Data reading method

1. Technical Field

This technique relates to a disk array apparatus that includes multiple hard disk drives constituting a RAID (redundant arrays of inexpensive disks) group for creating parity for each stripe and controls access requests to hard disk drives made by the host, a disk array apparatus control method, a disk array apparatus control program, and a disk array controller. In particular, the technique relates to a disk array apparatus, a disk array apparatus control method, a disk array apparatus control program, and a disk array controller that each improve the capability to perform read processes on hard disk drives using a simple process without running out of disk space.

2. Description of the Related Art

Large-scale computer systems have managed data using a dedicated data input/output unit provided independently of the host computer. Among such data input/output units, a disk array apparatus includes multiple hard disk drives (storage media), which constitute a RAID group. Thus, the reliability of data handled by the apparatus as well as the apparatus's capability to access the hard disk drives is improved.

RAID is classified into RAID levels corresponding to the levels of rapidity or fault tolerance. One of RAID levels being used commonly at present is RAID 5 . RAID 5 is intended to switch between hard disk drives that are assigned the storage of an error correcting code called “parity” and hard disk drives that are assigned the storage of data, for each stripe. Thus, a disk array apparatus in which RAID 5 is implemented improves fault tolerance, increases the capacity, and speeds up read processes.

As the amounts of data handled by computer systems are increased in recent years, the amounts of data held by disk array apparatuses have also been increased. Such a disk array apparatus is often accessed intensively. Although RAID 5 allows a disk array apparatus to speed up read processes, the apparatus's capability to perform read processes is naturally reduced if the disk array apparatus is intensively accessed. As the information technology progresses day by day, computer systems including a disk array apparatus are always required to enhance their capabilities. Under the circumstances, disk array apparatuses are required to sufficiently demonstrate their capabilities even if they are intensively accessed.

In view of the foregoing, a technology regarding an improvement in a capability of a disk array apparatus is disclosed in Laid-open Japanese Patent Application Publication No. 2003-150324. Specifically, a technology for improving the capability to perform read processes by providing a swap area in each RAID group and saving a frequently accessed data block in the swap area is disclosed in Japanese Laid-open Patent Publication No. 2003-150324.

However, the technology disclosed in Japanese Laid-open Patent Publication No. 2003-150324 has a problem that the disk space is pressed and a problem that a complicated process must be performed. Specifically, the technology disclosed in Japanese Laid-open Patent Publication No. 2003-150324 has a problem that since a swap area is provided for each RAID group, the disk space for storing user data is reduced and a problem that since the access frequency is monitored periodically by counting the number of accesses for each data block, the number of processes is increased so that the processing capability as a whole may be reduced.

In view of the foregoing, a significant challenge to a disk array apparatus in which RAID 5 being used commonly at present is implemented is to improve the capability to perform read processes using a simple process without running out of disk space.

According to an aspect of an embodiment, a method for controlling a controller connected to a plurality of storage units which are arranged in a redundant configuration, the controller reading data stored in the plurality of storage units in accordance with requests received from a host computer, the method comprising the steps of: receiving requests to read data successively from one of the storage units from the host computer; reading a part of requested target data from said one of the storage units; reading associated data and parity data stored in other storage units corresponding to other part of requested target data; generating other part of requested target data on the basis of the associated data and the parity data read out from the other storage units; and transmitting the part of the target data and the other part of the target data to the host computer.

FIG. 1 is a diagram showing an example of read processes performed by a related-art disk array apparatus;

FIG. 2 is a diagram showing an example of read processes performed by a disk array apparatus according to an embodiment;

FIG. 3 is a diagram showing an example of read processes performed by the related-art disk array apparatus;

FIG. 4 is a diagram showing an example of read processes performed by the disk array apparatus according to this embodiment;

FIG. 5 is a functional block diagram showing a configuration of the disk array apparatus according to this embodiment;

FIG. 6 is a flowchart showing steps of a read process performed by the disk array apparatus according to this embodiment;

FIG. 7 is a diagram showing an example of read processes performed by the related-art disk array apparatus in which RAID 6 is constituted;

FIG. 8 is a diagram showing an example of read processes performed by the disk array apparatus according to this embodiment in which RAID 6 is constituted;

FIG. 9 is a diagram showing an example of read processes performed by the related-art disk array apparatus in which RAID 6 is constituted;

FIG. 10 is a diagram showing an example of read processes performed by the disk array apparatus according to this embodiment in which RAID 6 is constituted; and

FIG. 11 is a diagram showing a computer for executing a disk array apparatus control program.

Now, a disk array apparatus, a disk array apparatus control method, a disk array apparatus control program, and a disk array controller according to an embodiment will be described in detail with reference to the accompanying drawings. First, read processes performed by a related-art disk array apparatus and read processes performed by the disk array apparatus according to this embodiment will be described using two examples for each apparatus. Then, a configuration of the disk array apparatus according to this embodiment and process steps performed by the disk array apparatus will be described.

Embodiment

In order to clarify the features of the disk array apparatus according to this embodiment, read processes performed by the related-art disk array apparatus that has received multiple read requests to obtain data from an identical hard disk drive will be described. FIG. 1 is a diagram showing an example of read processes performed by the related-art disk array apparatus.

FIG. 1 shows hard disk drives 11 to 14 included in the related-art disk array apparatus. The hard disk drives 11 to 14 constitute RAID 5 . The storage areas of the hard disk drives 11 to 14 are each divided into multiple logic blocks serving as units from or to which the host reads or writes data. In an example shown in FIG. 1 , it is assumed that data D 11 , data D 12 , and parity P 10 are each one logic block. The example shows that data is stored in logic blocks starting with “D,” such as data D 11 and data D 12 , and parity is stored in logic blocks starting with “P,” such as parity P 10 and parity P 20 .

Also, the hard disk drives 11 to 14 physically successively store data and parity in the order shown in FIG. 1 . For example, the hard disk drive 11 physically successively stores data D 11 to D 31 , parity P 40 , data D 51 to D 71 , parity P 80 , and data D 91 .

As shown in FIG. 1 , RAID 5 divides the hard disk drives into a hard disk drive that is assigned the storage of parity and hard disk drives that are assigned the storage of data, for each stripe. For example, in the first stripe (hereafter, an n-th stripe will be referred to as “stripe N”), the hard disks units 11 to 13 are assigned the storage of data and the hard disk drive 14 is assigned the storage of parity. In stripe 2 , the hard disks units 11 , 12 , and 14 are assigned the storage of data and the hard disk drive 13 is assigned the storage of parity.

Parity is created by performing an exclusive OR operation (XOR) on pieces of data stored in other hard disk drives in an identical stripe. For example, parity P 10 is created by performing an exclusive OR operation on data D 11 , D 12 , and D 13 . Thus, even if one of the hard disk drives constituting RAID 5 suffers a failure, data stored in the failed hard disk drive is restored by previously storing parity created in the above-described way and operating an exclusive OR operation on other pieces of data stored in the same stripe and the parity. If the related-art disk array apparatus configured as described above receives an access request for data read or data write made by the host as a higher-order unit, the disk array apparatus executes the access request. For example, if the related-art disk array apparatus receives read requests from the host, the disk array apparatus obtains data stored in the hard disk drives 11 to 14 according to the read requests and then transmits the obtained data to the host.

At that time, depending on the contents of the read requests received from the host, the disk array apparatus may have to obtain pieces of data only from an identical hard disk drive. One of such cases is a case where, in the example shown in FIG. 1 , the disk array apparatus successively receives a read request to obtain data D 1 , a read request to obtain data D 31 , a read request to obtain data D 61 , and a read request to obtain data D 91 from the host. Note that, in FIG. 1 , logic blocks from which the disk array apparatus is to obtain data are shown using oblique lines.

The disk array apparatus, in which RAID 5 is implemented, typically obtains pieces of data in a parallel manner from the hard disk drives 11 to 14 . As a result, read processes are speeded up. However, as shown in FIG. 1 , if the disk array apparatus successively receives read requests to obtain pieces of data from the hard disk drive 11 , it cannot read the pieces of data in a parallel manner. This causes a problem that the apparatus's capability to perform read processes is reduced.

On the other hand, if the disk array apparatus according to this embodiment successively receives read requests to obtain data from an identical hard disk, the disk array apparatus performs control such that it obtains pieces of data from other hard disk drives in order to obtain target data stored in the identical hard disk drive, that is, it performs read processes in a parallel manner.

FIG. 2 is a diagram showing an example of read processes performed by the disk array apparatus according to this embodiment. In the example shown in FIG. 2 , like in FIG. 1 , it is assumed that multiple read requests received from the host are requests to obtain data D 11 , D 31 , D 61 , and D 91 stored in the hard disk drive 11 .

As shown in FIG. 2 , when the disk array apparatus according to this embodiment executes the first read request (a read request to obtain data D 11 ), it obtains the data 12 , D 13 and parity P 10 from stripe 1 of the hard disk drives 12 to 14 . As described above, the hard disk drives 11 to 14 constitute RAID 5 . Therefore, by performing an exclusive OR operation on data D 12 , D 13 and parity P 10 , the disk array apparatus obtains data D 11 .

When the disk array apparatus according to this embodiment executes a subsequent read request (a read request to obtain data D 31 ), it obtains data D 31 from stripe 3 of the hard disk drive 11 . Further, when the disk array apparatus executes a subsequent read request (a read request to obtain data D 61 ), it obtains data D 62 , parity P 60 , and data D 63 from stripe 6 of the hard disk drives 12 to 14 and performs an exclusive OR operation on the obtained data D 62 , parity P 60 , and data D 63 . Thus, the disk array apparatus obtains data D 61 . Further, when the disk array apparatus executes a subsequent read request (read request to obtain data D 91 ), it obtains data D 91 from stripe 9 of the hard disk drive 11 .

As is understood from the above-description, if the disk array apparatus according to this embodiment successively receives read requests to obtain data from an identical hard disk drive, it alternately performs read processes in which data is obtained by performing an exclusive OR operation and a read process in which data is obtained as specified by a read request each time the disk array apparatus executes one of the read requests. Thus, as shown in FIG. 2 , the respective frequencies with which the hard disk drives 11 to 14 are accessed are equalized. As a result, the access load imposed on a particular hard disk drive is distributed. Also, the disk array apparatus obtains pieces of data in a parallel manner from the hard disk drives 11 to 14 ; therefore, the apparatus's capability to perform read processes is improved.

In this embodiment, each time the disk array apparatus executes “one” read request, it alternately performs read processes in which data is obtained by performing an exclusive OR operation and a read process in which data is obtained as specified by a read request. Alternatively, each time the disk array apparatus executes “multiple” read requests, it may perform read requests while switching between read processes in which data is obtained by performing an exclusive OR operation and read processes in each of which data is obtained as specified by a read request.

While read processes according to this embodiment are applied if the disk array apparatus receives the read request to obtain data D 11 , read request to obtain data D 31 , read request to obtain data D 61 , and read request to obtain data D 91 in this order, read processes according to this embodiment are applied even if the disk arrays apparatus successively receives read requests to obtain pieces of data from an identical hard disk drive in any order.

Hereafter, read processes performed by the related-art disk array apparatus that has received read requests to obtain pieces of data from multiple hard disk drives will be described. FIG. 3 is a diagram showing an example of read processes performed by the related-art disk array apparatus. In FIG. 3 , like in FIG. 1 , the related-art disk array apparatus includes the hard disk drives 11 to 14 , which constitute RAID 5 . FIG. 3 shows an example in which the disk array apparatus has received multiple read requests to obtain all pieces of data stored in stripes 1 to 9 of the hard disk drives 11 to 14 . Specifically, the disk array apparatus has successively received a read request to obtain data D 11 , a read request to obtain data D 12 , a read request to obtain data D 13 , . . . , and a read request to obtain data D 93 . If data with a size larger than the data capacity of a logic block is stored in the hard disk drives 11 to 14 , the disk array apparatus often successively receives read requests to obtain all pieces of data stored in a particular range of stripes as described above.

If the disk array apparatus receives read requests as described above, the respective frequencies with which the disk array apparatus accesses the hard disk drives 11 to 14 are equal to one another, as shown in FIG. 3 . Thus, the access load is distributed. However, the disk array apparatus is allowed to access pieces of data physically successively stored in each hard disk drive with respect to only three stripes.

In the example shown in FIG. 3 , the disk array apparatus obtains data D 11 , D 21 , and D 31 stored in stripes 1 to 3 with respect to the hard disk drive 11 , but it does not obtain parity P 40 . Subsequently, the disk array apparatus obtains data D 51 , D 61 , and D 71 stored in stripes 5 to 7 . That is, the disk array apparatus accesses only pieces of data corresponding to a maximum of three strips with respect to the hard disk drives 11 to 14 . If the disk array apparatus accesses pieces of data that are not physically successively stored, the hard disk drives 11 to 14 may have to re-calculate a seek process. This causes a problem that the apparatus's capability to perform read processes is reduced.

On the other hand, if the disk array apparatus according to this embodiment successively receives read requests to obtain all pieces of data stored in a particular range of stripes, it performs control such that the respective frequencies with which the disk array apparatus accesses the hard disk drives 11 to 14 are equalized and such that the respective frequencies with which the disk array apparatus may access the hard disk drives 11 to 14 are increased.

FIG. 4 is a diagram showing an example of read processes performed by the disk array apparatus according to this embodiment. In the example shown in FIG. 4 , like in FIG. 3 , it is assumed that the disk array apparatus has received read requests to obtain all pieces of data stored in strips 1 to 9 of the hard disk drives 11 to 14 .

In this case, as shown in FIG. 4 , the disk array apparatus according to this embodiment performs control such that it obtains pieces of data from the hard disk drives 12 to 14 with respect to stripes 1 to 3 , obtains pieces of data from the hard disk drives 11 , 13 , and 14 with respect to stripes 4 to 6 , and obtains pieces of data from the hard disk drives 11 , 12 , and 14 with respect to stripes 7 to 9 .

Then, the disk array apparatus performs an exclusive OR operation on data D 12 , D 13 , and parity P 10 obtained from the hard disk drives 12 to 14 with respect to stripe 1 . Thus, the disk array apparatus obtains data D 11 , and then transmits data D 11 , D 12 , and D 13 to the host. As such, the disk array apparatus performs exclusive OR operations with respect to stripes 2 to 9 . Thus, the disk array apparatus obtains pieces of data specified by the read requests, and then transmits the obtained pieces of data to the host.

As a result, the number of stripes with respect to which the disk array apparatus according to this embodiment successively accesses one hard disk drive is increased. In the example shown in FIG. 4 , the disk array apparatus successively accesses the hard disk drive 14 with respect to nine stripes. In this case, the frequency with which the disk array apparatus according to this embodiment accesses one hard disk drive is “triple” that in the related-art example shown in FIG. 3 . Also, if the disk array apparatus accesses pieces of data with respect to stripe 9 and later in the same way, it successively accesses the hard disk drives 11 to 13 with respect to nine stripes, although not shown in FIG. 4 . Thus, the respective frequencies with which the disk array apparatus successively accesses the hard disk drives 11 to 14 are increased so that the read processes performed by the disk array apparatus are speeded up.

Unlike in the read processes performed by the related-art disk array apparatus shown in FIG. 1 and FIG. 3 , the read processes performed by the disk array apparatus according to this embodiment involve the exclusive OR operations. Therefore, it can be considered that the processing load imposed on a central processing unit (CPU) in the disk array apparatus is increased. However, the additional processing load imposed on the CPU due to performance of the exclusive OR operations is small. Therefore, even if such an additional processing load is imposed on the CPU, the apparatus's capability to perform read processes is improved by distributing the access load imposed on each hard disk drive and increasing the frequency with which the disk array apparatus successively accesses each hard disk drive.

As is understood from the above-description, the disk array apparatus according to this embodiment performs read processes while performing control such that if the disk array apparatus successively receives read requests to obtain pieces of data from an identical hard disk drive, the disk array apparatus obtains pieces of data from other hard disk drives in order to obtain target data stored in the identical hard disk drive and such that if the disk array apparatus successively receives read requests to obtain all pieces of data stored in a particular range of stripes, the respective frequencies with which the disk array apparatus accesses the hard disk drives are equalized and the respective frequencies with which the disk array apparatus successively access the hard disk drives are increased. As a result, the apparatus's capability to perform read processes on the hard disk drives constituting RAID 5 is improved. Also, when the disk array apparatus according to this embodiment performs read processes, it does not additionally provide a save area such as a swap area. Therefore, the hard disk drives do not run out of disk space. Further, when the disk array apparatus according to this embodiment performs read processes, it only controls logic blocks from which the disk array apparatus is to obtain data. With such a simple process, the apparatus's capability to perform read processes is improved.

Next, the disk array apparatus according to this embodiment will be described. FIG. 5 is a functional block diagram showing a configuration of the disk array apparatus according to this embodiment. As shown in FIG. 5 , a disk array apparatus 300 is coupled to a host 400 and includes a DE (device enclosure) 10 and a CM (controller module) 310 . While an example in which the disk array apparatus 300 includes only one CM 310 is shown in FIG. 5 , the disk array apparatus may have a redundant configuration in which it includes two or more CMs. Also, in the example shown in FIG. 5 , the disk array apparatus 300 includes only one DE; however, the number of DEs is not limited to one.

The host 400 is coupled to the disk array apparatus 300 via CHs (channel) 410 a and 410 b and transmits a request to read or write data to the disk array apparatus 300 .

The DE 10 is an apparatus including multiple hard disk drives, that is, the hard disk drives 11 to 14 . While the DE 10 includes four hard disk drives in the example shown in FIG. 5 , the number of hard disk drives that the DE includes is not limited to four.

The hard disk drives 11 to 14 are storage devices such as magnetic disks. In the example shown in FIG. 5 , it is assumed that the hard disk drives 11 to 14 constitute RAID 5 . Note that the hard disk drives are not limited to magnetic disk drives and may be other storage devices such as thermomagnetic disk drives or optical magnetic disk drives.

Since the hard disk drives 11 to 14 constitute RAID 5 , the host 400 makes an access request while recognizing the hard disk drives 11 to 14 not as individual hard disk drives but as one logic disk. For example, if the host 400 transmits a read request to the disk array apparatus 300 , it specifies a logic block number that is the number of a logic block (also referred to as a “logic block address” or “logic address”) in which target data is stored.

The CM 310 includes CAs (channel adapters) 311 a and 311 b , DAs (device adapters) 312 a and 312 b , a read request reception unit 313 , and a storage unit 314 , and a controller 315 .

The CAs 311 a and 311 b are adapters for controlling communications between the disk array apparatus 300 and the host 400 as a higher-order apparatus and are coupled to the CHs 410 a and 410 b . DAs 312 a and 312 b are adapters for controlling communications between the CM 310 and DE 10 .

The read request reception unit 313 is a storage device for sequentially storing read requests received from the host 400 and is, for example, a queue. The controller 315 to be described later takes out a read request stored in the read request reception unit 313 and performs a process corresponding to the read request.

The storage unit 314 is a storage device such as a hard disk or a nonvolatile memory and includes an address conversion table 314 a related to this embodiment. The address conversion table 314 a stores physical block numbers indicating the physical storage positions of pieces of data stored in the hard disk drives 11 to 14 in such a manner that the physical block numbers are associated with the logic block numbers of the hard disk drives 11 to 14 .

The controller 315 is a controller for controlling the CM 310 as a whole and includes a read command creation unit 315 a and a read control unit 315 b related to this embodiment.

The read command creation unit 315 a is a processor for analyzing read requests stored in the read request reception unit 313 and creating read commands according to the read requests. Specifically, the read command creation unit 315 a reads read requests stored in the read request reception unit 313 . Then, the read command creation unit 315 a obtains physical block numbers corresponding to logic block numbers included in the read requests from the address conversion table 314 a.

Then, the read command creation unit 315 a determines whether there are two or more successive read requests to obtain pieces of data from an identical hard disk drive, according to the obtained physical block numbers. If there are such read requests, the read command creation unit 315 a alternately creates read commands to obtain data by performing an exclusive OR operation and a read command to obtain data as specified by a read request with respect to each of the read requests.

Hereafter, the example shown in FIG. 2 will be used for explanation. If the read command creation unit 315 a determines that the four read requests to obtain data D 11 , D 31 , D 61 , and D 91 are successive, first, it creates three read commands to obtain data D 12 and D 13 and parity P 10 . Then, the read command creation unit 315 a outputs the created three read commands to the read control unit 315 b and instructs the read control unit 315 b to execute the three read commands and then perform an exclusive OR operation on the obtained data D 12 and D 13 and parity P 10 so as to obtain data D 11 .

Subsequently, the read command creation unit 315 a creates a read command to obtain data D 31 , outputs the created read command to the read control unit 315 b , and instructs the read control unit 315 b to obtain data D 31 . Also, the read command creation unit 315 a creates three read commands to obtain data D 62 , parity 60 , and data D 63 , outputs the created three read commands to the read control unit 315 b , and instructs the read control unit 315 b to obtain data D 61 . Further, the read command creation unit 315 a creates a read command to obtain data D 91 , outputs the created the read command to the read control unit 315 b , and instructs the read control unit 315 b to obtain data D 91 .

On the other hand, if there are no two or more successive read requests to obtain pieces of data from an identical hard disk drive, the read command creation unit 315 a determines whether there are successive multiple read requests to obtain all pieces of data stored in a particular range of stripes.

Here, “a particular range of stripes” refers to a range of successive stripes such as stripes 1 to 9 shown in FIG. 4 . That is, the multiple read requests to obtain all pieces of data stored in a particular range of stripes refer to multiple read requests to obtain all pieces of data physically successively stored in the hard disk drives 11 to 14 .

If there are such successive read requests, the read command creation unit 315 a creates read commands such that the respective frequencies with which the hard disk drives 11 to 14 are accessed are made equal and the respective frequencies with which the hard disk drives 11 to 14 are successively accessed are increased.

Specifically, first, the read command creation unit 315 a determines the “unit number of read requests to be processed” used when processing the multiple read requests to obtain all pieces of data stored in a particular range of stripes.

The unit number of read requests to be processed is determined by multiplying a value (hereafter referred to as a “data storage number”) obtained by subtracting the “number of hard disk drives storing parity in one stripe” from the “number of hard disk drives constituting a RAID group,” by a value obtained by subtracting “1” from the “number of hard disk drives constituting a RAID group.”

That is, the unit number of read requests to be processed is determined by (X−Y)×(X−1) where X is the “number of hard disk drives constituting a RAID group” and Y is the “number of hard disk drives storing parity in one stripe.”

For example, as shown in FIG. 4 , if four hard disk drives constitute RAID 5 , X is “4” and Y is “1.” Therefore, “9” that is the operation result of (4−1)×(4−1) is obtained as the unit number of read requests to be processed.

When the read command creation unit 315 a processes read requests corresponding to the obtained unit number of read requests to be processed, it creates read commands to obtain pieces of data from identical hard disk drives corresponding to the data storage number. In other words, the read command creation unit 315 a data handles a hard disk drive(s) corresponding to the “number (one in the case of RAID 5 ) of hard disk drives storing parity in one stripe,” as a hard disk drive(s) from which no data is to be read.

Subsequently, when the read command creation unit 315 a processes read requests corresponding to the subsequent unit number of read requests to be processed, it also creates read commands to obtain pieces of data from identical hard disk drives corresponding to the data storage number. At that time, the read command creation unit 315 a handles, as a hard disk drive from which no data is to be read, a hard disk drive adjacent to the hard disk drive handled as a hard disk drive from which no data is to be read when previously processing the read requests corresponding to the unit number of read requests to be processed.

In the example shown in FIG. 4 , the unit number of read requests to be processed is “9”. Therefore, with respect to nine read requests to be processed first, the read command creation unit 315 a creates read commands to obtain pieces of data from the hard disk drives 12 to 14 and handles the hard disk drive 11 as a hard disk drive from which no data is to be read. With regard to nine read requests to be processed next, the read command creation unit 315 a creates read commands to obtain pieces of data from the hard disk drives 11 , 13 , and 14 while handling, as a hard disk drive from which no data is to be obtained, the hard disk drive 12 adjacent to the hard disk drive 11 previously handled as such a hard disk.

As is understood from the above-description, each time the read command creation unit 315 a processes read requests corresponding to the unit number of read requests to be processed, the read command creation unit 315 a creates read commands while handling, as a hard disk drive from which no data is to be read, a hard disk drive adjacent to a hard disk drive previously handled as such a hard disk drive. The read command creation unit 315 a creates read commands for each stripe and outputs the created read commands to the read control unit 315 b . For example, with regard to strip 1 , the read command creation unit 315 a creates three read commands to obtain data D 12 and D 13 and parity P 10 , outputs the created three read commands to the read control unit 315 b , and instructs the read control unit 315 b to obtain data D 11 , D 12 , and D 13 .

Hereafter, processes performed by the read command creation unit 315 a in a case where there are successive read requests to obtain pieces of data stored in a particular range of stripes will be described using the example shown in FIG. 4 . When the read command creation unit 315 a determines that 27 read requests to obtain data D 11 to D 13 , D 21 to D 23 , . . . , D 91 to D 93 are successive, it first obtains “9” as the unit number of read requests to be processed. Subsequently, with regard to nine read requests (nine read requests to obtain data D 11 to D 13 , D 21 to D 23 , and D 31 to D 33 ) to be processed first, the read command creation unit 315 a creates read commands to obtain pieces of data from the hard disk drives 12 to 14 .

At that time, the read command creation unit 315 a creates read commands for each stripe; therefore, first, the read command creation unit 315 a creates three read commands to obtain data D 12 and D 13 and parity P 10 in order to obtain data D 11 to D 13 stored in stripe 1 . Then, the read command creation unit 315 a outputs the created three read commands to the read control unit 315 b and instructs the read control unit 315 b to obtain data D 11 to D 13 .

As such, the read command creation unit 315 a creates three read commands to obtain data D 22 , parity P 20 , and D 23 in order to obtain data D 21 to D 23 stored in stripe 2 . Then, the read command creation unit 315 a outputs the created three read commands to the read control unit 315 b and instructs the read control unit 315 b to obtain data D 21 to D 23 . Also, the read command creation unit 315 a creates similar read commands with respect to stripe 3 and instructs the read control unit 315 b to obtain data D 31 to D 33 .

With regard to nine read requests (nine read requests to obtain data D 41 to D 43 , D 51 to D 53 , and D 61 to D 63 ) to be processed subsequently, the read command creation unit 315 a handles, as a hard disk drive from which no data is to be obtained, the hard disk drive 12 adjacent to the hard disk drive 11 handled as such a hard disk when previously processing the nine read requests. In other words, the read command creation unit 315 a creates read commands to obtain pieces of data from the hard disk drives 11 , 13 , and 14 .

Like when processing the first nine read requests, with respect to stripes 4 to 6 , the read command creation unit 315 a creates read commands to obtain pieces of data stored in each stripe. As for nine read requests to be processed subsequently, the read command creation unit 315 a creates read commands to obtain pieces of data from the hard disk drives 11 , 12 , and 14 while handling the hard disk drive 13 as a hard disk drive from which no data is to be obtained.

In this way, the read command creation unit 315 a creates read commands with respect to the successive read requests to obtain all pieces of data stored in a particular range of stripes.

On the other hand, if there are no two or more successive read requests to obtain data from an identical hard disk drive and if there are no successive read requests to obtain all pieces of data stored in a particular range of stripes, the read command creation unit 315 a creates read commands including physical addresses corresponding to logic blocks specified by the read requests.

Subsequently, the read command creation unit 315 a outputs the created read commands to the read control unit 315 b . The read control unit 315 b receives the read commands and executes the received read commands to obtain pieces of data from the hared disk units 11 to 14 . Then, the read control unit 315 b transmits the obtained pieces of data to the host 400 via the CA 311 a and 311 b.

If the read control unit 315 b is given an instruction for performing an exclusive OR operation by the read command creation unit 315 a , it performs an exclusive OR operation according to the instruction so as to obtain data.

For example, if the read control unit 315 b are given an instruction for obtaining data D 11 as well as three read commands to obtain data D 12 and D 13 and parity P 10 in the example shown in FIG. 2 , it executes the given three read commands so as to obtain data D 12 and D 13 and parity P 10 . Then, the read control unit 315 b performs an exclusive OR operation on the obtained data D 12 and D 13 and parity P 10 so as to obtain data D 11 and transmits the obtained data D 11 to the host 400 .

Also, for example, if the read control unit 315 b are given an instruction for obtaining data D 11 to D 13 as well as three read commands to obtain data D 12 and D 13 and parity P 10 in the example shown in FIG. 4 , it executes the received three read commands so as to obtain data D 12 and D 13 and parity P 10 . Then, the read control unit 315 b performs an exclusive OR operation on the obtained data D 12 and D 13 and parity P 10 so as to obtain data D 11 and transmits the obtained data D 11 to D 13 to the host 400 .

Hereafter, a read process performed by the disk array apparatus 300 according to this embodiment will be described. FIG. 6 is a flowchart showing steps of a read process performed by the disk array apparatus 300 according to this embodiment.

As shown in FIG. 6 , the read command creation unit 315 a of the disk array apparatus 300 reads read requests stored in the read request reception unit 313 (step S 101 ). Subsequently, the read command creation unit 315 a obtains physical block numbers corresponding to logic block numbers included in the read requests from the address conversion table 314 a (step S 102 ).

Subsequently, the read command creation unit 315 a analyzes the read requests according to the obtained physical block numbers. If there are two or more successive read requests to obtain pieces of data from an identical hard disk drive (YES in step S 103 ), the read command creation unit 315 a alternately creates read commands to obtain data by performing an exclusive OR operation and a read command to obtain data as specified by a read request, so that disk accesses are distributed (step S 104 ).

On the other hand, if there are no two or more successive read requests to obtain pieces of data from an identical hard disk drive (NO in step S 103 ) and if there are successive read requests to obtain all pieces of data stored in a particular range of stripes (YES in step S 105 ), the read command creation unit 315 a creates read commands such that the respective frequencies with which the hard disk drives 11 to 14 are accessed are equalized and such that the respective frequencies with which the hard disk drives 11 to 14 are successively accessed are increased (step S 106 ).

On the other hand, if there are no two or more successive read requests to obtain pieces of data from an identical hard disk drive (NO in step S 103 ) and if there are no successive read requests to obtain all pieces of data stored in a particular range of stripes (NO in step S 105 ), the read command creation unit 315 a creates read commands to obtain pieces of data as specified by the read requests (step S 107 ).

The read control unit 315 b receives the read commands from the read command creation unit 315 a , executes the received read commands so as to obtain pieces of data from the hard disk drives 11 to 14 (step S 108 ), and transmits the obtained pieces of data to the host 400 via the CAs 311 a and 311 b (step S 109 ).

As described above, in the disk array apparatus 300 according to this embodiment, the read command creation unit 315 a analyzes read requests and if there are two or more successive read requests to obtain pieces of data from an identical hard disk drive, the read command creation unit 315 a alternately creates read commands to obtain data by performing an exclusive OR operation and a read command to obtain data as specified by a read request. Also, if there are successive read requests to obtain all pieces of data stored in a particular range of stripes, the read command creation unit 315 a creates read commands such that the respective frequencies with which the hard disk drives 11 to 14 are accessed are equalized and such that the respective frequencies with which the hard disk drives 11 to 14 are successively accessed are increased. As a result, the apparatus's capability to perform read processes on the hard disk drives constituting RAID 5 is improved.

In this embodiment, if there are two or more successive read requests to obtain pieces of data from an identical hard disk drive, pieces of data are obtained from hard disk drives other than the identical hard disk drive. In addition thereto, if there are two or more successive read requests to obtain pieces of data from two hard disk drives, pieces of data may be obtained from hard disk drives other than the two hard disk drives.

Also, in this embodiment, the example in which the hard disk drives 11 to 14 constitute RAID 5 is shown. However, the hard disk drives may constitute RAID 6 that stores horizontal parity (P) and weighting parity (Q) for each stripe.

FIG. 7 is a diagram showing an example of read processes performed by the related-art disk array apparatus in which RAID 6 is constituted. In the example shown in FIG. 7 , the hard disk drives 11 to 14 constitute RAID 6 , which stores horizontal parity P and weighting parity Q for each stripe. In this case, if the related-art disk array apparatus receives read requests to obtain data D 11 , D 51 , D 61 , and D 91 from the host 400 , it cannot read pieces of data in a parallel manner. As a result, the apparatus's capability to perform read processes is reduced.

FIG. 8 is a diagram showing an example of read processes performed by the disk array apparatus according to this embodiment in which RAID 6 is constituted. Here, like in the example shown in FIG. 7 , it is assumed that the disk array apparatus according to this embodiment has received read requests to obtain data D 11 , D 51 , D 61 , and D 91 from the host 400 . As shown in FIG. 8 , the disk array apparatus performs control so that the respective frequencies with which the hard disk drives 11 to 14 are accessed are equalized. As a result, access loads imposed on the hard disk drives 11 to 14 are made uniform.

FIG. 9 is a diagram showing an example of read processes performed by the related-art disk array apparatus in which RAID 6 is constituted. In the example shown in FIG. 9 , it is assumed that the disk array apparatus has successively received read requests to obtain all pieces of data stored in stripes 1 to 9 . In such a case, the disk array apparatus successively accesses each of the hard disk drives 11 to 14 with respect to a maximum of two stripes.

FIG. 10 is a diagram showing an example of read processes performed by the disk array apparatus according to this embodiment in which RAID 6 is constituted. Here, like in the example shown in FIG. 9 , it is assumed that the disk array apparatus has successively received read requests to obtain all pieces of data stored in stripes 1 to 9 . As shown in FIG. 10 , the disk array apparatus performs control so that the respective frequencies with which the hard disk drives 11 to 14 are accessed are equalized and so that the respective frequencies with which the hard disk drives 11 to 14 are successively accessed are increased, like in the case (see FIG. 4 ) where RAID 5 is constituted. In the example shown FIG. 10 , the disk array apparatus successively accesses each of the hard disk drives 11 to 14 with respect to a maximum of six stripes. Accordingly, the frequency with which each hard disk drive is successively accessed becomes triple that in the related-art example shown in FIG. 9 .

While the example in which the hard disk drives 11 to 14 constitute RAID 5 is shown in this embodiment, the hard disk drives may constitute RAID 3 or RAID 4 in which a hard disk drive for storing parity is not changed for each stripe or may constitute RAID 2 in which hamming code is stored instead of parity. Various processes described in this embodiment are realized by executing a previously prepared program using a computer such as a personal computer or a workstation. Referring now to FIG. 11 , an example of a computer for executing a disk array apparatus control program having functions similar to those of this embodiment will be described. FIG. 11 is a diagram showing a computer for executing a disk array apparatus control program.

As shown in FIG. 11 , a computer 1000 includes a CPU 1010 as an operation means for performing various operation processes, a ROM (read only memory) 1020 as a nonvolatile memory, and a RAM (random access memory) 1030 for temporarily storing various types of information. The CPU 1010 , ROM 1020 , and RAM 1030 are coupled to one another via a bus 1040 .

The ROM 1020 stores a disk array apparatus control program 1021 having functions similar to those of the controller 315 shown in FIG. 5 and address conversion data 1022 corresponding to the address conversion table 314 a shown in FIG. 5 .

The CPU 1010 reads the disk array apparatus control program 1021 from the ROM 1020 and then loads the read program into the RAM 1030 . Thus, the disk array apparatus control program 1021 serves as a disk array apparatus control process 1031 . Then, the disk array apparatus control process 1031 loads information or the like read from the address conversion data 1022 into an area of the RAM 1030 assigned to the disk array apparatus control process 1031 as appropriate. Then, the disk array apparatus control process 1031 processes various types of data according to the loaded information or the like.

The disk array apparatus control program 1021 need not always be stored in the ROM 1020 . For example, the disk array apparatus control program 1021 may be previously stored in a “transportable physical medium” inserted into the computer 1000 , such as a flexible disk (FD), a CD-ROM, an MO disk, a DVD disk, an optical magnetic disk, or an IC card, a “fixed physical medium” provided inside or outside the computer 1000 , such as a hard disk drive (HDD), or “another computer (or server)” coupled to the computer 1000 via a public line, the Internet, LAN, WAN, or the like, and then may be read and executed by the computer 1000 .

Hereafter, appendixes with respect to the above-described embodiment will be disclosed.