JPH10320134A - Optical disk library device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accelerate the data transfer speed on the average, to secure high minimum performance with a small number of drives, and to reduce cost by writing data a half of the drives from the outer peripheries of optical disk media and writing data to the rest half of the drives from the inner peripheries. SOLUTION: A next writable block position on an optical disk medium set on a drive is read out (procedure 201). It is decided which of the outer and inner peripheries of the disk medium data are written from (procedure 202). For example, when the data are written from the inner periphery, blocks which are written from the write block position are counted (procedure 208). The number of blocks is calculated to obtain a constant read/write time without depending upon each read/write block position. The block position where the data are actually written is recalculated (procedure 203). A writable block position is calcuated (procedure 204). A processor issues an instruction having the calculated parameter to the drive (procedure 207).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a storage device using a removable storage medium, and more particularly to a plurality of drive devices for backing up a large amount of data at one time, and simultaneously writing a plurality of the storage media. Alternatively, it relates to a device for reading.

[0002]

2. Description of the Related Art In recent years, computer processing of important data such as online services of banks has become widespread, and at the same time, the demand for reliability of storage devices has become more and more severe. There are roughly two types of responses to this demand, and one is to increase the reliability of the storage device that is always used. For this purpose, for example, the same data is simultaneously written to two magnetic disk devices, and even if one device fails, the other magnetic disk device has a mirroring function for maintaining the data read / write function and a RAID function described later. Disk array devices are often used.

On the other hand, in addition to increasing the reliability of the storage device itself, a backup method of copying stored data to another storage device at specific time intervals has also become a countermeasure. When a storage device that is constantly used fails, the data that has been backed up by replacing the storage device with a normal storage device is returned. It takes a lot of time to replace with a normal storage device or return data, so it is not suitable for work that requires non-stop operation, but storage device that is always used due to disaster such as earthquake or fire This is effective in recovering data in the case where the data has been fatally broken or the storage device has not failed, but the data has been lost due to an operation error.

[0004] The problem of data backup is that it is unnecessary if there is no failure, but frequently accesses the storage device that is always in use, which is a burden and puts a burden on processes essentially required for business. It is to do. For this reason, backup is often performed during a period when business processing does not access the storage device. In recent years, however, various services have been available at any time and have become convenient. However, the processing time of business has been lengthened, and backup is required in a short time. However, magnetic tapes and optical disks have conventionally been mainly used for storage devices used for backup in terms of cost. These storage devices have inferior write performance to magnetic disk devices, and are not suitable for a demand for short-time backup.

To improve the write performance of a low-speed storage device, it is necessary to reduce the amount of data to be written. To reduce the amount of data to be written, you can compress it to reduce the amount of data.
There is a method in which a plurality of storage devices are arranged in parallel to reduce the amount of data written to one storage device. The latter method uses a RAID (Red
undant Arrays of Inexpensive Discs).

[0006] The disk array is composed of a plurality of magnetic disk devices.
There are types from AID1 to RAID5. For example, R
AID3 is composed of five magnetic disk devices, writes data in a distributed manner to four magnetic disk devices, and writes parity to one magnetic disk device. Except for various overheads, four times the performance of a single magnetic disk device can be theoretically obtained. In addition, even if one magnetic disk device fails due to the parity, data read / write can be continued, so that not only performance can be improved but also reliability of the storage device can be improved as described above. Such a disk array is described in IEICE Journal Vol. 79 No. 11 pp. 1109-1115 and the like. In this example, the amount of data written at a time (striping size) is described for a case where all the magnetic disk devices have the same 200 KB and a case where the data amount is 50 KB.

FIG. 4 shows the concept of a distributed data write method using ordinary RAID. Although the storage medium or disk into which parity is written changes depending on the type of RAID, it is omitted in FIG. 4 because it is not necessary to explain the striping size. Write data 401 to optical disc 403
Data. For the sake of explanation of data distribution, names from “A1” to “D2” are given for convenience. Data 402 is obtained by converting data 401 from small data “A1” to “D”.
2 ". “A1” to “D2” are the same size, and “A1”, “B1”, “C1” and “D1” are simultaneously written on the optical disc 403, and then “A2”
By writing “B2”, “C2”, and “D2” to the optical disc 403 at the same time, the entire data 401 can be written to the optical disc 403 in about twice the writing time of “A1” alone.

[0008]

The prior art of the above-mentioned disk array is extremely excellent in realizing high-speed data reading / writing and high reliability. However, since the technology was originally developed on the premise that a magnetic disk was used, processing for the characteristics of the optical disk was lacking.

In an optical disk, despite the fact that the recording density on the outer and inner circumferences of the medium is constant, the rotational speed is also constant, so that the outer circumference has a higher tangential speed than the inner circumference.
The data transfer speed is higher on the outer circumference than on the inner circumference. Further, in the optical disk drive device, since the head has a laser light emitter and a lens, it is heavier than the magnetic disk device head and has a lower seek speed. For this reason, generally, a magnetic disk is provided with file management data such as FAT (File Allocation Table) in one place, and a random read and write of data while the head moves between the position of the management data and the position of the actual data. In contrast to the access method, the optical disk often employs sequential access by providing file management data at a plurality of locations as the actual data is written.

If the processing for the characteristics of the optical disk as described above is omitted, even if a disk array is configured with optical disks for high-speed backup, access occurs such that the heads of all the drive devices are located on the inner circumference, and the performance is reduced. In order to ensure the minimum performance, it is necessary to read and write in parallel with an unnecessary number of drive devices on the outer periphery to secure the minimum performance,
The cost of the device increases.

[0011]

In order to achieve the above object, the present invention provides a means for dividing data to be written into a plurality of pieces of data, a means for collectively restoring the read plurality of pieces of data, Optical disk drive device, and using the plurality of optical disk drive devices,
A means for simultaneously reading and writing data on a plurality of optical disc media is used.

Also, half of the plurality of optical disk drive devices use means for writing data from the outer circumference of the optical disk storage medium, and the other half drive device uses means for writing data from the inner circumference of the optical disk storage medium.

When data is written to the outer periphery of the optical disk storage medium, means for increasing the amount of data to be written at one time to the amount of data to be written to the inner periphery is used.

Further, a means for keeping the writing time constant by changing the amount of data to be written at once regardless of the position where the data is written on the optical disk storage medium is used.

When writing data to the optical disk library device for backup, the divided data is sent to a plurality of optical disk drive devices provided by means for dividing the data to be written into a plurality. Further, by means of reading and writing data to a plurality of optical disk media at the same time, the amount of data to be written to the optical disk medium set in each optical disk drive can be reduced as compared with writing all data to one optical disk drive. Time is shortened. If the track position on the optical disk is the same and the overhead of the division processing is ignored, performance in proportion to the number of optical disk drive devices can be obtained.

However, in a normal file system of an optical disk, data is written sequentially from the outer periphery of the optical disk medium, so that the writing performance is initially high. However, as the data is written, the writing moves to the inner circumference of the optical disk medium. Decrease. In particular, when data is written to the inner track of all the optical disks, the performance is most deteriorated. Here, at the beginning of data writing, half of the plurality of optical disk drive devices can write data at a higher speed than the inner circumference by means of writing data from the outer circumference of the optical disk storage medium, and the other half of the drive devices By means of writing data from the inner circumference of the optical disk storage medium, writing is performed at a lower speed than writing to the outer circumference. On the other hand, as the data writing progresses, half of the plurality of optical disk drive devices decrease the writing speed by the means for writing data from the outer periphery of the optical disk storage medium, but the other half of the drive devices store the data in the optical disk storage medium. The writing speed is improved by means for writing data from the periphery, and an average writing speed can be secured.

Furthermore, when the data to be written is divided by a plurality of means, if the amount of data to be divided is equalized, the writing speed on the inner circumference is low, so that the writing on the outer circumference is completed first and the writing time is reduced. It depends on writing to the circumference. Therefore, when writing data to the outer circumference of the optical disk storage medium, by using means for making the amount of data to be written at one time larger than the amount of data to be written to the inner circumference, the writing time to the inner circumference can be shortened, and the entire writing time can be reduced. shorten. However, if the amount of data written to the inner periphery is excessively reduced, the time required for writing to the outer periphery becomes longer than the time required to write to the inner periphery, and the time depends on the time required for writing to the outer periphery. For this reason, regardless of the position where the data of the optical disk storage medium is written, by changing the amount of data written at a time, a means for keeping this writing time constant is used.
The writing time to the outer circumference and the writing time to the inner circumference are made to coincide with each other, so that the writing time as a whole can be minimized.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to FIG. FIG. 1 is a configuration diagram of an optical disk library device.

The processor 101 is a central processing unit (CPU) that manages the entire apparatus, writes data to each unit in the apparatus according to instructions stored in the memory 102,
It reads data from each part. The memory 102 is a semiconductor storage device for temporarily storing instructions of the processor 101, temporary work data, and data to be backed up on an optical disk. The LAN 103 is an adapter device for connecting to the local network. When data of another device is backed up via the network, the data is input from the LAN 103 and the
02 is temporarily stored. When restoring data backed up on an optical disk,
Output to other devices. The magnetic disk 104 has a small capacity in the memory 102 and loses its stored contents when the power of the apparatus is turned off. Therefore, data from the LAN 103 having a capacity that cannot be stored in the memory 102 or the capacity of the processor 101 when the power is turned off. Stores instructions and the like. SCSI10
Reference numerals 5, 106, and 107 denote adapter devices for inputting and outputting data to and from peripheral devices, and are SCSI (Small Computer).
er System Interface). These SCSI 105, 106, 10
7 is an optical disk drive 108, 109, 110, 1
11 is used for data exchange between the optical disk 11 and an accessor 112 which is an optical disk transport device.

Drives 108, 109, 110, 111
Is for reading and writing data on the inserted optical disk medium. In FIG. 1, drives 110 and 111 are SCSI
105 and drives 108 and 109 are SCSI
106, but up to seven devices can be connected in a daisy chain in the SCSI specification.
Without I106, drives 108, 109, 110,
It is also possible to connect four units 111 to one SCSI 105. Also, the drives 108, 109, 110, 11
Four SCSI adapter devices can be connected one-to-one to four devices. Such a connection form is determined by the performance ratio between the data read / write performance of the drive 110 and the data transfer performance of the SCSI 105. Here drive 11
The read / write performance of 0 is half of the data transfer performance of the SCSI 105, and the two drives 110 and 111 are connected. The accessor 112 is a robot that inserts a medium of an optical disk on a shelf into the drives 108, 109, 110, and 111. An operation of returning the optical disk medium from the drive to the shelf or extracting the optical disk medium in the drive once and inverting the front and back is also performed.

The process of backing up data of another device via a network is as follows. First, in accordance with the instruction stored in the memory 102, the processor 101
Sends a command to accessor 112 via CSI 107,
The optical disk medium used for the backup is taken out of the shelf and inserted into the drives 108, 109, 110 and 111 one by one. Next, the processor 101 drives the drives 108, 109, 110, and 11 via the SCSIs 105 and 106.
1 to read control data from the optical disk in each drive and confirm that the correct optical disk is set. Make the above preparations and wait for data.

The data to be backed up is the LAN 103
Through the memory 102. Further, when there is much data, the data is temporarily stored on the magnetic disk 104. The procedure for writing this data to the optical disk medium will be described with reference to FIG.

FIG. 2 shows a procedure for writing data to the optical disk by the processor 101 stored in the memory 102. Here, drive 1 via SCSI 106
A procedure for writing data to the data 08 will be particularly described, but the drives 109, 110, and 111 also perform a similar procedure. First, in step 201, the next writable block position on the optical disk medium set in the drive 108 is read. This position is written to a fixed position on the optical disk, and the magnetic disk 10
4 in some cases. The next writable block position is empty in which data is not always written itself. Next, whether the optical disk medium set in the drive 108 in step 202 writes data from the outer circumference,
It is determined whether data is to be written from the inner circumference. Here, it is assumed that the optical disk medium set in the drive 108 is for writing from the inner periphery. In this case,
The optical disk medium set in the drive 109 is written from the outer circumference, the optical disk medium set in the drive 110 is written from the inner circumference, and the optical disk medium set in the drive 111 is written from the outer circumference. Whether the writing is performed from the outer circumference or the inner circumference is performed by formatting the optical disk medium, and when one group is set with four disks, two disks are determined in advance. The process proceeds to step 208 according to the determination result of step 202. In step 208, the number of blocks to be written is calculated from the position of the block to be written.
The number of blocks is obtained by referring to the table in FIG. The read / write block number table in FIG. 7 is a table for obtaining the number of blocks from the block position to be read / written, and records the number of blocks for each block position when starting from the outer circumference and when starting from the inner circumference. The table in FIG. 7 indicates that the read / write position has the same number of blocks from 0 to 9, and the value of the block position increases, that is, the number of blocks decreases as going from the outer circumference to the inner circumference. The number of blocks is calculated so as to have a constant read / write time regardless of the position of each read / write block. Here, it is assumed that the data transfer speed of the outermost circumference and the innermost circumference is twice different, and the number of blocks of the outermost circumference and the innermost circumference is also twice as different. Also, when writing is started from the outer periphery, the writing is performed from the designated block position toward the inner periphery. Therefore, the value of the block position starts from 0 and the value of the block position is written up to 999. Is started from 1000, the write start position is once converted to a write position from the outer circumference in step 203, and then write is started. In this case, if the block position is 100, the number of blocks is 20 because the block position is 80 and the number of blocks is 20. On the other hand, when the block position is 100 when starting from the outer circumference, the number of blocks is written 19 from the block position 100, and the number of blocks of the item having the block position 100 is 19 and 20. Next, the procedure proceeds to step 203.
In step 203, the block position where data is actually written is calculated again. The data write instruction to the drive 108 requires three types of parameters: the starting block position to write data, the number of blocks of data to be written, and the contents of data to be written. Then, in accordance with these parameters, data for the number of blocks from the start block position is written from the outer periphery to the inner periphery. Since the drive 108 writes data from the inner circumference to the outer circumference, a calculation is made in step 203 to move the start block position to the outer circumference by the number of blocks. In step 204, the next writable block position is calculated for the next data write. Finally, in step 207, the SCSI
A command having the calculated parameter is issued to the drive 108 via the command 106.

The drive 109 writes data from the outer circumference to the inner circumference. In step 209, the number of blocks to be written is calculated from the position of the block to be written. The method of obtaining the number of blocks with reference to the table of FIG. Next, the procedure proceeds to step 205. Steps 205 and 206 are steps 203 and 20
4, but is a simple calculation because it is in accordance with the parameters of the write command of the drive 109.

FIG. 6 shows the structure of a data write position management table for each optical disk medium. Items in the table include a shelf number for storing optical disk media, a group number for identifying a group of four optical disk media as one group, a disk number for identifying optical disk media within a group, and each disk medium being There is a start position indicating whether writing is performed from the periphery or from the outer periphery, and a writing position indicating from which block position the writing is currently started. In the table of FIG. 6, for example, when data is read / written on the optical disk medium having the group number 2, the optical disk medium contained in the shelf numbers 5, 6, 7, and 8 is set in the drive, and the optical disk medium having the disk number 1 is set. Starts writing from the write start block position 600 from the inner circumference to the outer circumference. The optical disk medium of disk number 2 starts writing from the write start block position 399 from the outer circumference to the inner circumference. In this way, all the optical disk media on the shelf are managed by the data writing position management table.

FIG. 5 is a conceptual diagram of the data writing method of the present invention, which is in contrast to the conventional RAID data writing method of FIG. The optical disks 503, 504 of FIG.
The size of each frame from “A1” to “D2” of the data 501 to be written to 505 and 506 represents the size of each data. In the present invention, the data amount is smaller when writing to the inner circumference than when writing to the outer circumference. In FIG. 5, the medium A 503 and the medium C 505 are written from the outer circumference, and the medium B 5
04 and the medium D506 are written from the inner circumference, so that A1 and C1 to be written first in the data 501 to be written have a large data amount, and B1 and D1 have a small data amount. However, as the data is written, the medium A 503 and the medium C 505 write the inner circumference, and the medium B 504 and the medium D 506 write the outer circumference. D2 has a large data amount.

By the above procedure, the data write position on the optical disk medium changes as shown in FIG. When the length of the outer circumference and the inner circumference of the optical disk medium are twice different, the data transfer speed is also twice different between the outer circumference and the inner circumference. If the inner circumference is 1 MB / sec, the outer circumference is 2 MB / sec. Four drives 108,
When two of 109, 110, and 111 are on the outer circumference, the other two are on the inner circumference, so that a total data transfer rate of 6 MB / sec can be obtained.

As described above, according to this embodiment,
Since the data transfer speed of the other two drives is high when the data transfer speed of the two drives is low, the data transfer speed of 6 MB / sec can be obtained with the four drives. Since the data transfer speed fluctuates from 8 MB / sec to 4 MB / sec, when a request of at least 6 MB / sec is required, 6 drives are required, and a high minimum performance can be secured with a small number of drives.

[0029]

According to the present invention, when the data transfer speed of half of the drives is low, the data transfer speed of the other half of the drives is high, so that the data transfer speed can be increased on average. High minimum performance can be ensured, and the cost of the optical disk library device can be reduced.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an optical disk library device.

FIG. 2 is a flowchart of a data write procedure.

FIG. 3 is a conceptual diagram of data writing on an optical disk medium.

FIG. 4 shows a conventional RAID data distribution writing method.

FIG. 5 is a data distributed writing method of the present invention.

FIG. 6 shows the structure of a data write position management table.

FIG. 7 is a read / write block number table.

[Explanation of symbols]

101: Processor, 102: Memory, 103: L
AN, 104: magnetic disk, 105: SCSI, 10
6 SCSI, 107 SCSI, 108 drive, 109 drive, 110 drive, 111
... drive, 112 ... accessor.

Claims (3)

[Claims] 1. An optical disk storage device comprising a plurality of optical disk storage media and a plurality of drive devices for reading and writing data from and to the storage medium, wherein half of the drive devices write data from an outer periphery of the optical disk storage medium. An optical disk library device, wherein the other half of the drive device writes data from the inner periphery of the optical disk storage medium. 2. An optical disk library apparatus according to claim 1, wherein when writing data to the outer circumference of said optical disk storage medium, the amount of data to be written at once is larger than the amount of data to be written to the inner circumference. . 3. The optical disk library device according to claim 2, wherein the write time is made constant by changing the amount of data to be written at one time regardless of the position where the data on the optical disk storage medium is written. Optical disk library device.