JPS6139656A - Distributed network data processing system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a distributed data processing system using a network, and in particular to a distributed data processing system that performs automatic non-stop switching at low cost and with high reliability when having a redundant system. The present invention relates to network data processing systems.

[Background of the invention]

In recent years, with the development of microcomputers and local area networks, distributed data processing systems are becoming mainstream (Japanese Patent Application Laid-Open No. 129967/1983).
. In particular, by combining inexpensive microcomputers or personal computers with large-sized computers, a total system is realized, which makes it possible to achieve low cost and to construct the system in stages.

However, such a distributed data processing system has a problem in that as the number of computers increases, the probability that the computers will malfunction increases.

In general, when it comes to such problems, N) Even if one computer goes down, it only means that some functions are lost, and the impact is small, or (jl) The computer itself Since it is cheap, it is better to prepare a spare machine and replace it with the spare machine in case of failure.
It is said that the TTR (mean time to repair) is short. However, in reality, in most cases, even if one computer goes down, the entire system will stop operating, and in addition, even if the computer is replaced, the internal state (database) within the computer will be restored. It was an enormous task to do.

Against this background, in a distributed network data processing system, it is possible to configure a redundant system and perform non-disruptive switching, so that even if one computer goes down, the functionality of the entire system will not be lost. There is a need to realize highly reliable systems.

[Objective of the Invention] The present invention has been made in view of the above-mentioned problems, and its object is to automatically and non-disruptively change the functions of one computer when it goes down, thereby allowing continuous processing. It is an object of the present invention to provide a distributed network data processing system that guarantees performance, maintains system expandability, is inexpensive, can be used for general purposes, and does not add load to transmission paths.

[Summary of the invention]

In order to achieve the above object, the present invention stores the same functional module in a plurality of computers, allows each computer and terminal to receive data using a broadcast communication method, and one of the plurality of identical functional modules is the main system and the rest are the slave systems.The main function module takes in data from the transmission line, processes it internally, and outputs the result to the transmission line.On the other hand, the slave function module Similar to the main system, the input data is taken in and processed internally, but it is processed so as not to be output to the transmission path, and each function module transmits the priority level data given to it (hereinafter referred to as priority data) on the transmission path periodically. At the same time, it is configured to transmit data periodically, and at the same time, r(a) If it receives priority data higher than its own priority in one period, it becomes the subordinate system, and (b) one station period. If no data with a higher priority than mine is received, I will be made the master system.'' This check is performed to automatically determine whether you are the master system or the slave system.

By using the present invention, when an abnormality occurs in the main system (that is, when the highest priority data is no longer sent),
The slave system with the next highest priority becomes the master system and continues functioning.

In addition, regarding the amount of data flowing through the transmission line, although priority data is output from all computers, other data is only sent by the main system, so the line load will decrease. I can say that.

The above is an explanation of the outline of the present invention and its operation.

Next, the technology on which the present invention is based will be explained based on the drawings.

FIGS. 5 to 8 (1) and (IT) are system diagrams shown to explain the technology that is the basis of the present invention.

As shown in FIG. 5, a network having a broadcast communication function is used as the transmission path 10, and the transmission path 10 includes a plurality of computers 2OA1.
2OA, 20A, 20B.

20G, respectively, and a computer 2OA.

The same function module A is stored in 2OA, 2OA, function module B is stored in the computer 20B, and function module C is stored in the computer 20G. For example, data from the function module B stored in the computer 20B is D
If 1 is sent, the computer 2OA,,2OA,,2
The data D□ is input to the functional module A in the computer 20A1. o, , D, respectively (data D, , D, , D have the same function code and the data contents are also the same).

Now, considering the case where the function module C in the total W machine 20C requires this function code, the function module C has three pieces of data D (data D2, D, , D,), but this must be recognized as one piece of data.

Such data processing is called 0TING processing, and a specific processing method will be explained using FIG.

As shown in FIG. 8(I), the computer 20C stores the number of identical functional modules on the network (in this example, there are three functional modules A) in the counter 200, and As shown in FIG. 2, the number of counters 200 is decremented each time data D is input. The same data of the same number of cycles (If the source function code number stored in the received data frame, l, and the consistent number assigned to each data frame, I, match, it will be recognized as the same data frame) When it arrives, it is collated and made into one piece of data. At this point, data matching is performed using majority logic. Furthermore, since the data D4 arrives within the timeout period T, the counter 200 becomes zero. Therefore, each computer 2 OA,, 20A,.

20A3 is normal, and verification is started as is.

In addition, in order to deal with the case where one of the computers is dead and data is not sent, the timeout period (T shown in the figure) after receiving the first data is set as shown in Figure 8 (II). Even if the same data for the number of stored same function modules does not arrive, the verification process is started after this timeout.

As explained above, this basic technology originally focused on increasing the reliability of received data, but it also has the advantage that the same functional module runs in parallel on multiple computers, so Even in the event of a system failure, as shown in FIG. 6, system continuity is ensured without causing any malfunction.

On the other hand, the problem with this method is that advanced information processing called VOTING is required in each functional module, which increases the cost of the processor that performs this processing, especially when a terminal is installed on the transmission path, and the cost of the terminal itself increases. Can be more expensive (of course, data frame routing management and modification capabilities are also required).

Furthermore, if any of the same functional modules is down due to VOTING, the actual processing will not be performed until after the timeout, resulting in a time delay.

Furthermore, by connecting a redundant system to the transmission line I, the line load increases proportionally, increasing the line load which is always critical in a network system.

[Embodiments of the invention]

The present invention solves the problems of such basic technology, and embodiments thereof will be described with reference to FIGS. 1 to 4.

FIG. 1 is a system diagram of a system to which an embodiment of the present invention is applied. As shown in Figure 1, the computer 20A1.2OA2
．． Function module A is stored in computer 20A3, and function module B is stored in computer 20B.

Further, 22 is a terminal. Computer 2OA.

2OA2.2OA, , 20B and the terminal 22 are connected to the transmission path 1
Connected to 0.

FIG. 2 is a configuration diagram showing the internal structure of each functional module A, H. 30 is the entire functional module, 31 is the main system 1
A subordinate automatic determination unit, 32 is a switch unit that indicates whether or not the output is possible;
33 indicates the software group in the corresponding function module 1
There are various cancer programs that are not covered by the present invention.
4 and a database 35, which performs desired processing based on input data and requests output.

Reference numeral 36 denotes a storage section for the priority level of this functional module, in which the level given in advance to each functional module is stored. In this configuration example, 37 is general input data, 38 is priority input data, 39 is general output data, and 4o is own priority output data.

The operation of the master system 1 slave system determination section 31 will be explained with reference to the flowchart shown in FIG.

As shown in FIG. 3, this part operates periodically and sends out the priority level from the storage unit 36 of its own function module (step 8100), and also
The highest priority level among those received so far is determined (step 5101), and the determined priority level L, 1. and the priority level of its own function module (step 8102), and if the priority level of the own function module is higher, the output switch 32 is turned ON (step 8103), and if the priority level of the own function module is lower, the output switch 32 is turned OFF (step 8104).

Now, the computer 2OA shown in Figure 1, 20A2゜2
Assume that the priority levels of each functional module A in 0A have the following relationship.

A of 20A1>A of 20A2>A of 2OA In this case,
By operating the master system 1 slave system determining section 31, the computer 20
Function module A of A is the main system (therefore, the output is ON)
So, the calculation is 4ii2OA.

And each functional module A of the computer 20A3 becomes a slave system (therefore, the output is 0FF).

(In the state shown in FIG. 1, A' represents the main system, and A represents the slave system). At this time, the general data D6 sent from the functional module B is sent to the computer 2 OA1.2OA.

The input is input to the functional module A of the computer 20A3, and the same processing is performed within each functional module A, but only the general data D7 from the functional module A of the computer 20A0 is output, and the same data is sent to the terminal 22. You won't be disappointed.

Next, from the state of FIG. 1 to the computer 20A as shown in FIG.
Consider the case where 1 goes down. At this time, the computer 2O
Since priority data is not sent from the functional module A of computer A, the internal functional module of computer 2oA is switched to the primary system by the function of the primary system 1 slave system determination unit 31 in computer 2oA2, and the subsequent output is Sadoru Tsukasa.

In this way, non-stop and automatic switching can be achieved even when the main system is down.

In the embodiment of the present invention, computers 2OA, 2OA.

It is possible to continue functioning completely even if the system goes down.

In addition, the terminal connected to the transmission path only needs to process the data of the registered function code, which is low cost, and high responsiveness is achieved because the time required for verification processing and the like is not wasted.

The load on the transmission path also decreases.

Furthermore, in this example, even if it is desired to make the functional module B redundant, it is only necessary to simply insert another computer for storing the functional module B onto the line, and there is no need to change the existing software.

〔Effect of the invention〕

As described above, according to the present invention, it is possible to perform non-stop switching when a computer goes down, which has been the biggest problem in distributed network computer systems.

Further, according to the present invention, the amount of data flowing on the network does not increase at all, so the network load does not increase.

[Brief explanation of drawings]

Fig. 1 is a system diagram showing a system to which an embodiment of the present invention is applied, Fig. 2 is a block diagram showing the configuration of functional modules, and Fig. 3 is a flowchart shown to explain the operation of the embodiment of the present invention. , Fig. 4 is a system diagram shown to explain the case when the computer goes down, Figs. 5 and 6 are system diagrams shown to explain the technology that is the basis of the present invention, and Figs. 7 and 8 are This figure is an explanatory diagram used to explain the operations in FIGS. 5 and 6. 10... Transmission line, 2OA, 2OA2.20A, 2
0B. 20G...Computer, 22...Terminal, D...Data, A. B, C, 30...Function module.

Claims (1)

[Claims] 1. A distributed network data processing system that has a transmission line with a broadcast communication function and a plurality of data processing devices that are connected to this transmission line and send and receive data through broadcast communication, to achieve a specified function. The functional modules to be processed are stored in the plurality of data processing devices, one of the data processing devices storing the functional modules is designated as the main system, the rest are the slave systems, and the main and slave processing devices are connected to the transmission path. Data sent via the main system is taken in, processed internally, and only the main data processing device sends the processing results onto the transmission path, so that the internal states of the main and slave data processing devices always match. Distributed network data processing is characterized in that, if an abnormality occurs in the main system, one of the remaining slave systems switches to the main system, and thereafter that slave system outputs to the transmission path. system.