JP3302769B2 - Data transfer system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION This invention relates to a distributed processing system and the like, in which a large number of computers are connected to a public line or a TCP / IP network.
The present invention relates to a data transfer method for transferring data from one computer to all other computers at high speed in an environment where the same communication means can be used.

[0002]

2. Description of the Related Art In order to transfer the same data to a large number of computers on a network, such as upgrading the operating system of a distributed processing system, the data must first be installed on a single computer.
This is performed by transferring data one by one from the computer. Therefore, when transferring the same data to β computers, assuming that the transfer time per computer is τ, it takes β * τ (* is a multiplication operator). Most of the installation time must be spent monitoring and operating. If one operator is added, the work can be theoretically completed in half the time. However, in the case of a distributed processing system that has recently become large-scale with the progress of technology, the number of computers existing on the same network is several. In some cases, data transfer by human intervention is reaching its limits.

Next, a conventional data transfer method will be described with reference to the drawings. FIG. 5 is a diagram showing a conventional data transfer method. In the figure, reference numeral 10 denotes a computer for performing data transfer. 20 to 54 are computers to which data is transferred from the computer 10. Also, in the figure, arrows indicate data transfer by a communication line. However, these data transfers do not occur simultaneously.

Next, the operation will be described. First, data is installed in the basic computer 10. After this is completely completed, the data is transferred to another computer 20,.
30,... 40, etc. are sequentially installed. Basically, since one operator operates one computer, it is difficult to simultaneously transfer data from the computer 10 to the computers 20,... 30,. Therefore, data transfer operations for the number of computers to be transferred are required.

[0005]

The conventional data transfer is as follows.
As described above, data to be transferred is input to a certain computer, and data is sequentially transferred to the computer to which the data is to be transferred.
In the case of transferring data to zero computers, it takes 20 hours to calculate simply. Further, as the distributed processing system has been developed as in the present, the number of computers has increased to several hundreds, and the method of manually transferring data one by one has reached the limit.

[0006] The time required for the data transfer can be reduced by increasing the number of workers by {the number of computers / the number of operators} * time, but it is impossible for a person to work 24 hours a day, 7 days a week. And there is a limit to human resources.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problem, and when transferring the same data to many computers, high-speed and efficient data transfer is completed in a short time. The purpose is to obtain a data transfer method that minimizes human intervention.

[0008]

A data transfer method according to the present invention is a data transfer method for expanding data transfer as shown by a tree structure to which a mouse operation is applied, and transfers the data to a computer at the top of the tree. A computer that has not completed the transfer with the data of the computer recognition code group by adding the data of the recognition code of the computer to be transferred together with the data to be transferred (a code capable of identifying the other computer) and transferring the data. While developing the transfer.

Also, when a computer transfers data,
The recognition code group is equally divided by adding 1 to the number of communication lines that can be opened by the computer, and a computer to be transferred next to any one of the equally divided code groups is determined. , And the code group and the data to be transferred are simultaneously transferred to the target computer. With respect to only one code group remaining after the completion of this transfer, the computer performs the same processing again. By repeating this process, the computer that has completed the transfer does not enter the instruction waiting state and continues the transfer until there is no recognition code group. Computers located in the upper part of the tree structure perform this repetitive processing more. However, in order to obtain the effect of increasing the number of computers in the same layer, the number of computers performing data transfer at the same time increases, and the processing of lower-level computers is reduced. In other words, the number of hierarchies in the tree structure is reduced, which means a reduction in time over simple mouse calculation.

In theory, a computer that has received data to be transferred once from the time when the highest-level computer starts data transfer to the time when the lowest-level computer receives the data transfers data as long as the recognition code group data remains. In addition, since the time at which the top computer finishes all the repetitive transfer processing and the time at which the end computer finishes receiving data are theoretically almost the same, useless time is eliminated. .

[0011]

According to the data transfer method of the present invention, when transferring the same data of a large number of computers on a network, one data transfer method is used.
By developing data transfer from one computer to a mouse formula, it has a dramatic reduction in time, artificial deletion, reliability of work, and efficiency compared to the conventional method of transferring data one by one. Realize data transfer.

The computer which has finished receiving data repeats data transfer until the recognition code group reaches the last one, thereby reducing the hierarchy of the tree structure and further increasing the efficiency. In addition, only one operator is required. During the processing, only the work of monitoring is sufficient, and data transfer can be performed at night when the operation rate of each computer is low. Leads to savings.

[0013]

[Embodiment 1] An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a tree structure diagram showing one embodiment of data transfer according to the present invention. In the figure,
a, b,..., s, t are recognition codes of 20 computers, and are a1, a2, a3, b1, b2, b3, g1, g
2, g3, 11, 12, and 13 are computers 1 respectively.
0, 20, 30, and 40 communication lines. Here, it is assumed that each computer can simultaneously open three communication lines as shown in the figure.

Each square drawn by a solid line represents a computer regardless of its size. In each square, the upper row shows the recognition code of the computer, and the lower row shows the recognition code group of the untransferred computer. The broken-line square drawn at the tip of the broken-line arrow from each computer indicates the same computer that is repeatedly performing the transfer process after the data transfer through the communication line is completed. For example, in the tree structure in the figure, the computer 10 indicated by a broken line at the right side of the second layer is the computer 2
The computer 10 is performing the next transfer process after the data transfer to 0, 30, 40 is completed. Therefore, the computer with the broken line is physically the same as the computer with the same numbered solid line. In addition, each layer indicates a computer that is working simultaneously in time. The solid arrows indicate data transfer by the communication means. Furthermore, the dashed arrows indicate the passage of time.

When the data to be transferred is installed in the computer 10 and the data transfer is started, the computer 10 divides the recognition code group of the computer to which the data is to be transferred into four equal parts, and transfers the next data for each of the three groups. The computer to be used is determined, and three communication lines are opened to start data transfer. When this is completed, the computer 10 performs the same process using the data of one remaining recognition code group among the recognition code groups divided into four. By using a structure in which transfer is performed many times so that each computer does not return to the instruction waiting state, it helps to reduce the hierarchy of the tree structure. Further, since the processing is continued until the computer completes all data transfer, efficiency is improved.

By this repetition, the recognition code group continues to be divided into four equal parts, and the process ends when there is no more recognition code data for data end transfer. The computer to which the data has been transferred from the computer 10 performs the same processing using the recognition code group transferred at the same time. Therefore, the number of computers to which data is transferred is increased by a power of 4 for each layer, and when transferring the same data to many computers,
It has a dramatic time reduction effect.

Next, the operation of the above embodiment will be described.
FIG. 2 is a flowchart showing the flow of the data transfer operation according to the present invention. First, it is checked whether or not the number of recognition codes initially given by the computer 10 is larger than the number of communication lines that can be opened simultaneously (step 101). Here, if the number of recognition codes is smaller than the number of communication lines that can be opened at the same time, and if the number of communication lines and the number of recognition codes are equal, the process proceeds to step 104, and data is sent to those computers. Perform a transfer. If the number of recognition codes present in the recognition code group is larger than the number of communication lines that can be opened at the same time, the process proceeds to step 102, where the number of recognition codes of the remaining recognition code group is simultaneously opened. Divide by the number obtained by adding 1 to the number of possible communication lines (step 1
02). Next, an arbitrary computer is determined from each of the divided recognition code groups except for one recognition code group and recognition code groups of the number of communication lines that can be simultaneously opened (step 103). Next, this determined computer 2
A communication line is opened for 0, 30, and 40 to perform data transfer (step 104).

When the transfer is completed, it is checked whether there is any remaining recognition code group divided in step 102. If there is a recognition code group, the recognition code group is used again, and the process returns to step 101, and the same processing is performed. Perform processing. If there is no remaining recognition code group, this processing is terminated (step 105).

The computers 20, 30, and
40 also starts processing similar to the processing of the computer 10 described above,
The above operation is repeated. Thereafter, all the computers that have received the data repeat the same operation.

In this embodiment, as described above, data is installed in a certain computer, and the data is transferred to the computer of the number of communicable lines + 1 at the same time, and these computers similarly develop data transfer. As a result, data transfer is terminated in a mouse formula, and high-speed data transfer for a large number of computers is realized.

If a group of computers to which data is to be transferred is in an environment in which the same physical communication line can be used, when the time is theoretically calculated, α is set to the number of communication lines that can be opened by each computer at the same time. Assuming that the total number of groups is β, the hierarchy x of the tree structure according to the present embodiment can be represented by (α + 1) ^(x−1) <β <(α + 1) ^x . For example, when data is to be transferred to 1000 computers, if the conventional method is used, 1000 data transfer operations are required, and if the transfer time per unit is τ, it takes 1000τ. Will be. However, according to the present invention, even if the number of lines that can be opened at the same time is 2, since (2 + 1) ^(7-1) <1000 <(2 + 1) ⁷ is all completed, seven layers, that is, seven data transfer times , All the data transfer can be completed, and the data transfer to 1000 computers can be completed in a theoretically 7τ time. From this, the conventional about 140
This means that the speed has been doubled.

Further, in this embodiment, since the mouse arithmetic is applied, the number of computers operating at the same time increases in the lower layers. Therefore, the effect of the present invention becomes more remarkable as the number of computer groups increases. For example, even if data is transferred to 2,000 computers, the hierarchy is the same as in the case of 1000 computers, and data can be completed in seven layers. Therefore, the time to complete all data transfer is theoretically 7τ. . At this time, a dramatic time reduction of about 280 times compared with the conventional method can be realized.

This embodiment is different from the conventional data transfer.
By repeating the mouse formula data transfer, a remarkable effect of shortening the transfer time can be obtained, the number of computers that keep waiting for instructions is minimized, and the computer automatically sends a command to all computers. Data transfer, there is almost no human intervention, and no matter how much data is processed 24 hours a day, a highly efficient data transfer method can be obtained.

As described above, in this embodiment, of the computers to which data is to be transferred, for example, by activating the computer 10, data is transferred from this computer 10 to several computers 20, 30 and the like. Computer 2
0, by repeating the same data transfer from the computer 30 to each other computer, the data transfer is started in a mouse formula.

Also, among the computers to be transferred, the total number of computers to be transferred is divided by the number of communication lines that can be opened simultaneously by that computer + 1, and the same data transfer is equally requested to the next computer to be transferred. Then, the remaining number of computers processes themselves again, thereby reducing the number of hierarchies in the mouse operation, increasing the efficiency, and terminating the data transfer more quickly.

As described above, the computer automatically repeats the data transfer until the data transfer to all the computers is completed, so that even if the number of the computers to which the data is to be transmitted is the minimum time and Since data can be transferred to all computers with minimum human intervention, human intervention can be almost eliminated as compared with the conventional method.

Embodiment 2 FIG. In the first embodiment, a case has been described in which the computer that has completed the transfer performs the transfer process again. However, the computer that performed the transfer may perform the transfer once and then do not perform the transfer again. FIG. 3 is a diagram showing an example of this embodiment. The difference from FIG. 1 is that the computer that has finished transferring does not transfer again. That is, the broken-line square and the broken-line arrow shown in FIG. 1 do not exist. According to the example shown in FIG. 3, since the computer that has completed the transfer does not perform the transfer process again, there is a possibility that the end of the data transfer may be delayed, but the computer that has completed the data transfer once performs another process. Because of this, the disadvantage of having to spend a lot of time for data transfer is eliminated.

Embodiment 3 FIG. In the first and second embodiments, the description has been made on the assumption that each computer can simultaneously open three communication lines. However, each computer can open only one communication line. This may be the case. For example, FIG. 4 is a diagram showing an example in which one computer can open only one communication line. In FIG. 4, since only one communication line can be opened, the total number of computers to be transferred is divided into two equal parts, and the divided first half is transferred using the communication line.
After the transfer of the first half is completed, the second half is transferred again using its own computer. In this way, even when only one communication line can be opened, as shown in the figure, if the transfer time is five times, it is possible to transfer data to all the twenty computers, Dramatically high-speed data transfer becomes possible as compared with the conventional case.

Embodiment 4 FIG. In the above embodiment, the case of transferring data between computers has been described.
The device for transferring data is not limited to a computer, but may be any device having a communication function.

Embodiment 5 FIG. In the above embodiment, the type of network is not particularly described, but the network may be a local area network, a network using a public line, or another private network. The protocol used for this network is not particularly limited, and any protocol may be used.
The data communication method may be an analog method or a digital method. Alternatively, the network may be a communication system in which different types of networks or different types of protocols are mixed.

Embodiment 6 FIG. In the above embodiment, an example in which one or three communication lines can be opened has been described. However, the number of communication lines may be one or more. In addition, the present invention is not limited to the case where each computer opens the same number of communication lines, but may be the case where each computer can open a different number of communication lines. Communication lines are public and private lines, for example,
The computer 10 may be connectable to the computers 20 and 30 by TCP / IP, and may be connectable to the computer 40 by a telephone line. However, the speed of subsequent data transfer depends on the data transfer speed of the communication means,
It is also conceivable that there is a significant difference in the time when the data transfer to the terminal computer ends. Further, the content of the data may be anything as long as it can be transferred by the communication means.

Embodiment 7 FIG. In the above embodiment, the description is made in consideration of the necessity of version upgrade and the like for the computer such as the distributed processing system.For example, the sales situation of the chain store and the like are calculated by the computer of each store to all the computers of the chain store. The present invention can also be applied to a case in which data is mutually transferred and a high-speed data collection is required, for example, in a case where sales are mutually notified.

Embodiment 8 FIG. In the above embodiment, the case where the identification code of the computer is transferred to identify the computer that has not been transferred has been described. For example, a table having items corresponding to each computer may be transferred. A flag is set for an item corresponding to each computer in the table, and an operation may be performed such that a computer to which data has already been transferred turns on the flag and a computer to which data is to be transferred turns off the flag. Absent.

[0034]

According to the present invention, when transferring the same data to a large number of computers, such as upgrading the entire distributed processing system, the data transfer to all the computers can be completed in a short time. The effect increases as the number of computers increases. In addition, the automation of data transfer makes it possible to minimize human intervention, and even if the number of computers to which data is to be transferred becomes large, the computers can operate 24 hours a day. Efficient data transfer can be realized, for example, by performing data transfer during the late night hours when the operating rate of the computer is low.

[Brief description of the drawings]

FIG. 1 is a diagram showing the development of data transfer according to the present invention.

FIG. 2 is a flowchart showing processing of each computer at the time of data transfer in FIG. 1;

FIG. 3 is a diagram showing the development of data transfer according to another embodiment of the present invention.

FIG. 4 is a diagram showing the development of data transfer according to another embodiment of the present invention.

FIG. 5 is a diagram showing a conventional data transfer method.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 10 Computer which transfers data first 20 Computer which receives data from computer 10 21-24 Computer which receives data from computer 20 30 Computer which receives data from computer 10 31-34 Computer which receives data from computer 30 40 Data which is received from computer 10 Receiving computers 41-44 Computers receiving data from computer 40 51-54 Computers receiving data from computer 10 a1, a2, a3 Communication lines b1, b2, b3 of recognition code b (computer 20) of recognition code a (computer 10) Communication lines g1, g2, g3 Communication lines of recognition code g (computer 30) 11, 12, 13 Communication lines of recognition code 1 (computer 40)

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) H04L 12/28 H04L 12/44-12/46 H04L 12/44

Claims (3)

(57) [Claims] 1. A data transfer system for transferring data to a plurality of processing devices, each of said plurality of processing devices, a communication line n line (n ≧ 2, 3, 4, ...) of n-number using
Means for connecting the processing apparatus, when transferring data, and means for transferring are given the list of retransmission to be processing apparatus to the data, when it is transferring data, shown above list Retransfer
The processing unit to be divided into n units and one or more processing units
N processing device groups including the calculated n processing device groups
Determine one optional processing unit from each,
N processing units determined from each of the n processing device groups
One of the above n lines is assigned to each of the
And assigned to the determined n processing devices.
Using the n lines, the transferred data and the n
Identify the processing devices included in each processing device group of the processing device group.
And a list for each of the n processing units.
Means for transferring to the n processing devices determined from each of them.
Data transfer system, characterized in that was e. 2. The method according to claim 1 , wherein each of the plurality of processing devices sets another processing device that retransmits the transferred data to n + 1.
By dividing, a (n + 1) processing device group is obtained.
The remaining processing equipment excluding one processing equipment group from the processing equipment group
One of the above n lines is assigned to each
Processing device to be re-transferred to the transferred data.
2. A data transmission system according to claim 1, wherein said data transmission system executes means for adding and transmitting a list of data. 3. Each of the plurality of processing devices stores a list of processing devices to be re-transferred to the transferred data.
After executing the means for attaching and transferring
Processing equipment included in one processing equipment group excluded from the equipment group
Until the number of data is less than or equal to n, the transferred data
For transferring a list of processing units to be re-transferred to the
3. The data transfer system according to claim 2, wherein the process is repeatedly performed .