JP3052877B2 - Disk array device - Google Patents

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 system in which a plurality of storage devices are physically linked to each other to be regarded as a larger-capacity virtual storage device, and data is recorded on the virtual storage device. The present invention relates to a disk array device capable of performing reproduction.

[0002]

2. Description of the Related Art First, terms frequently used in this specification are defined below.

[0005] Data access: Data read and write operations. -Throughput: The amount of data that can be accessed per unit time. Client: A program that accesses data. Response time: The time required from the issuance of a data access request from a client to the end of processing. Virtual storage device: Physically composed of a plurality of storage devices, which are logically viewed as one storage device.

Next, a conventional technique will be described.

[0005] In a disk array device (striping disk) that forms a virtual storage device by combining a plurality of storage devices (hereinafter referred to as secondary storage devices), data is distributed to each secondary storage device in block units. Stored. In this case, a method is used in which data is allocated uniformly in block units in the order of the identifiers allocated to the respective secondary storage devices (reference: Japanese Patent Laid-Open No. Hei 8-
No. 30400). In this method, for example, when n blocks stored in the virtual storage device are actually stored in the secondary storage device t, n + 1 blocks are stored in the secondary storage device t.
+1 (n and t are natural numbers). If the virtual storage device is composed of m secondary storage devices and block a is stored in secondary storage device m, block a + 1 is stored in secondary storage device 1 (m and a are natural numbers). ).

In some cases, the disk array device constitutes a plurality of virtual storage devices from a plurality of secondary storage devices. FIGS. 17 and 18 show two typical examples, respectively.

In FIG. 17, a plurality of secondary storage devices are divided into several sets, and a virtual storage device is constructed for each set. In this method, one secondary storage device is not shared by a plurality of virtual storage devices. Therefore, even if a plurality of applications access different virtual storage devices, they are not mutually affected by accesses from other applications.

[0008] However, in this method, each virtual storage device:
Only a part of the secondary storage device constituting the disk array device is used. Therefore, there is a problem that the upper limit of the throughput at the time of access is lower than that of the virtual storage device using all the secondary storage devices of the disk array device, and the access concentration is weak.

On the other hand, in FIG. 18, a plurality of secondary storage devices are used transversely and divided into a plurality of virtual storage devices. In this method, each virtual storage device can use all the secondary storage devices constituting the disk array device. Therefore, the throughput that the disk array device can originally obtain is obtained as the upper limit of the throughput of each virtual storage device.

However, in this method, a secondary storage device is shared between virtual storage devices. Therefore, even when a plurality of applications access different virtual storage devices, each application may be affected by the access of another application. Also, as described above, the data blocks are stored in the order of the secondary storage device (1) → (2) → (3) → (4) → (1)... In FIG. Therefore, in a situation where a plurality of clients sequentially access data at the same time interval such as video-on-demand, at a certain point in time, client A and client B access the same secondary storage device in different virtual storage devices. In this case, thereafter, the clients A and B always access the same secondary storage device at the same timing. In such a case, in the secondary storage device in which accesses are concentrated, the waiting time until each access is processed increases,
Response time deteriorates. As a result, it seems that the response time of the entire disk array device has also deteriorated. In addition, when a secondary storage device in which accesses are sparse at the same time as access concentration occurs, the throughput of the entire disk array device also deteriorates.

[0011]

As described above, each conventional example has advantages and disadvantages, and there is an inconsistent inconvenience that it is not possible to simultaneously improve the throughput and the response time.

[0012]

SUMMARY OF THE INVENTION It is an object of the present invention to provide a disk array device which can solve the disadvantages of the prior art and, in particular, can improve the throughput and the response time at the same time. I do.

[0013]

To achieve the above object, according to the first aspect of the present invention, a plurality of storage devices are linked to form a virtual storage device and a storage device group assigned to the virtual storage device. An address conversion table in which the correspondence between the virtual address and the real address assigned to the storage device is set in advance, and the virtual address of the virtual storage device input from the outside is converted to the real address of the storage device by referring to the address conversion table. Address conversion means for converting and accessing the real address of the storage device. Then, if the specified sequence of the plurality of virtual addresses is in advance clear, it uniform access them for each storage device
Access to the storage device group by a plurality of clients.
In this configuration, a table rewriting means for setting the contents of the address conversion table in advance is provided so that the processes do not overlap the same storage device . Here, the components of the “address” include a block offset, an identifier, and the like.

According to the present invention, the contents of the address conversion table are changed by operating the table rewriting means so that accesses to the storage device group by a plurality of clients do not overlap the same storage device.

According to the second aspect of the invention, the storage device group includes a plurality of virtual storage devices, and the address translation table manages the correspondence between the virtual address and the real address for each virtual storage device. ing. In the present invention,
By operating the table rewriting means, the contents of the address conversion table are optimized for each virtual storage device.

According to a third aspect of the present invention, the address translation table has a configuration in which a correspondence between a virtual address and a real address is defined by a function. In the present invention,
The parameter of the function set in the address conversion table is changed by operating the table rewriting means. The virtual address input from the client is associated with a predetermined real address by a function.

[0017]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to FIG.
It will be described based on.

The disk array device shown in FIG. 1 includes a storage device group 4 in which a plurality of storage devices cooperate to constitute a virtual storage device, a virtual address V assigned to the virtual storage device, and a real address assigned to the storage device. The address conversion table 5 in which the correspondence relationship with the address T is set in advance, and the virtual address V of the virtual storage device input from outside are converted to the real address T of the storage device by referring to the address conversion table 5 and And an address conversion means 2 for accessing the real address T. Further, a table rewriting means 3 capable of freely changing the contents of the address conversion table 5 is provided. Reference numeral 1 denotes a virtual address input unit located at a higher level of the disk array device.

When the virtual address V for accessing the virtual storage device is output from the virtual address input means 1, the address translation means 2 refers to the address translation table 5 and corresponds to the virtual address V. Of the storage device to be specified. Then, access is made to the real address of the storage device.

Here, a plurality of virtual addresses V may be designated by the virtual address input means 1. In such a case, if the real addresses T corresponding to the plurality of designated virtual addresses V access different storage devices, the address conversion means 2 accesses the storage devices almost simultaneously. On the other hand, a plurality of designated virtual addresses V
If the real address T corresponding to the above is an access to the same storage device, one of the virtual addresses is processed, and then the other virtual address is processed. Therefore, the response time decreases.

Therefore, if the designation procedure of a plurality of virtual addresses V is known in advance, the table rewriting means 3 is operated in advance so that they do not access the same storage device. Optimize the content of Thereby, the response time of the disk array device can be improved. In addition, by changing the address conversion table so that all storage devices are accessed uniformly at the same time, it is possible to improve the throughput of the disk array device.

[0022]

Next, a first embodiment of the present invention will be described with reference to FIG.

In FIG. 2, the client 101 corresponds to the virtual address input unit 1 in FIG. The access request issuing unit 104 and the address conversion unit 106 correspond to the address conversion unit 2 in FIG. The data distribution database 107 corresponds to the address conversion table 5 in FIG. The arrangement registration unit 109 corresponds to the table rewriting unit 3 in FIG. The storage devices 11 to 1N (N is a natural number)
This corresponds to the storage device group 4 in FIG.

When operating the disk array device 103,
In the data distribution database 107, arrangement related information of the virtual storage device formed by the disk array device is registered in advance using the arrangement registration unit 109.

The client 101 issues an access request 102 for designating a block offset of a virtual storage device to the disk array device 103. The access request 102 arriving at the disk array device 103 is sent to the access request issuing means 104.

The processing of the access request issuing means 104 is shown in FIG.
Shown in

The access request issuing means 104 waits for the arrival of the access request (S101). When the access request 102 arrives, the process proceeds to S102, where the block offset is extracted from the access request 102, and the address translation unit 10
An address translation request 105 specifying the extracted block offset is issued in step 6. Then, it waits for the completion of the address conversion (S103). When the address translation response 106 is notified by the address translation unit 106, the process proceeds to S104, and the identifier of the storage device in the address translation response 108 is extracted. Then, the storage device corresponding to the identifier is specified from the storage devices 11 to 1N, and an access request 110 specifying the physical block offset in the address translation response 108 is issued to the storage device.

FIG. 4 shows the processing of the address conversion unit 106.

The address conversion means 106 waits for an address conversion request from the access request issuing means 104 (S11).
1). When the address translation request 105 arrives, the process proceeds to S112, where the data distribution database 107 is searched to obtain the placement-related information. Next, based on the placement-related information obtained in S112, the block offset specified by the address conversion request 105 is converted into an identifier of a storage device storing the corresponding data block and a physical block offset on the storage device. (S113). Upon completion of the conversion, the flow shifts to S114, where the access request issuing unit 10
4, the address response 108 is notified by designating the storage device identifier and the physical block offset obtained in S113. The storage devices 11 to 1N store access request issuing means 1
When the access request 110 arrives from the client 10, the access request is processed, and the access response 111 is sent to the client 10.
Return to 1.

FIG. 5 shows arrangement related information registered in the data distribution database 107 in a table format. Each row in the figure corresponds to each entry in the database.

The data distribution database has entries for the number of block offsets (16 in the figure) of the virtual storage device. The entry of each row is a block offset A of the virtual storage device (i (i = 1 to 16) in the figure), an identifier B of the storage device in which the corresponding data is allocated (% i (i = 1 to 1 in the figure)
4)), and each field of a physical block offset C (#i (i = 1 to 4 in the figure)) in which data is stored in the storage device. In this example, the data of block offset 1 of the virtual storage device is stored in physical block offset 1 of storage device 1, the data of block offset 2 is stored in physical block offset 2 of storage device 2, and the data of block offset 10 is stored in storage device 2. It is located at physical block offset 3 of device 4.

The operations of S112 and S113 of the address conversion means 106 in this format are as follows.

First, in the database search (S112), the data distribution database 107 searches for an entry in which the field of the block offset A in the address translation request 105 matches the field of the block offset A. Also,
In the address conversion (S113), the identifier B of the storage device and the value C of the physical block offset field are obtained from the searched entry.

According to this, it is possible to improve the throughput and the response time by variously setting the data arrangement in the disk array device and making the access of the client uniform for each storage device.

Here, the disk array device 103 may be provided with a plurality of access request issuing means 104 and address conversion means 106. In such a case, not only the same effect as in the first embodiment can be obtained, but also there is an advantage that a plurality of access requests can be processed simultaneously.

Next, a second embodiment of the present invention will be described with reference to FIG.

In FIG. 6, the client 101 corresponds to the virtual address input means 1 in FIG. The file access request issuing unit 120 and the file address conversion unit 123 correspond to the address conversion unit 2 in FIG. The data distribution database 124 corresponds to the address conversion table 5 in FIG. The file arrangement registration unit 126 corresponds to the table rewriting unit 3 in FIG. Storage devices 11 to
1N (N is a natural number) corresponds to the storage device group 4 in FIG.

When operating the disk array device 103,
In the file data distribution database 125, the arrangement related information of each of the plurality of virtual storage devices formed by the disk array device is registered in advance by using the file arrangement registration unit 126.

The client 101 sends a file access request 121 specifying the identifier of the virtual storage device and the block offset in the virtual storage device to the disk array device 1.
Issue on 03. The file access request 121 arriving at the disk array device 103 is sent to the file access request issuing means 120.

FIG. 7 shows the processing of the file access request issuing means 120.

The file access request issuing means 120
It waits for the arrival of the file access request (S121). When the file access request 121 arrives, the process proceeds to S122, where the identifier of the virtual storage device and the block offset are extracted from the file access request 121, and a file address conversion request 122 specifying them is issued to the file address conversion means 123. Then, it waits for the completion of the address conversion (S123). When the file address conversion response 125 is notified from the file address conversion means 123, the process proceeds to S124, where the file address conversion response 1
An access request 110 specifying the physical block offset in the file address conversion response 125 is issued to the storage device corresponding to the storage device identifier in the storage device 25.

FIG. 8 shows the processing of the file address conversion means 123.

The file address conversion means 123 waits for a file address conversion request from the file access request issuing means 120 (S131). When the file address conversion request 122 arrives, the process proceeds to S132, where the search is performed from the distribution related information file data distribution database 124 of the virtual storage device designated by the file address conversion request 122. Next, the storage-related information obtained in S132 is converted into an identifier of a storage device corresponding to the block offset specified by the file address conversion request 122 and a physical block offset on the storage device (S133).
When the conversion is completed, the process moves to S134, where the file access request issuing means 120 is designated with the storage device identifier and the physical block offset obtained in S133, and the file address conversion response 125 is notified. Storage devices 11 to 1
N receives the access request 110 from the file access request issuing unit 120, processes the requested access, and returns an access response 111 to the client 101.

FIG. 9 shows the format of the placement-related information registered in the file data distribution database 124. In the present embodiment, the data distribution database 1 in the first embodiment is used.
07 is provided for each virtual storage device.
That is, in this example, the file data distribution database 12
4 holds a plurality of tables simultaneously, and each table corresponds to each virtual storage device.

In this example, the data of the block offset 1 of the virtual storage device 1 is stored in the physical block offset 1 of the storage device 1, the data of the block offset 2 is stored in the physical block offset 2 of the storage device 2, and the data of the block offset 10 is stored. The data is located at the physical block offset 3 of the storage device 4, and the data at the block offset 1 of the virtual storage device 2 is stored at the physical block offset 5 of the storage device 3.
The data of the block offset 2 is located at the physical block offset 6 of the storage device 2, the data of the block offset 10 is located at the physical block offset 7 of the storage device 4, and the data of the block offset 1 of the virtual storage device 3 is The data at the physical block offset 12 of the storage device 4, the data at the block offset 2 are allocated at the physical block offset 10 of the storage device 3, and the data at the block offset 10 are allocated at the physical block offset 9 of the storage device 3.

The operations of S132 and S133 of the file address conversion means 123 in this format are as follows.

In searching the database (S132), the identifier of the virtual storage device is extracted from the file address conversion request 122, and the table of the virtual storage device that matches this identifier is set as the search target. Then, the block offset specified by the file address conversion request 122 and the field A
Search for an entry that matches. In the address conversion (S133), the identifier B of the storage device and the value of the field of the physical block offset C are obtained from the entry searched in S132.

FIG. 10 shows a modification of the table format in the second embodiment.

The file data distribution database has entries for the number of block offsets (16 in the figure) of all virtual storage devices. Each entry is an identifier D of the virtual storage device.
(＄ i (i = 1 to 4) in the figure), the block offset A of the virtual storage device (i (i = 1 to 4) in the diagram), the identifier B of the storage device in which the corresponding data is located (% in the figure) i (i = 1
4)), each field of a physical block offset C (#i (i = 1 to 4) in the figure) in which data is stored in the storage device.

In this example, the data at the block offset 1 of the virtual storage device 1 is located at the physical block offset 1 of the storage device 1, and the data at the block offset 2 is located at the physical block offset 3 of the storage device 2. The data of block offset 1 of device 2 is stored in storage device 2
Block offset 3 to physical block offset 2
Is placed at the physical block offset 1 of the storage device 2, the data at the block offset 2 of the virtual storage device 3 is placed at the physical block offset 2 of the storage device 4, and the data at the block offset 4 is placed at the physical block offset 1 of the storage device 1. The data at the block offset 3 of the virtual storage device 4 is located at the physical block offset 4 of the storage device 4, and the data at the block offset 4 is located at the physical block offset 1 of the storage device 4.

The operations of S132 and S133 of the file address conversion means 123 in this format are as follows.

In the database search (S 132), the identifier of the virtual storage device and the block offset are extracted from the file address conversion request 122, and an entry that matches these is searched from the file data distribution database 124. In the address conversion (S133), the identifier B of the storage device and the value of the field of the physical block offset C are obtained from the searched entry.

According to this, since the data arrangement of a plurality of virtual storage devices constituting the disk array device can be set variously for each virtual storage device, the access of the client to each storage device can be made uniform, thereby improving the throughput. And the response time can be improved.

Here, in the second embodiment, the disk array device 103 can have a plurality of file access request issuing means 120 and a plurality of file address conversion means 123. In this case, the same effects as those of the above embodiment can be obtained, and a plurality of file access requests can be processed simultaneously.

Next, a third embodiment of the present invention will be described with reference to FIG.
This will be described with reference to FIG. The configuration of this embodiment is the same as that of the above-described first embodiment except for the processing of the data distribution database 130, the address conversion unit 131, and the arrangement registration unit 132.

When operating the disk array device 103,
In the data distribution database 130, an arrangement function of a virtual storage device formed by the disk array device is registered in advance using the arrangement registration unit 132.

The allocation function of the virtual storage device is f (x), g
It consists of two equations (x). f (x) is a function that returns the identifier of the storage device in which the data block corresponding to the block offset x of the virtual storage device is stored. Also,
g (x) is a function that returns a physical block offset in the storage device corresponding to the block offset x of the virtual storage device.

FIG. 12 shows the processing of the address conversion means 131.

The address conversion means 131 waits for an address conversion request from the access request issuing means 104 (S14).
1). When the address translation request 105 arrives, the process proceeds to S142, where the data distribution database 130 is searched to obtain an allocation function. Next, assuming that the block offset specified by the address translation request 105 is x, equation (2) is calculated to obtain the identifier y of the storage device storing the corresponding data block (S143).

Y = f (x) (1)

Next, the block offset specified by the address translation request 105 is x, and the equation (2) is calculated.
The physical block offset z of the corresponding data block in the storage device is obtained (S144).

Z = g (x) (2)

When the conversion is completed, the flow shifts to S145, where the access request issuing means 104 is notified of the address conversion response 108 specifying the storage device identifier y and the physical block offset z obtained by the above-described operation. here,
It is also possible to perform the processing of S143 and S144 in parallel.

FIG. 13 shows the format of the placement-related information registered in the data distribution database 130.

In this format, the disk array device has several types of layout function templates in advance. #I in the figure
(I = 1) is the identifier E of the function template used as the placement function. ＄ i (i = 1, 2,...)
n) is a parameter F of the function template. In this format, a table storing a function template identifier E and a parameter F is used as a data distribution database, and various arrangement functions are realized by using a plurality of function templates E and parameters F of the function template E.

The address conversion means 131 in this format
The operations of S142, S143, and S144 are as follows.

In the database search (S142), the data distribution database 130 is searched for the identifier E and the parameter F of the function template. In the storage device identifier calculation (S143), the parameter f and the physical block offset in the address translation request 105 searched in S142 are substituted into the function f (x) corresponding to the function template identifier E searched in S142, and the calculation is performed.

In the storage device identifier calculation (S144), S
The function g (x) corresponding to the function template identifier E retrieved in 142 is substituted with the parameter F retrieved in S142 and the physical block offset in the address translation request 105 to perform the operation.

In this format, the placement function is registered in the data distribution database 130 by the placement registering means 132 as an identifier E and a parameter F of a function template of the placement function to be set. Expressions (3) and (4) are function templates of the placement function registered in the data distribution database 130. Where mod
(M, n) is a function that returns the remainder of dividing m by n, a, b, c, and d are parameters for determining the arrangement, and D is the number of storage devices constituting the disk array device. is there.

F (x) = mod (ax + b, D) (3)

G (x) = c [x / D] + d (4)

As another function template of the placement function registered in the data distribution database 130, there are Expressions (5) and (6). Here, am, A, and B are
D is a parameter for determining the arrangement, and D is the number of storage devices constituting the disk array device.

F (x) = mod (ax + b, D) (x <A) mod (cx + d, D) (A ≧ x <B) (5) mod (ex + f, D) (B <x)

G (x) = h [x / D] + i (x <A) j [x / D] + k (A ≧ x <B) (6) l [x / D] + m (B < x)

In this way, it is possible to reduce the storage capacity of the data distribution database from the table format of the first embodiment.

Next, a fourth embodiment of the present invention will be described with reference to FIG.
This will be described with reference to FIG.

The configuration of the present embodiment includes a file data distribution database 140, a file address conversion means 141,
This is the same as the above-described second embodiment except for the processing of the arrangement registration unit 142.

When operating the disk array device 103,
Each allocation function of a plurality of virtual storage devices formed by the disk array device is registered in advance in the file data distribution database 140 using the file allocation registration unit 142.

The allocation function of the virtual storage device is f (x), g
It consists of two equations (x). f (x) is a function that returns the identifier of the storage device in which the data block corresponding to the block offset x of the virtual storage device is stored. Also,
g (x) is a function that returns the physical block offset in the storage device of the data block corresponding to the block offset x of the virtual storage device.

FIG. 15 shows the processing of the address conversion means 141.

The address conversion means 141 waits for a file address conversion request from the file access request issuing means 120 (S151). File address conversion request 12
When S2 arrives, the process proceeds to S152, in which the file data distribution database 140 is searched to obtain an allocation function of the virtual storage device corresponding to the identifier specified by the file address conversion request 122. Next, the block offset specified by the file address conversion request 120 is x,
Equation (7) is calculated to obtain an identifier y of the storage device in which the corresponding data block is stored (S153).

Y = f (x) (7)

Next, the file address conversion request 120
The block offset specified by is set to x, and equation (8) is calculated to obtain the physical block offset z of the corresponding data block in the storage device (S154).

Z = g (x) (8)

When the conversion is completed, the flow shifts to S155, where the file access request issuing means 120 is notified of the file address conversion response 125 specifying the storage device identifier y and the physical block offset z obtained by the above operation. Here, the processing of S153 and S154 can be performed in parallel.

FIG. 16 shows the format of the placement-related information registered in the file data distribution database 140.

In this format, the disk array device has several types of layout function templates in advance. Each row in the figure corresponds to each entry in the database.

Each entry has an identifier D of the virtual storage device.
(% I (i = 1 to m) in the figure) is included. Also, #i (i = 1 to m) in the figure is an identifier E of the function template used as the placement function. Further, in FIG.
(I = 11 to mn) are parameters F of the function template.

In this format, a table storing function template identifiers E and parameters F is used as a data distribution database, and various arrangement functions are realized by using a plurality of function templates E and parameters of function templates F.

The operations of S152, S153, and S154 of the file address conversion means 141 in this format are as follows.

In the database search (S152), an entry corresponding to the virtual storage device identifier in the file address conversion request 122 is searched from the file distribution database 140. In the storage device identifier calculation (S153), the parameter F searched in S152 and the physical block offset in the file address conversion request 122 are substituted into the function f (x) corresponding to the function template identifier E searched in S152, and the calculation is performed. Execute In the storage device identifier calculation (S154), the function g (x) corresponding to the function template identifier E searched in S152 is added to the parameter F and the file address conversion request 1 searched in S152.
The operation is performed by substituting the physical block offset in 22.

In this format, the operation of registering the allocation function in the file data distribution database 140 by the file allocation registration unit 142 is performed by registering the identifier D of the virtual storage device to be set, the identifier E of the function template of the allocation function, and the parameter F. Done.

The function template of the placement function registered in the file data distribution database 140 is the third function template described above.
This is the same as the embodiment.

In this manner, the storage capacity of the file data distribution database can be reduced as compared with the table format of the second embodiment.

[0095]

Since the present invention is constructed and functions as described above, according to the present invention, since the table rewriting means is provided in the address conversion table for associating the virtual addresses with the real addresses, the access status of the storage device group is improved. The address conversion table is optimized in advance in accordance with the above, and it is possible to simultaneously avoid concentration of accesses to a specific storage device and simultaneous access to the same storage device, thereby simultaneously improving the throughput and the response time. Is possible.

According to the second aspect of the present invention, since the address conversion table is managed for each virtual storage device, the contents of the table can be changed quickly and accurately without error.

According to the third aspect of the present invention, since the correspondence between the virtual address and the real address is defined by a function, the correspondence between a plurality of addresses can be changed at once by changing the function format and parameters. Can be. Therefore, the address correspondence can be changed quickly. Further, it is possible to provide an unprecedented excellent disk array device in which the storage capacity of the memory for storing the address change table can be greatly reduced.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of the present invention.

FIG. 2 is a block diagram of a first embodiment of the present invention.

FIG. 3 is a flowchart showing a flow of processing of an access request issuing unit in the first embodiment.

FIG. 4 is a flowchart showing a flow of processing of an address translation unit in the first embodiment.

FIG. 5 is a configuration diagram of arrangement-related information on a data distribution database in the first embodiment.

FIG. 6 is a block diagram of a second embodiment.

FIG. 7 is a flowchart showing a processing flow of a file access request issuing unit in a mud 2 embodiment;

FIG. 8 is a flowchart showing a flow of processing of a file address conversion means in the second embodiment.

FIG. 9 is a configuration diagram of arrangement-related information on a file data distribution database in a second embodiment.

FIG. 10 is a configuration diagram of arrangement-related information on a file data distribution database according to a modification of the second embodiment.

FIG. 11 is a block diagram of a third embodiment.

FIG. 12 is a flowchart showing a flow of processing of an address conversion unit in the third embodiment.

FIG. 13 is a configuration diagram of arrangement-related information on a data distribution database in a third embodiment.

FIG. 14 is a block diagram of a fourth embodiment.

FIG. 15 is a flowchart showing a flow of processing of a file address conversion means in the fourth embodiment.

FIG. 16 is a configuration diagram of arrangement-related information on a file data distribution database in a fourth embodiment.

FIG. 17 is an explanatory diagram showing a first conventional method of dividing a disk array device into a plurality of virtual storage devices.

FIG. 18 is an explanatory diagram showing a second conventional method of dividing a disk array device into a plurality of virtual storage devices.

[Description of sign]

 DESCRIPTION OF SYMBOLS 1 Virtual address input means 2 Address conversion means 3 Table rewriting means 4 Storage device group 5 Address conversion table T Real address V Virtual address

Claims (3)

(57) [Claims] A storage device group in which a plurality of storage devices cooperate to form a virtual storage device, and a correspondence relationship between a virtual address assigned to the virtual storage device and a real address assigned to the storage device is determined in advance. A set address translation table, address translation means for translating a virtual address of the virtual storage device input from the outside into a real address of the storage device by referring to the address translation table, and accessing the real address of the storage device; In the disk array device provided with, when the designation order of the plurality of virtual addresses is known in advance, they become uniform access to each storage device, and the plurality of clients access the storage device group.
A table rewriting means for setting the contents of an address conversion table in advance so that accesses of the same type do not overlap with the same storage device. 2. The storage device group includes a plurality of virtual storage devices, and the address conversion table manages the correspondence between the virtual addresses and the real addresses for each virtual storage device. 2. The disk array device according to 1. 3. The disk array device according to claim 1, wherein the address conversion table defines a correspondence between the virtual address and the real address by a function.