JPH09160726A - Data processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To vary the line size and to suppress the read of unnecessary data, and to improve the preread effect by detecting an access pattern and varying the number of blocks constituting a line according to the detected access pattern. SOLUTION: When an external request is made, a CPU 1 reads out the request from an external I/O 2. When the request is a line size variation command for varying the line size of a device, a cache is used to detect distance. When one data D (logical block address Ad) is accessed, the minimum value Lmin=MIN(Ad-Am) of the distance to data Cn is found from the logical block address An (Ad>An) of the data Cn on the cache while retrieving whether the data D is on the cache or not retrieved, and the line size is determined according to the trend of the value Lmin. Consequently, the preread effect is improved. Further, the line size is dynamically varied to cope with changes in access pattern.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data processing device which is a mass storage device including optical disks as a plurality of portable information storage media.

[0002]

2. Description of the Related Art In recent years, the amount of data managed in an office is increasing, and in order to manage such a large amount of data, a large capacity portable optical disk (rewritable magneto-optical disk or CD-ROM) is used. Many proposals use a collective optical disc device (autochanger) that includes a storage slot unit that stores a plurality of optical discs, a plurality of optical disc drive units, and an accessor unit that moves an optical disc between the optical disc drive unit and the storage slot units. Has been done.

Conventionally, there are the following three techniques in a data processing device for storing information by using such a collective optical disc device. First, in order to handle a large amount of data as shown in FIG. 13, one using an optical disc and an autochanger is a LASER DIS manufactured by Item Co.
Many products, such as the K series, can handle several optical disks as one hard disk.

The optical disk is made to look like a hard disk as such an external interface because a hard disk is generally used as a method for connecting an external storage device of a computer so that the computer can easily handle the autochanger and the optical disk. It is for doing.

Further, in order to reduce the number of times of exchanging optical disks in an autochanger, as an optimum data arrangement, for example, Japanese Patent Application No. 5-238437, Japanese Patent Application No. 6-207771, Japanese Patent Application No. 7-59499. No. etc. have been proposed.

This optimal arrangement makes the location of physically storing the data variable according to the frequency of access to the data specified by the logical block address of the external interface. By storing the optical disk in the drive, the probability that this optical disk having a high access frequency is present in the drive is increased, and the number of times of exchange in the autochanger can be reduced.

Secondly, a file system that allocates one file or the like to consecutive blocks, shortens seek time during access of one file, and improves file access performance (for example, MS-
DOS, etc.) and applications (for example, Microdata's Nostradamus, Semantec's speed disk, etc.) and data of files with similar access frequency are stored in physically close places to shorten the seek time between files. (For example, USP53
33311), or a collection of files accessed at the same time as the access frequency (for example, Japanese Patent Application No. 7-
No. 60633).

Third, in the UNIX file system,
The file system uses the main memory in the computer as a cache to read data from the secondary storage device and transfer the data to the application software. When there is no data in the cache, the data is transferred from the secondary storage device to the cache. At this time, there is a method called "prefetch" in which not only the requested block of data but also the next block is read into the cache. This is because the number of accesses to the secondary storage device, which has a large transfer overhead due to seek or rotation wait, is reduced by increasing the size of one transfer in the expectation that accesses will continue. It improves the performance of the entire system.

In order to reduce the number of optical disk exchanges by using the above-mentioned first technique, it is desirable to make the area of the data for measuring the access frequency small, but only the management table becomes large in block units. However, even if the logical block addresses are consecutive, the possibility that they are physically consecutive is reduced,
The second technology on the computer cannot be used.

For this reason, as shown in FIG. 14, areas for actually storing data of a certain fixed number of continuous logical block addresses are also arranged continuously. In the case of (a) of FIG. 14, optimal allocation is performed in units of 16 blocks according to access frequency while maintaining block continuity. Data that maintains this continuity is called a line. The size of the line is called the line size.

FIG. 14B shows an example in which the optimum arrangement is performed with a line size of 32 blocks. When data is transferred to the cache, the data in the secondary storage device can be "read ahead" by always transferring in line units.

However, if the line size is unnecessarily increased, the data that may not be accessed may be read into the cache and the data that may be accessed in the near future may be expelled from the cache. Therefore, from a statistical pattern of access, when a block is accessed, it is necessary to find out how far away the block that can be expected to be accessed in the near future. This distance is the line size.

[0013]

As described above, the optimum placement is performed according to the access frequency while keeping the continuity of blocks. The data that maintains the continuity is called a line, and the size of the line is called the line size. Calling the data on the cache can be done by "read ahead" by always transferring it in units of lines. However, increasing the line size meaninglessly reads data that may not be accessed into the cache, and it is possible to use it in the near future. There was a problem that the possibility of expelling data that may be accessed from the cache increases.

Therefore, the present invention suppresses the reading of useless data by dynamically changing the line size,
It is an object of the present invention to provide a data processing device capable of improving the "look-ahead" effect.

[0015]

A data processing apparatus of the present invention uses a plurality of portable storage media in which data is stored in block units, and a certain number of blocks are stored in these portable storage media. A line as continuous data of, and in the data processing device accessed in units of this line, when the portable storage medium is accessed, detection means for detecting this access pattern, and detection means for detecting this access pattern It comprises a changing means for changing the number of blocks forming the line according to an access pattern.

The data processor of the present invention uses a plurality of portable storage media in which data is stored in block units, and blocks stored in these portable storage media are lined as a certain number of continuous data. And in the data processing device that is accessed in line units,
A plurality of read / write means for loading the portable storage medium and reading / writing data on / from the loaded portable storage medium line by line, and storage means for storing cache data for the portable storage medium. And receiving means for receiving an access request from the outside, and the portable type corresponding to the block address of the access request data when the access request data received by the receiving means is not stored in the storage means. It comprises a block address of the storage medium and a determination means for determining the number of lines to be read from the portable storage medium to the storage means.

The data processing apparatus of the present invention uses a plurality of portable storage media in which data is stored in block units, and blocks stored in these portable storage media are lined as a certain number of continuous data. And in the data processing device that is accessed in line units,
A plurality of read / write means for loading the portable storage medium and reading / writing data on / from the loaded portable storage medium line by line, and storage means for storing cache data for the portable storage medium. And receiving means for receiving an access request from the outside, and the portable type corresponding to the block address of the access request data when the access request data received by the receiving means is not stored in the storage means. Determining means for determining the number of lines to be read from the portable storage medium to the storage means from the block address of the storage medium, and data for the number of lines determined by the determining means is stored in the portable storage means using the read / write means. A processor that reads out from the medium, stores it in the storage means, and performs processing according to the access request received by the reception means. It is composed of a.

The data processing apparatus of the present invention uses a plurality of portable storage media in which data is stored in block units, and blocks stored in these portable storage media are lined as a certain number of continuous data. And in the data processing device that is accessed in line units,
A plurality of read / write means in which the portable storage medium is loaded, and data is read / written on a line-by-line basis with respect to the loaded portable storage medium, and a fixed number of blocks configuring the line for each read / write means. And a management means for managing the number of portable storage media corresponding to the number.

The data processing device of the present invention uses a plurality of portable storage media in which data is stored in block units, and blocks stored in these portable storage media are lined as a certain number of continuous data. And in the data processing device that is accessed in line units,
The conversion unit is configured to convert a logical block address into a physical block address for each minimum line unit that constitutes the above-mentioned fixed number of continuous blocks.

The data processing apparatus of the present invention uses a plurality of portable storage media in which data is stored in block units, and blocks stored in these portable storage media are lined as a certain number of continuous data. And in the data processing device that is accessed in line units,
The portable storage medium is loaded, and a plurality of read / write means for reading / writing data in line units from / to the loaded portable storage medium, and a logical block address and a cache for reading / writing by the read / write means are stored in the cache. It comprises a recording means for obtaining the minimum value of the address difference from the logical block address of the existing line for each reading and writing and recording the appearance frequency of the minimum value.

The data processing apparatus of the present invention uses a plurality of portable storage media in which data is stored in block units, and blocks stored in these portable storage media are lined as a certain fixed number of continuous data. And in the data processing device that is accessed in line units,
A plurality of read / write means for loading / unloading the portable storage medium and reading / writing data to / from the loaded portable storage medium on a line-by-line basis, receiving means for receiving an access request from the outside, and this receiving means Processing means for reading or writing data using the read / write means in response to the access request received by the means, and the access request received by the receiving means is read / written by the read / write means. When the number of blocks forming a line is changed, changing means for changing the number of blocks forming the line according to a request, and optimal placement according to the access frequency in line units according to the change of the changing means And the optimal placement means for performing.

[0022]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows a schematic configuration of a data processing device according to the present invention. That is, this data processing device includes a CPU 1, an external I / O 2,
RAM3, NVM4, ROM5, timer 6, I / O7,
It is composed of an I / O 8, an I / O 9, an auto changer 10, and a battery 30.

CPU 1, external I / O 2, RAM 3, NV
M4, ROM5, timer 6, I / O7, I / O8, I /
The O9 is connected by the internal bus 20. Further, the battery 30 is connected to the NVM 4 and the timer 6.

The CPU 1 is a processor that controls all the processes of the data processing device. The external I / O 2 as a transfer means (reception means) receives a request from the outside on the internal bus 20.
To transfer data between the internal bus 20 and the outside.

The RAM 3 is a storage means for storing software control data such as a data segment and a stack segment when executing a program, a cache of an optical disk (secondary storage means) and its cache management table 3a.

The NVM (nonvolatile RAM) 4 is a non-volatile memory backed up by the battery 30, and does not lose the data on the NVM 4 even if a power failure occurs.
The NVM 4 is adapted to store important management data that is changed with the operation and cannot be lost. As a table in which management data as this management means is stored, an OD management table 4a, a device management table 4b, a logical-physical conversion table 4c, and an access pattern characteristic table 4d are provided.

The ROM 5 stores data that does not change, such as programs. When power is turned on, CP
U1 starts executing the program from a specific address in ROM5. A program for processing a command requested from the outside is also stored in the ROM 5.

The I / Os 7 to 9 are, for example, SCSI controllers, which control devices to transfer data. The auto changer 10 includes a drive 11, ...
Accessor 12, storage slot 13, optical disk 14,
.., which is composed of the input / output port 15.

The drives 11, ... Are read / write means for reading / writing the optical disk 14, and are composed of, for example, two units. The drive 11 includes an optical disk 14
There is a loading / unloading port for loading and unloading, and the optical disk 14 is loaded and unloaded by the accessor 12. Loading into the drive 11 means taking in the optical disk 14 at the loading / unloading port and rotating the optical disk 14 so that reading / writing is possible. Such a state is called loaded. The removal is to stop the rotation of the optical disk 14 and take it out of the loading / unloading slot.

The storage slot 13 includes optical disks 14, ...
It is a shelf for storing. The accessor 12 is used when the optical disc 14 is transported between the storage slot 13, the drive 11 and the loading / unloading port 15, and has the following two functions. A moving function of moving the accessor 12 to the front of the storage slot 13, the drive 11 or the loading / unloading port 15, and the accessor 12 by sliding the optical disk 14.
It has a sliding function of pulling out the optical disk 14 or inserting the optical disk 14 of the accessor 12 into the storage slot 13 or the like.

The optical disk 14 is a portable, large-capacity, inexpensive storage medium. Block (1K bytes in this embodiment)
It is a medium that can be randomly accessed in units. An optical disk ID is attached to each optical disk 14.

The loading / unloading port 15 is used for the auto changer 10.
It is an entrance / exit for inserting / removing the auto changer 10 and the optical disk 14 into / from the storage slot 13 in which the optical disk 14 is inserted from the outside or the optical disk 14 in the storage slot 13 is taken out. This includes
There are a door, a mechanism for opening the door, and a sensor (not shown) for inspecting whether the optical disk 14 is present inside. Although the door is normally closed, the autochanger 10 automatically opens when the door is inserted and when it is taken out.

The battery 30 holds the data of the NVM 4 when the power supply to this system is stopped, and
This is for moving the timer 6. During normal operation, this data processing device uses an external power source (not shown), and the battery 30 is charged by the external power source.

In this embodiment, the block size will be described as 1 Kbyte. FIG. 2 shows an example of the OD management table 4a recorded in the NVM 4. An optical disk ID is attached to the optical disk 14 in the data processing apparatus of the present invention. The OD management table 4a stores the optical disk 14 in the storage slot 1 in the order of the optical disk ID.
The location information indicating where in 3 or where in the drive 11 it is and the device ID to which it belongs are recorded. The position information S represents the storage slot 13, and D represents the drive 11 of the optical disc 14. For example,
The optical disc 14 having the optical disc ID "0" indicates that it is in the model number 0 of the drive 11 of the optical disc in the position information "D0", and the device ID to which it belongs is "0".

FIG. 3 shows an example of the device management table 4b recorded in the NVM 4. The device management table 4b represents the optical disk 14 that constitutes the device that the data processing apparatus is showing to the outside and the size of the line expressed in block units.

FIG. 4 shows an example of the cache management table 3a recorded in the RAM 3. The cache is managed by dividing it into minimum line size units and assigning cache IDs. In this embodiment, the minimum line size is 1
It has 6 blocks. Cache management table 3a
Stores the cache ID, the device ID of the data stored in the cache, the leading logical block address of 16 blocks, and the flag.

The flag represents the state of the cache, "free" indicating that there is no data in the cache,
The data in the cache is the optical disk (secondary storage device) 14
“Clean” indicating that no changes have been made while being read from the cache, the data in the cache has changed and is different from the data in the secondary storage device, so it is necessary to write it from the cache to the secondary storage device. There are three types of "dirty" to indicate that.

The cache management table 3a is arranged in the order in which the cache is accessed. The cache at the head of the cache management table 3a is the most recently accessed cache, and the cache at the end indicates that the access is older.

FIG. 5 shows an example of the logical-physical conversion table 4c recorded in the NVM 4. The logical / physical conversion table 4c is a table for converting the logical block address in the device at the external interface as the conversion means to the optical disk 14 which is actually physically stored and the block address in the optical disk 14.

In this embodiment, the minimum line size is 16
Since blocks are used, it is possible to convert a logical block address to a physical block address in units of 16 blocks, and it is possible to convert all logical block addresses.

FIG. 6 shows an example of the access pattern feature table 4d recorded in the NVM 4. The access pattern feature table 4d is used to record, for each device, an outline of the distance between the block address and the logical block address of the block that has been accessed most recently when the block is accessed.

For example, if the accessed logical block address is 18 and the closest logical block address of the recently accessed block is 30, the distance is 12. If the distance is 0 to 15, the item of access counter A is added, and if the distance is 16 to 31, the item of access counter B is added, and the distance is 32.
From 63 to 63, add the items of access counter C,
If the distance is from 64 to 127, the item of the access counter D is added. If the distance is 128 or more, the access counter is not added.

FIG. 7 shows a software module configuration. The data processing device of the present invention is connected to another computer or the like, the command receiving unit 40 receives a command from the external device, and activates a module for processing the command according to the content of the command.

The data transfer processing unit 41 is a module that executes a process of reading or writing data to this data processing device as a processing means. In addition, the optimum placement unit 43 as the optimum placement unit is notified of the data access.

The line size change processing unit 42 is a module for changing the line size determined for each device as a changing means. The optimal rearrangement unit 43 is notified of the rearrangement of data due to the line size change.

The optimum arrangement according to the access frequency is disclosed in Japanese Patent Application No. 5-238437 and Japanese Patent Application No. 6-2077.
No. 71 and Japanese Patent Application No. 7-59499. First, the command receiving unit 40 will be described.

When the power of the data processor is turned on and the power is turned on, the CPU 1 executes the program on the ROM 5. First, the cache management table 3a is set. All flags are set to "free" and other parts are set to "-1".

When there is a request from the outside, the external I / O2
Interrupts the CPU 1 and the command receiving unit 40
Is executed. The CPU 1 reads the request from the external I / O 2. The request includes a "read command" that transfers data from the data processing device to the outside, a "write command" that transfers data to the data processing device from the outside, and a "line size change command" that changes the line size of the device. There are three types.

In the case of "read command" or "write command", the data transfer processing unit 41 is activated, and in the case of "line size change command", the line size change processing unit 42 is executed.

The "read command" includes a device ID for storing data to be read by the outside (this ID is called a request device ID) and a logical block address at the head of the data (this is called a request start logical block address). , Size of data (called number of required transfer blocks)
Consists of

Similarly, for the "write command", the request device ID for storing the data to be externally written, the request start logical block address, the number of request transfer blocks, and
It consists of write data.

The "line size change command" consists of the device ID and line size whose line size is desired to be changed. In such a configuration, the present invention uses a cache for distance detection. In other words, the fact that the data existing in the cache is the data accessed within a fixed time is used, and a certain data D (logical block address A
When d) is accessed, it is searched whether the data D is in the cache and at the same time the data Cn (n
Is the cache block number) and the minimum value Lmin of the distance from the data Cn is obtained from the logical block address An.

Lmin = MIN (Ad-An) Ad> An The line size is determined from the tendency of the value of Lmin. As a result, it is possible to suppress the reading of useless data and improve the "look-ahead" effect. Moreover, by dynamically changing the line size, it is possible to deal with the case where the access pattern changes.

Next, the operation of the data transfer processing unit 41 will be described with reference to FIG.
This will be described with reference to the flowchart of FIG. First, CPU1
Starts processing of "read command" or "write command" (ST1).

In step ST2, the CPU 1 obtains the line size of the device indicated by the requested device ID from the device management table 4b. The block address and the number of lines of the logical block of the line to be accessed are obtained from the requested start logical block address and the requested transfer block number. The quotient obtained by dividing the required start logical block address by the line size is the leading logical block address of the first line. The quotient obtained by adding the requested transfer size to the requested start logical block address, subtracting 1, and dividing the value by the line size is the leading logical block address of the last line.

For example, as shown in FIG. 9, when a device with a line size of 32 blocks requests a transfer block number of 3 from the start logical block address 31, lines and logical blocks starting from logical block address 0 are stored in the cache. The two lines starting at address 32 are the lines to be accessed.

In step ST3, the CPU 1 sets the first flag indicating the search for the first line on the RAM 3. In step ST4, the CPU 1 checks the device ID and the logical block address of the cache management table 3a to see if the line is in the cache, and if they are equal to the device ID in the command and the logical block address obtained in step ST2, the CPU 1 determines that the line is in the cache. (Cache hit), and if there is no equivalent, it is not in the cache (cache miss). If the initial flag is set, at the same time as this search, if the logical block address of the cache is smaller than the requested start logical block address with respect to all the caches having the same device ID and forming the head of the line, the difference is calculated. Then, the value obtained by adding the transfer size to these minimum values is set as the distance, and the initial flag is reset.

In step ST5, the CPU 1 rearranges the row of the cache management table 3a so that the row of the cache management table 3a becomes the head of the cache management table 3a when a cache hit is found by searching the cache management table 3a.

In step ST6, if not in the cache, the CPU 1 requests the device ID, the logical block address at the head of the line, and the line size (block unit).
Is set in the RAM 3 and passed to a cache load program (described later) to read data from the optical disk 14 into the cache.

In step ST7, the CPU 1 transfers the data to the outside when one line is prepared in the cache. In case of read, external I / from cache
Transfer to the outside through O2, and in case of write, external I / O
Transfer from 2 to cache. In the case of a write, the cache management table 3a of the cache that writes the data
Set the flag to "dirty".

In step ST8, the CPU 1 checks whether there is a line to be accessed, and if there is a next line, the process returns to step ST4. In step ST9, when the access to all the lines is completed, the CPU 1 increments the access counter by "1" as described in the access pattern feature table 4d in FIG. In addition, when the counters have reached the maximum value so that the counters do not overflow, set the values of all counters to 1/2.
To

The CPU 1 finally notifies the end of the transfer to the outside and ends the processing. Next, the operation of the cache load program will be described with reference to the flowchart of FIG.

In step ST11, the CPU 1 first uses the cache load program to set "0" to the total victim line size on the RAM 3.
In step ST12, the CPU 1 obtains the device ID of the cache at the end of the cache management table 3a, and further obtains the line size from the device management table 4b. This line is called the victim line. Add the victim line size to the total victim line size.

In step ST13, the CPU 1 must have all victim lines in the cache, and the caches forming the lines are set in the cache management table 3a.
Then, it is checked whether the flag has "dirty", and if there is "dirty", the data is stored in the optical disk 14 which is the secondary storage.

To store data on the optical disk 14, the logical block address of the data on the victim line is converted into the optical device ID and the block address on the optical disk 14 by the logical / physical conversion table 4c. The OD management table 4a is searched, and the optical disk 14 is the drive 11
If not, remove the optical disk 14 from the drive 11 and return it to the empty storage slot 13, set the number of the storage slot 13 stored in the position of the OD management table 4a of the removed optical disk 14, The required optical disk 14 is removed and loaded into the drive 11, and the drive number is set at the position of the OD management table 4a of this optical disk 14. The victim line is transferred to the block of the converted block address on the optical disc 14.

If the total victim line size is less than the line size passed to the cache load program in step ST14, the CPU 1 determines in step ST14.
Returning to step 12, if the line size or more, step ST1
Go to 5.

In step ST15, the CPU 1 converts the logical block address given to the cache load program into the optical device ID and the block address by using the logical / physical conversion table 4c, and similarly to step ST13, the optical disk of this optical device ID is converted. 14 is loaded in the drive 11, and data is transferred from the optical disc 14 to the victim line.

In step ST16, the CPU 1 sets the device ID and the logical block ID in the cache management table 3a corresponding to the cache forming the victim line after the transfer, sets the flag to "clean", and sets this tuple. The management tables are rearranged so as to be the head of the device management table 4b and arranged in the order of access.

Then, the process returns to the called processing unit. next,
The operation of the line size changing process will be described with reference to the flowchart of FIG.

In step ST21, the CPU 1 reads a request from the external I / O 2 and obtains a device ID (hereinafter referred to as requested device ID) and a line size (hereinafter referred to as requested line size).

CPU1 in steps ST22 and ST23
If the requested line size is "0", the new line size is determined from the largest counter of the four access counter values of the requested device ID in the access pattern feature table 4d. For example, in the case of FIG. 6, the line size is 64K.

At steps ST22, 24 and 29, C
If the requested line size is not “0”, the PU 1 determines that the requested size corresponds to the corresponding line size (16 in the case of this embodiment,
32, 64, 128), and if different, an error is returned to the host.

CPU1 in steps ST25 and ST27
If the requested line size is equal to or smaller than the line size of the device management table 4b (hereinafter referred to as the current line size), the line size of the device management table 4b is changed to the requested line size in step ST27.

CPU 1 in steps ST26 and ST27
Needs to change the data arrangement when making the line size larger than the current line size. This is because blocks within a line are arranged in physically continuous areas. For this reason, once the optimum placement is stopped, the data is rearranged and rearranged as shown in FIG. 12A so that the physical block addresses correspond to the logical block addresses in order. The logical-physical conversion table 4c is as shown in (b) to (c) of FIG. Then, in step ST27, the line size of the device management table 4b is changed.

In step ST28, the CPU 1 notifies the external I / O 2 of the end of the line size change and sends an external end signal. As described above, according to the embodiments of the present invention, a plurality of devices are provided inside, and the physical arrangement of data can be changed according to the access frequency without changing the address on the external interface of the data processing device. In the data processor for optimal placement, the optimal placement is performed in units of a fixed number of consecutive blocks (lines), and the transfer of cache in / out to the device and the internal cache is also performed in units of lines (read ahead). The line size can be automatically determined and dynamically changed by the access characteristics of.

[0076]

As described in detail above, according to the present invention,
By dynamically changing the line size, it is possible to provide a data processing device capable of suppressing the reading of useless data and improving the "look-ahead" effect.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic configuration of a data processing device according to the present invention.

FIG. 2 is a diagram showing an example of an OD management table recorded in NVM.

FIG. 3 is a diagram showing an example of a device management table recorded in NVM.

FIG. 4 is a diagram showing an example of a cache management table recorded in a RAM.

FIG. 5 is a diagram showing an example of a logical-physical conversion table recorded in NVM.

FIG. 6 is a diagram showing an example of an access pattern feature table recorded in NVM.

FIG. 7 is a diagram showing a software module configuration.

FIG. 8 is a flowchart for explaining the operation of the data transfer processing unit.

FIG. 9 is a diagram for explaining access in a device having a line size of 32 blocks.

FIG. 10 is a flowchart for explaining the operation of a cache load program.

FIG. 11 is a flowchart for explaining the operation of line size change processing.

FIG. 12 is a diagram for explaining the line size change of the logical-physical conversion table.

FIG. 13 is a diagram showing a data processing device using a conventional optical disc and an autochanger.

FIG. 14 is a diagram for explaining a logical block address and a physical block address.

[Explanation of symbols]

 1 ... CPU 2 ... External I / O 3 ... RAM 4 ... NVM 5 ... ROM 10 ... Autochanger 11 ... Drive 12 ... Accessor 13 ... Storage slot 14 ... Optical disk 15 ... Loading / unloading port 20 ... Internal bus

Claims (7)

[Claims] 1. A plurality of portable storage media in which data is stored in block units are used, blocks stored in these portable storage media are configured as a certain fixed number of continuous data, and a line is formed. In a data processing device accessed in units, when the portable storage medium is accessed, detection means for detecting the access pattern, and the number of blocks forming the line according to the access pattern detected by the detection means A data processing device comprising: a changing unit for changing the. 2. A plurality of portable storage media in which data is stored in block units are used, blocks stored in these portable storage media are configured as a certain number of continuous data, and a line is formed. In a data processing device accessed in units, the portable storage medium is loaded, and a plurality of read / write means for reading / writing data in line units from / to the loaded portable storage medium; Storage means for storing cache data for the medium, receiving means for receiving an access request from the outside, and data of the access request received by the receiving means is stored in the storage means, From the block address of the data and the corresponding block address of the portable storage medium, from the portable storage medium to the above The data processing apparatus according to claim determination means for determining the number of reading lines in 憶 means, by comprising a. 3. A plurality of portable storage media in which data is stored in block units are used, blocks stored in these portable storage media are configured as a certain number of continuous data, and a line is formed. In a data processing device accessed in units, the portable storage medium is loaded, and a plurality of read / write means for reading / writing data in line units from / to the loaded portable storage medium; Storage means for storing cache data for the medium, receiving means for receiving an access request from the outside, and data of the access request received by the receiving means is stored in the storage means, From the block address of the data and the corresponding block address of the portable storage medium, from the portable storage medium to the above Determining means for determining the number of lines to be read into the storing means, and data of the number of lines determined by the determining means is read from the portable storage medium using the reading / writing means and stored in the storage means, and the receiving means A data processing apparatus comprising: a processing unit that performs processing according to the access request received in. 4. A line is constructed by using a plurality of portable storage media in which data is stored in block units, and blocks stored in these portable storage media are arranged as a certain number of continuous data, and the line is formed. In a data processing device accessed in units, a plurality of read / write means for loading / unloading the portable storage medium and reading / writing data in line from / into the loaded portable storage medium, and for each of the read / write means And a management unit that manages the portable storage medium corresponding to a certain number of blocks forming the line, and a data processing apparatus. 5. A plurality of portable storage media in which data is stored in block units are used, and blocks stored in these portable storage media constitute a line as a certain number of continuous data, and this line A data processing device accessed in units, comprising a conversion means for converting a logical block address to a physical block address for each minimum line unit that constitutes the above-mentioned fixed number of continuous blocks. 6. A plurality of portable storage media in which data is stored in block units are used, and blocks stored in these portable storage media constitute a line as a certain number of continuous data, and this line In a data processing device accessed in units, a plurality of reading / writing means for loading / unloading the portable storage medium and reading / writing data in line units from / to the loaded portable storage medium, and the reading / writing means A recording unit that obtains the minimum value of the address difference between the read / write logical block address and the logical block address of the line existing in the cache for each read / write, and records the appearance frequency of the minimum value. Data processing device. 7. A line is constructed by using a plurality of portable storage media in which data is stored in block units, and a block stored in these portable storage media is formed as a certain number of continuous data, and the line is formed. In a data processing device accessed in units, a plurality of read / write means for loading / unloading the portable storage medium and reading / writing data in line from / to the loaded portable storage medium, and access from the outside The receiving means for receiving the request, the processing means for reading or writing the data using the read / write means according to the access request received by the receiving means, and the access request received by the receiving means If there is a change in the number of blocks that make up the line on which data is read or written by the read / write means, the line will be And changing means for changing the number of blocks constituting the data processing apparatus being characterized in that anda optimal placement means for performing optimum arrangement in accordance with the access frequency of the line unit in accordance with the change of the changing means.