JP2018163450A - Data storage control device, data storage system, data storage method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily recover data for an error caused by retention. SOLUTION: A horizontal code generation unit 110 that generates a first error detection code between a plurality of storage media for data written in each of the divided areas in which the storage areas of the plurality of storage media are divided into predetermined areas. The vertical code generation unit 120 that generates a second error detection code for each of the plurality of storage media and the first horizontal code generation unit 110 that generate the second error detection code for the data written in the divided regions at different timings. It has a code writing unit 130 that writes each of the error detection code and the second error code generated by the vertical code generating unit 120 into the divided area in which the data is not written. [Selection diagram] Fig. 1

Description

  The present invention relates to a data storage control device, a data storage system, a data storage method, and a program.

  In recent years, SSD (Solid State Drive: using non-volatile media such as a silicon disk, flash memory, etc.) has become more popular than HDDs (Hard Disk Drive: magnetic disk devices) as the storage capacity increases. It is coming. Although HDD capacity is expected to increase in the future, it is predicted that the SSD unit price will be lower than the HDD unit price. It is expected to be used for archive for preservation. The use of these archives is characterized in that access characteristics are additionally written (write data is written in the form of being added to existing data without rewriting data), and the frequency of rewriting is low. In addition, for the purpose of archiving, reading is frequently performed immediately after writing, but as the time passes, the frequency of reading decreases rapidly, and the older the data stored, the less the reading is performed. Also have.

  Further, the flash memory used for the SSD has a large capacity. However, the principle of storing charge in a small capacitor is that flash memory has a characteristic that the amount of stored charge decreases as the density increases, and the charge is discharged when left for a long period of time, and data is easily lost. . This charge retention period is called retention time and has a value between several years and several decades. In order to re-accumulate the charge, the retention time is reset by overwriting the flash memory with the same data for the cell (unit of charge accumulation).

  Further, in a normal SSD usage method (where rewriting occurs), the level of cell writing, which is another drawback of flash memory, is called Leveling, which is another level of flash memory. There is a method of extending the number of times). Under the condition in which WearLeveling occurs, a cell that has been written and has elapsed time is preferentially rewritten in order to level the writing to the cell. Therefore, problems related to retention (data loss) are very unlikely to occur. However, as in the case of archiving, in most cases, the number of writes is already leveled (one time), so the WearLeveling algorithm does not work and the data stored in the old days is written. The cell is not updated, and there is a possibility that data in the cell may be lost.

  In conventional RAID and similar redundancy techniques, redundancy is performed with data between storage element units of each SSD or flash memory at substantially the same time. Therefore, there is a high possibility that data loss will occur almost at the same time. This means that with the existing RAID method, there is a possibility that data may be lost simultaneously in a redundant SSD or flash memory. Thereby, there is a high possibility that the data cannot be reproduced beyond the redundancy set by RAID or the like. This means that existing RAID cannot protect data against data loss due to retention such as SSD using flash memory. This does not make sense as a means of data protection for an application where access is hardly generated once written, such as archive data.

  Depending on the storage device, SSDs and flash memories (hereinafter referred to as storage media) are two-dimensionally arranged, and the arrangement of data on the storage media is determined based on the error status of the storage media and the write / erase status. And a method for taking measures to prevent data loss due to an error due to the life of the storage medium has been considered (for example, see Patent Documents 1 to 3). Further, a method is considered in which the storage medium is arranged three-dimensionally and the RAID configuration is tripled using parity or the like (see, for example, Non-Patent Document 1).

  An error detection method using horizontal parity bits for the same row and vertical parity bits for the same column in data of a plurality of rows and columns is conceivable (see, for example, Patent Document 4).

Patent No. 5848353 JP-A-2006-105320 Special table 2013-539132 gazette Japanese Patent Laid-Open No. 2002-246917

Satoshi Uehara, “Multiple Parity Orthogonal RAID on Large Virtual Disks”, IPSJ Research Report, Vol. 2011-DPS-149 No.4, November 24, 2011

  However, in the techniques described in Patent Documents 1 to 3 and Non-Patent Document 1, data protection using parity is performed at almost the same time, and data is protected against occurrence of errors due to retention. It's hard to say. Furthermore, although the probability of data loss is reduced by performing parity in a triple or triple manner, it is difficult to say that countermeasures against errors due to retention are taken.

  Further, the technique described in Patent Document 4 has a problem that data cannot be transmitted unless the horizontal parity bits and the vertical parity bits are aligned.

  As described above, the above-described technique has a problem that data recovery cannot be easily performed for an error caused by retention.

  An object of the present invention is to provide a data storage control device, a data storage system, a data storage method, and a program that solve the above-described problems.

The data storage control device of the present invention
A horizontal code generator for generating a first error detection code between the plurality of storage media for data written in each of the divided areas obtained by dividing the storage area of each of the plurality of storage media into a predetermined area;
A vertical code generation unit that generates a second error detection code for each of the plurality of storage media for data written at different timings to each of the divided regions;
Each of the first error detection code generated by the horizontal code generation unit and the second error code generated by the vertical code generation unit is divided into portions in which the data is not written. A code writing unit for writing in the area.

The data storage system of the present invention is
A plurality of storage media;
A horizontal code generator for generating a first error detection code between the plurality of storage media for data written in each of the divided areas obtained by dividing the storage area of each of the plurality of storage media into a predetermined area;
A vertical code generation unit that generates a second error detection code for each of the plurality of storage media for data written at different timings to each of the divided regions;
Each of the first error detection code generated by the horizontal code generation unit and the second error code generated by the vertical code generation unit is divided into portions in which the data is not written. A code writing unit for writing in the area.

Further, the data storage method of the present invention includes:
A process of generating a first error detection code between the plurality of storage media for data written in each of the divided areas obtained by dividing the storage area of each of the plurality of storage media into a predetermined area;
A process of writing the first error detection code in a divided area of the divided area where the data is not written;
A process of generating a second error detection code for each of the plurality of storage media for data written at different timings to each of the divided areas;
A process of writing the second error code in a divided area of the divided area where the data is not written is performed.

The program of the present invention is
A program for causing a computer to execute,
A procedure for generating a first error detection code between the plurality of storage media for data written in each of the divided areas obtained by dividing the storage area of each of the plurality of storage media into a predetermined area;
Writing the first error detection code in a divided area of the divided area where the data is not written;
A procedure for generating a second error detection code for each of the plurality of storage media for data written at different timings to each of the divided regions;
And a step of writing the second error code in a divided area of the divided area where the data is not written.

  As described above, in the present invention, data recovery can be easily performed for an error caused by retention.

It is a figure which shows 1st Embodiment of the data storage control apparatus of this invention. 3 is a flowchart for explaining a data storage method in the data storage control device shown in FIG. 1. It is a figure which shows 2nd Embodiment of the data storage control apparatus of this invention. It is a figure which shows an example of the format of the management table for this data storage system to manage the storage area of a storage medium. It is a figure which shows one form of the data storage control apparatus used for the data storage system shown in FIG. It is a figure for demonstrating the whole operation | movement in this form. 6 is a flowchart for explaining a data storage method in the data storage control device shown in FIG. 5. It is a figure for demonstrating an example of the process sequence when an error generate | occur | produces at the time of reading and writing of data.

Embodiments of the present invention will be described below with reference to the drawings.
(First embodiment)
FIG. 1 is a diagram showing a first embodiment of a data storage control device of the present invention. As shown in FIG. 1, the data storage control device 100 in this embodiment includes a horizontal code generation unit 110, a vertical code generation unit 120, and a code writing unit 130. FIG. 1 shows an example of main components related to the present embodiment among the components included in the data storage control device 100 according to the present embodiment.

  The horizontal code generation unit 110 generates a first error detection code between the plurality of storage media for data written in each of the divided areas obtained by dividing the storage areas of the plurality of storage media into predetermined areas.

  The vertical code generation unit 120 generates a second error detection code for each of a plurality of storage media, with respect to data written at different timings in each divided area.

  The code writing unit 130 writes the first error detection code generated by the horizontal code generation unit 110 in a divided region in which data is not written in the divided region. In addition, the code writing unit 130 writes the second error code generated by the vertical code generation unit 120 in a divided region where data is not written in the divided region.

  Hereinafter, a data storage method in the data storage control device 100 shown in FIG. 1 will be described. FIG. 2 is a flowchart for explaining a data storage method in data storage control device 100 shown in FIG.

  First, the horizontal code generation unit 110 generates a first error detection code between a plurality of storage media for data written in each of the divided areas obtained by dividing the storage areas of the plurality of storage media into predetermined areas. (Step S1). Then, the code writing unit 130 writes the first error detection code generated by the horizontal code generation unit 110 in a divided region in which data is not written in the divided region (step S2).

  In addition, the vertical code generation unit 120 generates a second error detection code for each of the plurality of storage media for data written at different timings in each divided region (step S3). Then, the code writing unit 130 writes the second error code generated by the vertical code generation unit 120 in a divided region in which data is not written in the divided region (step S4).

As described above, the data storage control device 100 generates error detection codes in the horizontal direction and the vertical direction of a plurality of storage media, and writes each generated error code in a divided area where data is not written. Therefore, data recovery can be easily performed for errors associated with retention.
(Second Embodiment)
FIG. 3 is a diagram showing a second embodiment of the data storage control device of the present invention. FIG. 3 shows an example of a data storage system provided with the data storage control device in this embodiment. The data storage system shown in FIG. 3 includes host request processing means 201, arrangement map search means 202, arrangement determination means 203, and a plurality of storage media 204-1 to 204-4. FIG. 3 shows an embodiment in which there are four storage media, but the number is not limited.

  The host request processing unit 201 issues a command for reading or writing from a host computer which is a host device.

  The arrangement map search unit 202 searches for information on the data area corresponding to the address (host address) indicated by using the command output from the host request processing unit 201.

  The arrangement determining unit 203 uses the information on the management table shown in FIG. 4 determined by the arrangement map searching unit 202 to write data in which storage area of the storage media 204-1 to 204-4. Make a decision. Further, the arrangement determining unit 203 writes data in the determined data area.

  The storage media 204-1 to 204-4 are storage media configured from silicon disks that store information.

  FIG. 4 is a diagram showing an example of the format of a management table for the data storage system to manage the storage areas of the storage media 204-1 to 204-4. As shown in FIG. 4, in the management table 300, a head host address 301 and an erased flag 302 are stored in association with each other. The management table 300 shown in FIG. 4 is a table on which the arrangement map search unit 202 searches.

  The head host address 301 indicates the head address of the host address specified when a read / write command is sent from the host computer. The erased flag 302 is a flag indicating whether or not the data area starting from the head host address 301 has been erased. Data storage areas on the storage media 204-1 to 204-4 having the start host address 301 as a start address are a plurality of divided areas obtained by dividing the storage areas of the storage media 204-1 to 204-4 into predetermined areas. It may be configured to have an erased flag 302 for each of the divided areas. It is conceivable that the erased flag 302 is represented by various representation methods such as a bitmap representation method.

  Furthermore, the top host addresses 301 in the management table 300 are arranged in ascending order, and have a fixed value (for example, 128 kbytes) area from a certain top host address 301.

  In addition, the erased data area managed by the management table 300 may take time until data is transferred from the host after the device is initialized and the data is actually written. Therefore, as described above, there is a possibility that the contents of data may change due to retention. For this reason, when parity data is generated using actual data (data written to the storage media 204-1 to 204-4) stored in the data area that has been erased at the time of initialization, the second parity is used thereafter. When performing data recovery, data recovery is performed using data different from the data at the time of initialization, so normal data cannot be recovered. Therefore, the actual data in the erased data area cannot be used in the calculation of parity at initialization. Therefore, in the present invention, a fixed data string is virtually generated for the erased data area, and parity data is generated. This fixed data string is uniquely determined for each data area, may be different for each data area, or may be determined for each apparatus.

  FIG. 5 is a diagram showing an embodiment of a data storage control device used in the data storage system shown in FIG. The data storage control device 101 used in the data storage system shown in FIG. 3 includes a horizontal code generation unit 111, a vertical code generation unit 121, a code writing unit 131, and a data writing unit 141, as shown in FIG. Have. FIG. 5 shows an example of main components related to the present embodiment among the components included in the data storage control device 101 according to the present embodiment.

  The horizontal code generation unit 111 applies a plurality of storage media 204-1 to 204- to the data written in each of the divided areas obtained by dividing the storage areas of the storage media 204-1 to 204-4 into predetermined areas. A first error detection code (parity code) is generated between the four. The horizontal code generation unit 111 generates a first error detection code for data in which addresses of a plurality of storage media 204-1 to 204-4 are written in the same divided area. Details of this will be described later.

  The vertical code generation unit 121 generates a second error detection code (parity code) for each of a plurality of storage media with respect to data written at different timings in each divided area. The vertical code generation unit 121 generates a second error detection code for data written in a predetermined divided area among the divided areas for the plurality of storage media 204-1 to 204-4. Details of this will be described later.

  The code writing unit 131 writes the first error detection code generated by the horizontal code generation unit 111 in a divided region in which data is not written in the divided region. In addition, the code writing unit 131 writes the second error code generated by the vertical code generation unit 121 in a divided region in which data is not written in the divided region. The code writing unit 131 uses the first error detection code generated by the horizontal code generation unit 111 as the address of the divided area where the data for which the first error detection code is generated is written. Write to the divided area.

  The data writing unit 141 writes data to the plurality of storage media 204-1 to 204-4 in a predetermined order.

  The data storage control device 101 shown in FIG. 5 is a function that the arrangement determining means 203 shown in FIG. 3 plays. The function of the data storage control device 101 shown in FIG. 5 may be assigned to the host request processing means 201, the arrangement map search means 202, and the arrangement determination means 203 shown in FIG.

Below, the whole operation | movement in this form is demonstrated. FIG. 6 is a diagram for explaining the overall operation in the present embodiment.
[Initialize]
First, a process for enabling read / write to a storage medium in the initial stage will be described.

  At the time of initialization, data redundancy processing between the storage media 204-1 to 204-4 is performed. For this, the horizontal code generation unit 111 is the first parity data (error detection code) for the data stored in the areas A1, A2, and A3 shown in FIG. A parity part is generated. Since all the data areas (divided areas) have been erased at the time of initialization, the parity calculation is performed on the fixed data string as described above. In this case, the horizontal code generation unit 111 calculates the XOR (exclusive logic) of the fixed data sequence of A1, the fixed data sequence of A2, and the fixed data sequence of A3, and uses the result as A parity data. Then, the code writing unit 131 writes the parity data in the divided area of A parity on the storage medium. At this time, writing is not performed on the divided areas for storing normal data (in the example shown in FIG. 6, the divided areas A1 to A3), and the SSD or the flash memory is maintained in the erased state. Similarly, the horizontal code generation unit 111 generates parity for B1 to B3, the code writing unit 131 writes parity data in the B parity divided region, and the first RAID (Redundant Arrays of Inexpensive Disks). ).

  Subsequently, the vertical code generation unit 121 generates second parity data (error detection code) (a divided region that stores parity shifted in the time axis direction). As shown in FIG. 6, parity data is generated for the divided areas of the storage medium 204-1 such as A1 and E1. Similar to the generation of the first parity data, since all the data areas (divided areas) have been erased at the time of initialization, the parity calculation is performed on the fixed data string as described above. . In this case, the vertical code generation unit 121 calculates XOR (exclusive logic) of the fixed data sequence of A1, the fixed data sequence of E1, and the like, generates parity data, and the data of “Parity of A1 + E1 +. To do. At this time, if there is a portion of the first parity data, the vertical code generation unit 121 reads the parity data and sets it as XOR target data. Then, the code writing unit 131 writes the parity data in an area indicated by “Parity of A1 + E1 +...” Illustrated in FIG.

  The horizontal code generation unit 111 sequentially generates parity data for other areas of the storage media 204-1 to 204-4 (B1, C1, etc. in the example shown in FIG. 6), and the code writing unit 131 The generated parity data is written in the areas of the storage media 204-1 to 204-4. Similar to the first parity generation, at this time, writing is not performed to an area (divided area) in which normal data is written, and the SSD or flash memory is kept in an erased state.

  Similarly, the vertical code generation unit 121 generates the second parity data for A2 and E2 for the storage medium 204-2, and the second parity data for all the storage media constituting the apparatus. Generate data. The code writing unit 131 writes the generated parity data. Here, the generation target of the second parity data includes the first parity data.

  Although the redundancy system using the parity code has been described, the present invention can be similarly applied to a system using ECC (Error Correction Code) or the like.

Also, the management table 300 shown in FIG. 4 is initialized. At the time of initialization, all the data storage areas that can be referred to from the host are in an erased state in the parity generation stage, so that all erased flags are set in the erased flag 302. For example, as shown in the above-described configuration, an expression method using a bitmap or the like can be used. Further, for the head host address 301, the head address of the address designated from the host side of the data stored in A1 and A2 shown in FIG. 6 is stored. Furthermore, for the data storage area indicated by A1 and the like shown in FIG. 6, when the fixed-length data and arrangement can be determined regularly, the first host address 301 can be omitted. In this case, the location of the first parity data, the location of the second parity data, the location where the data indicated by the host address is stored in which storage medium is uniquely determined without using a table or the like. It only has to have a means to decide.
[writing]
When a write command is issued from the host device and the host request processing unit 201 receives the command, the arrangement map search unit 202 searches the head host address 301 of the management table 300 for the host address indicated by the command. The arrangement map search means 202 searches the management table 300 corresponding to the host address indicated by the write command, and the arrangement determination means 203 sets the erased flag 302 at the same time as the storage medium 204-1 to 204-. Data is written to the corresponding part of 4. The actual writing procedure to the storage media 204-1 to 204-4 will be described in more detail with reference to FIG.

  When writing to A1 is performed from the host, the arrangement map search means 202 searches the head host address 301 of the management table 300. When the corresponding address is found, the arrangement determining means 203 drops the erased flag 302 of the management table 300 and sets it as written.

  Subsequently, the data writing unit 141 writes data to the location A1 of the storage medium 204-1. Further, the horizontal code generation unit 111 reads A Parity data, and generates new parity data from the data and A1 data. The code writing unit 131 writes the parity data generated by the horizontal code generation unit 111 in the area of A Parity. Thus, the arrangement determining unit 203 generates and writes the first parity data.

  Further, in order to generate the second parity data, the vertical code generation unit 121 reads E1 of the storage medium 204-1 and “parity of A1 + E1 +...”, And receives the data from the host A1. XOR is calculated and second parity data is generated. The code writing unit 131 writes the parity data generated by the vertical code generation unit 121 to “parity of A1 + E1 +. Thus, the arrangement determining unit 203 completes the generation and writing of the second parity.

  Generation of such parity data is the same as in normal RAID 5 and RAID 6, and detailed description thereof is omitted here. In this embodiment, the address from the host is sequentially determined in the order of data storage in the order of A1, A2, A3, and B1 shown in FIG. 6. Therefore, when additional writing such as archiving is performed, A1 Writing is the oldest, and new data is stored as needed in the order of A2, A3, and B1. For this reason, the data to be protected by the second parity data can generate parity for data whose time axis is shifted. In the example shown in FIG. 6, the second parity data is generated for data whose time axis is shifted such that the writing timing of A1 is the oldest and the writing timing of E1 is next.

  Furthermore, when writing to the erased data area in the above-described processing sequence, the arrangement determining unit 203 checks the data in order to confirm whether the data area has been erased. Read once. The arrangement determining means 203 may perform writing if the data in the data area is erased, and may perform writing after erasing if the data is not erased. As another method, the arrangement determining means 203 may erase the data area and write it without reading it once.

Note that the addresses (physical addresses and logical addresses) of the divided areas in which A1 to A3 and A Parity are written may be the same among the storage media 204-1 to 204-4. The address of the divided area where B1 to B3 and B Parity are written, the address of the divided area where C1 to C3 and C Parity are written, the address of the divided area where D1 to D3 and D Parity are written, and the address of the divided area where E1 to E3 and E Parity are written The same applies to the address.
[reading]
When a read command is issued from the host device and the host request processing means 201 receives the command, the host address indicated by the command is read out on the storage media 204-1 to 204-4, and data is sent to the host. Forward. At that time, a process for determining whether there is data loss due to retention may be provided. The arrangement determining unit 203 reads data for generating second parity data such as E1 and “parity of A1 + E1 +...” Simultaneously with reading of the data of A1. The arrangement determining unit 203 compares the parity generated from the data such as A1, E1, and the “parity of A1 + E1 +. As a result of the comparison, the arrangement determining means 203 transfers the data to the host if the two match, and if not, tries the data recovery shown next.

  The data storage method in the data storage control device 101 shown in FIG. 5 will be described below. FIG. 7 is a flowchart for explaining a data storage method in the data storage control apparatus 101 shown in FIG.

  The data writing unit 141 writes data in a predetermined order in the divided areas of the storage media 204-1 to 204-4 (step S11). Then, the horizontal code generation unit 111 applies the data written in the divided areas obtained by dividing the storage areas of the storage media 204-1 to 204-4 to the predetermined areas, between the storage media 204-1 to 204-4. The first error detection code (parity data) is generated (step S12). Then, the code writing unit 131 writes the first error detection code generated by the horizontal code generation unit 111 in a divided region in which data is not written in the divided region (step S13).

  In addition, the vertical code generation unit 121 generates a second error detection code for each of the storage media 204-1 to 204-4 for data written at different timings in each of the divided areas (step S14). Then, the code writing unit 131 writes the second error code generated by the vertical code generation unit 121 in a divided region where data is not written in the divided region (step S15).

The data recovery operation in this embodiment will be described below.
[Data recovery]
FIG. 8 is a diagram for explaining an example of a processing procedure when an error occurs during reading or writing of data.

  In reading data from the host device, the parity check using the second parity data described above, or an error from the SSD at the time of reading, or a redundant code such as ECC provided as a countermeasure against an error in the flash memory A processing procedure when an error or an error during writing occurs will be described. Here, it is assumed that a data read or write error has occurred when accessing the divided area A1 shown in FIG.

  As a first procedure, a procedure for recovering data using the first parity data will be described. This procedure is suitable when an error occurs during single reading or writing. As shown in FIG. 8, it is assumed that an error occurs during reading or access in A1. In a general RAID, when an error occurs, all corresponding storage media are regarded as defective, and data generation from parity is performed on the entire storage media. On the other hand, the data storage system of the present invention assumes an error in the SSD using the flash memory or the flash memory itself, and tries to recover the data on the assumption that many errors occur. That is, the data storage system of the present invention reads A2, A3, and A Parity when the A1 data is restored when the above-described error occurs in A1, and reproduces the A1 data. Write to A1. At this time, if an error occurs in A2, A3, and A Parity other than A1, the data storage system tries the procedure using the second parity. At this time, the data storage system does not perform the data recovery operation after B1 of the storage medium 204-1.

  As a second procedure, a procedure for attempting data recovery using the second parity data will be described. This procedure is suitable when reading is performed sequentially in the order of addresses or when an error occurs in collation of the second parity data during reading. As processing, it is assumed that data is simultaneously read from a plurality of storage media, such as when data is simultaneously read from the storage media 204-1 and 204-2 in a sequential read-ahead operation. For example, this procedure is used when a read error is detected at A1 of the storage medium 204-1 and a read error is detected at A2 of the storage medium 204-2. If no error is detected, it is preferable to attempt data recovery using the first procedure described above.

  When an error occurrence is detected in A1 of the storage medium 204-1 and A2 of the storage medium 204-2, there is a high possibility of an error due to retention. At this time, the data storage system performs reading of data other than A1 constituting “parity of A1 + E1 +...” Such as E1 and parity of “A1 + E1 +. The data of A1 is regenerated and written again in the area of A1. At this time, if an error occurs in the procedure for regenerating data from the second parity data, the data storage system attempts to regenerate data using the first parity data in the second procedure. If the regeneration is successful, the data storage system writes the data to the divided area of the corresponding storage medium 204-1, tries to regenerate the data using the second parity data again, and reproduces the data of A1. Perform a writeback.

During such a recovery operation, when an erased data area indicated by the management table 300 is read, the data is recovered using data of a fixed data string determined in the data area.
These recovery procedures can be selected based on the error occurrence status in other accesses that operate in parallel, and an appropriate data processing method can be selected to recover error data by retention.

  As described above, the data is recorded at different times between the storage media such as the SSD using the flash memory or the flash memory, in addition to the RAID prepared for the failure of the storage medium, or the redundancy means (such as Erasure Coding) according to the RAID. RAID or a redundancy means equivalent to RAID is taken between data. As a result, it is possible to recover the data even with respect to the retention error.

  Further, when the configuration shown in FIG. 6 is applied to the storage device, the redundancy processing using the first parity data described above is performed on the storage device side, and the redundancy processing using the second parity data is performed. It is also conceivable to perform it inside the storage media 204-1 to 204-4 shown in FIG. 6 (SSD internal processing).

  The present invention relates to a storage device, and more particularly, an SSD (silicon disk) using a flash memory or the like, or a storage device using an SSD, and a device for storing data, such as a computer or server, incorporating a flash memory. It is possible to apply to

  As described above, each function (process) is assigned to each component, but this assignment is not limited to the above. In addition, the configuration described above is merely an example, and the present invention is not limited to this.

  The processing performed by each component provided in each of the data storage control devices 100 and 101 described above may be performed by a logic circuit that is produced according to the purpose. Further, a computer program (hereinafter referred to as a program) in which processing contents are described as a procedure is recorded on a recording medium readable by each of the data storage control devices 100 and 101, and the program recorded on the recording medium is subjected to data storage control. The devices 100 and 101 may be read and executed. The recording media readable by each of the data storage control devices 100 and 101 are a floppy (registered trademark) disk, a magneto-optical disk, a DVD (Digital Versatile Disc), a CD (Compact Disc), and a Blu-ray (registered trademark) Disc. In addition to a transferable recording medium such as, a ROM (Read Only Memory), a RAM (Random Access Memory), a HDD (Hard Disc Drive), or the like built in each of the data storage control devices 100 and 101. The program recorded on this recording medium is read by the CPU provided in each of the data storage control devices 100 and 101, and the same processing as described above is performed under the control of the CPU. Here, the CPU operates as a computer that executes a program read from a recording medium on which the program is recorded.

100, 101 Data storage control device 110, 111 Horizontal code generation unit 120, 121 Vertical code generation unit 130 Code writing unit 141 Data writing unit 201 Host request processing unit 202 Arrangement map search unit 203 Arrangement determination unit 204-1 to 204-4 Storage medium 300 Management table 301 First host address 302 Deleted flag

Claims (8)

A horizontal code generator for generating a first error detection code between the plurality of storage media for data written in each of the divided areas obtained by dividing the storage area of each of the plurality of storage media into a predetermined area;
A vertical code generation unit that generates a second error detection code for each of the plurality of storage media for data written at different timings to each of the divided regions;
Each of the first error detection code generated by the horizontal code generation unit and the second error code generated by the vertical code generation unit is divided into portions in which the data is not written. A data storage control device comprising a code writing unit for writing in an area. The data storage control device according to claim 1,
The horizontal code generation unit generates the first error detection code for data in which addresses of the plurality of storage media are written in the same divided area,
The code writing unit includes the first error detection code generated by the horizontal code generation unit, the address of the divided region in which the data for which the first error detection code is generated is written. A data storage control device for writing in the divided area having the same address as the address. The data storage control device according to claim 1 or 2,
The vertical code generation unit is a data storage control device that generates the second error detection code for data written in a predetermined divided area of the divided areas for each of the plurality of storage media. The data storage control device according to any one of claims 1 to 3,
A data storage control device comprising a data writing unit for writing the data in a predetermined order to the plurality of storage media. In the data storage control device according to any one of claims 1 to 4,
The data storage control device, wherein the plurality of storage media are silicon disks. A plurality of storage media;
A horizontal code generator for generating a first error detection code between the plurality of storage media for data written in each of the divided areas obtained by dividing the storage area of each of the plurality of storage media into a predetermined area;
A vertical code generation unit that generates a second error detection code for each of the plurality of storage media for data written at different timings to each of the divided regions;
Each of the first error detection code generated by the horizontal code generation unit and the second error code generated by the vertical code generation unit is divided into portions in which the data is not written. A data storage system having a code writing unit for writing in an area. A process of generating a first error detection code between the plurality of storage media for data written in each of the divided areas obtained by dividing the storage area of each of the plurality of storage media into a predetermined area;
A process of writing the first error detection code in a divided area of the divided area where the data is not written;
A process of generating a second error detection code for each of the plurality of storage media for data written at different timings to each of the divided areas;
A data storage method for performing a process of writing the second error code in a divided area of the divided area where the data is not written. On the computer,
A procedure for generating a first error detection code between the plurality of storage media for data written in each of the divided areas obtained by dividing the storage area of each of the plurality of storage media into a predetermined area;
Writing the first error detection code in a divided area of the divided area where the data is not written;
A procedure for generating a second error detection code for each of the plurality of storage media for data written at different timings to each of the divided regions;
A program for causing the second error code to be written in a divided area of the divided area where the data is not written.

Images