JPH07191813A - Data writing method - Google Patents

Abstract

PURPOSE:To reduce seek time and the rotation waiting time of a disk. CONSTITUTION:When a data control means 3 receives data to be written into auxiliary storage devices 10a-10c of disk structure, the received data are divided in the same size. The rotation waiting time is grasped by calculating the write starting positions of the auxiliary storage devices 10a-10c of storage means at the time point when a data writing instruction is outputted from the data control means 3. The rotation waiting time and the queue of data inside respective drivers 2a-2c of managing means corresponding to the respective disk auxiliary storage devices 10a-10c are confirmed, and these divided data are distributed to the respective auxiliary storage devices 10a-10c. The data control means 3 manages which of auxiliary storage devices 10a-10c the equally divided data are written in. The respective drivers 2a-2c fetch the data distributed from the data control means 3, write them in the auxiliary storage devices 10a-10c corresponding to the respective drivers 2a-2c and preserve those data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data writing method in a system including a plurality of auxiliary storage devices having a disk structure.

[0002]

2. Description of the Related Art A conventional data writing method provided with a plurality of disk-structured auxiliary storage devices is shown in FIG.
In 3), each management device (driver 1, 2, 3) manages and controls data. That is, the disk device to which the data is to be written is determined and the management device corresponding to the disk device manages the data. For example, when writing data to a magnetic disk, the head of the magnetic disk device is moved to the next cylinder and the data is written after waiting for the magnetic disk to reach the write position.

[0003]

As described above, according to the data writing method having the conventional auxiliary storage device having a plurality of disk structures, the seek of the head is performed at the time of writing the data. There is a problem that a rotation waiting time (average) of 1/2 of the required time is required. SUMMARY OF THE INVENTION It is an object of the present invention to eliminate the above-mentioned drawbacks of the prior art and to provide a data writing method in which seek time and disk rotation waiting time are reduced.

[0004]

In order to achieve the above object, in the present invention, a data storage means having an auxiliary storage device having a plurality of disk structures, and an auxiliary storage device having a plurality of disk structures of the data storage means. The management means for managing the access time and queue of the data for each and the write data to be stored in the auxiliary storage device of the previous disk structure are equally divided, and the auxiliary storage of the plurality of disk structures is prepared based on the management information of the previous management means. It is an object of the present invention to provide a data writing method characterized by comprising a data control means for controlling transmission of write data to be stored in the device.

[0005]

In the structure thus constructed, when the data to be written in the auxiliary storage device having the disk structure is received by the data control means, the received data is divided into the same size. The write start position of the auxiliary storage device having the disk structure of the storage device at the time of outputting the data write command from the data control device is calculated, and the rotation waiting time is grasped. The queue of data in each driver of the management means corresponding to the rotation waiting time and the auxiliary storage device of each disk structure is confirmed, and this divided data is distributed to the auxiliary storage device of each disk structure. The data control means manages in which disk structure the evenly divided data is written in the auxiliary storage device. Each driver of the management means fetches the distributed data from the data control means, writes the data in an auxiliary storage device having a disk structure corresponding to each driver of the management means, and saves the data.

[0006]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a configuration for carrying out a data writing method according to an embodiment of the present invention. Storage means 1 including a plurality of magnetic disks 10a to 10c having the same block length and the same rotation speed for storing data, drivers 2a to 2c of a management means for managing a queue for each magnetic disk 10a to 10c and an access time, A plurality of magnetic disks 10a to 10c are treated as one virtual disk, and are composed of a virtual driver 3 of a data control means for distributing data writing to the plurality of disks based on the management information of the management means 2.

Next, the operation of the data writing method of the present invention will be described with reference to FIGS. 1, 2 and 3. FIG. 2 is a diagram showing the rotation waiting time of each magnetic disk in the embodiment of the present invention. FIG. 3A shows a block 0 of the magnetic disk.
FIG. 3B is a diagram showing the state of the head at the start of reading, and FIG. 3B is a diagram showing the state of the head at the completion of reading block 0 of the magnetic disk. First, when the computer is started up, each disk (10a to 10c) is sent from the virtual driver 3 in order to know in which block the head is located when data is actually read or written.
The read command is output to block 0 of the outermost cylinder of the block, and block 0 is output after the read command is output.
Is stored as management information of each of the drivers 2a to 2c. In this case, the magnetic disks (10a to 10c) have the same speed, but the reading start position 5 (block 0) when the disks start to rotate when the computer starts up is not always located at the same position. However, the time required for reading one block is the same, but the time required for the head 4 to reach the block 0 in which the read data is stored is different for each disk (10a to 10c). On the other hand, the time from the output of the read command to the end of the read differs for each disk (10a to 10c). For example, in the disk 10a, 3 milliseconds after the read command is output to the block 0, 10b is 5
It is assumed that the reading is completed in 6 milliseconds in milliseconds and 10c. The management unit 2 stores the read end time. At this time, each driver (2a to 2c) manages information of each disk (10a to 10c), such as the driver 2a for the information of the disk 10a and the driver 2b for the information of the disk 10b. At the same time, at the end of reading of this block 0, based on the address that manages the write status, which is one of the management information of the drivers 2a to 2c of the management means 2,
Next, the head 4 is moved to the writable cylinder. When starting up the computer, each disk (10a-10
Since the read command is output to the block 0 in (c), as shown in FIG.
When viewed from the head 4 (0a to 10c), the read start position 5 (the head of block 0) is one block ahead. Then, after a certain time, write commands are simultaneously output to the respective disks (10a to 10c). The elapsed time from the end of reading for block 0 to the output of this write command can be calculated from the read end time stored in each driver (2a to 2c). For example, if the elapsed time until the write command is output is 10 milliseconds on the disk 10a, the elapsed time from the read end time of block 0 of the other disks 10b and 10c to the output of this write command. Considering the read end time of the disk 10a of 3 milliseconds as a reference, the read end time of the disk 10b is 5 milliseconds, so
This means that the reading is completed with a delay of millisecond, and the elapsed time of the disk 10a is 10 milliseconds, which is 8 milliseconds obtained by subtracting the time difference of 2 milliseconds. Similarly, the read end time of block 0 of the disk 10c is 6 milliseconds, which is 3 milliseconds later than the end time of the disk 10a, so the elapsed time of the disk 10c is 7 milliseconds, which is 10 milliseconds of the disk 10a minus 3 milliseconds. Seconds. Each driver (2a to 2c) grasps the read start position 5 (head of block 0) at the time when the write command is output from the virtual disk 3 from the elapsed time thus obtained. Since the number of blocks in one track and the rotation time of one disk (10a to 10c) can be calculated in advance, the virtual disk 3
Based on the position of the read start position 5 (block 0) at the time when the write command for each driver (2a to 2c) is output from, the block to be written on each disk (10a to 10c) is directly under the head 4. The time comparison can be understood from the remainder calculated by the following formula.

(Elapsed time + data transfer time) ÷ 1 rotation time of magnetic disk The relationship between the remainder calculated by the above formula and the rotation waiting time is shown below. Too much ... Large rotation waiting time ... Small Too little ... 0 rotation waiting time ... 0 For example, assuming that the data transfer time per block is 1 ms and the magnetic disk rotation time is 10 ms, in the above example, the disks 10a ... (10 ms + 1 ms) / 10 ms = 1 Not much 1 disk 10b ... (8 ms + 1 ms) ÷ 10 ms = 0 too much 9 disks 10 c ... (7 ms + 1 ms) ÷ 10 ms = 0 too much 8 and in this case the rotation waiting time of disk 10 b 6b will be the shortest. In consideration of the rotation waiting time 6 and the data transfer time of the queue of the data temporarily stored in each driver (2a to 2c), the virtual driver 3 newly receives the data on the disk capable of the fastest data processing. Distribute data. Since the data of each driver (2a to 2c) is evenly divided in block length units, the data transfer time can be grasped from the number of data.
The virtual driver 3 manages division of data in block length units, calculation of rotation waiting time, management of which block of which disk the divided data is stored, and the like. Similarly, when the one-rotation time of each magnetic disk (10a to 10c) is different, the block length, the number of blocks per track, one rotation time of the magnetic disk, and the data transfer time are similarly obtained, and the rotation waiting is performed. You can figure out time 6. In this way, by dividing the write data evenly and distributing the data to each disk, it is possible to process the data at the same time, which makes it possible to speed up the process and grasp the length of the rotation waiting time to grasp the data. By performing the distribution, the speedup is superior.

[0009]

According to the present invention, since the write data is evenly divided and the data control means for collectively managing the states of a plurality of magnetic disks is provided, the data can be distributed, so that one data can be recorded in a plurality of disks. Simultaneous processing is possible. Further, since the rotation waiting times of the respective disks can be compared, and the disks can be selected in consideration of the rotation waiting time, it becomes possible to perform faster file access (data storage) while reducing the rotation waiting time.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration for carrying out a data writing method according to an embodiment of the present invention.

FIG. 2 is a diagram showing a rotation waiting time of each magnetic disk in the embodiment of the present invention.

FIG. 3 is a diagram showing a state of a head at the start of reading and at the time of completion of reading of block 0 of the present invention.

FIG. 4 is a diagram showing a configuration for implementing a conventional data writing method.

[Explanation of symbols]

 1 ... Virtual disk 10a, 10b, 10c ... Magnetic disk 2a, 2b, 2c ... Driver 3 ... Virtual driver 4,4a, 4b, 4c ... Head 5,5a, 5b, 5c ... Writing start position 6a, 6b, 6c ... Rotation Waiting time

Claims (1)

[Claims] 1. A data storage means having a plurality of disk-structured auxiliary storage devices, a management means for managing a data access time and a queue for each of the plurality of disk-structured auxiliary storage devices of the data storage means; Data control means for equally dividing write data to be stored in the auxiliary storage device having the disk structure and controlling transmission of write data to be stored in the plurality of auxiliary storage devices having the disk structure based on the management information of the previous management means. A data writing method characterized by the above.