JP3513537B2 - Fault-tolerant computer system - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fault tolerant computer system, and more particularly to a fault tolerant computer system using a boundary scan element as a communication element.

[0002]

2. Description of the Related Art Computers used for building monitoring systems, various databases, cash dispensers of financial institutions, etc. are required to perform non-stop and continuous processing for 24 hours or a predetermined time. Therefore, a so-called fault-tolerant computer system has been proposed in order to avoid interruption of processing due to a failure of the computer. A fault-tolerant computer system is a computer that is duplicated or multiplexed, and the same processing as that of the main computer is virtually executed by another computer at all times, and when the main computer goes down, the other computer's processing It is a system that allows inheritance. Therefore, according to this system, even when the main computer that is actually executing the predetermined processing is down, the processing is immediately continued by the other calculation beach, and the interruption of the processing can be avoided.

Here, in such a fault-tolerant computer system, it is important for another computer to operate when the main computer is down, so that a series of processes can be transferred to another computer without delay. The purpose is to detect the down time quickly and to identify in which process it went down. In this regard, conventionally, for example, an identification method using a watchdog timer has been proposed. The identification method by the watchdog timer is a method in which the main computer supplies a signal to another computer at a fixed timing, and the other computer operates the main computer normally while the signal is being supplied. If the signal is not supplied within a certain period of time, it is assumed that the main computer is down and the processing is inherited.

[0004]

However, when the watchdog timer is used as a reference for identification, the signal supply interval becomes a problem. In other words, the shorter the signal supply interval, the faster the downtime of the main computer can be detected.
Moreover, although the position of the interrupted process can be easily specified, the process of supplying the signal is frequently performed, and the process that should be originally performed by the computer is delayed. On the contrary, if the interval of supplying the signal is long, the processing originally supposed to be performed by the computer will not be delayed, but the detection of the down of the main computer will be delayed, and the position of the interrupted processing will be inaccurate. Therefore, inheritance of processing may not be successful. Further, since the processing for the watchdog timer is executed asynchronously with the processing originally performed by the computer, it is difficult to accurately specify the position of the interrupted processing.

Therefore, an object of the present invention is to provide a fault-tolerant computer system which can detect the down of the main computer quickly without delaying the original processing of the computer and accurately specify the position of the interrupted processing. To provide.

[0006]

Here, the inventor of the present application is
Regarding a boundary scan element for performing a wiring test of an electronic circuit board or an operation test of an IC on the board, the boundary scan element is not used as an element merely for connection inspection of wiring, but various processing targets such as a CCD camera are controlled. Focusing on the usefulness as a communication element for doing so, a communication device applying this as a communication element has been proposed so far (International Publication No. WO98 / 55925 etc.). Then, the present inventor adopts the boundary scan element as a communication element in the fault tolerant computer system,
The above problems have been solved by the following means.

That is, according to the present invention, a main computer,
Auxiliary computer that virtually executes the same processing as the main computer at normal times, a communication element connected to the processing target, and a connection between the communication element and either the main computer or the auxiliary computer Switching means for switching, the communication element is a boundary scan element, and the switching means performs the connection switching based on the presence or absence of a clock signal supplied from the main computer to the communication element. A fault tolerant computer system is provided.

According to this means, in a normal state, the switching means connects the main computer and the communication element as a boundary scan element, and the main computer processes the processing target. Then, communication of various data between the main computer and the processing target is performed via the communication element which is a boundary scan element. At this time, the auxiliary computer virtually executes the same process as the process performed by the main computer on the processing target.

If the main computer goes down,
The clock signal sent to the communication element by the main computer is interrupted. At this time, the switching means functions to cause the auxiliary computer to process the processing target instead of the main computer. That is, the switching means disconnects the main computer and the communication element and connects the auxiliary computer and the communication element. As a result, the auxiliary computer actually executes the processing of the processing target via the communication element. As a result,
This avoids the interruption of the series of processes for the process target.

Due to this action, this means has the following unique effects. First, it is determined whether or not the main computer is down based on the presence or absence of a clock signal supplied from the main computer to the communication element, and a down identification process is performed between the main computer and the auxiliary computer. Is not performed, the processing that the main computer should originally perform will not be delayed regardless of the cycle of the clock signal. Further, since the communication element, which is the boundary scan element, executes the processing in synchronization with the clock signal, it is extremely easy to identify the processing that was being performed by the main computer immediately before the down.

In the present invention, the main computer means a computer that communicates with the processing target in normal times, and the auxiliary computer is the processing target instead of the main computer when the main computer goes down. It means a computer that performs communication processing. In addition, the above-mentioned “virtually execute the same processing”
Means that the auxiliary computer is aware of at least the processing performed by the main computer, and is in a state in which the processing can be continued if necessary. As the switching means, a combination of logic ICs, a PLD, or the like can be used.

The boundary scan element in the present invention will be described after explaining the conventional technology of the boundary scan element and the function as a communication element. FIG. 5A is a block diagram of a general boundary scan element 100. The boundary scan element 100 is
An input-side boundary cell 103 provided individually to the input-side terminal 101, an output-side boundary cell 104 provided individually to the output-side terminal 102, and a TDI terminal 105 for inputting / outputting data to / from the boundary cells 103 and 104.
And TDO terminal 106, TAP circuit 107, TAP
The package 11 includes a TCK terminal 108 for supplying a clock signal to the circuit 107 and a TMS terminal 109 for supplying an operation mode switching signal to the TAP circuit 107.
It is built in 0.

The input-side boundary cell 103 and the output-side boundary cell 104 are connected in series in a beaded manner, and the cells at both ends thereof are connected to the TDI terminal 105 or the TDO terminal.
The terminal 106 is connected. And the TDI terminal 10
By serially shifting the boundary data 103 and 104 from the serial data input from the cell 5, the data can be set in all the boundary cells 103 and 104. In addition, all boundary cells 10
The data set in 3 and 104 can be sent as serial data from the TDO terminal 106 to the outside by shifting these in order. The input-side boundary cell 103 can take in the data from the input-side terminal 101, and can output the set data to the input-side terminal 101. This is the same between the output-side boundary cell 104 and the output-side terminal 102.

The TAP circuit 107 synchronizes with the clock signal supplied from the TCK terminal 108, and the TMS terminal 10
Predetermined processing regarding the boundary cells 103 and 104 is executed in accordance with the operation mode switching signal supplied from the controller 9. For example, a process of fetching data from the TDI terminal 105 and shifting or setting the data to each of the boundary cells 103 and 104, a process of outputting the data set in each of the boundary cells 103 and 104 from the TDO terminal 106, an input side terminal 101 and an input side boundary. Processing for inputting / outputting data to / from the cell 103 and inputting / outputting data between the output-side terminal 102 and the output-side boundary cell 104 is executed.

If the TCK line and the TMS line are very long, the TMS signal may be delayed with respect to the clock signal. Therefore, as shown in FIG. 5D, two TMS terminals 109 and two TCK terminals 108 are provided, one of which is an input side terminal and the other of which is an output side terminal. Once TAP
If it is taken into the circuit 107, latched, and output,
The delay can be eliminated.

When using the boundary scan element 100 as a communication element, for example, the input side terminal 101 is connected to the input terminal of the communication target (processing target), and the output side terminal 102 is connected to the output terminal of the communication target. Also, the TDI terminal 105 and the TDO terminal 10 are connected to the host computer.
6, the TCK terminal and the TMS terminal 109 are connected respectively. Then, the host computer sends a clock signal and a TMS signal, causes the TAP circuit 107 to execute a predetermined process, and performs data communication between the host computer and the communication target. For example, the control data is transmitted to the TDI terminal 105, is set in the input-side boundary cell 103, and is further transmitted from the input-side terminal 101 to the communication target to control the communication target. Alternatively, the data acquired or analyzed by the communication target is fetched from the output side terminal 102 to the output side boundary cell 104, and this is transferred to the TDO terminal 106.
Output from the host computer. In this way, the boundary scan element 100 can function as a communication element.

As another example of the boundary scan element 100, an RST terminal for supplying a reset signal to the TAP circuit 107, or a TDI terminal 105 and a TD is added to the structure of the boundary scan element 100 described above.
Bypass line that short-circuits the O terminal 106, or TD
In some cases, an ID code register or the like is added to the I terminal 105 and the TDO terminal 106. Further, as another example of the boundary scan element 100, as shown in FIG. 5B, an input side boundary cell 103 ′ is provided. TDI terminal 1 without connecting all of the output side boundary cell 104 'and the output side boundary cell 104' in series.
There is also one connected in parallel between 05 'and the TDO terminal 106'. The reference numeral of FIG. 5 (b) corresponds to the reference numeral of FIG. 5 (a).

Furthermore, the boundary scan element 100
As another example, as shown in FIG. 5C, the input side boundary cell 103 ″ and the output side boundary cell 104 ″ are connected in parallel, and each boundary cell 103 ″ and 1
04 "TDI terminal 105" or TDO terminal 1 respectively
There is also one in which the input side terminal 101 ″ and the output side terminal 102 ″ are connected to the TAP circuit 107 ″.

As described above, although various kinds of boundary scan elements themselves have been proposed, the present invention uses the boundary scan element as a communication element and requires a clock signal. Therefore, the boundary scan element of the present invention is not limited to the boundary scan element. at least a boundary cell, a TAP circuit, and TCK terminal, but requires that and a TMS terminal, connection method between additional configurations and boundary cells mutual described above, both for the number of bits or the like of the boundary cells <br I can't.

[0020]

BEST MODE FOR CARRYING OUT THE INVENTION (Embodiment 1) Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings. Figure 1
FIG. 1 is a block diagram of a fault tolerant computer system A according to the present invention. The fault-tolerant computer system A is a system for continuously and continuously processing a communication operation with a communication target or a processing target (hereinafter, represented by “processing target etc.” in the present embodiment). is there. In this embodiment, the fault-tolerant computer system includes a main computer 1, an auxiliary computer 2, and a switcher 4 that performs connection switching between a processing target and either the main computer 1 or the auxiliary computer 2. Prepare The processing target and the like include various devices such as a communication unit and a processing device for transmitting and receiving data to and from the main computer 1, various I / O (input / output devices), and the like.

The main computer 1 and the auxiliary computer 2 are storage means.
As a dual port ram 3 is shared. The main computer 1 and the auxiliary computer 2 can freely write / read data to / from the dual port RAM 3 and can share these data with each other. Therefore, when the main computer 1 is operating normally, the auxiliary computer 2 grasps the processing executed by the main computer 1 on the basis of the data written in the dual port ram 3, so to speak, it is virtually the same. It is possible to execute the processing of.

The processing object and the like are controlled by receiving data from the main computer 1 or the auxiliary computer 2 via the switch 4, and the data obtained by the processing operation of the processing object or the like is used as the main computer 1 or the auxiliary computer. Send to 2.

The switch 4 is a TCK terminal 1d of the main computer 1.
The output terminal 4e, the input terminal 4f, the TMS output terminal 4g and the TCK based on the presence / absence of the clock signal sent from the
The output terminal 4h is connected to the input terminal 4a and the output terminal 4 respectively.
b, TMS input terminal 4c and TCK input terminal 4d, or output terminal 4e, input terminal 4f, TMS output terminal 4g and TCK output terminal 4h, respectively, input terminal 4a ', output terminal 4b', TMS input. It is configured by a PLD that performs either one of the connection operation with the terminal 4c ′ and the TCK input terminal 4d ′. When the clock signal is sent from the main computer 1, the input terminal 4a, the output terminal 4b, the TMS input terminal 4c, and the TMS input terminal 4c.
When the clock signal is interrupted, it is connected to the CK input terminal 4d and connected to the input terminal 4a ', the output terminal 4b', the TMS input terminal 4c 'and the TCK input terminal 4d'.

The switch 4 has a selection signal output terminal 4i connected to the selection signal input terminal 1e of the main computer 1 and a selection signal output terminal 4j connected to the selection signal input terminal 2e of the auxiliary computer 2. When the fault-tolerant computer system is operating, the selection signal input terminal 1e of the main computer 1 or the selection signal input terminal 2e of the auxiliary computer 2 is notified of which computer is selected. While the selection signal is transmitted, the non-selection signal is transmitted to the other. Specifically, when a clock signal is sent from the TCK terminal 1d of the main computer 1, a selection signal is sent to the main computer 1, a non-selection signal is sent to the auxiliary computer 2, and the clock signal is interrupted. Sends a non-selection signal to the main computer 1 and a selection signal to the auxiliary computer 2, respectively. The main computer 1 and the auxiliary computer 2 can judge which one actually executes the process with respect to the process target or the like according to these selection and non-selection signals. It suffices to know the time when the auxiliary computer 2 actually executes the processing, and it is not necessary to send these selection / non-selection signals to the main computer 1, so
It is also possible to send the non-selection signal only to the auxiliary computer 2. However, after the main computer 1 goes down,
When performing the restoration work, it is useful to have a configuration in which a selection / non-selection signal is sent to the main computer 1 and the auxiliary computer 2.

Next, the operation of the fault tolerant computer system A of FIG. 1 having the above configuration will be described below. First, if the main computer 1 is operating normally,
That is, when the TCK input terminal 4d of the switch 4 is continuously output from the TCK terminal 1d of the main computer 1, the switch 4 outputs the output terminal 4e and the input terminal 4
f, TMS output terminal 4g and TCK output terminal 4h, input terminal 4a, output terminal 4b, TMS input terminal 4c and T
While being connected to the CK input terminal 4d, a selection signal is sent from the selection signal output terminals 4i and 4j to the main computer 1 and a non-selection signal is sent to the auxiliary computer 2. As a result, the main computer 1 actually performs communication processing with the processing targets 5a and 5b via the communication elements 100a and 100b, and the auxiliary computer 2 can recognize that the main computer 1 is operating normally.

The main computer 1 writes the obtained data such as a result of the communication processing sequential dual port tram 3, the data is shared between the main computer 1 and the auxiliary computer 2. As a result, the auxiliary computer 2 can virtually execute the same processing as that of the main computer 1. -However, if an abnormality occurs in the main computer 1 and the processing stops, TCK
The clock signal from the terminal 1d will be lost. At this time, the switching unit 4 recognizes that the clock signal has been interrupted, and sends the non-selection signal to the main computer 1 and the selection signal to the auxiliary computer 2 from the selection signal output terminals 4i and 4j, respectively. Further, the switching device 4 includes an output terminal 4e and an input terminal 4
f, TMS output terminal 4g and TCK output terminal 4h, input terminal 4a ', output terminal 4b', TMS input terminal 4c '
And TCK input terminal 4d '.

Based on the received selection signal, the auxiliary computer 2 identifies that the main computer 1 has gone down, and functions to perform the processing virtually performed up to now with the processing target. The operation stop is avoided. At this time, the processing of the main computer 1 is inherited by the auxiliary computer 2 extremely accurately. That is, since the processing target or the like performs communication processing in perfect synchronization with the clock signal, the point where the clock signal is interrupted can be grasped as it is as the point where the communication processing by the main computer 1 is interrupted. Therefore, the auxiliary computer 2 shifts the processing virtually performed up to that point to the actual processing from the time when the selection signal is received from the switch 4 while referring to the data recorded in the dual port RAM 3. For example, it is possible to continue the processing of the processing target without any interruption. Further, the determination as to whether or not the main computer 1 is down is made based on the presence or absence of the transmission of the clock signal from the main computer 1 as described above.
Since the recognition process between the computer and the auxiliary computer 2 is unnecessary,
It also avoids processing delays.

(Second Embodiment) Next, a second embodiment of the fault-tolerant computer system according to the present invention will be described. FIG. 2 is a block diagram of a fault tolerant computer system A ′ according to the second embodiment of the present invention. In this embodiment, the main computer 1, the auxiliary computer 2, the dual port ram 3, and the switch 4 which compose the fault-tolerant computer system have the same configurations as those in the first embodiment. The fault-tolerant computer system A'is a system for continuously and non-stop processing the processing targets 5a and 5b, and is connected to the main computer 1, the auxiliary computer 2, and the processing targets 5a and 5b, respectively. Communication element 10
0a and 100b, and a switching unit 4 that switches connection between the communication elements 100a and 100b and either the main computer 1 or the auxiliary computer 2.

The main computer 1 and the auxiliary computer 2 share the dual port RAM 3. The main computer 1 and the auxiliary computer 2 can freely write / read data to / from the dual port RAM 3 and can share these data with each other. Therefore, when the main computer 1 is operating normally, the auxiliary computer 2 grasps the processing executed by the main computer 1 on the basis of the data written in the dual port ram 3, so to speak, it is virtually the same. It is possible to execute the processing of.

The processing targets 5a and 5b are, for example, cash dispensers, financial lines, communication lines, monitoring devices, etc., which receive data from the main computer 1 or the auxiliary computer 2 via the communication element 100a or 100b. The data controlled and detected is sent to the main computer 1 or the auxiliary computer 2.

Communication element 100a and communication element 100b
Are all boundary scan elements, and in particular, in the present embodiment, all of them employ the boundary scan element 100 of FIG. 5A. In the following description and drawings, the communication element 100a and the communication element 100 will be described.
The reference numeral b corresponds to the reference numerals of the boundary scan element 100 and its components. Further, in the present embodiment, two communication elements 100a and 100b are provided corresponding to the number of processing targets 5a and 5b, but the present invention is not limited to this and may be more. ,
It may be less than this.

The communication elements 100a and 100b are connected to the switch 4 in series. Specifically, the TDO terminal 106a of the communication element 100a is the communication element 100b.
Of the communication element 100a, the TDI terminal 105a of the communication element 100a to the output terminal 4e of the switch 4,
The TDO terminal 106b of 0b is the input terminal 4f of the switch 4.
, Respectively. The communication element 100a
And 100b, the TCK terminals 108a and 108b, and the TMS terminals 109a and 109b are the TCKs of the switching unit 4.
The output terminal 4h and the TMS output terminal 4g are respectively connected.

The switch 4 is a TCK terminal 1d of the main computer 1.
The output terminal 4e, the input terminal 4f, the TMS output terminal 4g and the TCK based on the presence / absence of the clock signal sent from the
Output terminal 4h, input terminal 4a, output terminal 4b, TMS
Input terminal 4c and TCK input terminal 4d, or input terminal 4a ', output terminal 4b', TMS input terminal 4c 'and T
The PLD is connected to either one of the CK input terminals 4d '. In addition, when the clock signal is sent from the main computer 1, it is connected to the input terminal 4a, the output terminal 4b, the TMS input terminal 4c and the TCK input terminal 4d, and when the clock signal is interrupted, the input terminal 4a ′, Output terminal 4
b ', TMS input terminal 4c' and TCK input terminal 4d '
Connect to.

The switching unit 4 has a selection signal output terminal 4i.
4j, and sends a selection signal or a non-selection signal notifying which computer is selected to the selection signal input terminal 1e of the main computer 1 or the selection signal input terminal 2e of the auxiliary computer 2. Specifically, when a clock signal is sent from the TCK terminal 1d of the main computer 1, a selection signal is sent to the main computer 1 and a non-selection signal is sent to the auxiliary computer 2, and the clock signal is interrupted. Sends a non-selection signal to the main computer 1 and a selection signal to the auxiliary computer 2. Based on these signals, the main computer 1 and the auxiliary computer 2 can determine which of the processing targets 5a and 5b actually executes the process. In addition,
It is not necessary to send these selection signals to the main computer 1 because it is sufficient to know the time when the auxiliary computer 2 actually executes the process. However, when the main computer 1 goes down and the recovery work is performed, Is beneficial.

Next, the main computer 1 and the communication elements 100a, 1
The communication processing between 00b and the processing targets 5a and 5b is performed as follows, for example. Main computer 1 is T
When the clock signal is constantly transmitted from the CK output terminal 1d and the data is transmitted from the main computer 1 to the processing targets 5a and 5b, first, the communication element 10 is transmitted from the TMS output terminal 1c.
0a and 100b TAP circuits 107a and I07b
TMS via switch 4 (not shown in FIG. 2)
A signal is transmitted and the operation mode of the TAP circuits 107a and 107b is switched to the setting of data to the input side boundary cell 103a (not shown in FIG. 2).

Subsequently, the main computer 1 has a data output terminal 1a.
To send the predetermined data in serial form to the TDI terminal 105a of the communication element 100a via the input terminal 4a and the output terminal 4e of the switch 4. In the communication element 100a, the data input from the TDI terminal 105a is sequentially transferred to each of the boundary cells 103a and 104a, and further, the TDI of the communication element 100b is transmitted from the TDO terminal 106a.
This is sent to the terminal 105b, and in the communication element 100b,
Similarly, these data are transferred to each boundary cell 103b and 104b. Then, the input-side boundary cells 103a and 103 of the communication elements 100a and 100b, respectively.
When necessary data is spread in b, this is set.

Next, the main computer 1 sends the TMS signal again to switch the operation mode of the TAP circuits 107a and 107b, and thereby the communication elements 100a and 100b.
Sends the data set in each of the input-side boundary cells 103a and 103b from the input-side terminal 101a or 101b to the processing target 5a or 5b. The processing target 5a or 5b executes processing according to the content of the data. On the other hand, when sending data from the processing target 5a or 5b to the main computer 1, the main computer 1 first issues a data output command to the processing target 5a or 5b by the procedure described above. as a result,
The data output from the processing target 5a or 5b is the output side terminal 102a or 1 of each of the communication elements 100a and 100b.
It will be transmitted to 02b.

Subsequently, the main computer 1 sends a TMS signal to switch the operation mode of the TAP circuits 107a and 107b, whereby the communication elements 100a and 100b are
The data transmitted to each output side terminal 102a or 102b is set in each output side boundary cell 104a or 104b. The main computer 1 sends a TMS signal again to send TA
Switching the operation mode of the P circuits 107a and 107b,
The data set in the output-side boundary cells 104a and 104b are sequentially transferred to the TDO of the communication element 100b.
It is sent as serial data from the terminal 106b.

The main computer 1 can take in the transmitted data from the data input terminal 1b via the input terminal 4f and the output terminal 4b of the switch 4 and perform processing such as writing to the dual port RAM 3 or the like. it can. The main computer 1, the communication elements 100a and 100b, and the processing target 5
Since the communication processing between a and 5b is performed in exactly the same way as the communication processing between the auxiliary computer 2 and these questions when the main computer 1 goes down, the description thereof is omitted.

Next, the operation of the fault-tolerant computer system A'of FIG. 2 having the above configuration will be described below. First, when the main computer 1 is operating normally, that is, when the TCK terminal 1d of the main computer 1 continuously outputs the TCK input terminal 4d of the switch 4 to the switch 4, the switch 4 outputs The terminal 4e, the input terminal 4f, the TMS output terminal 4g and the TCK output terminal 4h,
Connect to the input terminal 4a, the output terminal 4b, the TMS input terminal 4c and the TCK input terminal 4d, and send the selection signal from the selection signal output terminals 4i and 4j to the main computer 1 and the non-selection signal to the auxiliary computer 2, respectively. To do. As a result, the main computer 1 actually performs communication processing with the processing targets 5a and 5b via the communication elements 100a and 100b, and the auxiliary computer 2 can recognize that the main computer 1 is operating normally.

The main computer 1 sequentially writes the data and the like obtained as a result of the communication processing to the dual boot RAM 3, and the data is shared between the main computer 1 and the auxiliary computer 2. As a result, the auxiliary computer 2 can virtually execute the same processing as that of the main computer 1. -However, if an abnormality occurs in the main computer 1 and the processing stops, TCK
The clock signal from the terminal 1d will be lost. At this time, the switching unit 4 recognizes that the clock signal has been interrupted, and sends the non-selection signal to the main computer 1 and the selection signal to the auxiliary computer 2 from the selection signal output terminals 4i and 4j, respectively. Further, the switching device 4 includes an output terminal 4e and an input terminal 4
f, TMS output terminal 4g and TCK output terminal 4h, input terminal 4a ', output terminal 4b', TMS input terminal 4c '
And TCK input terminal 4d '.

The auxiliary computer 2 discriminates from the received selection signal that the main computer 1 has gone down, and executes the processing virtually performed up to now on the communication elements 100a and 100.
It functions so as to actually perform the processing with the processing targets 5a and 5b via b, and avoids stopping the processing targets 5a and 5b. On this occasion,
The process of the main computer 1 is inherited by the auxiliary computer 2 extremely accurately. That is, the communication elements 100a and 100
This is because in b, since the communication processing is performed in perfect synchronization with the clock signal, the point where the clock signal is interrupted can be grasped as it is as the point where the communication processing by the main computer 1 is interrupted. Therefore, the auxiliary computer 2 shifts the processing virtually performed up to that point to the actual processing from the time when the selection signal is received from the switch 4 while referring to the data recorded in the dual port RAM 3. For example, processing target 5a
The processing of 5 and 5b can be continued without any interruption. Further, the determination as to whether or not the main computer 1 is down is performed based on the clock signal sent from the main computer 1 as described above, and the recognition process between the main computer 1 and the auxiliary computer 2 is not necessary. Because of this, there is no delay in processing.

(Third Embodiment) Next, a third embodiment of the fault-tolerant computer system according to the present invention will be described. FIG. 3 is a block diagram of the fault tolerant computer system B. The fault-tolerant computer system B, like the fault-tolerant computer system A ', is a system for continuously and non-stop processing the processing targets 5a and 5b. It is characterized in that the computers 11, 11 ′ and 11 ″ are provided, and that the switch 14 has a function of switching and driving any of the main computers 11, 11 ′ or 11 ″. That is, the fault-tolerant computer system B includes a plurality of main computers 11,
11 ', 11 ", one auxiliary computer 12, communication elements 100a and 100b connected to the respective processing targets 5a and 5b, communication elements 100a and 100b and main computer 11, 11', 11" or auxiliary computer And a switching device 14 for switching the connection between any one of the two.

As described above, the main computers 11, 11 'and 11 "are computers that execute different processes, and any one computer is selected according to the processing contents of the processing targets 5a and 5b. This selection is made from the switching device 14 to the selection signal terminal 1 of each of the main computers 11, 11 'and 11 ".
Recognized by the selection signal sent to 1b, 11b 'and 11b ". Also, each main computer 11, 11' and 1
1 "is a communication terminal group 11a, 11a 'and 11 corresponding to the data output terminal 1a, the data input terminal 1b, the TMS output terminal 1c and the TCK output terminal 1d of the main computer 1 in the fault tolerant computer system A'described above.
a ”and performs the same communication processing as the above-described main computer 1 through this communication terminal group. In addition, although three main computers are provided in this embodiment, the present invention is not limited to this.

The auxiliary computer 12 is the main computer 11, 11 '.
And 11 "can execute the respective processing contents executed by the auxiliary computer 12 and the auxiliary computer 12 sends a signal for identifying which main computer 11, 11 'or 11" is actually executing the process. Identification signal terminal 1 for receiving
2b, and the main computer 11 which is actually performing the process by the identification signal sent from the switch 14,
11 'or 11 "is identified. Further, the main computer 11 is provided with a selection signal terminal 12c for receiving a selection signal as to whether or not the selected main computer 11, 11' or 11" is down. , 11 'or 11 "is working properly

On the other hand, the auxiliary computer 12 has the data output terminal 2a, the data input terminal 2b, and the TMS of the auxiliary computer 2 in the fault-tolerant computer system A'described above.
The communication terminal group 12a corresponding to the output terminal 2c and the TCK output terminal 2d is provided, and the same communication processing as that of the auxiliary computer 2 described above is performed through this communication terminal group. Each main computer 11,
11 ′ and 11 ″ and the auxiliary computer 12 share the corresponding dual port rams 13, 13 ′ and 13 ″ , respectively. This dual port ram 13, 13 'and 1
For 3 ", the corresponding main computers 11, 11 'and 1
1 ”and the auxiliary computer 12 can freely write and read data, and can share these data with each other. Therefore, the auxiliary computer 12 is actually executing any process. When the main computer 11, 11 'or 11 "is operating normally, based on the data written in any of the corresponding dual port rams 13, 13' or 13", the main computer 11, 11 'or It is possible to grasp the processing that 11 ″ is executing and, so to speak, virtually execute the same processing.

The switch 14 is connected in parallel with the communication terminal groups 11a, 11a 'and 11a "of the main computers 11, 11' and 11", and is an input terminal of the switch 4 of the fault tolerant computer system A'described above. 4a, output terminal 4
b, the communication terminal group 14a corresponding to the TMS input terminal 4c and the TCK input terminal 4d, and the communication terminal group 12a of the auxiliary computer 12, and also the input terminal 4a ′ of the switching device 4,
The communication terminal group 14a ′ corresponding to the output terminal 4b ′, the TMS input terminal 4c ′, and the TCK input terminal 4d ′, and the output terminal 4e, the input terminal 4f, the TMS output terminal 4g, and the TCK output terminal 4h of the switching device 4 are also provided. Corresponding communication terminal group 14
e, and any selected main computer 1
It is a PLD that connects the communication terminal group 14e and either the communication terminal group 14a or 14a 'depending on the presence or absence of a clock signal sent from 1, 11' or 11 ".

Further, the switching unit 14 has a selection signal terminal 14b for sending a selection signal to any one of the main computers 11, 11 'or 11 ", and any one of the main computers 1 to the auxiliary computer 12.
An identification signal terminal 14c for transmitting an identification signal for identifying whether 1, 11 'or 11 "is selected, and identifying whether the selected main computer 11, 11' or 11" is down A selection signal terminal 14d for transmitting a selection signal for The relationship between the switch 14 and the communication elements 100a and 100b is the same as in the case of the fault-tolerant computer system A'described above.
The description is omitted.

Next, the operation of the fault tolerant computer system B having the above configuration will be described. First,
If any of the main computers 11, 11 'or 11 "is selected and, for example, the main computer 11 is selected, the switching device 14 sends a non-selection signal from the selection signal terminal 14b to the main computers 11' and 11". Then, these processes are stopped. At the same time, an identification signal is sent from the identification signal terminal 14c to the auxiliary computer 12 to notify that the main computer 11 is selected.

The auxiliary computer 12 receives the identification signal,
In preparation for the down of main computer 11, dual port ram 13
Virtually the same processing as that of the main computer 11 is performed while referring to the data of 1. The switching device 14 connects the communication terminal group 14a and the communication terminal group 14e as long as the clock signal is sent from the main computer 11, and the main computer 11 connects the processing target 5a through the communication elements 100a and 100b. And 5b
And communication processing. Since the communication processing method is the same as that of the fault tolerant computer system A ′ described above,
The description is omitted.

Here, when switching the main computer 11,
For example, when causing the main computer 11 'to execute processing, the main computer 11 sends a selection signal from the selection signal terminal 11b to the switch 14 so as to select the main computer 11'. Switch 1
In response to this, the main computer 1 receives the selection signal from the selection signal terminal 14b.
A non-selection signal is sent to 1 and 11 ", and a selection signal is sent to the main computer 11 '. At the same time, the switch 14 is
An identification signal is sent from the identification signal terminal 14c to the auxiliary computer 12 to notify that the main computer 11 'is selected. With this identification signal, the auxiliary computer 12 prepares for the down of the main computer 11 'and refers to the data of the dual port 13', and virtually performs the same process as the main computer 11 '.

The switching device 14 is provided with the communication terminal group 14a as long as the clock signal is sent from the main computer 11 '.
And the communication terminal group 14e are connected to each other, and the main computer 11 ′ performs communication processing with the processing targets 5a and 5b via the communication elements 100a and 100b. In this way, the main computer 11,
Switching of 11 ′ or 11 ″ is performed.

On the other hand, if the main computer 11 'goes down,
That is, when the clock signal from the main computer 11 'is interrupted, the switch 14 sends a non-selection signal from the selection signal terminal 14b to all the main computers 11, 11' and 11 ". At the same time, from the selection signal terminal 14d. The selection signal is sent to the auxiliary computer 12, and the communication terminal group 14e is set to the communication terminal group 14
Connect with a '. Based on this selection signal, the auxiliary computer 12 recognizes that the main computer 11 'has gone down, and actually executes the processing that has been virtually performed up to that point.
Inherit the process of 1 '. In this way, the auxiliary computer 12
Corresponds to the down of the main computer 11, 11 ′ or 11 ″.

(Fourth Embodiment) Next, a fourth embodiment of the fault-tolerant computer system according to the present invention will be described. FIG. 4 is a block diagram of the fault tolerant computer system C. The fault-tolerant computer system C has a plurality of main computers 11, 11 'and 1 that execute different processing contents, as in the fault-tolerant computer system according to the third embodiment.
However, in the fourth embodiment, the processing targets 5c, 5d, and 5e in which the main computers 11, 11 ′, and 11 ″ execute different processing contents are
One-to-one via communication elements 100c, 100d, 100e
The third embodiment is different from the third embodiment in that it is provided corresponding to. The main computers 11, 11 ′ and 11 ″ can process the corresponding processing targets 5 c, 5 d and 5 e simultaneously, and the entire system can process the processing targets 5 c.
It is possible to process c, 5d and 5e without stopping and continuously. In addition, the switching device 14 is the main computer 11
And the processing target 5c, the main computer 11 ′ and the processing target 5d, and the main computer 11 ″ and the processing target 5e are connected to each other and driven independently of each other. Main computer 11, 11 'and 11 "
And the respective processing targets 5c, 5d, and 5e, or in order to realize connection switching, switching units 14-1, 14-2,
It is divided into 14-3.

It should be noted that the main computers 11, 11 'and 11 "do not process the corresponding processing targets 5c, 5d, 5e at the same time, but any of the main computers and their corresponding processing targets are paired. In this case, a time-sharing operation may be performed using a time-sharing method, or a master computer or a slave computer may be used between the main computers 11, 11 ', and 11 ". There is a method such as setting up a master-slave relationship and letting the entire system be processed and operated under the control of the master computer. That is, the fault tolerant computer system C includes a plurality of main computers 11, 11 ', 11 ", one auxiliary computer 12, and communication elements 100c, 100d, 100e connected to the respective processing targets 5c, 5d, 5e.
And a switching device 14. The switch 14 is the communication element 10
0c, 100d, or 100e, the corresponding main computer 11, 11 ', 11 "or auxiliary computer 1
The connection is switched between any one of the two.

As described above, the main computers 11, 11 'and 11 "are computers that execute different processes, and operate independently of each other according to the processing contents of the processing objects 5c, 5d, 5e. In addition, the processing operations can be performed simultaneously.
1, 14-2 and 14-3 to the instruction signal terminals 11c, 11c 'and 11c "of the main computers 11, 11' and 11".
It is recognized by the operation instruction signal sent to. Also,
Each of the main computers 11, 11 'and 11 "is the main computer 1 in the fault tolerant computer system A'described above.
The data output terminal 1a, the data input terminal 1b, the TMS output terminal 1c, and the TCK output terminal 1d are provided with communication terminal groups 11a, 11a ′, and 11a ″, and the communication terminal groups 11a, 11a ′, and 11a ″ have been described above. Communication processing similar to that of the main computer 1 is performed. Although three main computers are provided in this embodiment, the number of main computers is not limited to this.

The auxiliary computer 12 is the main computer 11, 11 '.
And 11 "can execute the respective processing contents executed by the auxiliary computer 12 and the auxiliary computer 12 sends a signal for identifying which main computer 11, 11 'or 11" is actually executing the process. The identification signal terminal 12b for receiving from the switching device 14 is provided, and the main computer 11, 11 'and / or 11 "that is actually executing the process is identified by the identification signal sent from the switching device 14. To do.
From the switch 14, the main computers 11, 11 'and / or
Depending on the status of processing execution of 11 ″, an identification signal based on the truth table shown in FIG.
Identifies the execution or non-execution of the processing of the main computers 11, 11 ′ and / or 11 ″ by this identification signal. In FIG. 6, “1” in the truth table indicates the processing execution,
“0” indicates that processing is not executed. That is, for example, the switch 1
When the identification signal of "110" is output from No. 4, this corresponds to the data on the second row in the truth table, and the main computer 11
And 11 ′ are executing processing, and the main computer 11 ″ is not executing processing. Further, the auxiliary computer 12 indicates that the main computer 11, 11 ′ or 11 ″ which is executing processing (a plurality of main computers It is equipped with a notification signal terminal 12d for receiving a notification signal of whether or not one of them is down when processing is being executed, and the main computer 11, 11 'or 11 "is normal. The notification signal is, for example, the main computer 11 or 1.
A unique code number is assigned to each of 1'or 11 ", and when there is a down main computer, the switch computer 14 outputs the code number.
To notify.

On the other hand, the auxiliary computer 12 has the data output terminal 2a, the data input terminal 2b, and the TMS of the auxiliary computer 2 in the fault-tolerant computer system A'described above.
The communication terminal group 12a corresponding to the output terminal 2c and the TCK output terminal 2d is provided, and the same communication processing as that of the auxiliary computer 2 described above is performed through this communication terminal group. Each main computer 11,
11 ′ and 11 ″ and the auxiliary computer 12 share the corresponding dual port rams 13, 13 ′ and 13 ″ , respectively. This dual port ram 13, 13 'and 1
For 3 ", the corresponding main computers 11, 11 'and 1
1 "and the auxiliary computer 12 can freely write and read data, and can share these data with each other. Therefore, the auxiliary computer 12 is actually executing any process. Or multiple main computers 1
When 1, 11 ′ or 11 ″ is operating normally, any one or more of the corresponding dual port rams 13,
Based on the data written in 13 'or 13 ", any one or more of the main computers 11, 11' or 1 above
It becomes possible to understand the processing that 1 "is executing and virtually execute the same processing. In other words, when a plurality of main computers 11, 11 'or 11" are executing processing concurrently. The virtual execution of the same processing in the auxiliary computer 12 is performed by a plurality of dual port RAMs 13, 13 'or 1
Frequent and simultaneous execution may occur between 3 ", but in the case of, it can be dealt with by time sharing or temporary save processing using a buffer memory. Also, a plurality of main computers 11, 11 When'or 11 'are being processed concurrently, a plurality of main computers 1
When 1, 11 'or 11 "are down at the same time, a priority order is set in advance among a plurality of main computers 11, 11' or 11" according to a predetermined condition (eg importance), and this priority order is set. Accordingly, a notification signal can be sent to the auxiliary computer 12 so that the processing can be inherited.

The switch 14 is connected to the communication terminal groups 11a, 11a 'and 11a "of the main computers 11, 11' and 11", and has the input terminal 4a of the switch 4 of the fault tolerant computer system A'described above. Output terminal 4b, TM
The communication terminal groups 14a-1, 14a-2 and 14a-3 corresponding to the S input terminal 4c and the TCK input terminal 4d, and the communication terminal group 12a of the auxiliary computer 12 are connected to the input terminal 4a 'of the switcher 4, An output terminal 4b ', a TMS input terminal 4c', and a communication terminal group 14a 'corresponding to the TCK input terminal 4d' are provided. Further, the switching device 14 includes an output terminal 4e, an input terminal 4f, and a TMS output terminal 4g of the switching device 4 of the fault tolerant computer system A'described above.
And a communication terminal group 14e- corresponding to the TCK output terminal 4h.
1, 14e-2 and 14e-3. The communication terminal groups 14a-1, 14a-2, and 14a-3 are described above.
Are provided corresponding to the switching units 14-1, 14-2 and 14-3 of the switching device 14, respectively. The communication terminal groups 14e-1, 14e-2 and 14e-3 are also the switching units 14-1, 14-2 and 14-3 of the switch 14, respectively.
It is provided corresponding to. The switching device 14 having such a communication terminal configuration has communication terminal groups 14e-1 and 14e-2 depending on the presence or absence of a clock signal transmitted from any one of the selected main computers 11, 11 'or 11 ".
14e-3 and communication terminal groups 14a corresponding to these,
14a 'for connection with the main computer to execute processing, or communication terminal groups 14e-1, 14e-2 and 14
It has a function as a PLD that realizes any one of connecting any one of e-3 and 14a 'to execute processing on the auxiliary computer 12 side.

Further, the switching unit 14 has instruction signal terminals 14b-1, 14b-2 and 14b-3 for transmitting an operation instruction signal to the main computer 11, 11 'or 11 "to be processed.
And an identification signal terminal 14c for sending an identification signal for identifying which main computer 11, 11 'or 11 "is executing the process to the auxiliary computer 12, and the main computer 11, 11' executing the process. Or a notification signal terminal 1 for sending a notification signal for notifying whether or not 11 "is down
4f. The instruction signal terminals 14b-1 and 14
b-2 and 14b-3 are provided corresponding to the switching units 14-1, 14-2 and 14-3 of the switch 14, respectively. The switch 14 and the communication elements 100c, 100
The relationship between d and 100e is the same as in the case of the fault-tolerant computer system A'described above, and thus the description thereof is omitted.

Next, the operation of the fault tolerant computer system C having the above configuration will be described. First,
When any one or more of the main computers 11, 11 'or 11 "is selected and, for example, the main computers 11 and 11' execute processing, the main computers 11 and 11 'have respective instruction signal terminals 11c, 11c. Then, an instruction signal is sent from "to the switch 14 to select the main computers 11 and 11". The switch 14 receives this instruction signal terminal 14b-
1 and 14b-2 from the main computer 11 and 11 "to the operation instruction signal, the instruction signal terminal 14b-3 from the main computer 11"
A non-operation instruction signal is sent to each of them to control their operation. At the same time, the switching device 14 has the identification signal terminal 14c.
Sends an identification signal composed of the corresponding data of the truth table shown in FIG. 6 from the auxiliary computer 12 to the main computers 11 and 11 'to notify that they are performing the processing.

The auxiliary computer 12 receives the identification signal from
In preparation for the down of the main computer 11 or 11 ', the same processing as that of the main computers 11 and 11' is performed virtually while referring to the data of the dual port rams 13 and 13 '. Further, the switching device 14 keeps the communication terminals 14a-1 and 14a as long as the clock signals are sent from the main computers 11 and 11 '.
a-2 and communication terminal groups 14e-1 and 1 corresponding to these
4e-2 is connected, and the main computers 11 and 11 'perform communication processing with the processing targets 5c and 5d via the communication elements 100c and 100d, respectively. The communication processing method is
Since the fault tolerant computer system A'is the same as that described above, description thereof will be omitted.

Here, for example, when the main computer 11 'out of the main computers 11 and 11' goes down, that is, when the clock signal from the main computer 11 'is interrupted, the switching device 14 outputs from the instruction signal terminal 14b-2. A non-operation instruction signal is sent to the main computer 11 '. On the other hand, the switcher 14 continues to output the operation instruction signal from the instruction signal terminal 14b-1 to the main computer 11. At the same time, a notification signal for notifying that the main computer 11 'is down is sent from the notification signal terminal 14f to the auxiliary computer 12, and the communication terminal group 14e-2 is sent to the auxiliary computer 12.
It is connected to the communication terminal group 14a 'on the side. Auxiliary computer 12
Recognizes that the main computer 11 'has gone down by this selection signal, actually executes the process virtually performed for the main computer 11' until then, and inherits the process of the main computer 11 '. In this way, the auxiliary computer 12 corresponds to the down of the main computer 11 or 11 '.

Here, when switching the main computer 11,
For example, in the case where only the main computer 11 'executes processing,
The main computer 11 sends an instruction signal from the instruction signal terminal 11c to the switch 14 so as to select only the main computer 11 '.
In response to this, the switch 14 sends a non-operation instruction signal from the instruction signal terminals 14b-1 and 14b-3 to the main computers 11 and 11 ", and an operation instruction signal from the instruction signal terminal 14b-2 to the main computer 11 '. At the same time, the switching unit 14 sends an identification signal from the identification signal terminal 14c to the auxiliary computer 12 to notify that the main computer 11 'is executing the process. Due to
In preparation for the down of the main computer 11 ', this time, virtually referring to the data of the dual port 13', the same processing as the main computer 11 'is performed.

Further, the switch 14 is provided with the communication terminal group 14a as long as the clock signal is sent from the main computer 11 '.
-2 and the communication terminal group 14e-2 are connected, and the main computer 1
1'communicates with the processing target 5d via the communication element 100d. In this way, the main computer 11, 11 'or 11 "is switched.

On the other hand, if the main computer 11 'goes down,
That is, when the clock signal from the main computer 11 'is cut off, the switch 14 sends out the non-operation instruction signal from the instruction signal terminal 14b-2 to the main computer 11'. At the same time, a notification signal for notifying that the main computer 11 'is down is sent from the notification signal terminal 14f to the auxiliary computer 12, and the communication terminal group 14e is sent.
-2 is connected to the communication terminal group 14a 'on the auxiliary computer 12 side. Based on this selection signal, the auxiliary computer 12 recognizes that the main computer 11 'has gone down, actually executes the process virtually performed up to that point, and inherits the process of the main computer 11'. In this way, the auxiliary computer 12 corresponds to the down of the main computer 11, 11 ′ or 11 ″. In FIG. 4 showing the fourth embodiment, the boundary scan elements 100c, 100d and 100e are respectively of it respectively is one by one connected to a main computer 11, 11 'or 11 "may be connected several double not limited thereto in the present invention. In addition, a plurality of main computers 11, 1
A plurality of boundary scan elements 1 for 1 ′ and 11 ″
00c, 100d, 100e and a plurality of processing targets 5c,
Corresponding to the connection of 5d and 5e, a plurality of auxiliary computers can be provided.

[0067]

As described above, according to the present invention,
The fault-tolerant computer system, the main computer,
The auxiliary computer that virtually executes the same processing as the main computer at normal times, the I / O processing unit that is the processing target, and the I / O processing unit
/ O processing unit and switching means for switching connection between either the main computer or the auxiliary computer, so that the same data processing of the main computer is virtually executed on the auxiliary computer. Then, when the main computer is used, the auxiliary computer can immediately take over. As a result, the process of the main computer 1 is inherited by the auxiliary computer 2 very accurately, and the process can be continued without any interruption.

According to another aspect of the present invention, a fault-tolerant computer system is connected to a main computer, an auxiliary computer that virtually executes the same processing as the main computer at normal times, and a processing target. A communication element, and a switching means for switching the connection between the communication element and any one of the main computer or the auxiliary computer, the communication element is a boundary scan element, the switching means, Based on the presence / absence of a clock signal supplied from the main computer to the communication element, the connection is switched, so in a normal state, the switching means is the main computer and the communication element as a boundary scan element. Are connected, and the processing target can be processed by the main computer. Then, communication of various data between the main computer and the processing target is performed via the communication element which is a boundary scan element, during which,
The auxiliary computer can virtually execute the same process as the process performed by the main computer for the process target.

Under this circumstance, if the main computer goes down and the auxiliary computer executes the process of the main computer, the inheritance of this process is performed very accurately without interrupting the process of the process target. You can continue.
Further, the determination as to whether or not the main computer 1 is down is performed based on the presence or absence of the clock signal output from the main computer as described above, and the recognition process between the main computer and the auxiliary computer is unnecessary. Therefore, the effect of avoiding the delay in processing can be obtained.

[Brief description of drawings]

FIG. 1 is a block diagram of a fault tolerant computer system A according to a first embodiment of the present invention.

FIG. 2 is a block diagram of a fault tolerant computer system A ′ according to a second embodiment of the present invention.

FIG. 3 is a block diagram of a fault tolerant computer system B according to a third embodiment of the present invention.

FIG. 4 is a block diagram of a fault tolerant computer system C according to a fourth embodiment of the present invention.

5A is a block diagram of the boundary scan element 100. FIG. (B) A block diagram of another boundary scan element 100 ′. (C) A block diagram of still another boundary scan element 100 ″. (D) A block diagram of another aspect of the boundary scan element 100.

FIG. 6 is a diagram showing a truth table for generating an identification signal in the fourth embodiment of the present invention.

[Explanation of symbols]

1 Main computer 1a, 2a Data output terminal 1b, 2b Data input terminal 1c, 2c TMS output terminal 1d, 2d TCK output terminal 2 Auxiliary computer 3 Dual port RAM 4 Switching device 5a, 5b Processing target 100a, 100b Communication element (boundary scan element)

Claims (5)

(57) [Claims] 1. A plurality of main computers capable of executing processing concurrently, a single auxiliary computer that virtually executes the same processing as the main computer during normal operation, and each of the main computers or connected to the auxiliary computer. A plurality of processing targets for transmitting and receiving data, and an arbitrary one of the plurality of processing targets and an arbitrary one of the plurality of main computers, while the processing target is either the main computer or the auxiliary computer. A switching means for switching connection between the main computer and the auxiliary computer, and the main computer and the auxiliary computer are arranged between the main computer and the auxiliary computer to write and read data. The storage unit shares data between computers, and the switching unit transmits data between the main computers and processing targets. , Having a plurality of switching units that realize any connection to be executed simultaneously, the switching means, the data transmission and reception between the plurality of main computers and a plurality of processing targets are executed simultaneously, In addition, it outputs the identification signal of the main computer that actually performs the transmission / reception operation to the auxiliary computer, and the auxiliary computer virtually executes the same processing as the plurality of main computers that simultaneously perform the data transmission / reception operation. In this case, the same processing is virtually executed by time sharing, and the auxiliary computer executes the virtual processing when at least one of the main computers that actually perform transmission / reception stops. A fault-tolerant computer system, wherein the processing that has been performed is transferred to actual processing while referring to the data recorded in the storage means. 2. The auxiliary computer, when virtually executing the same process as a plurality of main computers that simultaneously perform data transmission / reception, performs the virtual process of the same process by time sharing. The fault tolerant computer system according to claim 1, wherein the fault tolerant computer system is performed by a temporary saving process using a buffer memory instead of. 3. A plurality of main computers capable of executing processes concurrently, a single auxiliary computer that virtually executes the same process as the main computer at normal time, and each of the main computers or connected to the auxiliary computer. A plurality of processing objects for transmitting and receiving data, and a set of a plurality of communication elements that are connected to each of these plurality of processing objects and are connected to the plurality of main computers or auxiliary computers to transmit and receive data, Switching means for connecting any one of the plurality of communication elements and any one of the plurality of main computers, and for switching the connection between the communication element and one of the main computer or the auxiliary computer And the main computer and the auxiliary computer are arranged between each of the main computers and the auxiliary computer to write and read data to and from the main computer. Comprising a plurality of storage means that data is shared between a computer and an auxiliary computer, the communication element is a boundary scan element, the switching means, between the plurality of main computers and a plurality of communication elements It has a plurality of switching units that realize arbitrary connections so that data transmission and reception can be performed simultaneously, and the switching unit can simultaneously transmit and receive data between the plurality of main computers and the plurality of communication elements. The communication is performed based on the presence or absence of a clock signal supplied to the communication element from the main computer that performs the transmission / reception operation and outputs the identification signal of the main computer that actually performs the transmission / reception operation to the auxiliary computer. Switching the connection between the element and one of the main computer or the auxiliary computer, the auxiliary computer is the same as a plurality of main computers that simultaneously perform data transmission / reception operations. When virtually executing the processing, the virtual execution of the same processing is performed by time sharing, and when the auxiliary computer stops at least one of the main computers that actually perform transmission / reception operation. To
A fault tolerant computer system, wherein the virtually executed process is transferred to an actual process while referring to the data recorded in the storage means. 4. The auxiliary computer, when virtually executing the same processing as a plurality of main computers that simultaneously perform data transmission / reception operations, performs the virtual execution of the same processing by time sharing. The fault-tolerant computer system according to claim 3 , wherein the fault-tolerant computer system is performed by a temporary saving process using a buffer memory instead of. 5. A plurality of auxiliary computers are provided for a plurality of main computers, a plurality of processing targets, and a set of a plurality of communication elements connected to each of the plurality of processing targets. The fault tolerant computer system according to claim 3.