JPS5823365A - Positioning mechanism controlling system - Google Patents

Abstract

PURPOSE:To take a maximum stroke of a next new positioning as 1/2 or below of an interval of inner and outer tracks, by using a vacant time of a positioning mechanism and performing positioning to a center track with an independent operation of a device in advance. CONSTITUTION:The content of a track address register 21 is revised at each positioning and the track address is displayed at the present time. A start monitor circuit 4 monitors an I/F circuit 8, detects the start of a new input/output operation, then informs the result to a discrimination control circuit 7. When an operation end monitor circuit 5 detects the end of operation, the detection is informed to a delay circuit 6, which in turn informs the result to the circuit 7 with a suitable delay time. The circuit 7 controls the circuit group so that the positioning to the objective track can be made, when the operation signal is set and the start signal is also set, and controls the circuit group so that the positioning is made to the center track when the end signal is set and the start signal is reset.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a control method for a positioning mechanism that effectively speeds up the average positioning time of a magnetic disk device.Magnetic disk devices are frequently used for file storage in information processing systems, etc. However, with advances in magnetic recording technology, storage capacity is increasing. Larger capacity generally brings about an increase in the capacity of the positioning mechanism, which tends to become a bottleneck and reduce the throughput per unit capacity (the number of input/output operations that can be processed per unit capacity per unit time). It is in. For this reason, improvements have been made to drive mechanisms to increase the operating speed of conventional positioning mechanisms.
Currently, the average positioning time is around 20tlLS. However, there are problems such as it is becoming increasingly difficult to shorten the average positioning time, and even if the time can be shortened, it is difficult to increase the speed to an extent commensurate with the increase in the capacity of the positioning mechanism. One solution is to increase throughput by installing multiple positioning mechanisms on the disk medium without increasing the capacity of each positioning mechanism, but from the economic point of view, it is necessary to install multiple positioning mechanisms on the disk medium. There are problems with installing . Furthermore, recently MO5RAM
With the decline in the price of , it is used as buffer memory,
There is a method to shorten the effective input/output access time by so-called disk cache control, but access to tracks or cylinders on the disk surface is used randomly, that is, data access locality (Localit)
When y) is low, the disadvantage is that the system is generally less effective or requires a large buffer, resulting in an expensive system.

The present invention focuses on the usage characteristics of magnetic disk drives, and is characterized in that the positioning mechanism is controlled to be positioned in advance at the center of the movement range of the disk surface by utilizing idle time that is not used by software. The purpose is to reduce the maximum moving distance of the positioning mechanism by half and shorten the average positioning time, especially for applications in which tracks or cylinders of a disk are randomly accessed.

The present invention will be explained in detail below using figures. 1st
The figure shows the positioning mechanism and magnetic head of a magnetic disk device.
1 shows the relationship with the disk medium, 1 is the disk medium, 1-1 is the outermost track, 1-2 is the innermost track, 2 is the positioning mechanism, and 3 is the magnetic head. In the figure, the positioning mechanism 2 is configured to select any concentric track recorded between the outermost track 1-1 and the innermost track 1-2 according to a command from a host device (for example, a magnetic disk control device, etc.). Perform positioning operation on top of.

When the positioning operation is completed, the magnetic head 3 waits for the rotation of the disk, that is, waits for the start position of the record to be recorded in the track, and then performs a write operation to continue recording the target record. .

The positioning operation time is proportional to the moving distance from track to track (referred to as a stroke) (roughly proportional to the square root of the stroke), and the average positioning time is currently about 20 ymS. The rotation waiting time is determined by the disk medium 10 rotation time, which is on average 1/2 rotation time for random data access, and currently this average rotation wait time is 8 to 10+a5. Figure 2 shows the first
This figure schematically shows how the positioning mechanism 2 in the figure is actually used. In Figure 2 (α), the outer ~ n+1
During the second operation, the positioning mechanism 2 is held in the host device (or software) for the time required for [10 positioning rotations waiting + Rg ad/W'ri t g J operation, and each operation time is 'Af&, 'An+1, 'A
r,+! , """("-'-") are usually activated at each time interval of 'In, 'In++, ...... (n-1~axis), and the The actual positioning operation is as shown in FIG. 2 (6).

That is, the conventional device is characterized in that once a required positioning operation is completed, the device does not move from the positioned track until the next new positioning operation request is received. On the other hand, FIG. 3 has the same meaning as FIG. 2(c) except for the positioning operation to the center track located near the center between the base inner circumference and outer circumference tracks. However, the length of each operation time t't't' ・・・・・・
(n-1 to #) and each operation amount An, ^a+I,
Talk about An+2゜time interval t'glaze+"In++," 3.
...(n-1 to -) are different from those in FIG. 2 (α). That is, in the present invention, by utilizing the idle time of the positioning mechanism and performing positioning to the center track in advance by autonomous operation of the device, no matter what the address of the next new positioning track is, It is possible to obtain a remarkable effect of reducing the maximum stroke to 1/2 or less between the inner track and the outer track. In actual use of the apparatus, the effectiveness of the present invention is determined by two factors: (1) whether there is idle time, and (2) the ability to complete the positioning operation to the center track during the idle time. Regarding this point, when using an actual magnetic disk device, it is normal to set the upper limit of the device usage rate to about 0.4.
0- is an empty state, and the average input/output operation time (average positioning time + average rotation waiting time + average data transfer time (4
KE)) is currently about 50txS, so the average C
D operation time interval B O, 6/(30x10-'#*C)
-20x10-'sac=20*S, and positioning to the center track within this time is sufficiently possible even with current technology.

Regarding the case where the effect of the present invention is significantly impaired, the positioning is performed on a track near the inner or outer circumference, and after positioning on the center track during the operation time interval, the next operation is performed again near the same inner or outer circumference. There are cases where it is necessary to position the vehicle on the same track and the time interval between operations is extremely short. However, if we limit this point to applications where track selection is random, the probability of the above case is /J%. By providing a means for the software to contact the magnetic disk drive when the desired track is known, it is easy to interrupt positioning to the center track and prevent unnecessary operations. Considering that this can be done, this does not become a factor that makes the present invention lose its meaning.

Next, examples will be specifically described. FIG. 4 shows an embodiment in which the present invention is applied to a magnetic disk device, in which 1 is a disk medium, 2 is a positioning mechanism, 3 is a magnetic head, 4 is a startup monitoring circuit, 5 is an operation completion monitoring circuit, and 6 is a delay circuit. ,
7 is a judgment control circuit.

8 is an interface circuit, 9 is a target track address register, 10 is a center track address generation circuit, 11
12 is a switching circuit, 13 is a complement circuit, 14 is an addition circuit, 15 is a comparison circuit, 16 is a constant circuit, 17 is a buffer register, 18 is a positioning mechanism drive circuit, 19 is a positive/negative determination circuit, 20 2 is a switching circuit, and 21 is a current track address register. FIG. 4 shows only the circuit blocks and mechanisms according to the present invention. An example of operation will be explained according to the diagram. The contents of the current track address register 21 are updated every time a positioning operation is performed, and the track addresses that can be read/written by the magnetic head 3 are displayed in binary numbers. The activation monitoring circuit 4 monitors the interface circuit 8 and notifies the determination control circuit 7 when it detects activation of a new input/output operation. In parallel with this, the operation completion monitoring circuit 5 monitors the completion of the activated input/output operation.
When the end of the operation is detected, it is notified to the delay circuit 6, and the delay circuit 6 notifies the determination control circuit 7 of the end of the operation after an appropriate delay time, for example, about 0.5 seconds. When the operation signal is on and the start signal is on, the determination control circuit 7 controls the circuit group shown in FIG. 4 to position the circuit group to the target track according to the start, and when the end signal is on. Therefore, when the start signal is off, the circuit group shown in FIG. 4 is controlled to be positioned to the center track. The purpose of the delay circuit 6 is to prevent the center track positioning operation time from being added to the input/output access time when input/output access from software occurs continuously.
The delay time can be varied depending on the actual usage status of the magnetic disk device.

Now, assume that the determination control circuit 7 accepts and accepts the activation of a new crowd movement operation. A binary destination track address sent from a host device (such as a magnetic disk control device) through an interface circuit 8 is stored in a destination track address register 9, and in this case, the switching circuit 12 selects the output of the destination track address register 9. Then, a signal is sent to the adder circuit 14, and in synchronization with this operation, the binary number output from the current track address register 21 is converted into a two-complement number by the complement circuit 15, and the signal is sent to the adder circuit 14 as the other input. Upon completion of these two inputs, the adder circuit 14 performs modulo-2 addition, functionally calculating the difference between the target track address and the current track address, and sends the result to the comparator circuit 15 and buffer register 17. Comparison circuit 1
5 compares the magnitude of the output of the adder circuit 14 and the binary output of the constant circuit 16; If the output value is small, the purpose is to output an output that stops positioning to the center track, and if it is large, the purpose is to avoid adding the operation time for positioning to the center track, and this setting value of the constant circuit 16 is also based on the actual It can be set variably depending on the usage status of the magnetic disk device.The operating condition of the comparison circuit 15 is that the output of the startup monitoring circuit 4 is off and the output of the delay circuit 6 is on. This is limited to the time interval of the judgment control circuit 7.
controlled by the control output signal of Therefore, currently,
Since we are in the process of positioning operation due to a new activation, the output of the adder circuit 14 is set in the buffer register 17,
The output of the buffer register 17 is the positioning mechanism drive circuit 1.
At the same time, a positive/negative determining circuit 19 of the adding circuit 14 determines whether the number is positive or negative and transmits it to the positioning mechanism drive circuit 18. Through these circuit operations, the number of tracks to be moved and the direction of movement (positive/negative correspond to the inner circumferential direction/outer circumferential direction, respectively, but the correspondence is not an issue here) are determined, and the positioning mechanism The drive circuit 18 drives the positioning mechanism 2 to position it on the target track.

Simultaneously with the start of the positioning operation, the switching circuit 2o selects the output of the target track address register 9 in this case.
The contents of the current track address register 21 are output to be updated, and the current track address register 21 is updated. After these series of operations, the startup is completed and the process moves on to the first operation. The operation end monitoring circuit 5 monitors the end of the started input/output operation, and when it detects the end, turns off the output and notifies the judgment control circuit 7 through the delay circuit 6. 4 will be described below with reference to the operation when positioning to the center track is performed during the time interval of the operation. In this case, the switching circuit 12 supplies the output of the center track address generating circuit 10 to the adding circuit 14. MaraSimilar to the above explanation, when the output of the current track address register 21 is complemented by the complement circuit 15 and the input to the adder circuit 14 is completed, the adder circuit 14 calculates the difference and stores the calculation result in the buffer register. 17 and the comparison circuit 15. In this case, the comparison circuit 15 performs a comparison with the output of the constant circuit 16 to determine whether or not positioning to the center track is to be performed. If the determination result is that it is OK to proceed with positioning to the center track, the positioning operation according to the contents of the buffer register 17 and the output of the positive/negative determination circuit 19 is performed in the same manner as in the above case. However, in this case, the current track address register 21 is updated by the output of the center track address generation circuit 10. Further, if the determination result is that the positioning to the center track is interrupted, the subsequent operation does not proceed and the process returns to the initial state of waiting for activation.

In addition, at the time of starting as described above, if the next track address to be positioned is determined in advance by the host device, the next track display circuit 11 is set to ON, and after the input/output operation accompanying the starting is completed, The higher-level device is notified of the output of the next track display circuit 11 via the interface circuit 8, and requests the destination track address register 9 to send the next track address. Positioning control after receiving the next track address is performed in the same manner as at startup. Although the method of receiving the next track address is not shown in FIG. 4, it is conceivable to provide a second target track address register and receive the next track address in advance at the first startup.

In the embodiments described above, the present invention is realized by a circuit function within a magnetic disk device, but many well-known magnetic disk control devices employ a microprogram control method, which has a high level of sophistication and flexibility. If we focus on the high control capability of the magnetic disk controller, it is possible to realize most of these control functions with a magnetic disk controller. Although this embodiment does not specifically show how to receive the next track address from the software, the positioning address information length of the input/output command (usually called a command) to the magnetic disk controller is expanded, and the first This can be easily realized by making it possible to specify the target track address of the first target track and the next target track address.

As detailed above, the present invention is a positioning mechanism control system having the following percentage characteristics. That is, 1. Detecting the vacant state of the positioning mechanism, and using the detected time as an opportunity to position the wrinkle positioning mechanism at or near the center of the movable range corresponding to the track to be read/written.

2. After detecting the vacant state of the positioning mechanism, a request for use of the positioning mechanism from a host device is monitored during the arbitrary fixed time interval, and when activation is detected, priority processing is given to the activation.

& Start the positioning operation to the center of the movable range of the positioning mechanism upon detection of the idle state of the positioning mechanism, if the difference between the track address currently supported by the positioning mechanism and the track address at or near the center is arbitrary. Limiting the process to stop when the value is below a certain value.

4. Means is provided for notifying the host device in advance of an instruction to position the positioning mechanism next to the center of its movable range upon detection of an empty state of the positioning mechanism, and the notifying means receives the instruction. Sometimes, the robot should be stopped and the robot should perform positioning operations according to instructions given by the robot.

Furthermore, as described in detail above, the present invention is a system in which the positioning mechanism can be controlled by combining any two or all of the features described in sections 2 and 3.4 above. it is obvious.

Next, calculation of the average positioning time reduction effect corresponding to the actual device usage state when the present invention is applied will be described.

Such calculations are extremely difficult in terms of analysis and must be obtained by means such as simulation. However, by adjusting the conditions, the maximum effect that can be obtained can be easily determined. The average number of strokes of the positioning mechanism under conditions where access to tracks on the disk is random can be calculated using the following formula il+ (References: J. Martin, translated by Yasushi Kitahara, "Real Time" 9, Japanese Management Edition) Society, pp. 545-548).

Parameter 2N: Number of tracks on the disk - 2Average number of strokes On the other hand, in the operation model to which the present invention is applied, ■ Access to tracks on the disk is random, ■ Positioning operation time to the center track is within the operation time interval. Let us consider the formula for deriving the average stroke number under the following two conditions: (In other words, the positioning operation time to the center track is not added to the normal input/output operation time). However, N and Z follow the above definitions.

According to the present invention and condition (1), the positioning operation related to the required input/output operation always starts from the center track. Also, from the center track.

The distribution of positioning operations toward the outer or inner circumference is symmetrical about the center track. Therefore, determining the average number of strokes in the moving range i from the center track and determining the average number of strokes over all tracks are expensive. Based on the condition (2), the average number of strokes - can be determined as follows, assuming that the probability of accessing each track from the center track toward the inner circumference, for example, is uniform.

If the ratio of the average stroke of the conventional device between equation (1) and equation (2) is determined as R, equation (3) is obtained.

Now, if A'>>1, then R#i, and in the device to which the present invention is applied, the average number of strokes of the positioning mechanism is reduced by up to 3/4 times as compared to the conventional device.

In recent magnetic disk drives, the operating time of the positioning mechanism is generally proportional to the square root of the number of strokes. Therefore, in the device to which the present invention is applied,
Compared to conventional equipment, average positioning time is maximum A#0.87
It can be doubled.

For example, if the average positioning time is 20rnS, it can be expected to be shortened to 17IIIS by applying the present invention, and this value is recognized as sufficiently effective in the field of magnetic disk technology.

As explained above, when the present invention is applied to the positioning control of a magnetic disk device in which track access is random, it is possible to effectively and easily shorten the average positioning time, thereby improving throughput. It has the advantage of greatly contributing to

[Brief explanation of the drawing]

Fig. 1 shows an example of the relationship between a disk medium and a positioning mechanism in a magnetic disk device, Fig. 2 shows the operating state of positioning &* in a normal usage state, and Fig. 3 shows the positioning mechanism according to the present invention. The operating state is shown in FIG. 4, which is an embodiment to which the present invention is applied. 1...Disk medium, 1-1...Outermost track,
1-2... Innermost track, 2... Positioning mechanism,
3...Magnetic head, 4...Startup monitoring circuit, 5...
Operation completion monitoring circuit, 6... Delay circuit, 7... Judgment control circuit, 8... Interface circuit, 9... Target track address register, 10... Center track address generation circuit, 11... Next track display circuit, 12
...Switching circuit, 16...Complement circuit, 14...Addition circuit, 15...Comparison circuit, 16...Constant circuit, 17
...Buffer register, 18.. Positioning mechanism drive circuit, 19.. Positive/negative determination circuit, 20.. Switching circuit.
21...Current track address register. Patent applicant Nippon Telegraph and Telephone Public Corporation agent Patent attorney Gobe Tamamushi (3 others) Figure 1 M2 (α) Cb) Truck

Claims (1)

[Claims] In controlling a positioning mechanism of a magnetic disk device, detecting an empty state of the positioning mechanism, and using the detection time as a trigger to position the positioning mechanism at or near the center of a movable range corresponding to a track to be read/written. A positioning mechanism control method characterized by the following.