Embodiment of this invention. 2 is a block diagram showing a configuration of a flash memory module according to the same embodiment. FIG. FIG. 3 is a block diagram showing a configuration of a flash memory chip according to the same embodiment. It is a conceptual diagram explaining the content of the data concerning the embodiment. It is a conceptual diagram explaining the content of the page concerning the embodiment. It is a conceptual diagram explaining the outline | summary of the deduplication process concerning the embodiment. It is a conceptual diagram explaining the logical / physical page address concerning the embodiment. It is a chart which shows an example of the logical / physical address conversion table concerning the embodiment. It is a chart which shows an example of the reverse reference table concerning the embodiment. It is a flowchart which shows the flow of the deduplication process concerning the embodiment. It is a flowchart which shows the flow of the reclamation process and refresh process concerning the embodiment. It is a conceptual diagram explaining the outline | summary of the merge process concerning the embodiment. It is a conceptual diagram explaining the outline | summary of the merge process concerning the embodiment. It is a flowchart which shows the flow of the merge process concerning the embodiment. It is a flowchart which shows the flow of the read process concerning the embodiment. It is a conceptual diagram explaining the outline | summary of the deduplication process which concerns on the 2nd Embodiment of this invention. It is a conceptual diagram explaining the logical / physical page address concerning the embodiment. It is a chart which shows an example of the logical / physical address conversion table concerning the embodiment. It is a chart which shows an example of the reverse reference table concerning the embodiment. It is a flowchart which shows the flow of the deduplication process concerning the embodiment. It is a flowchart which shows the flow of the reclamation process and refresh process concerning the embodiment. It is a flowchart which shows the flow of the read process concerning the embodiment. It is a timing chart which shows the flow of the deduplication process which concerns on the 3rd Embodiment of this invention.

  Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

(1) First Embodiment (1-1) Hardware Configuration of Computer System First, the hardware configuration of the computer system 1 according to the present embodiment will be described. As shown in FIG. 1, the computer system 1 according to the present embodiment includes a storage device 11 and first and second hosts 12A and 12B (hereinafter, the first and second hosts are simply referred to as host 12). A management terminal 13 and a SAN (Storage Area Network) 14.

  The storage apparatus 11 interprets the command transmitted from the host 12 and executes read / write into the storage area of the storage apparatus 11. As shown in FIG. 1, the storage apparatus 11 includes first and second host interfaces 112A and 112B (the first and second host interfaces may be simply referred to as a host interface 112), first and second. Storage interfaces 113A and 113B (hereinafter, the first and second storage interfaces may be simply referred to as storage interfaces 113) maintenance interface (I / F) 114, CPU 115, memory 116, connectors 16A, 16B, 16C, 16D (hereinafter simply referred to as connector 16), flash memory storage 17, and the like.

  The host interface 112 is an interface that receives data, control commands, and the like from the host 12. The storage interface 113 is an interface that transmits data and control commands to the flash memory storage 17. The maintenance interface (I / F) 114 is an interface that is connected to the management terminal 13 and receives control commands and the like related to management and maintenance from the management terminal 13.

  The CPU 115 functions as an arithmetic processing device, and has a function of controlling the operation of the entire storage device 11 in accordance with various programs and arithmetic parameters stored in the memory 16. Specifically, the CPU 115 processes data input / output (data I / O) from the host 12 and issues an I / O command related to data input / output to the flash memory storage 17. In addition, a guarantee code is generated for the user data received from the host 12 and assigned to the user data. The memory 116 temporarily stores control information, management information, and data from the user in addition to various programs and calculation parameters.

  The flash memory storage 17 is composed of flash memory modules 21A to 21F (hereinafter sometimes referred to as flash memory module 21). In the present embodiment, the flash memory storage 17 includes six flash memory modules 21 to 21F. However, the number of flash memory modules is not limited to this example. The flash memory module 21 is connected to the storage interface 113 via the connector 16, and data and control commands received from the host 12 via the connector 16 are written into the flash memory module 21. The internal configuration of the flash memory module 21 will be described in detail later. In this embodiment, the flash memory storage 17 is integrated with the storage device 11. However, the present invention is not limited to this example, and the flash memory storage 17 may be a separate device from the storage device 11.

  The host 12 is connected to the storage apparatus 11 via the SAN, and transmits a data read / write request to the storage apparatus 11. The host 100 is a computer device provided with information processing resources such as a CPU (Central Processing Unit) and a memory, and includes, for example, a personal computer, a workstation, a main frame, and the like. For the communication between the host 12 and the storage apparatus 11, for example, a block protocol such as SCSI (Small Computer System Interface) is used. The host 20 includes an information input device such as a keyboard, a switch, a pointing device, and a microphone, and an information output device such as a monitor display and a speaker.

  The management terminal 13 is a computer device provided with information processing resources such as a CPU and a memory, and includes, for example, a personal computer, a workstation, a main frame, and the like. The CPU functions as an arithmetic processing unit, and controls the operation of the management terminal 13 according to programs, arithmetic parameters, and the like stored in the memory. The management terminal 13 includes an information input device such as a keyboard, a switch, a pointing device, and a microphone, and an information output device such as a monitor display and a speaker, and manages the storage device 11 and the like according to an input from an operator or the like. It is a device to do.

  The number of hosts 12 and flash memory modules 21 and the connectors 16 that connect them are not limited to the numbers shown in FIG. 1, but may be one or more. Similarly, the standards for each connector and interface are not limited to specific ones. For example, FC (Fibre Channel), SCSI, iSCSI (Internet Small Computer System Interface), SAS (Serial Attached SCSI), and the like can be exemplified as typical standards applied to the SAN 14.

  Here, the flash memory (flash memory module 21) will be described. The flash memory has a feature that the minimum erase unit is larger than the minimum write unit. Specifically, in a flash memory, a plurality of pages are provided in one block, data is erased in units of blocks, and data is read and written in units of pages.

  As described above, the flash memory cannot directly rewrite data due to its characteristics. That is, in the flash memory, when the stored data is rewritten, the stored effective data is saved in another block and the stored data is erased in units of blocks. The data is rewritten by writing the data in the block from which the data has been erased.

  As described above, the data rewrite processing in the flash memory is accompanied by erasure of data in block units. However, the time required for erasing one block of data in the flash memory takes about one digit longer than the time required to write one page of data. Therefore, if data erasure for one block is executed every time in order to rewrite data for one page, the data rewriting performance of the flash memory is deteriorated. That is, when a flash memory is used as a storage medium, it is necessary to write data using an algorithm that can hide the time for erasing data from the flash memory.

  Normally, when data is rewritten to the flash memory, data is not erased every time data is rewritten, and data is rewritten by a method in which data is added to an unused area where no data is recorded. Do. However, if the data rewrite process is executed repeatedly, the unused area in the flash memory will be exhausted, so unnecessary data written in the flash memory can be erased and the storage area can be reused. Need to be in the correct state.

  Accordingly, a block reproduction process (hereinafter referred to as a reclamation process) in which only valid data in a block including old data is copied to an unused area, and the copy source block is erased to make it reusable. .) Is an essential process for high-speed data rewriting of flash memory. This reclamation process is executed for a block in which invalid data increases.

  On the other hand, the flash memory has a feature that the number of times of erasing data is limited. For example, an erase count of up to 100,000 times per block is guaranteed. In this case, if data rewrite concentrates on one block and the erase count of the block increases, there is a problem that the data cannot be erased and the block becomes unusable. Therefore, when the flash memory is used as a storage medium, it is necessary to perform the erase count leveling process so that the data erase process for a specific block is not concentrated.

  Further, the flash memory has a feature that the read error rate increases with time for a page once written. In this way, an error that occurs just by holding data is called a retention error. To avoid this, a process (refresh process) that writes a page that has passed a certain time after writing to another page is executed. The Even when the refresh process is executed, it is necessary to equalize the number of erasures described above and to consider the influence on the performance.

  As described above, when the flash memory is used as a storage medium, the logical address provided to the host device and the physical address of the actual storage area are used to conceal the data erasure time or to equalize the number of data erasures. And a process of changing from a logical address to a physical address is performed when data is written. In other words, by sequentially changing the correspondence between the logical address provided to the host device and the physical address on the storage area, the host device only needs to write data into the logical address space. It is no longer necessary to be aware of the physical address change associated with.

  In a storage apparatus having a flash memory, it is considered to apply a deduplication technique when managing logical addresses and physical addresses in association with each other as described above. The deduplication technique is a technique for associating a plurality of logical address spaces storing the same data with a single physical address space storing the data. The deduplication technology can reduce the amount of data and save the storage data capacity.

  In particular, format data is an example of data that is considered effective for applying the deduplication technique. When the storage device 11 uses the storage device, a format process for writing data of a specific pattern in the entire area of the storage device for the purpose of detecting an abnormal part of the storage device or writing a guarantee code in advance. Execute. This formatting process has a problem of unnecessarily increasing the number of times of rewriting in a storage device such as a flash memory where rewriting is restricted.

  In addition, since the flash memory employs a method of appending to an unused area for data rewriting, data erasure processing must be frequently executed when the unused area is reduced. For this reason, there is a problem in that the data that does not include the entity is written by the formatting process, and the unused area is reduced and the performance stability is impaired. Therefore, by applying the above-described deduplication technology, it is possible to reduce the amount of data written and to increase the life of the flash memory and improve the performance stability in the storage device 11 using the flash memory as a storage medium. .

  Further, in the storage device 11, a guarantee code is assigned to the write data in order to prevent data stored in the data area from being garbled due to a hardware failure or writing to or reading from an illegal address due to an internal control error. The In general, this guarantee code has a combination of a part calculated from data contents for the purpose of providing redundancy to data and a part whose value can be freely set for each vendor. When deduplication is performed on data with such a guarantee code, the guarantee code part is different for each data, so even if the data content is the same, there is a difference in the guarantee code, so deduplication is performed. There was a problem that could not.

  Therefore, in the present embodiment, when deduplication processing is performed on data including a guarantee code, the guarantee code portion and the data portion of the target data are separated, and only the data portion is subjected to deduplication processing. The guarantee code portion is managed separately from the data, and each piece of data is stored and managed in the flash memory. In flash memory, as described above, since the minimum writing unit is a page unit, not only the guarantee code part is simply separated and stored in the flash memory, but also the writing unit is efficiently considered. A guarantee code is stored or a change from a logical address to a physical address is performed.

  Further, for simple pattern data such as the same value or increment value represented by the format data described above, the amount of data written can be reduced by further compressing the data portion. In this embodiment, the compressed data portion and the guarantee code related to the data are efficiently managed so as to conform to the write restriction in the flash memory.

  As described above, according to the present embodiment, it is possible to extend the life of the flash memory by deduplicating the data to which the guarantee code is assigned and reducing the amount of data written to the storage medium. it can. In addition, in storage media such as flash memory that employs the write-once type writing method, the unused area that is not used by the user is increased, and the unused area is used as a data write-on area. Can be improved.

(1-2) Internal Configuration of Storage Device Next, the internal configuration of the storage device 11 will be described. First, details of the internal configuration of the flash memory module 21 will be described. The flash memory module 21 includes a flash control device 210 and flash memory chips 31A to 31H (hereinafter sometimes referred to as flash memory 31). In the present embodiment, the flash memory chip 31 is composed of six flash memories as shown in FIG. 2, but is not limited to this example and may be one or more.

  As shown in FIG. 2, the flash control device 210 includes a storage interface 221, a CPU 212, a controller 213, a memory 214, a buffer 215, a flash memory interface 216, and the like.

  The storage interface 221 is an interface that receives data and control commands from the CPU 115 of the storage apparatus 11 main body. The CPU 212 has a function of controlling the operation of the entire flash control device 210 and operates based on various programs stored in the memory 214. For example, the CPU 212 reads / writes data to / from the flash memory chip 31 with reference to various tables stored in the memory 214 for data I / O received from the host 12 via the storage interface 211. . Further, depending on the usage status of the flash memory chip 31, execution of reclamation for reproducing the used area, wear leveling for evenly distributing the wear of the storage medium due to data rewriting to each element of the storage medium, and the like are performed.

  The controller 213 reads / writes data to the flash memory chips 31A to 31H under the control of the CPU 212. Also, error correction code generation, data comparison for deduplication, hash value generation, and the like are performed on the data. In the present embodiment, the controller 213 executes processing such as generation of an error correction code. However, the present invention is not limited to this example, and the processing of the controller 213 may be executed by the CPU 212.

  The memory 214 stores various programs executed by the CPU 212 and the controller 213, various tables, and control information. The internal configuration of the flash control device 210 is not limited to the configuration shown in FIG. 2, and each function may be executed by one or more devices.

  Next, the internal configuration of the flash memory chip 31 will be described. As shown in FIG. 3, the flash memory chip 31 has page buffers 312A and 312B (hereinafter also referred to as page buffers 312) that temporarily store the target data of the data I / O command issued from the flash control device 21. The physical blocks 313A to 313F (hereinafter also referred to as physical blocks 313), which are actual storage areas.

  The physical block 313 is further composed of pages 32A to 32C (hereinafter also referred to as page 32). As described above, the physical block 313 is an erase unit in the flash memory, and the page 32 is a write / read unit in the flash memory. Further, the page 32 is an area divided into a size of 2 Kbytes, 3 Kbytes, 8 Kbytes, and the block 313 is an area divided into a size of 128 pages, 256 pages, and 1 Mbytes. It has a size of 2 Mbytes.

  Next, a guarantee code for user data will be described. As shown in FIG. 4A, a code (guarantee code) 42 given to data (user data) 41 is given to the data 41 at every predetermined address delimiter. Generally, an 8-byte guarantee code is generated and assigned for each sector and 512-byte user data. Hereinafter, this data delimitation may be referred to as a sector.

  The code 42 includes a bit error correction code unit 421 generated from the data 31 and a UDT unit 422 that can be freely set by a vendor. For example, the UDT unit 422 stores the address of the data for the purpose of confirming the consistency of the address of the target data. In this case, the UDT unit 422 stores a unique value regardless of the value of the user data 41.

  Since the page size of the flash memory is usually larger than 512 bytes, a plurality of sectors are stored in one page. At this time, the page 32 has the configuration shown in FIG. 4B. That is, n sectors can be stored in the page 32 of the flash memory, and the data portions 41A to 41C of each sector are collectively used as the data portion 411 of the page 32. Further, the codes 42A to 42C of each sector are collectively used as a page code portion 421. The arrangement of the data and the guarantee code in the page is not limited to the configuration shown in FIG. 4B, and any configuration that can determine the position of the data and the guarantee code in each sector is acceptable. Further, for the sake of explanation, the code 42 is assumed to be other than the data 41 in the sector, but the present invention is not limited to this example, and the bit error correction code 421 generated from the user data is stored in the data part 41 as a part of the data. You may do it.

  Next, an overview of deduplication processing according to the present embodiment will be described. As shown in FIG. 5, a page group 43 including three pages: a page including a data portion 411A and a code portion 421A, a page including a data portion 411B and a code portion 421B, and a page including a data portion 411C and a code portion 421C. Can be stored in two pages, the page 44 including the data portion 441 and the page 45 including the code portion 451, by deduplication processing. That is, since the data stored in the data portions 411A, 411B, and 411C are data having the same content, the data portion is stored in the data portion 441 of the page 44. Since the guarantee codes stored in the code portions 421A, 421B, and 421C are guarantee codes having different contents, all the guarantee codes of the code portions 421A, 421B, and 421C are stored in the code portion 451 of the page 45.

  Then, each logical address of the page group 43 is associated with the physical addresses of the page 44 and the page 45. That is, for the data portion, the top of the logical address of the data portion 411A, 411B, 411C and the physical address of the page 44 are associated with each other, and for the code portion, each physical address of the page 45 and the physical address concerned Is associated with offset information from the head of the. As described above, when there are three or more pages including the data part and the code part and the data part has the same content, the page number can be reduced to 2 pages by deduplication, and the effect of reducing the number of pages can be obtained. it can. Note that free space 422 and 452 can be obtained in the page by deduplication, but the free space may be left unused or may be used for storing management information described later.

  Next, a specific example of deduplication will be described. FIG. 6A shows the correspondence between logical page addresses and physical page addresses. As shown in FIG. 6A, Addr 51 indicates a logical page address, and Data 52 indicates data stored in the logical page address space. Code 53 indicates a guarantee code stored in the logical page address space. Addr 54 indicates the physical page address associated with the logical page, and Data 55 indicates the content of the data stored in the physical page address space.

  The data 52 stored in the logical pages 511 to 514 indicated by the logical page addresses 0 to 3 stores the data “AAAA” having the same contents. In this case, the logical page addresses 0 to 3 of the logical pages 511 to 514 are associated with the physical page address 0 of the physical page 541 storing the data “AAAA”. However, since the physical page 541 is only user data that does not include the guarantee code, each logical page 511 to 514 is associated with the physical page address 1 of the physical page 542 that stores the guarantee code. Further, the position where the guarantee code corresponding to each logical page 511 to 514 is located from the head code of the physical page address 1 is associated as offset information.

  Next, a management table for realizing the association between the logical page and the physical page will be described. The management table is stored in the memory 214 of the flash memory module 21. In addition, the controller 213 of the flash memory module 21 executes deduplication processing to create various management tables. Under the control of the CPU 212 of the flash memory module 21, the controller 21 reads / writes data to / from the flash memory chip 31 with reference to the management table.

  The logical / physical address conversion table 56 manages the correspondence between logical addresses and physical addresses, and specifically manages in which physical page the data stored in the logical page address and the guarantee code exist. is there. As shown in FIG. 6B, the logical / physical address conversion table 56 includes a logical page address column 561, a physical page data address column 562, a physical page code address column 563, and a physical page code offset address column 564.

  The logical page address column 561 stores the start address of the logical page. The physical page data address column 562 stores the start address of the physical page in which the data portion corresponding to the address stored in the logical page address column 561 is stored. The physical page code address column 563 stores the start address of the physical page in which the code portion corresponding to the address stored in the logical page address column 561 is stored.

  The physical page code offset address 564 stores an address indicating the storage position of the guarantee code corresponding to the guarantee code of each logical page among the codes in the page indicated by the address stored in the physical page code address field 563. The As described above, the logical / physical address conversion table 56 can obtain the position where the physical page data and the guarantee code corresponding to the data and the guarantee code included in the logical page are stored.

  In FIG. 6B, the offset address of the guarantee code is stored in the physical page code offset address field 564, but the physical position of the guarantee code may be recognized without providing the physical page code offset address field 564. . That is, by providing a restriction that the address group of the page group to be de-duplicated is continuous, the physical position of the guarantee code can be recognized without a physical page code offset address. A page that is not subject to de-duplication can be de-duplicated by providing a flag for the page or by setting the physical page data address and physical page code address to the same value. It may be determined whether or not it is.

  Next, the reverse reference table 57 will be described. The reverse reference table 57 is a table that manages to which logical page a physical page is referenced. As shown in FIG. 6C, the reverse reference table 57 includes a physical page address column 571, a reference count column 572, and a logical page address column 573. Composed.

  The physical page address column 571 stores the start address of a physical page in which data and a guarantee code are stored. The reference count column 572 stores the number of logical pages that refer to data stored in the physical page. Therefore, when the logical page that referred to the physical page is updated and is no longer referred to, the reference count is decremented. The logical page address column 573 stores a logical page address as a reference source. When one physical page is referred to by a plurality of logical pages, a plurality of logical page addresses as reference sources are stored in the plurality of logical page address fields 573.

  For example, when a reclamation process is executed, the reverse reference table 57 can determine whether or not a physical page is referred to. From the logical page address 573 serving as a reference source, the logical page to be rearranged can be determined. The page address can be acquired.

  As described above, by storing the above management table in the memory 214 of the flash control device 210, it is possible to speed up processing such as data writing / reading. May be stored in the flash memory 31 when the capacity of the memory 241 is large. In this case, for example, empty spaces 442 and 452 in the physical page may be used.

  Further, as described above, the creation and management of the management table is executed by the flash control device 210. However, the present invention is not limited to this example, and the page structure of the flash memory is recognized, and the deduplication processing and management of the write data are performed. An apparatus capable of holding information may be used. For example, the CPU 115 of the storage apparatus may create and manage a management table. However, in the flash control device 21, since the association between the logical address and the physical address is indispensable, there is no need to newly establish a management table, but the association between the logical address and the physical address is managed by the CPU 115 of the storage device. In this case, the above management table is newly established.

(1-3) Details of Operation of Storage Device (1-3-1) Deduplication Processing Next, deduplication processing in the storage device 11 will be described. As described above, the controller 213 of the flash memory module 21 executes deduplication processing. As shown in FIG. 7, first, the controller 213 acquires a logical page list to be deduplicated (S11). The logical page list is composed of two or more pages in which the contents of the data part storing the user data in the page are the same and the contents of the code part storing the guarantee code are different.

  As a search method for the logical page, for example, when writing large-size data, or when relocating data such as reclamation processing or refresh processing, the logical page to be deduplicated from large-size data is searched. Search for combinations. Alternatively, the hash value of the data portion of each page may be calculated and managed, and the combination of pages that match the data portion may be dynamically searched based on the hash value.

  Subsequently, the controller 231 acquires a physical page storing the data portion of the logical page acquired in step S11 (S12). For example, the controller 231 may preferentially acquire a physical page with a low degree of wear from the viewpoint of extending the life of the flash memory. However, the physical page acquisition policy is not limited to this example, and is set by the administrator. The physical page may be acquired based on the set policy.

  Then, the controller 231 writes the user data stored in the data part of the logical page acquired in step S11 to the physical page acquired in step S12 (S13). However, the controller 231 does not necessarily have to write data in step S13. For example, data writing may be registered in a scheduler or the like.

  Next, the controller 231 acquires one page of the physical page that stores the code part of the logical page acquired in step S11 (S14). Here, the block of the physical page that stores the code part is different from the block of the physical page that is the write destination of the data part acquired in step S12. As a result, when data is read / written, the data part and the code part can be read separately, and the process for the data part and the process for the code part can be performed in parallel.

  Next, the controller 231 determines whether there is a required number of physical pages acquired in step S14 (S15). Specifically, the controller 231 determines whether the capacity of the guarantee code for the number of logical pages to be deduplicated is within the acquired physical page capacity. The necessary capacity for storing the guarantee code is the guarantee code size × the number of logical pages to be deduplicated. Further, the maximum number of guaranteed codes that can be stored in one physical page (maximum number of target pages) can be calculated by dividing the physical page size by the guaranteed code size. Therefore, the controller 231 determines whether there is a required number of physical pages by determining whether the number of target logical pages is smaller than the number obtained by multiplying the number of acquired physical pages by the maximum number of target pages.

  If it is determined in step S15 that the required number of physical pages is present, the controller 231 executes the process of step S16. On the other hand, when it is determined in step S15 that the required number of physical pages is not necessary, the controller 231 repeats the processes in steps S14 to S15.

  Subsequently, the controller 231 extracts and connects the code portions of the logical pages to be deduplicated (S16). Specifically, the controller 231 concatenates code portions extracted from each logical page so as to fit in the physical page size.

  Then, the controller 231 writes the guarantee code concatenated in step S16 into the physical page (S17). When the controller 231 has acquired a plurality of physical pages for writing the guarantee code, the controller 231 writes the guarantee code linked to each physical page.

  Subsequently, the controller 231 updates the logical / physical address conversion table 56 and the reverse reference table 57, which are management tables (S18). Specifically, with respect to the logical / physical address conversion table 56, the controller 231 sets the value of the physical page data address column 562 of the physical page address of the data storage destination for the entry corresponding to the logical page that is the object of deduplication. The value in the physical page code address field 563 is updated to the value of the physical page address of the guarantee code storage destination. In addition, the controller 231 stores the position of each guarantee code in the physical page that is the storage destination of the guarantee code in the physical page code offset address field 564.

  For the reverse reference table 57, the controller 231 updates the entry corresponding to the physical page associated with the logical page that is the deduplication target. That is, the value of the reference counter column 572 is decremented for the physical page that is no longer referred to by updating the address of the logical page that is the object of de-duplication. Then, the address of the logical page that is no longer referred to is deleted from the logical page address column 573.

  It should be noted that the logical / physical address conversion table 56 needs to be updated even in a normal write (write) process in which the deduplication process is not performed. That is, the address of the logical page is associated with the address of the physical page that is the new write destination, or the address of the logical page that is referenced from the entry of the physical page of the original storage destination is deleted.

(1-3-2) Reclamation Process and Refresh Process Next, details of the reclamation process and the refresh process will be described. As shown in FIG. 8, first, the controller 231 acquires a physical block list that is a target of reclamation or refresh (S21). For example, the controller 231 sets a physical block with many invalid pages as a target of reclamation processing, and sets a physical block that has not been accessed for a long time as a target of refresh processing.

  Subsequently, the controller 231 determines whether or not processing has been executed for all physical blocks included in the physical block list acquired in step S21 (S220). If it is determined in step S220 that processing has been executed for all physical blocks, the controller 231 executes processing subsequent to step S221. On the other hand, if it is determined in step S220 that the process has not been executed for all physical blocks, the controller 231 executes the processes in and after step S22.

  Subsequently, the controller 231 determines whether or not there is a physical page for which all physical pages in the physical block have a value greater than 0 in the reference count column 572 of the reverse reference table 57 (S22). If it is determined in step S22 that there is a physical page whose value in the reference count field 572 is greater than 0, the processes in and after step S23 are executed, and there is no physical page whose value in the reference count field 572 is greater than 0. If it is determined, the process after step S221 is executed.

  Subsequently, the controller 231 acquires one physical page X having a value of 0 or more in the reference count column 572 among the target physical pages (S23). Hereinafter, the physical page acquired in step S23 will be described as a physical page X. Then, the controller 231 acquires an entry corresponding to the physical page X of the reverse reference table 57 (S24). Further, the controller 231 acquires the address of the logical page referring to the physical page X from the reference source logical page address column 573 (S25). Thereafter, the controller 231 investigates the reference source logical page registered in the reference source logical page address field 573.

  Then, the controller 231 investigates whether or not the logical page acquired in step S25 still refers to the target physical page X (S26). It should be noted that the logical page that no longer refers to the physical page from the reference source logical page address column 573 of the dereference table 57 in the deduplication process or the management table update process (step S18) executed during the normal write process. If the address has been deleted, the process of step S26 is not necessary. On the other hand, if the reverse reference table 57 is stored in the flash memory and cannot be easily updated, it is necessary to check whether or not the physical page X is referenced in step S26.

  Next, the controller 231 adds a logical page referring to the physical page X to the copy target list (S27). The controller 231 finally sets the logical page referring to the physical page X to 0 so that the physical block can be erased. In step S27, the controller 231 only adds the logical page to be copied to the list, and the actual copying process of the logical page may be performed later on an arbitrary occasion.

  Subsequently, the controller 231 checks whether or not there is a reference to all logical pages that may refer to the physical page X (S28). In step S28, if there is a logical page whose physical page X has not been checked, the controller 231 repeats the processing from step S26. On the other hand, if it is determined in step S28 that the physical page X has been checked for all logical pages, the process returns to step S220.

  Then, the controller 231 tries deduplication processing for the logical page included in the copy target list created in step S27 (S221). Specifically, the controller 231 executes deduplication processing on logical pages that have the same data part among the logical pages included in the copy target list.

  Subsequently, after executing deduplication processing in step S221, the controller 231 copies the logical page remaining in the copy target list to the physical page that is the copy destination (S222). Then, the controller 231 erases the physical block that is the target of the reclamation process or the refresh process in step S21 (S223).

(1-3-3) Merge Process Next, the merge process will be described. Here, the merge process means a process of integrating logical pages or physical pages having the same data with respect to logical pages that have already undergone deduplication processing. The deduplication processing described above is executed when writing large-size data as described above, or during reclamation processing or refresh processing. Therefore, there may be a logical page for which deduplication processing has not yet been executed in the same flash memory module 21. In addition, when there are a plurality of flash memory modules 21 in the storage device 11, deduplication processing is executed in each flash memory module 21, so that there are logical pages including the same data in the plurality of flash memory modules. There is a case. In this case, by executing the merge process, it is possible to integrate the same data and further reduce the data capacity.

  For example, as shown in FIG. 9A, it is assumed that there are two sets of physical pages in which deduplicated logical pages are stored. It is assumed that the same data is stored in the two physical pages 61A and 61B storing the data part. Further, the two physical pages 62A and 62B storing the guarantee code respectively store the guarantee codes of the logical page list. For the two sets of these physical pages, the data portion can be further shared, and the four physical pages 61A, 61B, 62A, and 62B can be combined into three physical pages 61C, 62C, and 62D.

  As shown in FIG. 9A, when four physical pages are grouped into three physical pages, the physical pages 62A and 62B in which the guarantee code is stored are used as they are to be the physical pages 62C and 62D, and the physical in which the data is stored. One of the pages 61A and 61B is used as 61C, and the other is set as an invalid page. However, when management information such as the logical / physical address conversion table 56 is stored in the free area 442 of the physical page, it is necessary to newly allocate a physical page for data storage. When a new physical page is allocated for data storage, both the physical pages 61A and 61B storing data are invalid pages.

  Further, if there is an empty area in the physical page in which the guarantee code portion is stored in the set of physical pages in which the deduplicated logical page is stored, the guarantee code can also be merged. For example, as shown in FIG. 9B, the guarantee codes stored in the physical page 64A and the physical page 64B are extracted and copied together on the physical page 64C. Then, the physical page 64A and the physical page 64B become invalid pages and are discarded. Further, since the same data is stored for the physical pages 63A and 63B in which data is stored, one of the physical pages 63A and 63B is a physical page 63C and the other is the same as in FIG. 9A. Invalid page. Thus, by integrating not only the data part but also the code part, four physical pages can be combined into two physical pages.

  Next, the processing procedure of the merge process will be described. As shown in FIG. 10, the controller 231 first acquires a physical page list to be merged (S31). Specifically, the controller 231 acquires a physical page list that is a valid page in the physical page list, is a page that has been subjected to deduplication processing, and is a set of physical pages having the same data part. The physical page list is created, for example, by searching physical page data when writing large size data or processing large size data triggered by reclamation processing or refresh processing. The Alternatively, the hash value of the data portion of each physical page may be calculated and managed, and the physical page list may be created by dynamically searching for a combination of pages that match the data portion based on the hash value.

  Then, the controller 231 determines whether or not processing has been executed for all physical page lists acquired in step S31 (S320). If it is determined in step S320 that processing has been executed for all physical pages, the controller 231 executes processing in step S321 and subsequent steps. On the other hand, if it is determined in step S320 that the processing has not been executed for all physical blocks, the controller 231 executes the processing after step S32.

  Subsequently, the controller 231 determines whether there is a physical page having a value greater than 0 in the reference count column 572 of the reverse reference table 57 among the physical pages acquired in step S31 (S32). If it is determined in step S32 that there is a physical page whose value in the reference count field 572 is greater than 0, the processing after step S33 is executed, and there is no physical page whose value in the reference count field 572 is greater than 0. If it is determined, the processing after step S321 is executed.

  Subsequently, the controller 231 acquires one physical page X having a value of 0 or more in the reference count column 572 among the target physical pages (S33). Hereinafter, the physical page acquired in step S33 will be described as a physical page X. Then, the controller 231 acquires an entry corresponding to the physical page X of the reverse reference table 57 (S34). Furthermore, the controller 231 acquires the address of the logical page referring to the physical page X from the reference source logical page address column 573 (S35). Thereafter, the controller 231 investigates the reference source logical page registered in the reference source logical page address field 573.

  Then, the controller 231 investigates whether or not the logical page acquired in step S35 still refers to the target physical page X (S36). It should be noted that the logical page that no longer refers to the physical page from the reference source logical page address column 573 of the dereference table 57 in the deduplication process or the management table update process (step S18) executed during the normal write process. If the address has been deleted, the process of step S36 is not necessary. On the other hand, if the reverse reference table 57 is stored in the flash memory and cannot be easily updated, it is necessary to check whether or not the physical page X is referenced in step S36.

  Next, the controller 231 adds the logical page referring to the physical page X to the list to be merged (S37). Then, the controller 231 checks whether or not there is a reference to all logical pages that may refer to the physical page X (S38). In step S38, if there is a logical page whose physical page X has not been checked, the controller 231 repeats the processing from step S36. On the other hand, if it is determined in step S38 that the physical page X has been referenced for all logical pages, the process returns to step S320.

  Then, the controller 231 acquires a physical page for storing the guarantee code stored in the logical page included in the merge target list created in step S37 (S321). Specifically, the controller 231 acquires the number of physical pages corresponding to the number of logical pages to be processed.

  Then, the controller 231 merges the guarantee codes of the logical pages to be merged, and writes the guarantee code on the physical page acquired in step S321.

  Then, the controller 231 updates the logical / physical address conversion table 56 and the reverse reference table 57, which are management tables (S323). Specifically, with respect to the logical / physical address conversion table 56, the controller 231 uses the value stored in the physical page data address column 562 for the entry corresponding to the logical page to be merged, The page address value is updated, and the value in the physical page code address field 563 is updated to the physical page address value of the guarantee code storage destination. In addition, the controller 231 stores the position of each guarantee code in the physical page that is the storage destination of the guarantee code in the physical page code offset address field 564.

  For the reverse reference table 57, the controller 231 updates the entry corresponding to the physical page associated with the logical page that is the duplication target. That is, the address of the logical page that is no longer referenced from the logical page address list column 573 is decremented by decrementing the value of the reference counter column 572 with respect to the physical page that is no longer referred to due to the update of the address corresponding to the logical page. Is deleted.

  It should be noted that the logical / physical address conversion table 56 needs to be updated even in a normal write (write) process in which the deduplication process is not performed. That is, the address of the logical page is associated with the address of the physical page that is the new write destination, or the address of the logical page that is referenced from the entry of the physical page of the original storage destination is deleted.

(1-3-4) Read Processing Next, the read processing procedure will be described. As shown in FIG. 11, the controller 231 acquires a logical page address to be read (S41). Specifically, the controller 231 acquires a logical command address to be read at the time of copying a read command from the host 12 or internal control such as reclamation processing or refresh processing.

  Next, the controller 231 acquires the target logical / physical address information from the logical / physical address conversion table 56 (S42). Subsequently, the controller 231 determines whether the logical page acquired in step S41 is a page subjected to deduplication processing (S43). Specifically, the controller 231 determines that the address stored in the physical page data address column 562 of the logical / physical address conversion table 56 and the address stored in the physical page code address column 563 are the same address. Alternatively, it may be determined that the logical page is not deduplicated.

  If it is determined in step S43 that the logical page is a page from which deduplication has been performed, the controller 231 executes processing subsequent to step S44. On the other hand, if it is determined in step S43 that the logical page is a page that has not been deduplicated, the controller 231 executes the process of step S47.

  If the controller 231 determines that the logical page is a page that has been deduplicated, the controller 231 acquires a physical page data address that stores the data portion of the logical page (S44). Subsequently, the controller 231 acquires a physical page code address in which the guarantee code of the logical page is stored (S45). Further, the controller 231 acquires a physical page code offset address indicating the physical position of the guarantee code of the logical page (S46).

  If it is determined in step S43 that the logical page is a normal page that has not been deduplicated, the controller 231 acquires the address of the physical page corresponding to the logical page (S47). Specifically, the controller 231 acquires the physical page data address stored in the physical page data address column 562 of the logical / physical address conversion table 56.

  Then, the controller 231 reads data stored in the physical page based on the physical page address associated with the logical page to be read (S48).

(1-4) Effects of this Embodiment As described above, in the computer system 1 according to this embodiment, among a plurality of data written to the flash memory chip, a plurality of data having a common data portion including user data. The data and the guarantee code are separated, one data is stored in the physical page, and the guarantee codes corresponding to a plurality of data are concatenated and stored in the physical page. As a result, it is possible to extend the life of the flash memory by deduplicating the data to which the guarantee code is assigned and reducing the amount of data written to the storage medium. In addition, in storage media such as flash memory that employs the write-once type writing method, the unused area that is not used by the user is increased, and the unused area is used as a data write-on area. Can be improved.

(2) Second Embodiment (2-1) Hardware Configuration of Computer System The hardware configuration of the computer system 2 according to the present embodiment is the hardware configuration of the computer system 1 according to the first embodiment. Therefore, detailed description is omitted. This embodiment is common to the first embodiment in that deduplication of a plurality of logical pages is performed, but further differs from the first embodiment in that the data portion to be deduplicated is compressed. Yes. Also in this embodiment, the search method for the logical page to be deduplicated is, for example, when writing large data or when relocating data such as reclamation processing or refresh processing. Search for a combination of logical pages to be deduplicated. Alternatively, the hash value of the data portion of each page may be calculated and managed, and the combination of pages that match the data portion may be dynamically searched based on the hash value. In the present embodiment, in order to facilitate the description, the following description will be made with respect to logical pages that are divided at regular intervals and have continuous addresses.

(2-2) Internal Configuration of Storage Device With respect to the internal configuration of the storage device 11 according to this embodiment, a detailed description of the same configuration as that of the first embodiment will be omitted, and the first embodiment will be omitted. A configuration different from the embodiment will be described in detail. First, the outline of deduplication in the present embodiment will be described with respect to differences from the first embodiment.

  As shown in FIG. 12, in the first embodiment, a page group 43 including data portions 411A, 411B, and 411C in which data of the same content is stored includes a page 44 including a data portion 441 and a code portion 451. It is stored in two pages of page 45. In the present embodiment, the data “Data 0 to Data n−1” of the page 44 is further compressed and converted into compressed data (Comp Data) 72.

  Here, the compressed data is composed of format data, for example. As described above, the format data is data of a specific pattern in the entire area of the storage device for the purpose of detecting an abnormal part of the storage device or writing a guarantee code in advance when using the storage device. Is written. However, in a storage device such as a flash memory in which rewriting is restricted, there is a problem that the number of rewrites is unnecessarily increased by writing the format data.

  Therefore, in this embodiment, as the format data, normally, data without an entity such as “0” is written, so the page group 43 in which “0” is written in the data portions 411A, 411B, and 411C is used as the page. Deduplication 44 and 45 are performed. Further, the data portion in which format data such as “0” is written is compressed and collected in the page 71.

  Also, as shown in FIG. 12, the guarantee code “C 0 to C n−1” is stored in the empty area of the page 71 in which the compressed data 72 is stored. As a result, the two pages of page 44 and page 45 are combined into one page of page 71. Note that all of the page 71 may be used to store the compressed data 72 and the guarantee code 73, or management information such as a management table may be stored in a free area (reserve).

  Next, a specific example of deduplication in this embodiment will be described. FIG. 13A shows the correspondence between logical page addresses and physical page addresses. As shown in FIG. 13A, Addr 81 indicates a logical page address, and Data 82 indicates data stored in the logical page address space. Code 83 indicates a guarantee code stored in the logical page address space. Addr 84 indicates a physical page address associated with the logical page, and Data 85 indicates the content of data stored in the physical page address space.

  Data “AAAA” having the same content is stored in the data 82 stored in the logical pages 811 to 813 indicated by the logical page addresses 0 to 2. In this case, the logical page addresses 0 to 2 of the logical pages 811 to 813 are associated with the physical page address 0 of the physical page 841. The physical page 841 stores data “A” obtained by compressing the data “AAAA” and guarantee codes C0 to C2 corresponding to the logical pages 811 to 813 serving as reference sources. On the other hand, a logical page 814 that is not deduplicated is associated with a physical page 842, and data and a guarantee code are stored in the physical page 842 as usual.

  Next, a management table for realizing the association between the logical page and the physical page will be described. The logical / physical address conversion table 86 is a table for managing the association between logical addresses and physical addresses, and includes a logical page address column 861, a physical page address column 862, and a flag column 863 as shown in FIG. 13B. Is done.

  The logical page address column 861 stores the start address of the logical page. The physical page address column 862 stores the start address of the physical page associated with the logical page address column 861. The flag column 863 stores a flag for determining whether the corresponding logical page is deduplicated.

  For example, when “1” is stored in the flag column 863, it is determined that the corresponding logical page is a deduplicated page, the compressed data is decompressed, or the position of the guarantee code in the physical page is It is calculated. The position of the guarantee code in the physical page is calculated from, for example, an offset position from the logical page address serving as a reference source. By setting a flag for determining whether or not the logical page address corresponding to the flag field 863 is compressed, it is not necessary to hold the actual position of the physical page in which the guarantee code is stored.

  Furthermore, in order to reduce the amount of data stored in the physical page, a flag for determining whether or not deduplication is performed in units of a plurality of arbitrary logical pages instead of in units of logical pages may be set. For example, if the address sequence of the logical page group to be duplicated is continuous and the data is guaranteed to fit within a certain size after the data is compressed, for example, if it is format data Assume that the number of logical pages in the deduplicated physical page is fixedly determined in advance. In this case, the table size of the logical / physical address conversion table 86 can be reduced by setting a flag for determining whether or not deduplication is performed in units of a plurality of logical pages. However, in this case, a flag for deduplication is set for each logical page in an arbitrary range, so that logical pages other than one logical page among the logical pages in a specific range are deduplicated. Requires separate information for logical pages that are not deduplicated.

  The reverse reference table 87 is a table for managing which logical page the physical page is referred to. As shown in FIG. 13C, the physical page address column 871, the reference count column 872, and the logical page representative address. It consists of a column 873.

  The physical page address column 871 stores the start address of a physical page in which data and a guarantee code are stored. The reference count column 872 stores the number of logical pages referring to data stored in the physical page. The logical page representative address column 872 stores the representative address of the reference logical page. The representative address of the logical page is the address of the first logical page of the plurality of logical pages that are deduplicated. That is, the position of the guarantee code of an arbitrary logical page pointing to the physical page can be obtained from the offset from the representative address of the logical page.

  Further, in the reverse reference table 87, the information corresponding to each physical page is not stored, but the information corresponding to an arbitrary number of physical page units is stored, thereby reducing the amount of information stored in the reverse reference table 87. You may do it. For example, the reference count number may be managed collectively in units of physical blocks serving as erase units. In this case, the logical address and the physical address are converted based on the reference count number of the erasure unit and the representative address of the reference source logical page, and the reference source logical page is confirmed, so that each physical page unit The reference count number can be restored.

(2-3) Details of Operation of Storage Device (2-3-1) Deduplication Processing Next, deduplication processing in the storage device 11 according to this embodiment will be described. In the following, processing different from that of the first embodiment will be particularly described, and description of the same configuration as that of the first embodiment will be omitted. As shown in FIG. 14, first, the controller 213 acquires a logical page list to be deduplicated (S51). The logical page list is composed of at least three or more pages in which the contents of the data part in which the user data in the page is stored are the same, and the contents of the code part in which the guarantee code is stored are different.

  Next, the controller 213 determines whether or not the compressed data size of the data portion of the plurality of logical pages included in the logical page list acquired in step S51 is smaller than a predetermined threshold TH (S52). If the data size after compression is equal to or greater than the predetermined threshold value TH in step S52, the controller 213 performs normal write processing without performing deduplication processing (S58). If it is determined in step S52 that the compressed data size is smaller than the predetermined threshold value TH, the de-duplication processing after step S53 is executed.

  Here, the threshold value TH is set in advance in the management program or the like, and when the data size is smaller than the threshold value TH, the size of the data portion after compression of the logical page to be deduplicated and the size of the code portion are set. The total value is a value set so as to fit within the size of one physical page.

  If it is determined in step S52 that the data size after compression is smaller than TH, the controller 231 acquires a physical page storing the data portion of the logical page acquired in step S51 (S53).

  Then, the controller 231 compresses the data portion of the logical page to be deduplicated (S54). Then, the controller 231 extracts and connects the code portion of each logical page to be deduplicated (S55). Then, the data part compressed in step S54 and the code part connected in step S55 are connected and written to the physical page acquired in step S53 (S56).

  Subsequently, the controller 231 updates the logical / physical address conversion table 56 and the reverse reference table 57, which are management tables (S57). Specifically, with respect to the logical / physical address conversion table 86, the controller 231 sets the value of the physical page address field 862 to the value of the physical page address of the data storage destination for the entry corresponding to the logical page that is the object of deduplication. And “1” is stored in the deduplication flag column 863.

  For the reverse reference table 87, the controller 231 updates the entry corresponding to the physical page associated with the logical page that is the duplication target. That is, “1” is stored in the reference count column 872 of the reverse reference table 87, and further, the representative address of the target logical page is stored in the logical page representative address 873.

(2-3-2) Reclamation Process and Refresh Process The reclamation process and the refresh process according to the present embodiment are substantially the same as the reclamation process and the refresh process according to the first embodiment. This will be described in detail. In this embodiment, as shown in FIG. 15, the controller 231 obtains the representative address of the logical page referring to the physical page X after obtaining the physical page X to be processed in step S63 ( S64) is different from the first embodiment in the point.

  Then, using the representative address of the logical page acquired in step S64 as a starting point, in order to check the current reference destination of the logical page that may refer to the physical page X, the logical next to the representative address of the logical page is checked. An entry in the page logical / physical address conversion table 86 is acquired (S65). Then, it is checked whether the logical page acquired in step S65 still refers to the physical page X (S66). If the physical page X is referred to, the logical page is added to the copy target list. (S67).

(2-3-3) Read Processing The difference between the read processing according to the present embodiment and the read processing according to the first embodiment will be particularly described in detail. As shown in FIG. 16, the controller 231 acquires a logical page address to be read (S71).

  Next, the controller 231 acquires target logical / physical address information from the logical / physical address conversion table 86 (S72). Subsequently, the controller 231 determines whether the logical page acquired in step S41 is a page subjected to deduplication processing (S73).

  If it is determined in step S73 that the logical page is a page from which deduplication has been performed, the controller 231 executes processing in step S74 and subsequent steps. on the other hand. If it is determined in step S73 that the logical page is a page that has not been deduplicated, the controller 231 executes the process of step S731.

  If the controller 231 determines that the logical page is a page that has been deduplicated, the controller 231 acquires a physical page data address that stores the data portion of the logical page (S74). Subsequently, the controller 231 calculates the offset position of the guarantee code in the physical page from the logical page representative address of the reference source (S75). Then, the controller 231 reads the data stored in the physical page address associated with the logical page to be read (S76). Furthermore, the controller 231 decompresses the data acquired in step S76 (S77).

  On the other hand, if it is determined in step S73 that the logical page is a page that has not been deduplicated, the controller 231 acquires the physical page address storing the data part and the guarantee code part corresponding to the logical page. (S731). Then, the controller 231 reads the physical page stored in the physical page address acquired in step S731 (S732).

  If it is determined in step S73 that the logical page is a normal page that has not been deduplicated, the controller 231 acquires the address of the physical page corresponding to the logical page (S731). Specifically, the controller 231 acquires the physical page data address stored in the physical page data address column 562 of the logical / physical address conversion table 56.

  Then, the controller 231 reads the data stored in the physical page based on the physical page address associated with the logical page to be read (S732).

(2-4) Effects of this Embodiment As described above, in the computer system 2 according to this embodiment, in addition to the deduplication processing, the data portion to be deduplicated is further compressed. This makes it possible to extend the life of the flash memory by deduplicating the data to which the guarantee code is assigned and further compressing the data portion to reduce the amount of data written to the storage medium. it can. In addition, in storage media such as flash memory that employs the write-once type writing method, the unused area that is not used by the user is increased, and the unused area is used as a data write-on area. Can be improved.

(3) Third Embodiment (3-1) Hardware Configuration of Computer System The hardware configuration of the computer system 3 according to the present embodiment is the flash memory storage of the computer system 1 according to the first embodiment. 17 differs from the storage apparatus 11 in that it is configured as a separate apparatus. This embodiment is common to the first embodiment in that deduplication of a plurality of logical pages is executed, but differs in that deduplication is executed in cooperation with the storage apparatus 11 and the flash memory storage 17. ing. Specifically, the storage device 11 does not actually transmit data to the flash memory storage 17, but transmits data by sending a specific pattern for format data or the like written to the storage area in response to a request from the host 12. The transfer amount is reduced.

(3-2) Details of Operation of Computer System Next, deduplication processing executed between the storage apparatus 11 and the flash memory storage 17 will be described.

  As shown in FIG. 17, first, the storage apparatus 11 starts formatting the storage area in response to a request from the host 12 (S81). The storage device 11 formats the storage area by writing a specific pattern for each sector in all the target storage areas. The storage apparatus 11 also assigns a guarantee code when writing the specific pattern.

  Then, the storage apparatus 11 notifies the controller 213 of the format pattern generated in response to the request from the host 12 and the generation rule of the guarantee code to be added to the format data as pattern information (S82). Here, the pattern information is, for example, a fixed pattern such as all zeros, and the guarantee code is, for example, a code in a format to which logical address information is added in addition to an error correction code generated from the data part. It is. That is, the storage apparatus 11 does not actually write the format information in the storage area, but notifies the pattern information including the format pattern and the guarantee code generation rule. Thereby, the data transfer amount can be reduced.

  Then, the flash memory storage 17 receives the pattern information transmitted from the storage device 11 in step S82 (S83). Then, the flash memory storage 17 generates a guarantee code for each logical page corresponding to the storage area to be formatted based on the pattern information transmitted from the storage device 11 (S84). Subsequently, the flash memory storage 17 generates a format pattern based on the pattern information (85S).

  Then, the flash memory storage 17 performs deduplication processing on the logical page generated in step S85 (S86). The de-duplication process in step S86 is the same process as in the first embodiment, and a detailed description thereof will be omitted.

  Then, the flash memory storage 17 updates the target entry of the logical / physical address conversion management table 56 for the deduplicated logical table (S87). The update process of the logical / physical address conversion management table 56 in step S87 is the same process as in the first embodiment, and thus detailed description thereof is omitted.

  Then, the flash memory storage 17 notifies the storage device 11 that the formatting process has been completed (S88). Then, the storage device 11 receives a notification from the flash memory storage 17 that the formatting process has been completed, and notifies the host 12 that the formatting process has been completed.

(3-3) Effects of this Embodiment According to the computer system 3 according to this embodiment, data and a guarantee code are written in the storage area based on specific pattern information. Thereby, the data transfer amount in data writing can be reduced. Further, the life of the flash memory can be extended by deduplicating the data to which the guarantee code has been assigned to reduce the amount of data written to the storage medium. In addition, in storage media such as flash memory that employs the write-once type writing method, the unused area that is not used by the user is increased, and the unused area is used as a data write-on area. Can be improved.

(4) Other Embodiments In the above-described embodiment, the deduplication processing or the like is executed based on various programs stored in the flash control device 210. However, the present invention is not limited to this example. For example, the flash control device 210 may be provided in a separate device from the storage device 11, and various functions may be realized in cooperation with the flash control device 210. Further, various programs stored in the flash control device 210 may be provided in another device separate from the storage device 11, and various functions may be realized by calling the program to the flash control device 210. Good.

  The present invention can be applied to a storage apparatus that can improve data capacity efficiency by deduplicating data to which different codes are assigned for each data.

11 Storage Device 112 Host Interface 113 Storage Interface 114 Maintenance Interface 115 CPU
116 Memory 12 Host 13 Management terminal 14 SAN
16 Connector 17 Flash memory storage device 21 Flash memory module 210 Flash control device 211 Storage interface 212 CPU
213 Controller 213 Memory 215 Buffer 216 Flash memory interface 31 Flash memory chip

Claims (16)

Multiple flash memory chips,
A memory controller;
With
Each flash memory chip of the plurality of flash memory chips includes a plurality of blocks, each block includes a plurality of physical pages, each physical page is a unit for writing / reading data,
The memory controller manages a first data set and a first guarantee code set, a second data set and a second guarantee code set,
The first data set includes a plurality of first user data, the first guarantee code set includes a plurality of first guarantee codes, and each first guarantee code of the plurality of first guarantee codes is , Associated with one of the plurality of first user data,
The second data set includes a plurality of second user data, the second guarantee code set includes a plurality of second guarantee codes, and each second guarantee code of the plurality of second guarantee codes is , Associated with one of the plurality of second user data,
The memory controller stores the same data set in a first physical page when the first data set and the second data set are the same data set, and the first guarantee code set and the second data set A flash memory module, wherein the second guarantee code set is stored in a second physical page. The memory controller is
Copying the first guarantee code and the second guarantee code to a new physical page when there is a free space in the physical page in which the first guarantee code or the second guarantee code is stored. The flash memory module according to claim 1, wherein the first guarantee code and the second guarantee code are merged into one physical page. Each guarantee code of the plurality of first guarantee codes and the plurality of second guarantee codes is:
The flash memory module according to claim 1, comprising a bit error correction code code generated from the user data and an address of the user data. Each guarantee code of the plurality of first guarantee codes and the plurality of second guarantee codes is a different guarantee code.
The flash memory module according to claim 1, wherein: The position information of the first guarantee code and the second guarantee code stored in the second physical page is held as offset information indicating the position from the head address of the second physical page. The flash memory module according to claim 1, wherein: The offset information is stored in a logical / physical address conversion table.
The flash memory module according to claim 5, wherein: The flash memory module according to claim 1, wherein the first physical page and the second physical page are included in different blocks. The memory controller is
Get the logical page address to be read
Get the target logical / physical address information from the logical / physical address conversion table,
Determine whether the acquired logical page is a logical page that has been deduplicated,
If the logical page is a logical page that is not deduplicated, obtain the address of the physical page corresponding to the logical page;
If the logical page is a deduplicated logical page,
Obtaining a physical page data address of a physical page in which the data portion of the logical page is stored;
Obtaining a physical page code address of a physical page in which the guarantee code of the logical page is stored;
Obtaining a physical page code offset address indicating the position of the guarantee code in the physical page;
The flash memory module according to claim 1, wherein data stored in a physical page is read based on the physical page address. A storage controller;
Multiple flash memory modules;
With
The flash memory module includes a plurality of flash memory chips and a memory controller,
Each flash memory chip of the plurality of flash memory chips includes a plurality of blocks, each block includes a plurality of physical pages, each physical page is a unit for writing / reading data,
The memory controller manages a first data set and a first guarantee code set, a second data set and a second guarantee code set,
The first data set includes a plurality of first user data, the first guarantee code set includes a plurality of first guarantee codes, and each first guarantee code of the plurality of first guarantee codes is , Associated with one of the plurality of first user data,
The second data set includes a plurality of second user data, the second guarantee code set includes a plurality of second guarantee codes, and each second guarantee code of the plurality of second guarantee codes is , Associated with one of the plurality of second user data,
The memory controller stores the same data set in a first physical page when the first data set and the second data set are the same data set, and the first guarantee code set and the second data set A storage device, wherein the second guarantee code set is stored in a second physical page. The memory controller is
Copying the first guarantee code and the second guarantee code to a new physical page when there is a free space in the physical page in which the first guarantee code or the second guarantee code is stored. The storage apparatus according to claim 9, wherein the first guarantee code and the second guarantee code are merged into one physical page. Each guarantee code of the plurality of first guarantee codes and the plurality of second guarantee codes is:
The storage apparatus according to claim 9, comprising a bit error correction code code generated from the user data and an address of the user data. Each guarantee code of the plurality of first guarantee codes and the plurality of second guarantee codes is a different guarantee code.
The storage apparatus according to claim 9, wherein: The position information of the first guarantee code and the second guarantee code stored in the second physical page is held as offset information indicating the position from the head address of the second physical page. The storage apparatus according to claim 9, wherein: The offset information is stored in a logical / physical address conversion table.
The storage apparatus according to claim 13, wherein the storage apparatus is a storage device. The storage apparatus according to claim 9, wherein the first physical page and the second physical page are included in different blocks. The memory controller is
Get the logical page address to be read
Get the target logical / physical address information from the logical / physical address conversion table,
Determine whether the acquired logical page is a logical page that has been deduplicated,
If the logical page is a logical page that is not deduplicated, obtain the address of the physical page corresponding to the logical page;
If the logical page is a deduplicated logical page,
Obtaining a physical page data address of a physical page in which the data portion of the logical page is stored;
Obtaining a physical page code address of a physical page in which the guarantee code of the logical page is stored;
Obtaining a physical page code offset address indicating the position of the guarantee code in the physical page;
The storage apparatus according to claim 9, wherein data stored in a physical page is read based on the physical page address.

Images