JPS6162920A - Magnetic disk device system - Google Patents

Abstract

PURPOSE:To perform high-speed data transfer by arranging small-sized disks loosely and recording data sent from a host device loosely. CONSTITUTION:Data sent from high-priority equipment through a high-priority interface circuit 5 in the magnetic disk device system is set in an interface circuit input/output part 4 in byte form. This data is sent to a track buffer 2 composed of a shift register with the next clock and shift successively as shown by an arrow 6. When a disk driver 1 generates an index signal after the data is all stored in the buffer 2, the data is written on the disk drive 1. Disks on those nine disk drives 1 rotate asynchronously with one another, so write operation starts with the index of each disk and read instructions are generated for the disk drives 1 according to priority to start reading operation with the index. When all data are set up, they are transferred through said input/output part 4.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to a magnetic disk drive system, and particularly to a magnetic disk drive system capable of high-speed data transfer.

[Background of the invention]

Conventional magnetic disk devices transfer data at a rate of 3 MB/s, so data is transferred to a 14-inch disk at a speed of 15 MB/s.
,000 B P I, and the disc is 360 OR
It is rotated by PM. Therefore, the maximum reading frequency from the medium (disc) is 8 MHz, and the data discrimination window width is 21
Nanoseconds and broadband characteristics are required due to the circuit configuration. Furthermore, in order to increase the transfer rate to 6 MB/sec to 10 MB/sec, the recording/reproducing circuit system needs to be more sophisticated, which is becoming difficult to achieve.

A magnetic disk device for high-speed data transfer is described in, for example, Japanese Patent Laid-Open No. 56-118158.

[Purpose of the invention]

The objects of the present invention are (1) to realize a disk device system capable of high-speed transfer, and (2) to provide a disk device system with low focusing cost.

[Summary of the invention]

In order to perform high-speed transfer, large-sized disk devices require wide-band amplifiers, high-speed switch circuits, and high-speed logic circuits in the recording and reproducing circuitry, which is technically difficult and considerably disadvantageous in terms of cost. On the other hand, small disk devices have low transfer speeds, so the above.

Technology with less charge for circuits such as.

It is easy to use and advantageous in terms of cost. One aspect of the present invention is to arrange small disks separately and place them on top.

[Embodiment of the Invention] Fig. 1 shows a magnetic disk drive system according to an embodiment of the present invention. It is a diagram showing the upper interface “varnish circuit 5.
As a result, the data sent from the host device is set in the interface circuit input/output section 4 in the form of bytes.

This data is shown below.

The data is sent to traffic buffer 2 by the
Each time more data is sent, it is shifted in the direction of arrow 6. The track buffer 2 is a shift register that can sequentially store data for the maximum record written on the disk. After receiving data for one track from the higher order and storing it all in the track buffer 2, when the disk drive 1 generates an index signal, the data is written to one disk drive from the shift register.

At the interface circuit output section 4 for data transferred from the upper level, the data is arranged in bytes, but bits 0 to 7 of the bytes and the parity bit are stored separately on the track buffer 2. . This each separate bit.

0 to 7 and parity bits are each written to a disk drive 1 consisting of nine drives dedicated to each bit. Since the disk rotations of the nine disk drives 1 are mutually asynchronous, writing is started after waiting for the index of each disk to arrive. Since the disks rotate asynchronously, there will be a delay in track buffer 2 time and index waiting time, but in order to avoid time loss, when the index of disk drive 1 arrives,
That is, it goes without saying that it is also possible to take out the output of the track buffer 2 from the middle of the shift register and write it to the disk drive 1. Regarding reading, the higher level issues a read command to the disk drive 1, waits for an index, and then starts reading. Disk rotation is mutually asynchronous.

Therefore, the data is stored in track buffer 2.

will be ascribed. When all tracks have been read.

Now, all the data is on track buffer 2.

This is passed through the interface circuit input/output section 4.

The data is transferred to the upper level via the upper level interface circuit 5. Note that reading is the same as writing.

All drives have finished reading the entire track.

The slowest interface among the 9 drives.

When a loss occurs, the first part of the data of the other data bits that are entering the track buffer 2 is taken out from the middle of the shift register, connected to the input/output section 4 of the interface branch circuit, and the loss time is calculated. Needless to say, it can be made smaller.

Figure 2 outlines the time chart for writing and reading.

to show. 20 is the disk rotation corresponding to each bit.
4 represents a live 1 to 0 interloss signal, and shows, as an example, the case where the drive index of bit 0 is the fastest and the drive index of bit 1 is the slowest.

Reference numeral 21 indicates read and write commands given to the disk drive. When reading, bit 0 at the tip of 12
Reading of the portion starts, and storing to the track buffer 2 ends at the rear end. Another beep.

Similarly, track 1 ends at the rear end, all data is stored in track buffer 2, skew 1 is corrected, and transfer to the upper level becomes possible. Also,.

When writing, the data is aligned in the track buffer 2, and writing starts temporally from the beginning of data 12 for bit 0 and from the beginning of data 13 for bit 1, and writes at the rear end of each. Finish □゛.

By using the method described above, data that was conventionally written serially in one trough can be separated into bits of bytes.
Write to each drive.

As a result, the upper data transfer speed can be improved to eight times the transfer speed of the disk drive, and the transfer speed specifications of the disk drive can be significantly relaxed. - As an example, some 8" small disks currently have a transfer rate of 1.2MB/sec, and even the latest high-speed disks have a transfer rate of 3MB/sec at best, whereas the present invention allows transfers of 9.2MB/sec. 6MB/sec transfer can be easily performed. .

Also, in the above explanation, 1 byte of data is used.

Although it is possible to separate bits and perform parallel reading and writing, it is also possible to divide the data into multiple bits and transfer them to higher order; '□It is possible to effectively increase the speed using a similar method.

There is no book to tell you what I am capable of.

〔Effect of the invention〕

According to the invention, the transfer rate of the disk drive.

Eight times the transfer speed can be easily achieved without increasing the speed. Therefore, there is no need for any circuit means to increase the speed of the disk drive, which is advantageous in terms of cost. In addition, when comparing costs, small disks have lower transfer rates, and because they are mass-produced, the cost per bit, which is an indicator of storage devices, is also lower. It is quite advantageous to use.

[Brief explanation of drawings]

FIG. 1 shows a block diagram of one embodiment of the present invention, and FIG. 2 shows a schematic time chart of one embodiment of the present invention. DESCRIPTION OF SYMBOLS 1...Disk drive, 2...Track buffer, 3...Drive interface, 4...Upper transfer interface, input/output circuit section, 5...
...Upper transfer interface circuit, 10...bit 0
Equivalent drive index signal, 11... Drive index signal equivalent to bit 1, 12... Drive read/write signal equivalent to bit 0, 13... Drive successive write signal equivalent to bit 1, 20... Index signal, 21 ...
Read/write signal.

Claims (1)

[Claims] A means for arranging the byte data transferred from the upper level in parallel, a means for converting the bits constituting the byte into serial data, a means for writing the serialized data into each of the plurality of disk devices, and rotation of the plurality of disks. A magnetic disk device comprising: a data buffer for synchronously writing data; means for buffering mutually asynchronous data read from a plurality of disks; and means for eliminating skew of parallel data. system.