JPH1186452A - Disk storage device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve an access speed of the data without enlarging a device by providing a block data write-in/read-out means successively writing/reading the block data in/from a disk with plural heads fitted to a common access arm alienated from each other. SOLUTION: A disk mechanical part 25 is started, and a first head 14 is moved to a control area of a disk, and a file control table is read out by the first head 14, and free areas of respective head areas 14 are detected from the file control table. Then, a data write-in part 26 is started, and N pieces of block data among unwritten-in block data in disk caches 21 are written in the beforehand decided write-in position in respective answering head areas of the disk by respectively allocated N pieces of respective heads 14. Then, the storage positions of the control areas in the file control table are set/ updated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk storage device for storing various information, and more particularly to a disk storage device capable of simultaneously writing and reading information on a single disk with a plurality of heads.

[0002]

2. Description of the Related Art As storage means for temporarily storing data in an information processing system such as a computer, in addition to a main storage device such as a RAM or a ROM directly connected to a system bus, an auxiliary storage for storing a large amount of data. The device is installed. The auxiliary storage device includes a memory card and a disk storage device. Among them, a disk storage device using a disk-type storage medium such as an FD (floppy disk) storage device, an HD (hard disk) storage device, or an MO (magneto-optical disk) storage device is generally used.

However, while this disk storage device has a large storage capacity, it has mechanical movable members such as a disk rotation mechanism and a head moving mechanism, so that a main storage such as a RAM or ROM directly connected to a system bus. It has a disadvantage that the access processing speed as viewed from the CPU side is slower than that of the storage device. Therefore, when a certain amount of data such as file data is written to and read from the disk storage device, a large amount of processing time is required.

In order to eliminate such inconvenience, RAI
D (Redundant Arrays of Inexpensive Disks) Magnetic disk devices have been developed. In this RAID magnetic disk drive, as shown in FIG.
(Hard disk drive device) 1 is arranged, a series of file data 2 is divided into block data 3 in units of several bits, and each of the divided block data is simultaneously written to each HDD 1 for each installation number of HDDs 1. That is, a series of file data 2 is written in a distributed manner to a plurality of HDDs 1 at the same time. When reading the written data, the data is read from each HDD 1 at the same time. Therefore, the access processing time can be greatly reduced as compared with the case where all the data constituting the series of file data 2 are written to and read from one HDD.

[0005]

However, the RAID magnetic disk device shown in FIG. 9 also has the following problems to be improved. That is, in order to improve the data access processing speed to the RAID magnetic disk device, it is necessary to increase the number of HDDs 1 installed. As a result, an information processing system such as a computer in which the RAID magnetic disk device is incorporated becomes large. Further, there is a problem that the manufacturing cost is significantly increased.

The present invention has been made in view of such circumstances, and by simultaneously writing and reading data to and from a single disk with a plurality of heads, the information of a computer or the like in which the apparatus is incorporated. An object of the present invention is to provide a disk storage device capable of significantly increasing the data access processing speed without increasing the size of the processing system and without significantly increasing the manufacturing cost.

[0007]

In order to solve the above-mentioned problems, in a disk storage device according to the present invention, a plurality of heads attached to a common access arm at a distance from one another with respect to one disk simultaneously can simultaneously transmit data. A disk mechanism capable of writing and reading, file data dividing means for dividing input file data into a plurality of block data having a unit data length, and dividing the plurality of divided block data by the number of heads installed Block data writing means for sequentially writing data on the disk by each head, and a file management table stored on the disk and storing the writing position of each block data in the file data written on the disk by the block data writing means And each write position stored in the file management table in response to a read command specifying file data. Block data reading means for sequentially reading each block data by the number of heads installed in each head, and arranging each block data read simultaneously by each head in the order specified by the identification information of the read head of the corresponding block data. Data output means for sending the data to the outside instead.

In the disk storage device configured as described above, the input file data is divided into a plurality of block data having a unit data length. And
The plurality of divided block data are simultaneously written on a disk by each head according to the number of heads installed. Therefore, the time required to write a series of file data to the disk is approximately one-fourth the number of heads installed compared to a disk storage device having only one head.

The disk is provided with a file management table for storing the writing position of each block data of the file data written on the disk. Therefore, when reading out a series of file data stored in this disk, each head can simultaneously read out each block data by referring to the file management table and simultaneously read each block data in the head. By rearranging them in numerical order, the original file data can be reproduced. That is, even at the time of reading,
The time required for reading can be reduced to approximately one-half the number of heads installed.

Further, since data can be simultaneously accessed by a plurality of heads with respect to one disk, the entire apparatus can be formed smaller than an apparatus having a plurality of HDDs, and equipment costs can be reduced.

In the disk storage device according to another aspect of the present invention, when the file data is longer than a specified data length determined according to the type of the disk, a plurality of divided blocks may be used instead of the block data writing means in the above invention. Block data writing means is provided for sequentially writing data to the disk by each head according to the number of heads installed.

In the disk storage device configured as described above, each divided block data is written to the disk by a plurality of heads only when the file data is longer than the specified data length. That is, in the case of short file data, it is not particularly necessary to access using a plurality of heads.

Further, in the disk storage device of another invention according to the above-mentioned invention, when the file data is shorter than a prescribed data length determined according to the type of the disk, a plurality of divided block data are transferred to a plurality of heads. Among them, there is provided a block data writing means for sequentially writing data on the disk with one predetermined head.

That is, in the case of short file data,
Even if access is performed using one head, the processing efficiency of the entire disk storage device does not significantly decrease.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a schematic configuration of a disk storage device according to an embodiment, FIG. 2 is a diagram showing storage contents of a file management table formed on a disk, and FIG. 3 is a disk mechanism of the disk storage device. It is a schematic diagram which extracts and shows.

As shown in FIG. 3, in a disk mechanism unit incorporated in the disk storage device, a common access arm 12 is driven by a head drive mechanism 13 with respect to a disk 11 which is rotationally driven by a drive motor (not shown). 1
1 is controlled to move in the radial direction. The disk 11 is formed of a magnetic disk in this embodiment.

The common access arm 12 has a disk 11
N heads 14 are mounted at equal intervals in the radial direction. Each head 14 is assigned a head number from No. 1 to No. N in order from the tip direction of the common access arm 12. Each head 14 can simultaneously access the recording surface of the disk 11. Common access arm 12
Move in the radial direction, the heads 14 also move in the radial direction on the recording surface of the disk 11 while maintaining a constant interval from each other.

The maximum movement distance of the common access arm 12 is the distance between the heads 14. Therefore, as shown in FIG. 4, on the recording surface of the disk 11, each of the first to N-th heads 14 has a ring-shaped N-th to N-th head area 15 that can be exclusively accessed. Is formed.

A management area 16 is formed inside the first head area 15. In this management area 16, FIG.
The file management table 17 shown in FIG. The management area 16 is accessed only by the first head 14 with the common access arm 12 moved toward the innermost side of the disk 11.

FIG. 5 shows an address positional relationship between the management area 16 formed in the disk 11 and each of the N head areas 15 and a series of data written separately on each head area 15 of the disk 11. , B _N , B _{N + 1} ,..., B ₁ , B ₂ , B ₃ ,.
, _Bn-1 and _Bn .

[0021] That is, file data 18 having an input data length L _A is divided into a plurality of block data 19 for each unit data length L _S. As a result of the division, n having a unit data length L _S from B ₁ to B _n n
Block data 19 are formed.

Then, N pieces of block data 19 (n> N) from B ₁ to B _n are collected together by N heads 14 from No. 1 to No. The data is written into each of the head areas 15 up to the Nth. Therefore, the first head 14 writes each block data 19 of B ₁ , B _{N + 1} , B _{2N + 1} ,... _Into its own first head area 15.

Similarly, B ₂ , B
_Write each block data 19 of _{N + 2} , B _{2N + 2} ,... _Into its own second head area 15. In other words, first, B
N block data 19 from ₁ to B _N are simultaneously written to the disk 11, and then _N block data from B _{N + 1} to B _2N are written.
The block data 19 are simultaneously written on the disk 11.

The file management table 17 shown in FIG.
In the state shown in FIG. 5, the storage position of each file data 18 stored in each head area 15 of the disk 11 is stored.
That is, for each file name of each file data 18, 1
Head area 15 corresponding to head numbers N to N
Block data 1 of the corresponding file data 18 in
Nine storage positions (storage addresses) are stored.

In FIG. 1, the file data 18 shown in FIG. 5 input via the data input / output unit 20 is written to the disk cache 21 and the data length determination unit 2
2 is input. Data length determination unit 22 detects the data length L _A of the file data 18, the detected data length L _A
There determines whether less than the specified data length L _B predetermined specified data length L _B above, the head assignment unit 2 the determination result
Send to 4.

When the input file data 18 is written to the disk cache 21, the data dividing unit 23 converts the file data 18 into a unit data length L as shown in FIG.
_It is divided into a plurality of block data 19 having _S. The head allocating unit 24 allocates each of the divided block data 19 to each of the heads 14, and sends out the allocation result to the data writing unit 26.

The data writing unit 26 is a disk cache 2
Each block data 19 is sent to each head 14 of the disk mechanism 25, and each block data 19 is written to each head area 15 of the disk 11.

The address management unit 27 writes the result of writing each block data 19 to each head area 15 of the disk 11 to the file management table 17 using the first head 14.

Read command 2 specifying a file name from outside
8 is input, the data reading unit 29 sends the file management table 1 of the disk 11 via the address management unit 27.
Referring to FIG. 7, each block data 19 is simultaneously read from each head area 15 of the disk 11, assigned a head number, and written to the disk cache 21.

The data sort processing unit 30 stores the N pieces of block data 19 read into the disk cache 21 at the same time.
Are sorted in order of head number. The data input / output unit 20 sends N block data 19 rearranged in a correct order in the disk cache 21 to the outside as a part of the file data 18.

The operation at the time of data writing in the disk recording apparatus having such a configuration will be described with reference to the flowcharts of FIGS. In S (step) 1, when the file data 18 shown in FIG. 5 is input via the data input / output unit 20, the file data 18 is stored in the disk cache 2.
1 is written (S2). Next, the data length determination unit 22 is activated, the storage of the data length L _A of the file data 18 entered is equal to or higher than the predetermined data length L _B described above (S3), preset the recording medium (magnetic disk) determining a unit data length L _S of the optimal partition the capacitance (S4).

After that, the data dividing unit 23 is activated to divide the input file data 18 into a plurality of block data 19 having a unit data length L _S (S5). The head allocating unit 24 is activated, and the divided block data 1
_{9 (B 1, B 2,} B 3, ... B N, B N + 1, ..., B n-1,
B _n ), the first to Nth heads 14 are allocated (Sn).
6).

Next, the disk mechanism 25 is activated and
The first head 14 is moved to the management area 16 of the disk 11, and the first head 14
7 is read (S7). Then, an empty area of each head area 14 is detected from the file management table 17 (S
8). Next, the write position of each block data 19 in the head area 15 is determined (S9). Specifically, in each of the empty areas of the N head areas 15, the head position of the area that can be used in common is set as the writing position.

Next, the data writing unit 26 is activated, and the unwritten block data 19 in the disk cache 21 is read.
The N pieces of block data 19 are written into the previously determined write positions in the corresponding head areas 15 of the disk 11 with the N pieces of the allocated heads 14 (S10). Then, the new setting of the storage position for each head area 15 with respect to the corresponding file data 18 in the file management table 17 of the management area 16 of the disk 11,
Alternatively, the storage location already written is updated (S1).
1).

If there is unwritten block data 19 in the disk cache 21 (S12), S
9, the write position of each block data 19 in the head area 15 is determined. When there is no unwritten block data 19 in the disk cache 21, the data writing process ends.

Further, in S3, if it is less than the data length L _A of the file data 18 entered is above defined data length L _B, preset for example No. 1, etc. One of the file data 18 input Allocate the head 14 (S1
3). Then, at S14 in FIG. 7, to divide the file data 18 input into a plurality of block data 19 with the unit data length L _S.

Next, the file management table 17 in the management area 16 of the disk 11 is read (S15), and an empty area in the head area 15 corresponding to the head 14 previously allocated is detected (S16).

Next, the corresponding 1 of each block data 19
Write position in one head area 15 is determined (S
17). Next, the data writing unit 26 is activated, and one block data 19 among the unwritten block data 19 in the disk cache 21 is assigned to the corresponding head 14 of the disk 11 by the designated one head 14. Write to the previously determined write position of one corresponding head area 15 (S
18). Then, a new storage location for the corresponding file data 18 in the file management table 17 of the management area 16 is updated, or a previously written storage location is updated (S19).

If unwritten block data 19 exists in the disk cache 21 (S20), the process proceeds to S20.
17, the same head area 15 of the block data 19 is returned.
The writing position within the. Disk cache 2
When there is no unwritten block data 19 in 1, the data writing process is terminated.

Next, the operation at the time of reading data in the disk recording apparatus will be described with reference to the flowchart of FIG. At step Q1, a read command 2 specifying a file name from the outside
When 8 is input, the data reading unit 29 starts up, reads the file management table 17 via the address management unit 27 (Q2), and reads the storage position of each head area 15 in the file name to be read. (Q3).

The block data 19 written in each head area 15 of the disk 11 is simultaneously read out by the N heads 14 from 1 to N, and the head number is stored in the disk cache 21. Write with (Q
4). Thereafter, the data sort processing unit 30 is activated, and rearranges the N block data 19 temporarily stored in the disk cache 21 in the order of the head numbers from 1 to N assigned to each block data 19 (Q5). . Then, the N block data 19 temporarily rearranged in the normal order and stored in the disk cache 21 are transmitted from the data input / output unit 20 to the outside as a part of the file data 18.

If the reading process of all the block data 19 of the file data 18 to be read has not been completed (Q7), the process returns to Q4, where the unread block data 19 from each head area 15 of the disk 11 is read. Start reading.

When the reading process of all the block data 19 constituting the file data 18 is completed, the data reading process is completed. In the disk storage device thus configured, the file data 1 input from the outside
8 is divided into a plurality of block data 19 having a unit data length L _S as shown in FIG. Then, the plurality of divided block data 19 are collectively written N by the number of installed heads 14 and simultaneously written into the corresponding head areas 15 of the disk 11 by the N heads 14.

Therefore, the time required to write a series of file data 18 to the disk 11 is approximately 1 / N compared to a disk storage device having only one head. The storage position of each block data 19 in the file data 18 written on the disk 11 is stored in the file management table 17 formed in the management area 16 of the same disk 11.

Therefore, when reading out a series of file data 18 stored on the disk 11, the N heads 14 refer to the file management table 17 so that the N heads 14 store the N data stored in each head area 15. Block data 19 can be read simultaneously. Further, the head 14 that has read each block data 19 read at the same time
The original file data 18 can be reproduced by rearranging the files in the order of numbers. That is, even when reading data, the required reading time can be reduced to approximately 1 / N.

Further, in the disk mechanism 25,
As shown in FIG. 3, a plurality of heads 14 can simultaneously access data to one disk 11. Therefore, as compared with the RAID magnetic disk storage device having a plurality of HDDs shown in FIG. 9, the entire device can be configured to be smaller, and the equipment cost can be reduced.

[0047] Further, the data length L _A is even when less than the predetermined data length L _B determined in accordance with the type and storage capacity of the disk 11 as a recording medium, divided into a plurality of block data 19 which is the file data 18 input The same head area 15 of the disk 11 is determined by a predetermined one of the N heads 14 such as the first head 14.
Are sequentially written to For example, in a magnetic disk or the like having a large storage capacity, the specified data length L _B
Long, a floppy disk storage capacity is small conversely are set a short prescribed data length L _B.

In this case, when the file data 18 is read from the disk 11, the block data 19 is read only from one head area 15, so that it is not necessary to rearrange the block data 19 after reading.

That is, in the case of short file data 18, even if access is made using one head 14,
The processing efficiency of the entire disk storage device is not significantly reduced.

[0050] When writing a number Manual continuously shorter file data 18 defines the data length L _B, by going to be shifted sequentially write specified head 14 from No. 1 to No. N, each head area of the disk 11 15 so that the block data 19 is written almost evenly.

[0051] By controlling in this way, when writing a long file data 18 than the predetermined data length L _B in the disk 11 later, each block data 19 simultaneously at each head region 15 using N heads 14 Write
Since the empty area of each head area 15 is substantially the same, each head area 15 can be used effectively.

The present invention is not limited to the above embodiment. Disk 11 for storing various data
Is a magnetic disk, but may be a floppy disk or a magneto-optical disk.

[0053]

As described above, the disk storage device of the present invention is provided with a disk mechanism for simultaneously writing and reading data to and from a single disk with a plurality of heads, and a series of input files. The data is divided into a plurality of block data having a unit data length, and the block data is allocated to each head and simultaneously written to a disk.

Therefore, the time required for writing and reading a series of file data to and from the disk can be reduced in accordance with the number of heads installed. Therefore, the data access processing efficiency can be significantly improved without increasing the size of the information processing system such as a computer in which the disk storage device is incorporated and without significantly increasing the manufacturing cost.

[Brief description of the drawings]

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

FIG. 2 is a diagram showing storage contents of a file management table formed on a disk of the disk storage device.

FIG. 3 is a schematic configuration diagram of a disk mechanism unit of the disk storage device.

FIG. 4 is a diagram showing a storage area of a disk of the disk storage device.

FIG. 5 is a diagram showing a relationship between file data of the disk storage device and a storage area of the disk;

FIG. 6 is a flowchart showing a data write processing operation of the disk storage device.

FIG. 7 is a flowchart showing a data write processing operation of the disk storage device.

FIG. 8 is a flowchart showing a data read processing operation of the disk storage device.

FIG. 9 is a schematic diagram showing a write operation in a conventional RAID magnetic disk storage device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 ... Disk 12 ... Common access arm 14 ... Head 15 ... Head area 16 ... Management area 17 ... File management table 18 ... File data 19 ... Block data 20 ... Data input / output unit 21 ... Disk cache 22 ... Data length judgment unit 23 ... Data dividing unit 24 ... Head allocating unit 25 ... Disk mechanism unit 26 ... Data writing unit 27 ... Address management unit 29 ... Data reading unit 30 ... Data sorting processing unit

Claims (3)

[Claims] 1. A disk mechanism capable of simultaneously writing and reading data with a plurality of heads mounted on a common access arm at a distance from each other with respect to one disk, and a unit data unit for converting input file data into unit data. File data dividing means for dividing the data into a plurality of block data having a length; and block data writing means for sequentially writing the divided plurality of block data on the disk by the number of heads installed by each of the heads. A file management table stored in the disk and storing a write position of each block data in the file data written to the disk by the block data writing unit; and a read command specifying the file data. , Each block data from each writing position stored in the file management table by each head. Block data reading means for sequentially reading out the number of heads installed, and rearranging the block data read out simultaneously by each head in the order specified by the identification information of the read head of the corresponding block data and sending it out. A disk storage device comprising data output means. 2. A disk mechanism capable of simultaneously writing and reading data with a plurality of heads attached to a common access arm at a distance from one disk with respect to one disk; File data dividing means for dividing the plurality of divided block data into a plurality of block data having a length, and when the file data is equal to or longer than a specified data length determined according to the type of the disk, the number of the plurality of divided block data is set to Block data writing means for sequentially writing data to the disk by each of the heads; and writing of each block data in the file data stored in the disk and written to the disk by the block data writing means. According to a file management table for storing the position and a read command specifying the file data, the file is stored. Block data reading means for sequentially reading each block data by each head from the respective writing positions stored in the file management table according to the number of heads installed, and reading each block data read simultaneously by each head from a corresponding block. A data output means for rearranging the data in the order specified by the identification information of the data read head and sending the data to the outside; 3. When the file data is shorter than a specified data length determined according to the type of the disk, the divided plurality of block data is transferred to the disk by a predetermined one of the plurality of heads. 3. The disk storage device according to claim 2, further comprising block data writing means for sequentially writing data to the disk.