JPH08115172A - Disk array device - Google Patents

Abstract

PURPOSE: To average the time required for accessing by storing continuous data in plural. disk devices dividedly in each different transfer rate section. CONSTITUTION: This disk array device 15 constituted of an array controller 12 and five disk devices 13A to 135E Each disk device 13 is provided with the prescribed number of blocks for storing data with prescribed size and time necessary for data transfer is changed in accordance with a position on a disk medium. Five sections in which total time necessary for the transfer of data in the blocks are successively changed are set up and addresses to be identification information for accessing blocks are allocated to blocks in each section so as to be circulated in five disk devices 13A to 13E and circulated in respective sections. Consequently five continuous data can be stored in different sections in different disk devices and the time required for accessing can be averaged.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk array device, and more particularly to a disk array device having a disk device whose data transfer rate changes according to the position of the block on the disk medium.

[0002]

2. Description of the Related Art A disk array device is a storage device that is a combination of a plurality of disk devices and that can be handled as a single disk device by a host device. The storage form is used.

In a disk array device for the purpose of improving the reliability of data, for example, the parity which is the exclusive OR of the data is calculated for every several data, and the parity and the parity are calculated. The used data is stored in different disk devices.

In a disk array device storing such a parity, as shown in FIG. 10, the parity is always stored in one disk device.
As schematically shown in FIG. 3, there is one in which the parity is distributed and stored in all the disk devices (five in the figure) that constitute the disk array device. 10 and 11, the area surrounded by a frame indicates a block which is the minimum access unit of the disk device, and the blocks shown at the same position indicate blocks identified by the same address. ing. Further, the block indicated as DATAXY (Y = 1,2,3,4) is a block in which data is stored, and the block indicated as PARITY X is an exclusive logic of data comrades having the same X. The sum, that is, the block in which the parity is stored.

As described above, in any of the disk array devices, the parity which is the exclusive OR of the contents of the same addresses of the four disk devices is stored in the same address of the remaining one disk device. Therefore, even if a failure occurs in a certain disk device and the data in that disk device cannot be read, the contents stored in the other four disk devices are read and the exclusive OR operation is performed. As a result, the contents of the failed disk device can be restored, and the storage device has higher reliability than a single disk device.

Further, if the data can be accessed in parallel from N disk devices, it operates as a storage device having a data transfer rate N times the data transfer rate of a single disk device. The disadvantage of the disk device, which has a low access speed, can be compensated. For this reason, there is also a disk array device that does not calculate and store the parity but only increases the access speed as schematically shown in FIG.

[0007]

As described above,
A disk array device configured by combining disk devices can compensate for defects such as low reliability and slow response of individual disk devices. But,
It is also true that the characteristics of the disk array device are greatly affected by the characteristics of the individual disk devices, and it is desirable that the disk array device be configured by combining disk devices that can be accessed quickly.

There are various means for improving the access speed of a single disk device.
One is a technique called zone bit recording. In this recording method, as schematically shown in FIG. 13, the storage area of the disk medium is divided into several zones 32,
The blocks 33 are set in each track so that the number of blocks (sectors) on the outer peripheral side is larger and that the same number of blocks are included in one zone.

Even in this recording system, the disk medium is rotated at a constant speed as in the conventional disk device, but since the number of blocks increases toward the outer peripheral side, the blocks on the outer peripheral side are closer to the inner peripheral side. Block, or data can be transferred at a higher speed than the conventional disk device.

Therefore, in the disk device using the zone bit recording method, as described in Japanese Patent Laid-Open No. 3-34158, control such as storing a high priority file in the outer storage area is performed. By doing so, it is possible to improve the overall performance of the storage device.

The conventional disk array device is not premised on the use of the disk device adopting such a zone bit recording method, so that the disk device adopting the zone bit recording method can be used. However, there is a problem that the characteristics of the zone bit recording method are not fully utilized.

Therefore, it is an object of the present invention to provide a disk array device which makes the best use of the characteristics of the disk device of the zone bit recording system.

[0013]

According to the first aspect of the present invention,
(A) N, in which a predetermined number of blocks for storing data of a predetermined size are provided, the time required to transfer the data in each block changes according to the position of the block on the disk medium. N disk units and (B) N disk units each of which has the same number of blocks and in which the total time required to transfer data in the same number of blocks sequentially changes Setting means for setting partitions, and (c) N blocks are circulated through blocks in N partitions respectively provided in the N disk devices by this setting means, and N partitions are allocated. An allocating unit that allocates an address serving as identification information when accessing each block so as to cyclically travel.

That is, according to the first aspect of the present invention, the N disk devices constituting the disk array device are each provided with a predetermined number of blocks for storing data of a predetermined size. The time required for transfer is changed according to the position of the block on the disk medium, and the setting unit configures each disk device to have the same number of blocks. Allow N partitions to be set that sequentially change the total time required to transfer the data in the block. Then, N × set by the assigning means
An address that is identification information when accessing each block is allocated to the blocks in the N partitions so that the N disk devices are circulated and the N partitions are circulated. Block in each disk device is accessed.

As a result, N consecutive data can be stored in different partitions of different disk devices, and the time required for access can be averaged. The unit for receiving data from an external device such as a host device may be the same as the block size of the disk device, or N blocks of data may be received collectively.

According to a second aspect of the present invention, (a) a predetermined number of blocks for storing data of a predetermined size is provided, and the time required to transfer the data in each block exists in that block. (B) Necessary to transfer data in the same number of blocks, which is composed of the same number of blocks, to each of the N disk devices that change according to the position on the disk medium and (b) each of the disk devices. Setting means for setting N partitions in which the total time sequentially changes, and (c) N disk devices in blocks in the N partitions respectively provided in the N disk devices by this setting means. And a means for allocating an address serving as identification information when accessing each block so as to circulate in a predetermined order and circulate in N partitions,
(D) A block in one disk device designated for every N addresses by the address assigned to each block by the assigning means is used as a block for storing parity data, and N-1 other blocks are used. The block designated by the address is used as a block for storing N-1 data which is the source of the parity data, and a storage unit for storing the data in each disk device is provided.

That is, according to the second aspect of the invention, the data in each block is provided such that each of the N disk devices constituting the disk array device is provided with a predetermined number of blocks for storing a predetermined size of data. The time required for transfer is changed according to the position of the block on the disk medium, and the setting unit configures each disk device to have the same number of blocks. Allow N partitions to be set that sequentially change the total time required to transfer the data in the block. Then, the allocating unit circulates the N disk devices in a predetermined order to the blocks in the set N × N partitions, and circulates the N partitions.
An address that is identification information when accessing each block is assigned, and is designated for every N addresses 1
A block in one disk device is used as a block for storing parity data, and a block designated by another N-1 address is used as a block for storing N-1 data which is the source of the parity data. hand,
Allow data to be stored.

Thus, while storing the parity data in one disk device, the N that is the source of the parity is
It is possible to store -1 piece of data in different partitions of other disk devices, and it is possible to average the time required for access. The unit for receiving data from an external device such as a host device is
The block size may be the same as that of the disk device, or N-1 blocks of data may be collectively received.

According to the third aspect of the present invention, (a) a block is provided with a predetermined number of blocks for storing data of a predetermined size, and the time required to transfer the data in each block exists. (B) Necessary to transfer data in the same number of blocks, which is composed of the same number of blocks, to each of the N disk devices that change according to the position on the disk medium and (b) each of the disk devices. Setting means for setting N partitions in which the total time sequentially changes, and (c) N disk devices in blocks in the N partitions respectively provided in the N disk devices by this setting means. Of the N partitions set by the setting means, the blocks in the partition having the longest total time required for data transfer are designated for each N blocks. Bro Assigned to each block by the assigning means for allocating an address serving as identification information when accessing the memory, and (d) the address assigned to each block by this assigning means, necessary for most data transfer. A block in a partition having a long total time is used as a block for storing parity data, and a block specified by another N-1 address is stored as N-1 data which is a source of the parity data. And a storage unit for storing data in each disk device.

That is, according to the third aspect of the invention, the data in each block in which a predetermined number of blocks for storing data of a predetermined size are provided in each of the N disk devices constituting the disk array device The time required for transfer is changed according to the position of the block on the disk medium, and the setting unit configures each disk device to have the same number of blocks. Allow N partitions to be set that sequentially change the total time required to transfer the data in the block. Then, the allocating unit circulates the N disk devices to the blocks in the set N × N partitions, and for each N, the most of the N partitions set by the setting unit. An address that is identification information when accessing each block is assigned so that a block in a partition whose total time required for data transfer is long is designated, and it is specified every N addresses. A block in a partition having a long total time required for data transfer is used as a block for storing parity data, and other N-1
Data is stored by using the block specified by the address as a block for storing N-1 data which is the source of the parity data.

As a result, it becomes possible to store the parity data in the partition having the slowest data transfer rate and at the same time, store the N-1 pieces of data, which are the sources of the parity, in the other disk devices. , The time required for reading can be shortened. The unit for receiving data from an external device such as a host device may be the same as the block size of the disk device, or N-1 blocks worth of data may be received collectively.

[0022]

EXAMPLES The present invention will be described in detail below with reference to examples.

FIG. 1 shows the configuration of a disk array device according to an embodiment of the present invention. As shown in the figure, the disk array device 11 of the embodiment includes an array controller 12
And five disk devices 13 _A to 13 _E , the array controller 12 includes a CPU 14 and an RO.
M15, RAM 16, data transfer control circuit 15, parity generation circuit 17, interface 19 for connecting the array controller 12 to the host device 31, and disk interface (DI / F) for connecting each disk device 13 to the array controller 12. ) 20 _A to 20 _E.

Of the elements constituting the array controller 12, the CPU 14 is a control circuit that controls each unit in an integrated manner and responds to an access request from the host device 31, and a program that defines the operation procedure is ROM15
Is stored in The RAM 16 is a memory used by the CPU 14 as a work area, and a partition information table and a transfer mode table, which will be described later, are also read from a predetermined storage area of the disk device 13 when the power is turned on, and this memory is used. Remembered above.

The data transfer control circuit 17 is a circuit for controlling data transfer between the host device 31 and the array controller 12 and between the disk device 13 and the array controller 12 according to an instruction from the CPU 14. A buffer memory is provided in the data transfer control circuit 17, and the data transfer control circuit 17 is configured to be able to perform data transfer with a plurality of disk devices at the same time by utilizing the buffer memory. In addition, the data transfer control circuit 15
The data from the host device 31 or the data from the disk device is also supplied to the parity generation circuit 16,
The parity created by the parity generation circuit 16 is also transferred to the host device 31 or the disk device 17 via the data transfer control circuit 17.

The disk device 13 is a hard disk device adopting the zone bit recording method, and a continuous logical address is assigned to a block (a storage area unit which can be read and written by one access) in each disk device. ing. The allocation of this logical address is performed so that its value increases toward the inner circumference side.
A block with a smaller logical address is a block that can transfer data at a higher speed.

The operation of the disk array device of the embodiment will be described in detail below.

The disk array device of the embodiment is constructed so that various data storage forms can be adopted according to the setting contents at the time of initialization. First, the first operation mode, which is one of them, will be described.

In the first operation mode, the disk array device operates as a storage device whose minimum access unit is 4S (S is the minimum access unit of each disk device), and the data of size 4S requested to be written is The data is divided into four data of size S and written in four disk devices, respectively. Also, the parity is calculated from the four data, and the calculated parity is written in the remaining one disk device that was not used for writing the divided data.

When the first mode is selected, two types of tables as described below are set in the disk array device of the embodiment at the time of initialization.

One table (hereinafter referred to as partition information table) is a table for setting partitions having the same storage capacity in the storage area of each disk device, the partitions having the same storage capacity. The start address indicating the start position of is stored for the number of partitions (the number of disk devices). The range of the i-th partition is defined by the starting address of the partition i and the starting address of the partition i + 1, and when the number of accessible blocks of each disk device is 5M, as shown in FIG. In the form, each section identification information 2
The start address 22 is stored for 1. The range of the fifth section is configured to be defined by the start address of section 5 and the maximum address of the disk device.

In addition, another table (hereinafter, referred to as a transfer mode table) set at the time of initialization defines a transfer mode that defines a procedure for processing an access request from the host device. Information is stored. In the disk array device of the embodiment, the processing procedure for the access request changes according to the logical address of the data accessed by the access request, and in the transfer mode table, as schematically shown in FIG. The transfer mode information 24 and the threshold value 23, which is information for associating the access-requested logical address with the transfer mode information 24, are set.

In the first operation mode, the response process to the access request from the host device is performed in the following procedure based on these pieces of information. First, the response procedure to the write request will be described.

FIG. 4 shows the disk array device of the embodiment.
7 shows a flow of response processing to a write request in the first operation mode. As shown in the figure, the CPU, which has received the write request from the host device, selects the transfer mode according to the write-requested logical address based on the contents of the transfer mode table (step S101).
To do.

In this step, the logical address requested to be accessed is compared with the start address in the transfer mode table, and one transfer mode is selected. For example, the logical address requested to be accessed is X, and the X is between the first threshold “0” and the second threshold “M” in the transfer mode table (however,
(Except "M"), the transfer modes "A1, B1, C1, D stored in association with the first threshold value" are stored.
1, E5 "is selected.

After that, the CPU interprets the contents of the selected transfer mode, grasps the order of use of the disk devices and the partition used in each disk device, and specifies the logical address of each block in which data is to be written (step S102). The details of the processing performed in this step are as follows.

As described above, in the first operation mode,
Data of size 4S from the host device is
The data is divided into two pieces of data and stored in different disk devices, respectively, but the first four pieces of information in the transfer mode are information indicating the disk device and the partition in which each divided data is to be stored. The fifth information is information designating a disk device and a partition for storing the parity data created by the parity generation circuit based on each divided data. For example, transfer mode "A1, B
1, C1, D1, E5 ", the four data obtained by dividing the data of size 4S are divided into partition 1 of disk device A, partition 1 of disk device B, and disk device C, respectively. The partition 1 is written in the partition 1 of the disk device D, and the parity data is written in the partition 5 of the disk device E.

Since the block and the address have the same correspondence in all the disk devices connected to the disk array device of the embodiment, the address of the block to which the actually divided data is written is as follows. Is calculated. First, the CPU calculates the distance between the requested logical address X and the threshold value corresponding to the X. Then, the start address of the partition to which the data is to be written (its value is stored in the partition information table shown in FIG. 2) is added to the calculated distance, and the result is added to the block to which the data is to be written. It is a logical address.

Then, the CPU notifies the data transfer control circuit of the logical address and the like thus obtained, thereby instructing the data transfer control circuit to perform the operation (step S103). Upon receiving this instruction, the data transfer control circuit notifies each disk device of the block in which the data to be transferred is to be stored (step S
Then, the data received from the host device is divided into four blocks of the same size, and each block is transferred to the disk device and the parity generation circuit (step S105). Then, the parity generated by the parity generation circuit is transferred to the disk device designated to be used for storing the parity data (step S
106) and the process is terminated.

In response to the read request, as shown in FIG. 5, the CPU which has received the read request from the host device first makes the read request based on the contents of the transfer mode table. The transfer mode used for the existing data is specified (step S201). Then, the contents of the specified transfer mode are interpreted to specify the use order of the disk device and the logical address of the block to be read in the disk device (step S20).
2) Then, based on the specified logical address, the operation procedure is instructed to the data transfer control circuit (step S203).

The data transfer control circuit, which has received the instruction of the operation procedure, gives logical address information to each disk device,
After instructing to output the contents of the block specified by the logical address (step S204) and reading the four blocks in which the data is stored normally completed (step S205; Y), the four data are stored. By transferring the combined data to the host device, the read request is responded to (step S206).

As described above, in the disk array device of the embodiment, the storage of the four data obtained by dividing the data and the parity data created based on the divided data is as follows.
This is done for different disk devices according to the contents of the transfer mode table, but as shown in FIG. 3, the fifth information of all transfer modes in the transfer mode table is
Since the partition 5 is set, the parity data is always stored in the partition having the lowest data transfer rate.

Since the parity data is data required only when a failure occurs in the data, it is not necessary to access it during normal data reading as shown in FIG. Therefore, in the disk array device of the embodiment in which the storage area having the lowest data transfer rate (the storage area on the inner circumference side of the disk medium) is used for the parity data as described above, the data is read out. The average time required will be shorter than if the parity and data were stored in blocks of the same address.

In the disk array device of the embodiment,
Since the data transfer control circuit that can perform data transfer with each disk device in parallel is used, the transfer mode table as shown in FIG. 3 is adopted to maximize the average response speed. When operated in accordance with this transfer mode table, the response speed varies depending on the address requested for access. Although the fluctuation of the response speed is less likely to cause a problem, if it is desired to prevent the fluctuation of the response speed, the average response speed is slightly reduced, but during each transfer mode as shown in FIG. You may use the transfer mode table containing the information which designates.

However, when a data transfer control circuit that cannot process data from disk devices in parallel is used, that is, as shown in FIG. 7, individual data is read out from each disk device in order (step S304,
In the case of S305), even if a transfer mode table in which use of all partitions is designated for each transfer mode as shown in FIG. 6 is used, a response speed equivalent to that of the transfer mode table of FIG. 3 can be obtained. Therefore, it is better to use the transfer mode table as shown in FIG. 6 which can prevent the variation of the response speed for each access request.

That is, when the transfer mode table of FIG. 6 is used, the data requested to be written by the host device, for example, DATA 0 and DATA J, are DATA 01 to DATA 04, as shown schematically in FIG. It is divided into DATAJ1 to DATAJ4, which are stored in different transfer rate partitions of different disk devices. If the data from the disk device cannot be processed in parallel, reading the data will require the time that is obtained by adding the time required to read each divided data. In the disk array device in which the data is stored, the partition 1 to the partition 4 can be used for reading any data.
Therefore, the time required to read the data is added, and as a result, the response speed is averaged.

Similarly, when a data transfer control circuit that cannot process data from the disk device in parallel is used, as the transfer mode table, as shown in FIG. Even if a device stored in the device (disk device E in the figure) is used, it is possible to obtain a disk array device in which the variation in response speed to each access request is small.

Since the effect of suppressing the response speed fluctuation is naturally obtained even when the parity is not stored, for example, the data is simply divided without providing the parity generation circuit in the array controller. Thus, even when the disk array device is configured by using the array controller having only the function of sequentially supplying to N disk devices, if the above-mentioned data storage form is adopted,
It is possible to configure a disk array device, such as a single zone bit type disk device, in which the time required to read a file does not differ even if the file size is the same. .

In the first operation mode, the data requested to be written by the host device is divided into four pieces. However, the data is divided into 4K pieces, and the individual data is written to each disk device in order. It goes without saying that such a configuration may be adopted.

Next, the second operation mode of the disk array system of the embodiment will be described. In the second operation mode, in the disk array device, the minimum access unit is S (the minimum access unit of the disk device), and the number of accessible blocks is 20M (that is, the number of blocks accessible by one disk device). It operates as a storage device of 4 times "5M".

In the second operation mode as well, the partition information table and transfer mode table having the same contents as in the first operation mode are set at the time of initialization, but the method of using the transfer mode table is the first. It is different from the usage of modes.

In the first operation mode, the logical address notified by the host device at the time of an access request was directly compared with the threshold value in the transfer mode table, but in the second operation mode, the logical address received from the host device was compared. The lower 2 bits of the address X are ignored, the remaining information X _REST is compared with the threshold value, and the transfer mode is selected.

Then, when the lower 2 bits of the logical address to be accessed are "00", in the partition of the disk device designated by the first information of the information included in the transfer mode, It is determined that the data specified as the access target exists at the address obtained by adding the start address of the partition to X _REST . When the lower 2 bits are “01”, “10”, and “11”,
The second, third and fourth information in the transfer mode are used respectively.

For example, when the access request is a write request, the CPU specifies the address to write the data by the above-mentioned procedure and, from the last information of the transfer mode, the parity corresponding to the data. Specifies the address of the block in which is stored. Then, the data transfer control circuit is controlled to read the contents of the data block and the contents of the parity block, write the data requested to be written into the data block, and read the two data and the data to be written. Write the result of the exclusive OR operation to the parity block.

In the second operation mode, when the access request is a read request, the procedure described above is used.
Simply specify the block from which to read the data,
The data of the block is transferred to the host device.

Thus, even in the second operation mode,
Of the storage areas of the disk device, the area with a low data transfer rate is used for storing parity data that is not used during normal reading, so a disk device of the zone bit recording method was simply used. The average response speed to a read request is faster than that of a disk array device.

[0057]

According to the first aspect of the present invention, a disk array device is configured so that data that is continuously accessed for the number of disk devices can be stored in partitions of different disk devices having different data transfer rates. With this configuration, the access time can be averaged.

According to the second aspect of the present invention, while storing the parity data in one disk device, the N-1 pieces of data which are the basis of the parity are stored in different disk devices. By configuring the disk array device so that the data transfer rates are stored in different partitions, it is possible to obtain a highly reliable storage device in which access times are averaged.

Then, as in the third aspect of the present invention, while storing the parity data in the partition having the slowest data transfer rate, N-1 pieces of data which are the sources of the parity are different from each other. If the disk array device is configured to be stored in the device, it is possible to obtain a highly reliable storage device in which the time required for reading is shortened.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a disk array device according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing an outline of a partition information table set in the disk array device of the embodiment.

FIG. 3 is an explanatory diagram showing an outline of a transfer mode table set in the disk array system of the embodiment.

FIG. 4 is a flowchart showing a flow of response processing to a write request when the disk array system of the embodiment is operated in the first operation mode.

FIG. 5 is a flowchart showing a flow of response processing to a read request when the disk array device of the embodiment is operated in the first operation mode.

FIG. 6 is usable in the disk array device of the embodiment,
It is explanatory drawing which showed an example of the transfer mode table.

FIG. 7 is a flowchart showing a flow of response processing to a read request in a disk array device in which data is sequentially accessed.

FIG. 8 is an explanatory diagram schematically showing a data storage form of the disk array device of the embodiment.

FIG. 9 is usable in the disk array device of the embodiment,
It is an explanatory view showing an example of a transfer mode table in which parity data is stored in one disk device.

FIG. 10 is an explanatory diagram showing an example of a data storage mode using parity in a conventional disk array device.

FIG. 11 is an explanatory diagram showing another example of a data storage mode using parity in a conventional disk array device.

FIG. 12 is an explanatory diagram showing an example of a data storage mode that does not use parity in a conventional disk array device.

FIG. 13 is an explanatory diagram showing an outline of a zone bit recording method.

[Explanation of symbols]

11 ... Disk array device, 12 ... Array controller, 13 ... Disk device, 14 ... CPU, 15 ... RO
M, 16 ... RAM, 17 ... Data transfer control circuit, 18 ...
Parity generation circuit, 19 ... Interface, 20 ... Disk interface, 21 ... Partition information, 22 ... Start address, 23 ... Threshold, 24 ... Transfer mode, 31 ... Host device, 32 ... Zone, 33 ... Block (sector)

Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI Technical display location G11B 20/18 574 H 8940-5D

Claims (3)

[Claims] 1. A predetermined number of blocks for storing data of a predetermined size is provided, and the time required to transfer the data in each block changes depending on the position of the block on the disk medium. N disk units to be used, and N partitions each of which has the same number of blocks and in which the total time required to transfer data in the same number of blocks changes in order And a setting means for setting the number of blocks, and the block in each of the N partitions provided in each of the N disk devices by the setting means circulates the N disk devices and circulates the N partitions. To
A disk array device comprising: an allocation unit that allocates an address that is identification information when accessing each block. 2. A predetermined number of blocks for storing data of a predetermined size is provided, and the time required to transfer the data in each block changes depending on the position of the block on the disk medium. N disk units to be used, and N partitions each of which has the same number of blocks and in which the total time required to transfer data in the same number of blocks changes in order Setting means for setting the N disk units in a predetermined order through blocks in N partitions respectively provided in the N disk devices by the setting unit, and N partitions Allocating means for allocating an address serving as identification information when accessing each block so that each block is circulated, and the address allocated to each block by this allocating means. The block in one disk device specified for each N addresses is used as a block for storing parity data, and the block specified by the other N-1 addresses is used as the source of the parity data. N-
A disk array device comprising: storage means for storing data in each disk device by using it as a block for storing one piece of data. 3. A predetermined number of blocks for storing data of a predetermined size are provided, and the time required to transfer the data in each block changes depending on the position of the block on the disk medium. N disk units to be used, and N partitions each of which has the same number of blocks and in which the total time required to transfer data in the same number of blocks changes in order And a setting means for setting the number of the disk units in the N partitions provided in each of the N disk devices by the setting means, and the setting means for each N units. Identification information for accessing each block so that the block in the partition having the longest total time required for data transfer is designated among the N partitions set in The block in the partition having the longest total time required for data transfer, which is designated for every N addresses, is assigned by the assigning means for allocating the following address and the address assigned to each block by this assigning means. The parity block is used as a block for storing the parity data, and the block designated by the other N-1 addresses is used as a block for storing the N-1 data which is the source of the parity data, and is transferred to each disk device of data. And a storage unit for storing the disk array device.