JPH0256019A - Overlap recording and reproducing method for information using plural disks - Google Patents

Abstract

PURPOSE:To shorten time required for write and read-out by utilizing plural disks, recording the same information in each disk, and also, changing a recording position of the same information at every disk. CONSTITUTION:A disk device 1 rotates and drives a rewritable disk, and executes write and read-out through a head. In this case, two sets of disks are used as disks A, B, respectively, and simultaneously, the same information is brought to write and read-out in parallel with respect to both disks A, B. Also, the disk device concerned 1 is provided with a disk controller 11, a processing circuit 12, and a main memory device 13 of a central processor connected to the disk controller 11. This processing circuit 12 is provided with a path switching means 14, a virtual cylinder/real cylinder address converting means 16 and a disk input/output completion confirming means 17. In such a manner, the same information is recorded in respective disks A, B, and its recording position is changed at every disk. As a result, by executing an operation to the disk in which the moving amount of a movable head is relatively small, the time is shortened.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a method for recording and reproducing information on a rewritable recording medium such as a disk.

BACKGROUND ART As shown in FIG. 9, a movable head Ht (
In a rewritable disk device, cylinders, which are storage area units for recording information, extend in the circumferential direction and are arranged in the radial direction, and each cylinder is numbered from 1 to n.

Conventionally, information recording from a central processing unit is performed only at one location on a particular cylinder of a particular disk among the connected disks. The cylinder number in which the previously recorded information exists remains unchanged, and unless the information is forcibly relocated, the cylinder number continues to exist at the same position until the information is erased. Therefore, when reading recorded information, the seek distance of the head varies depending on the position of the head inside the disk at that time, and when the central processing unit exchanges information with the disk, it takes a considerable seek time on average. This results in a state of waiting for disk input/output operations to be completed, resulting in a phenomenon in which the performance of the central processing unit cannot be fully utilized.

Other prior art techniques use multiple rewritable disks to record information for so-called backup. In such prior art, no efforts have been made to shorten the time required to write to or read from a disk.

Problems to be Solved by the Invention An object of the present invention is to provide a method for recording and reproducing information on a rewritable disk-type recording medium that can reduce the head seek time required for inputting and outputting information.

Means for Solving the Problems The present invention provides an apparatus configured such that a central processing unit connected to a plurality of disks can execute the same write command or read command for each disk in parallel. By converting the disk internal address of the information writing position into a different value for each disk and writing or reading information, the seek distance of the head for the same information on each disk is different, and the same information is sent from the central processing unit. A plurality of rewritable disk type recording characterized in that write commands or read commands are executed simultaneously on each disk in parallel, and information on the disk processed in the shortest time is used by a central processing unit. This is a method of recording and reproducing information using a medium.

The present invention is configured to move a write and read head across the disk while the disk is rotating, and the disk has a plurality of circumferentially extending storage area units radially disposed therein. , the address of this storage area unit requested by the central processing unit when writing or reading information is converted into a different address, and the writing or reading is actually performed on the disk, and in this way information can be recorded and played back on the disk. The same information is recorded on each disk, and the recording location address of the same information is different for each disk, and the same information is written to or read from each disk at the same time. Processing to the disk that was completed in the shortest time is defined as the primary disk, processing to other disks is defined as secondary, and processing after disk input/output of the central processing unit is completed is connected based on the processing result to the primary disk. This is a method for repeatedly recording and reproducing information using multiple discs, which is characterized by:

Further, the present invention is characterized in that the addresses of each storage area unit on the disk are set in order from the outside in the radial direction or from the inside.

Furthermore, in the present invention, when converting addresses in units of storage areas, consecutive storage area units are grouped so that addresses can be changed in units of groups, and the number of groups is equal to the number of disks. It is characterized in that it is set at least twice.

According to the present invention, by using a plurality of disks that write to or read from a rewritable disk using a movable head, the same information can be recorded on each disk, and the same information can be recorded on each disk. By changing the recording position of the disk, information is written and read from each disk in parallel, and the amount of movement (seek distance) of the movable head at that point in time is relatively small. Since the information input/output process can be completed by a read operation, it is possible to shorten the writing and reading time to the disk.

That is, in the present invention, by using a plurality of disks, recording the same information on each disk, and changing the recording position of the same information for each disk, the movable head at the time of inputting and outputting information to and from the disks. The head seek time required for inputting and outputting information can be shortened by making write or read operations on a disk I\ whose movement amount is relatively small as primary and write or read operations on other disks as secondary.

Embodiment FIG. 1 is a simplified plan view of a disk device 1 according to an embodiment of the present invention, and FIG. 2 is a block diagram showing an electrical configuration connected to this disk device 1. As shown in FIG. With reference to these drawings, disc 2, which is a rewritable recording medium,
is driven to rotate around the rotating shaft 3 as shown by the arrow 4.

A helad H for writing and reading data on the disk 2 is fixed to a helad arm 5, and this arm 5
The head H is moved across the disk 2 by the drive means 6. The head H may be configured to be displaced in the radial direction of the disk 2, as shown in FIG. It may be configured to move along a route.

On this disk 2, store area units are formed concentrically from the radially inner side to the radially outer side, and addresses of the store area units are set in order.

In the block diagram of FIG. 2, two disks 2 are used, named disk A and disk B, respectively, and the same information is written to and read from disk A and disk B in parallel at the same time.

In this embodiment, a storage area unit is referred to as a cylinder, a cylinder reservation request from the central processing unit is assigned to a virtual cylinder, and actual disk cylinder assignment is assigned to a real cylinder. Accordingly, the cylinder seen from the disk control device 11 on the central processing unit side is a virtual cylinder, the number of this virtual cylinder is called a virtual cylinder number, the cylinder in the disk is called a real cylinder, and the number of this real cylinder is called a real cylinder number. Say.

-m has the same virtual cylinder number, but the real cylinder numbers of disk A and disk B are different.

Here, the total number of virtual cylinders n is 4 as shown in Table 1.
Divide into equal parts and determine "virtual cylinder sections" 1 to 4. The total number n of real cylinders of disks A and B is also divided into four equal parts, and "real cylinder sections" 1 to 4 are set respectively.

Both the "virtual cylinder section" and the "real cylinder section" include cylinders, and the "intra-section cylinder numbers" are assigned in ascending order from 1 to k, or in descending order from k to 1.

The correspondence between the "virtual cylinder partition" number and the "real cylinder partition" number of disk A is 1, 1.2, 2.3, 3.4, and 4.
It is.

The correspondence between the "virtual cylinder partition" number and the "real cylinder partition" number of disk B is 1, 3.2, 1.3, 4.4, and 2.

The "intra-partition cylinder numbers" of the "virtual cylinder partitions" are all in ascending order.

The "intra-partition cylinder numbers" of the "actual cylinder partitions" are all arranged in ascending order on disk A, and on disk B, partitions 2.3 are arranged in ascending order and partitions 1.4 are arranged in descending order.

The cylinders having actual cylinder numbers 1 to r1 extend circumferentially and are arranged radially, and although each cylinder is formed concentrically, in other embodiments these cylinders may be spirally arranged. may be formed.

The number m of disks recording the same information may be any number of disks, from two to more.

The number of "virtual cylinder partitions" and "real cylinder partitions" in one disk may be any number as long as they are equal across all disks, but according to the inventor of the present invention, in order to be effective, the number should be twice the number of disks (m). was found to be desirable.

The reason for this will be explained later.

The correspondence between the "virtual cylinder partition" number and the "real cylinder partition" number is such that different "virtual cylinder partition" numbers across all disks are arranged adjacent to each other on any disk.

In the case of Table 1, if the arrangement of virtual cylinders is tabulated in order of the real cylinders of each disk, it will be as shown in Table 2.

(Left below) As the actual cylinder numbers are assigned to distinguish the physical location of each storage area, they can be assigned in either order from the inside in the radial direction or from the outside in the radial direction. It may be numbered.

In connection with the disk device 1 shown in FIG. 1, the disk control device 11 and processing circuit 1 shown in FIG.
2, and a main storage device 13 of a central processing unit connected to the disk control device 11. In FIG. 2, arrow lines indicate the flow of information.The disk control device 11 functions to exchange information to be recorded and reproduced on the disk 2 with the main storage device 13 of the central processing unit. A path switching means 14 is connected to the device 11 for recording and reproducing information from the disk control device 11 on the disk 2 of the disk device l.

The disk input/output data buffer section 15 temporarily stores data that the path switching means 14 writes from the disk controller 11 to the disk device 1, or stores data that the path switching means 14 reads from the disk device 1 and transfers to the disk controller. It is used to temporarily store data to be sent back to W11.

Furthermore, this processing circuit 12 performs virtual cylinder/real cylinder address conversion that correlates and mutually converts the virtual cylinder number of the data stored in the disk input/output data buffer section and the real cylinder number according to instructions from the path switching means 14. After issuing a write or read command to each disk of the disk device 1 according to instructions from the means 16 and the path switching means 14, the arrival of an input/output completion signal from each disk is checked, and the first input/output signal that arrives is checked. A disk input/output completion confirmation means 17 is provided that validates information from the disk that has returned the information and invalidates information related to input/output signals that arrive after that.

Figure 3 is a disk control device! 11 is a flowchart showing a processing procedure for explaining the operations of the processing circuit 11 and the processing circuit 12. FIG. Moving from step n1 to step n2 in FIG. 3, an input/output command for writing or reading information is received from the disk controller 11.

In step n3, the receiver determines whether the command received is an output command, and if it is an output command, the process moves to step n4, and the output data from the disk controller 11 is stored in the disk input/output data buffer section 15.

In step n5, a counter i, which determines the disk on which the input/output command is executed, is initialized to 1.

In step n6, the real cylinder address necessary to execute an input/output command to the first disk 2 of the disk device 1 is determined based on the virtual cylinder address obtained from the disk controller. Calculated by address conversion means 16.

In step n7, the real cylinder address calculated in step n6 and the data accumulated in the disk input/output data buffer section are used to write or read information to or from the first disk 2 of the disk device 1. An input/output instruction is executed.

At step n8, the counter 1 is incremented and the process moves to step n9, where it is determined whether the value of the counter 1 exceeds the number m of disks 2.

If i does not exceed m, the process returns to step n6 and the input/output command for the next disk is executed.

If i exceeds m, the process moves to step nlO and waits for an input/output completion signal to arrive from one of the disks 2 of the disk device 1 that issued the input/output command.

In step nil, the disk input/output completion confirmation means confirms that the first input/output completion signal has arrived from one of the disks, and the process moves to step n12, where the disk input/output completion confirmation means confirms that the first input/output completion signal has arrived from one of the disks. and executes an instruction to refuse acceptance of the input/output completion signal.

In step n13, it is determined whether the input/output command is an input command, and if it is an input command, the process moves to step n14, where the data read from the disk that received the first input/output completion signal is transferred to the disk input/output data buffer. Accumulate in.

In step n15, the real cylinder address of the disk that received the first input/output completion signal is converted back into a virtual cylinder address by the virtual cylinder/real cylinder address conversion means.

In step n16, the disk input/output result is returned to the disk controller. In the case of reading, the disk data accumulated in the disk input/output data buffer section is also returned to the disk controller.

When the execution of the input/output command is completed, the process returns to step n2 for execution of the next input/output command.

Table 3 shows an example of cylinder arrangement when data is recorded on two disks each having 20 cylinders by the method of the present invention.

(Margin below) The real cylinder and virtual cylinder of disk A have the same cylinder number, partition number, and intra-partition number.

In disk B, virtual cylinder partition numbers 2 are arranged in real cylinder partition number 1 in descending order of the partition numbers,
Virtual cylinder partition numbers 4 are arranged in real cylinder partition number 2 in ascending order of the partition numbers, virtual cylinder partition numbers 1 are arranged in real cylinder partition number 3 in ascending order of the partition numbers, The virtual cylinder partition numbers 3 are arranged in the real cylinder partition number 4 so that the intra-block numbers are arranged in descending order.

For both disk A and disk B, file A is stored in virtual cylinder numbers 2 to 3, file B is stored in virtual cylinder numbers 4 to 8, file C is stored in virtual cylinder numbers 12 to 15, and file C is stored in virtual cylinder numbers 16 to 16. File D is recorded in .17, but since the actual cylinder numbers are different between disks A and B, the relative positional relationship of the actual data is different between disks A and B.

Virtual cylinder numbers 9.10, 11.18, and 19°20 are empty cylinder numbers in which no data is recorded.

The virtual cylinder number is a logical cylinder address for the disk 2 used by the disk control device 11 as described above.

The real cylinder number indicates the actual address of the cylinder of disk 2.

The manner in which information is recorded and reproduced in order from the state shown in Table 3 will be explained.

Table 4 shows a state in which a new file E is allocated and occupies virtual cylinder numbers 9, 10, 11.18.

Table 5 shows a state in which the secondary extension area of file C occupies virtual cylinder number 19.

Table 6 shows a state in which file B is deleted and virtual cylinder numbers 4 to 8 are released.

Table 7 shows a state in which the secondary expansion area of file D occupies virtual cylinder numbers 4 to 6.

Table 8 shows a comparison of head seek distances when data is searched under the conditions shown in Table 7.

The shortest seek distance in Table 8 indicates the seek distance in the case of the present invention, which employs the shorter of the seek distances of disk A and disk B.

The average value of the seek distance is 7.7 for disk A, 6.8 for disk B, and the shortest seek distance is 4.4, which shows that the method of the present invention is effective.

(Leaving space below) Table 4: Cylinder arrangement example for data Table 6: 1 example of cylinder vf for data In an information recording and reproducing method equipped with means similar to the above, using sectors instead, or blocks instead of cylinders in the case of block format,
An information recording and reproducing method is characterized in that it includes steps similar to those described above.

According to the present invention, the following further modifications are possible.

A recording medium is one that can record and reproduce information as a digital signal using magnetic, electrical, optical, magneto-optical, or other methods for the purpose of information processing. It is sufficient if the time information can be recorded and reproduced, that is, it can be rewritten.

A recording medium is one in which information can be written to and read from a disk in one of the following ways: cylinder format, track format, sector format, or block format. The track arrangement trajectory for the sector format, the sector arrangement trajectory for the sector format, and the block arrangement trajectory for the block format may be concentric circles or spirals, that is,
The shape of the storage area unit does not matter.

The disk may be one that can be attached and removed as a single unit from the rotational drive mechanism of the disk device, or one that cannot be attached or removed as a single unit.

If either the recording head, the reproducing head, or the recording/reproducing head for the disk of the disk device is movable by the head arm, the moving direction of the head may be either linear or curved, and when recording and reproducing information, It is possible to freely reciprocate on a fixed line connecting the end points of the parts, and the position of the head on this line, in combination with the rotational position of the disk, the cylinder number in the case of a cylinder type, and the track type. may be determined by associating it with the track number, the sector number in the case of a sector format, and the block number in the case of a block format.

The disk control device 11 on the central processing unit side may be a dedicated microcomputer incorporating a dedicated processing circuit (CPU), or a general CPU used in a word processor, a picture processing device, etc. .

Table 9 shows examples of efficient arrangement of virtual cylinder partition numbers for cases where the number of disks ranges from 2 to 8. In these examples, any two virtual cylinder partition numbers can be assigned to either disk. They are always arranged next to each other. Even if there are nine or more disks, as shown in Table 9, if you arrange any two virtual cylinder partition numbers so that they are always adjacent to each other on some disk, no matter where the head of each disk is located, one of the The disk seek distance can be shortened.

(Margin below) Table 9 Example of boardo cylinder compartment number arrangement Table 9 Example of boardo cylinder compartment number arrangement Table 9 Example of boardo cylinder compartment number arrangement Table 9 Example of virtual cylinder partition number arrangement The embodiment having such a configuration has the following advantages.

, the same data is recorded at different addresses on different disks, and the relative distance from the address of any other data is always small on either disk, so a single The seek distance of the head is shortened compared to when data is recorded only on the disk.

As a result, the average disk seek time required for one data input/output can be reduced to the millisecond level.

Note that the average execution time per instruction of the central processing unit (CPU) is on the order of microseconds or nanoseconds, so
This improvement in average seek time greatly contributes to shortening information processing time.

The following calculation shows that it is desirable that the number of "virtual cylinder partitions" and "real cylinder partitions" in one disk be twice the number m of disks recording the same data.

Figure 4 shows a situation where the disk head position changes as information is written or read from the disk. Here, the total number of cylinders is n, and the axis is the head position before the seek operation. , and the X axis represents the head position after the seek operation.

When the head is between tl and t2 before the seek operation and moves between xl and x2 after the seek operation, the shaded area in the figure corresponds.

A region where x<t indicates that the head position t before the seek operation is larger than the head position X after the seek operation, that is, it is on the right side, and a region where x>t indicates that the head position t before the seek operation Indicates that it is smaller than the later bet position ix, that is, it is on the left side.

If the cylinder access probabilities are equal and do not depend on the previous position of the head, the probability of the head existing at position t before the seek operation is equal to a constant value of 1/0, and from the position t due to the seek operation The transition probability of the head to position X is also equal to a constant value of 1/n.

Therefore, the average Radseek length of the 4121st shaded area can be obtained from the first equation.

The average rad seek length of the entire disk is the first equation in the interval 0.
By calculating in the area of ≦t≦n, 0≦X≦n, the second
It is given by Eq.

= -Saki 0.333n (2) Using this calculation method, when the number of cylinder partitions changes from 2 to 5 in the case of two disks, the average herad seek length is calculated by the method of the present invention. Shows the shortening effect.

Figure 5 shows that when the number of cylinder partitions is 2, disk A
FIG. 4 is a disk head position transition diagram when cylinders are arranged on two disks, disk B and disk B, according to the method of the present invention. 2 represents the "virtual cylinder section" number, and the arrow indicates the direction of the order in which the "virtual cylinder" numbers are assigned in ascending order.

a1 to a8 indicate area pots representing states when the head of disk A changes position due to a seek operation, and b1 to b8
indicates an area representing a state when the head of disk B changes position due to a seek operation.

In a1 to a8 and b1 to b8, the areas where the respective subscript numbers match are transition areas where heads simultaneously seek for the same data on each disk according to the present invention.

The symbols written in each area are compared with the same subscript number areas in a1 to a8 and b1 to b8, and the O mark is an area with a shorter seek length, and the X mark is an area with a longer theta length, The Δ marks indicate areas with equal seek lengths.

That is, comparing the area a1 and the area b1, both have the same seek length, so a Δ mark is written. In this case, the information input/output time to disk A and disk B is equal.

Comparing area a5 and area b5, area a5 is better than area b.
Since the seek length is shorter than 5, an O mark is written in area a5 and an X mark is written in area b5. In this case, since information input/output time is shorter on disk A than on disk B, the average seek length when inputting/outputting information to/from the disk using the method of the present invention can be calculated in area a5.

Table 10 shows the average seek length calculated for each area in this way, and the overall average seek length when the number of cylinder sections is 2 is 0.25n.

(Left space below) FIG. 6 is a disk head position transition diagram when the number of cylinder sections is three. Table 11 shows that the average seek length when the number of cylinder sections is 3 was calculated using the same calculation method as when the number of cylinder sections is 2, and the overall average seek length is 0.235n.

(Margins below) Table (Continued) Table When the number of cylinder sections is 4, the disk head position transition diagram shown in FIG. 7 is obtained, and the overall average seek length is 0.208n from the calculation results in Table 12.

(Margin below) Table (Continued) When the number of cylinder sections is 5, the disk head position transition diagram shown in FIG. 8 is obtained, and the overall average seek length is 0.211n from the calculation results in Table 13.

(Margins below) Table (Continued 2) Table (Continued 1) Table (Continued 3) If the number of cylinder compartments is 6 or more, calculate using the same method and gradually calculate the You can see it getting bigger.

From the above, it can be seen that when the number of disks is two, the average theta length is shortest and the effect of the present invention is greatest when the number of r cylinder sections J is twice the number of disks, that is, four.

If the number of disks is 3 or more, use the same calculation method.
It can be confirmed that the highest efficiency is achieved when the number of "cylinder partitions" is twice the number of disks (m).

Effects of the Invention As described above, according to the method for storing a rewritable disk, the amount of movement of the movable head can be minimized, and the head can write to or read from the disk in a short time. .

[Brief explanation of the drawing]

FIG. 1 is a simplified plan view of a disk device 1 according to an embodiment of the present invention, FIG. 2 is a block diagram showing the electrical configuration connected to the disk device 1, and FIG. 3 is a disk control device 11.
FIG. 4 is a flowchart showing a processing procedure for explaining the operation of the processing circuit 12, FIG. Number of cylinder compartments is 2
FIG. 6 is a disk head position transition diagram when cylinders are arranged on two disks, disk A and disk B, by the method of the present invention. FIG. 6 shows the disk head position il! when the number of cylinder sections is 3. Transition diagram, 7th rf! J is a disk head position transition diagram when the number of cylinder sections is 4, FIG. 8 is a disk head position transition diagram when the number of cylinder sections is 5, and FIG. 9 is a simplified plan view of the prior art. . DESCRIPTION OF SYMBOLS 1... Disk device, 2... Disk, 3... Rotating shaft, 5... Arm, 6... Drive means, 11...
Disk control device, 12... Processing circuit, 13... Main storage device, 14... Path switching means, 15... Disk input/output data buffer unit, 16... Virtual cylinder/
Actual cylinder address conversion means, 17...Disk input/output completion confirmation means Agent Patent attorney Number of strokes A Disc B Disk A To the disc, l toy stand! Transition diagram disk B

Claims (4)

[Claims] (1) In a device configured so that a central processing unit connected to multiple disks can execute the same write or read command for each disk in parallel, the disk internal address of the information write position on each disk is By writing or reading information by converting it into a different value for each disk, the seek distance of the head for the same information on each disk is different, and the same write or read instruction from the central processing unit is sent to each disk. A method for recording and reproducing information using a plurality of rewritable disk-type recording media, characterized in that information on a disk processed in the shortest time is used by a central processing unit. Method. (2) A write and read head is configured to move across the disk as the disk rotates, and the disk has a plurality of circumferentially extending storage area units radially disposed therein. , the address of this storage area unit requested by the central processing unit when writing or reading information is converted into a different address to actually write or read information on the disk, and in this way information can be recorded and played back on the disk. The same information is recorded on each disk, and the recording location address of the same information is different for each disk, and the same information is written to or read from each disk at the same time, so that the operation is as short as possible. Processing to the disk that is completed at the time is considered as the primary disk, processing to other disks is considered to be the secondary process, and processing after the completion of the disk input/output of the central processing unit is continued based on the processing result to the primary disk. A method for repeatedly recording and reproducing information using multiple discs. (3) The method for repeatedly recording and reproducing information using a plurality of disks according to claim 2, characterized in that the addresses of each storage area unit on the disk are set in order from the outside in the radial direction or from the inside. . (4) When converting addresses in units of storage areas, contiguous storage area units are grouped so that addresses can be changed in units of groups, and the number of groups is at least twice the number of disks. 3. A method for repeatedly recording and reproducing information using a plurality of disks according to claim 2, characterized in that: