JPH03224332A - Node equipment for indefinite form communication network - Google Patents

Abstract

PURPOSE:To release a fault between two node equipments automatically by locating an input/output means not receiving an active signal and a collision signal in excess of a prescribed time limit of a timer means monitoring an input means as a faulty means and outputting a specific signal to a relevant node equipment. CONSTITUTION:When a reply signal from a called terminal equipment does not reach other input channel of a node equipment 10 before lapse of a prescribed period after an input signal comes to an input channel of the node equipment 10 in which a monitor circuit 400 is loaded to a start control section 60, the monitor circuit 400 generates an ACK signal simulatingly to a sequence control section 90. A switching gate section 40 separates the input channel from other channels. Thus, a fault state in which the entire network is monopolied due to the consecutive broadcast is avoided.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to the O81 layer, network, and network control layer of a local area network, and particularly relates to a node device of an amorphous communication network.

Conventional technology Local area networks (LANs) and public line networks, which are especially applicable to multimedia communications, are known as communication networks that are particularly applicable to multimedia communications. A communication network has been proposed.

This takes the form of a communication network in which communication control elements for a large number of car output signals are connected as nodes in a multi-connection structure to form a communication network, and each node transfers digital signals on a first-come, first-served basis.

This grid-like communication network is especially excellent in the following points.

One is that the multi-connected structure provides a high degree of freedom in network topology. Therefore, fault tolerance
(Survivability) is high, that is, even if there is a failure in part of the network, communication is adaptively secured through other routes. Next, the best communication path is selected by first-come, first-served logic.

Furthermore, this system uses a multi-channel method in which multiple connection channels are established at the same time in a node to efficiently establish full-duplex communication. Such a grid communication network is effectively applied from the physical layer to the network layer of, for example, 03I (Open Systems Interconnection).

When fault tolerance is important in a grid communication network, it is important to reduce the influence of failures and to quickly detect failure locations. There are three types of serious obstacles. The first is a failure in the node itself, the second is a failure in the transmission transmission path, and the third is a failure in the reception transmission path. In the above-mentioned grid communication network, these first. The probability that communication will be inhibited by the second and third failures is very small.

In this way, if a node or terminal beyond a port exiting from a certain node is at fault, this can be detected.

However, if a broadcast is made to an address that does not exist on the network, or if a terminal or node makes a call, the broadcast state continues for all nodes in the network, thereby inhibiting new communications. In order to solve this problem, Japanese Patent Application No. 125811/1982 provides monitoring means for monitoring the broadcast time in the node equipment, detects the oscillation port where the broadcast is input for more than a predetermined period, and It is disconnected from the port and made into a dormant port to minimize failures.

Problems to be Solved by the Invention In conventional node devices, a down port that is determined to be down is in a dormant state, so there is a drawback that the port cannot be restored to a normal port unless the node device is initialized.

Means for Solving the Problem At least one output means to which each transmission line in a transmission line to a terminal or other node device is connected, and each reception line in the transmission line corresponding to the transmission line is connected. at least one input means, a connection means for connecting these input means and the output means, and a control means for controlling the connection means and selectively connecting the input means to the output means; The means includes a first-come-first-served input detecting means connected to the input means and identifying a first-come-first-served input means to which an outgoing signal has arrived first among these input means, and a time limit of a predetermined period from the identification of the first-come-first-served input detecting means. a time limit means for starting, and a monitoring means for monitoring the reception state of the return signal from the reception line of the input means according to the time limit set by the time limit means;
The signal output means outputs a predetermined active signal from the output means corresponding to the first input means, and the fault storage means stores a fault between the input means and the output means, and the control means controlling a connecting means to connect an input means in an idle state to at least all output means except for an output means corresponding to the input means with respect to a transmission path that is not included in the already set communication among the input means; In response to the identification of the first-come-first-served input detection means, the control means controls the connection means, and disconnects all input means except the first-come-first-served input means from the corresponding output means, and from the first-come-first-served input means. The outgoing signal is transferred to all output means except for the output means corresponding to this first-come-first-served input means, and the incoming signal is sent from the receiving line to the input means corresponding to the output means that transferred the outgoing signal among the input means. the input means that has received the return signal is connected to the output means corresponding to the first-come-first-served input means, and the first-come-first-served input means is connected to the output means that corresponds to the input means that has received the return signal. to fix the connection between these input/output means, and connect all other input means to all output means except at least the output means corresponding to this input means, and connect the output means to which the outgoing signal is transferred. If neither the active signal nor the collision signal is input to the corresponding input means for a predetermined period determined by the time limit means, the input/output means is regarded as a fault and is stored in the fault storage means by the control means, and the connection means is When a specific signal is input to the fault input means, the control means releases the memory of the fault storage means and controls the connection means to disconnect the fault input means from the other input means. The connection of the input/output means that was present is automatically restored to normal.

At this time, a collation means is provided in the control means to collate a specific signal and a signal inputted by the fault input means, and the specific signal is sent from the fault output means, and when the collation means detects a match. The connection of the failed input/output means was restored to normal.

Further, when the incoming signal is not returned to the other input means and the outgoing signal is received beyond the predetermined time limit, the signal of the input means separated by the other input means is outputted corresponding to this input means. An input/output return means for outputting to the means is provided.

On the other hand, an oscillation port detection means having an oscillation port detection timer and a clock frequency switching means are provided, and a port receiving a specific signal is determined to be an oscillation port, and the clock frequency of the oscillation port detection timer is immediately set to a higher one. I tried to change it.

Here, the specific signal is a down signal sent from the down port.

Further, the clock frequency switching means is configured to change the clock frequency of the oscillation port detection timer so as to detect the oscillation port as an oscillation port and return the input signal before receiving a return signal.

On the other hand, an abnormality detection means having an oscillation port detection timer is provided, and when a collision signal cannot be detected at the first input port within the collision signal detection time, this signal is determined to be abnormal and is immediately treated as an oscillation port. .

In addition, a clock frequency switching means is provided, and when a collision signal cannot be detected within the collision signal detection time at the first input port, this signal is determined to be abnormal, and the clock frequency of the oscillation port detection timer is immediately changed to a higher one. It is now treated as an oscillation port.

Here, the clock frequency switching means is configured to change the tarok frequency of the oscillation port detection timer so as to detect the oscillation port as an oscillation port and return the input signal before receiving a return signal.

According to the invention according to claims 1 to 3, the node device includes:
Detects the first input means to which the outgoing signal has arrived, transfers the outgoing signal by the output means, and then blocks the input/output means from which active signals and collision signals are not input beyond a predetermined time limit of the time limit means that monitors the input means. A specific signal is stored in the fault storage means, and a specific signal is sent to the corresponding node device by the fault output means. If the fault is removed, the corresponding node device will receive the signal from the input means for a predetermined period of time.
This input means is used as an oscillation port, and the received signal of the input means is returned to the output means and sent to the node device serving as the oscillation source. This node device compares the specific signal outputted by the output means and the input signal using the collation means, and in this case, since they match, the fault is removed, the signal transmission is stopped, and the node of the twin j The device returns to normal. With this signal control method, a fault between two node devices can be automatically cleared.

In addition, according to the invention described in claims 4 to 7, when performing 1i) of automatically restoring a down port, when a normal port receives a down signal and returns it as an oscillation port, it is possible to broadcast for a long time. It is possible to prevent this and improve the efficiency of the network, and since oscillation detection is completed before ACK reception is possible, it is also possible to prevent malfunctions such as fixing a wrong path as a down port due to noise or the like.

Further, according to the invention described in claims 7 to 9, even if an abnormal signal such as noise is input, the erroneous path can be fixed and continuous broadcasting can be performed by treating it as an oscillation port before the erroneous path is fixed. can be prevented.

Embodiment An embodiment of the invention recited in claims 1 to 3 will be explained based on FIGS. 1 to 10.

When the node device according to the present invention detects a down port to which a predetermined signal is not returned, it closes this port and sends out a down signal. If there is a first-arrival input port (oscillation port) that receives a specific signal or broadcast signal for a predetermined period of time or more, it is determined to be abnormal, and this oscillation port is disconnected as a down port and the input signal is returned to the transmission source node device. When this node device receives a signal identical to the specific signal, it stops transmitting the signal and returns to normal.

The receiving side node device also returns to normal because the abnormal signal stops.

The amorphous communication network to which the node device according to the present invention is applied is advantageously realized as a lattice communication network in which the node devices e10 are connected in a two-dimensional or three-dimensional lattice shape by transmission paths 12, as illustrated in FIG. The network structure is essentially amorphous. For example, other network configurations such as linear or loop configurations may be used.

The node device 10 is provided with a plurality of human output ports, eight in this example, to which other node devices 10 and/or terminals 14 can be connected via transmission paths 12. There is no limit to the number of input/output ports, as long as there is at least one. The node device 10 connects to the node device 1 via the transmission path 12 if it is within the capacity of the input/output port.
There is no limit to the number of terminals 0 or 14. Further, the mesh body may be formed by a single node device 10, or a plurality of node devices 10 may be mounted, for example, on a single printed wiring board, so that the whole is treated as one node device, and the Additional input/output port capacity may be increased.

In this embodiment, the terminal 14 is a terminal device capable of asynchronously transmitting and receiving data, and includes a processing system such as a personal computer, a service station such as a file station and a print station, and the like. Advantageously, the data is transferred in the form of message packets. As will be described later, if the terminal 14 is a full-duplex terminal, it is advantageous to use a system that immediately sends out a response signal upon receiving a packet addressed to itself.

The transmission path 12 is, for example, an optical transmission path using an optical fiber, or an electrical transmission path such as a twisted wire or coaxial cable, and in this embodiment, data is transmitted in analog or digital format. It also has a full-duplex configuration. Node device 10 and terminal 1
The transmission line 12 between the four channels may have a half-duplex configuration. Further, a plurality of transmission lines 12 between the node devices 10 may be provided depending on the traffic.

Referring to FIG. 1, the node device 10 includes an input port (input means) 20 to which a reception line from the transmission line 12 is connected, and an output port (output means) 30 to which a transmission line to the transmission line 12 is connected. Both are connected to each other via a switching gate section (connection means) 40. The input port 20 in this embodiment has eight receiving or input channels 1
0 to 17, and the output port 30 has correspondingly eight transmission or output channels oO to o7. As a result, up to eight other node devices 10 and terminals 14 in total can be connected to the node device 10 via the transmission path 12. Among the output channels 00 to o7, output channels having the same number, that is, "corresponding" to each of the input channels 10 to 17, are connected to the transmission line 12 of the same route.

The switching gate section 40 has input channels 1O-i7
any of the output channels. This is a gate circuit that selectively interconnects any one of O to 07. The input port 20 is also connected to a start control section 60 and an end control section 70 via a control gate section 5o.

The control gate section 50 sends a signal from the input port 20 to the start control section 60, and sends a control signal from the start control section 60, fault storage section (fault storage means) 210, and termination control section 70 to the switching gate section 40 and termination control section. This is a gate circuit that is appropriately connected to and controlled by the section 70. The start control unit 60 is a functional unit that identifies the input channel to which the input signal arrived first and also detects whether or not there is an input signal in each input channel. The termination control unit 70 is a circuit that detects that there is no longer an input signal in an input channel of an already set communication route and performs a process of terminating the communication. The switching gate section 40, the start control section 60, and the end control section 70 are
They are interconnected by a gate set bus 80.

The switching gate section 40 is also connected to an active signal output section (signal output means) 200, which is also connected to the start control section 60.

A failure storage unit 210 is also connected to the start control unit 60 and the termination control unit 70 for storing a channel in which a failure has occurred. The fault storage section 210 is also connected to the gate set bus 80.

Switching gate section 40, control gate section 5o, start control section 60, end control section 70, active signal output section 20
0 and the fault storage unit 210 are controlled by a sequence control unit 90 that controls the entire device including them.

The active signal output unit (AFU 200) includes circuits O to 3 corresponding to ports of channels 0 to 3, as shown in FIG.
has. The AND gate 211 receives a signal S A U representing the first input port of port O from the start control unit 60 .
(And when inputting H of the active signal sending signal MODEO, the pulse ENA is sent from the sequence control unit 90.
When BLEO is input, this pulse is outputted, outputted from the OR gate 212 to the switching gate section 40, and outputted from the output port O as an active signal.

When the NAND gate 214 is the first input port described above, when the output signal SAU and the H signal N0ACK output when the port O is an oscillation port are input, the input SB of the flip-flop (FF) 216 is input to the NAND gate 214. Output L,
Set FF216. FF216 is 1 from output Q
-( is output to the AND gate 218. The gate 218 outputs the input signal I -PO of the oscillation port O of the input port 20.
RT is output to the OR gate 212, and this OR gate 21
2 to the switching gate unit (SGU) 40, the signal is looped back to the output port O, and is sent to the oscillation source node device. When the transmission signal stops, the termination control section 7
The signal EAtJ from 0 becomes L, and the signals SAU and N
0ACK becomes L. Therefore, inverter 220 and N
AND gate 214 outputs H to NAND gate 222. The FF 216 is reset by inputting L from the key 1-222 to the input RB. When port O goes down, the signal DMU from the fault storage unit 210 becomes L, and the AND gate 226 outputs the specific signal CLKD of the signal generator 270 to the OR gate 212 via the inverter 224.
Output to. The signal CL K D is sent out from the output port O through the switching gate section 40 from the same gate.

The specific configuration of the start control unit 360 is as shown in FIG. 3 in the case of a four-channel input/output pot for simplicity. means) consisting of 60b.

The first input signal detection unit 60a detects input channels 10 to 13.
This is a functional unit that identifies the channel on which the input signal arrives first according to first-come, first-served logic. This corresponds to the number of input channels, that is, four flip-flops 62 and -
Group NAND gate 66 and 4-person NAND gate 68
and inverter 61, and four three-man powered NAND gates 6
3, a path buffer 65, and a mode changeover switch 67
are connected as shown in the figure.

The flip-flop 62 is a circuit that holds the state of the input channel through which the input signal has arrived. A group of NAND gates 66 provides priority between the outputs 64 of the flip-flops 62. The four-man power NAND gate 68 and the inverter 61 have a holding function that responds to the arrival of an input signal to any of the flip-flops 62 by setting the S terminals of all the flip-flops 62 to a low level and fixing their states. This is a circuit for notifying the sequence control unit 90 that the first outgoing signal has arrived.

The three-man power NAND gate 63 is the - group NAND gate 6
6 and the output of the input signal detection section 60b, and the logical sum output is outputted to the gate set bus 80 via the pass buffer 65. Note that the mode changeover switch 67 is always connected in this embodiment.

The input signal detection section 60b is a circuit that detects whether an input signal has arrived at the input port 20.

This consists of flip-flops 69 and 120 and four N
An AND gate 122 and a five-man OR gate 124 are connected as shown. Flip-flop 69 is a two-state circuit for holding the state of the input channel through which the input signal has arrived. flip flop 120
is a circuit for storing the output states of the flip-flops 69 and fixing the states by setting their S terminals to a low level. The NAND gate 122 is a gate circuit that controls the connection of the output of the flip-flop 69 to the first-arrival input detection section 60a. The OR gate 124 is a circuit for calculating the logical sum of the outputs of the flip-flop 69 and notifying the sequence control unit 90 that the first demodulated signal has arrived.

The start control section 60 has a monitoring circuit 400 shown in FIG. This is basically the NA of the first input signal detection section 60a.
It is connected between the signal line 402 of the output 5TART from the ND gate 68 and the remaining input 404 of the 5-way OR gate 124. If a response signal from a receiving terminal does not arrive within a predetermined period after an input signal arrives from the input channel of the node device 10 in which the monitoring circuit 400 is installed, the monitoring circuit 400 controls the input/output signal to the switching gate section 40. This is a circuit for separating a channel from other channels. This may be installed in all node devices 10 in the system, or may be selectively installed in specific node devices 10.

More specifically, as described below, input channels i0 to i17
When an input signal arrives first from a specific channel, it is detected by the first input signal detection section 60a, and a signal 5TART indicating this is outputted from the signal line 402 to the sequence control section 90. In addition, in the first-arrival input signal detection section 60a, which input channel the input signal arrived first is determined from the NAND gate 63 as signals 0 to 3 to the active signal output section 200 and the control gate section 50, and from the bus buffer 65 to the bus 80. Each is output to NAN
Channels corresponding to high level outputs O to 3 from the D gate 63 are connected to the switching gate section 50.
A link is set up, and other channels are disconnected from this.

For example, if only one of the outputs O to 3 from the NAND gate 63 is at a high level, the switching gate section 40 receives the control pulse WR from the sequence control section 90.
Even if it receives ITE O, the only channel corresponding to that high level will not be able to establish a link with any channel and will be isolated from other channels. In this embodiment, this is utilized, and the monitoring circuit 400 monitors the output signal 5TART of the NAND gate 68, and if the high level continues for a predetermined period, it forcibly sets an ACK sequence.

In response to this pseudo ACK signal, the switching gate section 4
0, only one output from the NAND gate 63 remains at a high level and is linked to the switching gate section 40. Therefore, that input channel is isolated from the others. This avoids abnormal conditions such as unnecessary network transmissions, continuation of broadcasts, and inability to release set links. The predetermined period will be detailed later.

The fault storage unit (DMU) 210 has circuits 0 to 3 corresponding to ports of channels O to 3, as shown in FIG. 5(a).

For example, in the case of port ○, pulse WRITE 2 from the sequence control unit 90 is a flip-flop (DFF).
242, if the signal from the start control unit (SAU) 60 input to the OR gate 240 is L, port O is not receiving an active signal or a collision signal. If the bus fixing signal EAU and node device startup signal 800T from the termination control unit 7o are received, the port o
is not integrated into the bus fixation.

In the above case, the DFF 242 inputs the OR gate 24
0 output is input, it is in the reset state, and the output Q is L
It is in. Output Q representing port 0 down is output to the monitor display (MON) and OR gate 246. gate 2
The switch input to 46 is in the closed state, and gate 2
46 outputs L to the control gate unit (CCU) 50. The output of the gate 246 is input to the inverter 248, and when the pulse ENABLE2 is input, the inverter 248 inverts the down port 0 to the gate set bus (GSB) 80.
Output at level. Port 0 is excluded from the GSB bus configuration. The verification circuit (verification means) 250 connects DF to the input EN.
It is enabled by inputting H of the output QB of F242.

Down port 0 is disconnected from other ports and outputs a specific signal CLKD from its output port.

If the signal continues for longer than the time limit set by the monitoring circuit (time limit means) 400, other node devices that have received the signal CLKD disconnect the port and return the received signal. Port O receives the return signal. In the verification circuit 250, the port O is connected to the input 11.
The received signal CLKD is input to the input 12 from the signal generator 270 shown in FIG. 5(b). Verification circuit 25
0 outputs H to the output CRY when a match is detected.

This is output from the OR gate 252 and input as a signal DEND to the OR gate 100 of the sequence control unit (SCU). Also, since the inverter 244 inputs input to the input RB of the DFF 242 and resets it, the port O
The problem caused by this will be removed.

The sequence control unit 90 includes a gate group for generating control signals necessary for controlling the present device, and a circuit for giving priority to the termination of communication when there is a conflict between the occurrence and termination of communication. The “active detection time constant” and “input signal detection time constant” are formed by the sequence control unit 90. The sequence control unit 90 also does not require any change in the hardware of the device itself in the case of full-duplex communication and in the case of including both full-duplex communication and half-duplex communication.

Referring to FIG. 4, a monitoring circuit 400 of the start control section 60
In this example, a timer circuit 406 that monitors the continuation of the high level state of the signal 5TART and outputs a pseudo ACK signal is connected as shown.

In this embodiment, the output channel of input port 20 is normally at a low level in an idle state with no signal. The timer circuit 406 starts counting the clock CLOCKI in response to the high level of the signal 5TART. If the high level continues for the predetermined period of time, ie, the broadcast condition continues, the timer circuit 406 sets its output QH to the high level. This is active signal sending section 2
00 as the N0ACK signal, and also gate 1
24 to the sequence control unit 90 as a pseudo ACK signal. In response to this, the switching cable
The 1.1 section 40 sets a link to the switching gate section 40 with the output from the NAND gate 63 at a high level by one, and isolates its input/output channel from the others. The input/output channels separated in this way are released and return to the idle state when the input signal disappears. This prevents abnormal conditions such as continuation of broadcasting. In addition,
By changing the frequency of the clock CLOCKI,
The time constant of the timer circuit 406, ie, the monitoring time required to disconnect an input channel that is in a continuous broadcast state, can be changed.

The aforementioned predetermined period is a period during which input of the first demodulated signal is guaranteed. In the case of full-duplex communication, number 1 ["
The period starts from the detection of the first input channel of the outgoing signal, and its length is determined by the round-trip propagation delay time over the maximum network length and the period from when the receiving terminal starts receiving the first outgoing signal to the time when the first outgoing signal is received. This is set substantially equal to the sum of the time required to start transmitting the return signal. If full-duplex communication and half-duplex communication are included, the period starts from the end of the first outgoing signal. Its length is the sum of the propagation delay time for round trip through the maximum network length and the time required from when the receiving terminal finishes receiving the first outgoing signal until it starts transmitting the first incoming signal. are set substantially equal.

Usually, some extra time is added to these.

The circuit example of the monitoring circuit 400 shown in FIG. 4 releases the disconnected channel as soon as the input signal, ie, the first outgoing signal, disappears.

However, the link may be configured not to be released until a reset instruction is given even if the human power signal is lost.

As shown in FIG. 6, the sequence control unit 90 includes five shift registers 91 to 95, a gate group 96 for appropriately combining their output states to generate necessary control signals, and a gate group 96 for controlling the occurrence of communication. A flip-flop 97 and an AND gate 221, a mode changeover switch 98, and a boot switch 99 are connected as shown in the figure to give priority to termination of communication when there is a conflict of termination.

A system clock CKI or CKO is connected to clock input terminals of shift registers 91 to 95. Note that in this embodiment, the flip-flop 95 is not used;
Further, the mode changeover switch 98 is always open. The boot switch 99 is an operation switch that is operated only when starting up the node device 10 and initializes all flip-flops in the node device 10. The sequence control unit 90 also does not require any change in the hardware of the device itself in the case of full-duplex communication and in the case of including both full-duplex communication and half-duplex communication. FIG. 10 shows the operation timing of the sequence control section 90. [Active detection time constant],
The timing of monitoring the input means according to the "input signal detection time constant" is determined by the sequence control l1 section 90.

An outline of communication control in the node device 10 will be explained.

Here, for convenience, the term "transmission terminal" refers to
The term "receiving terminal" refers to the terminal that sends the signal to the transmission line 12, and the term "receiving terminal" refers to the terminal that receives the signal from the transmission path 12. Also,
The term "originating terminal" refers to a terminal that starts transmitting information to a specific terminal from a state where it is not connected to other terminals, that is, an idle state, and "receiving terminal" refers to a terminal that starts transmitting information to a specific terminal from a state where it is not connected to any other terminal, that is, from an idle state. This refers to the destination terminal that returns the response.

A signal sent from a calling terminal is called an "outgoing signal," and a signal sent from a receiving terminal, particularly a signal sent back in response to an outgoing signal, is called a "returning signal."

In a certain node device 10, in an idle state where no connection is established between specific input/output channels, the connection gates of the switching gate unit 40 are in an open state, and all input channels except their corresponding output channels are in an idle state. Connected to all output channels.

When an input signal arrives at any of the input channels io"-i7 in the idle state, the first-arrival input signal detection section 60a of the start control section 60 selects the channel from which the input signal arrived first among the input channels IO"-i7. , that is, the "first-come-first-served input channel" is detected by first-come-first-served logic. A broadcast is performed in which the signal received from the first input channel is transferred to all output channels except the corresponding output channel.

First-come-first-served input signal detection unit 60 of starting side gU jclf 60
The sequence control unit 90 detects the first-arrival human power channel of a.
is activated, and the sequence control unit 90 starts time-limited monitoring using the activity detection time constant.

[Active detection time constant] is when the first outgoing signal from the same transmission source is received from an input channel other than the input channel that detected the input signal first, or when the first outgoing signal from the same transmission source is received. This is the time for receiving another first outgoing signal or for receiving an active signal.

The length of the active detection time constant is set to be substantially equal to the sum of the propagation delay time for reciprocating the maximum allowable distance between adjacent node devices 10 or the opposite terminal 14 and the time required for the active signal. Usually, some extra time is added to this. Within this time, the first diverted source from the same source
The first outgoing signal, another first outgoing signal from another transmission source, and the active signal arrive. This allows the detection of obstructions or body pressure channels.

The input signal detection section 60b of the start control section 60 detects the channel from which the input signal has arrived within the monitoring time limit of the active detection time constant.
is stored in a flip-flop.

When the period defined by the active detection time constant expires, the sequence control unit 90 clocks the failure storage unit 210 and determines whether any input signal has arrived within the period of the active detection time constant among the input channels 1o-i3. The input channel is stored in flip-flop 212 as a failed or dormant channel.

Subsequently, the sequence control section 90 performs time-limited monitoring of the input signal detection time constant. [Input signal detection time constant] is the time for detecting whether or not there is a signal after the period determined by the active detection time constant has elapsed. The length is, for example, 2 bits in the case of Manchester coding, and 7 bits or more in the case of the coding rule of inserting rQJ into 6 consecutive bits of rlJ in NRZI. Usually, some margin time is added to this and the time length is set to twice that amount, ie, 4 bits or 14 bits each. This applies to input channels other than the incoming channel that detected the input signal first, which received the first outgoing signal from the same transmission source or another first outgoing signal from another transmission source. This is the time required to distinguish and detect active signals.

The input signal detection unit 60 of the start control unit 60 selects channels from which an input signal has arrived within the monitoring time limit of this input signal detection time constant.
It is stored in the flip-flop of b. When this period ends, the switching gate section 40 switches the input signal detection section 6
When an input signal subsequently arrives from any of the input channels for which there was no input signal within the period of the input signal detection time constant stored in 0b, that input channel is connected to the output channel corresponding to the first input channel. .

When there is no input signal in any input channel included in the communication path and the time specified by the communication end detection time constant has elapsed, the end control section 7o starts the sequence control section 90.
, the sequence control section 90 resets the first input signal detection section 6a and the input signal detection section 60b of the start control section 60 to the initial state.

This end of communication may be detected by monitoring the input signal from the first-come-first-served input channel, detecting that it has disappeared, and performing recovery processing, or by monitoring the input signal from the first-come-first-served input channel and performing recovery processing. It may also be configured to monitor the input signals from both of the other input channels, detect that one of them is no longer available, and perform recovery processing.

The absence of an input signal is detected by detecting that the logic state of the signal is maintained at a predetermined state, for example rQJ, for a period of a communication end detection time constant.

In the above-described embodiment, the input channels to which no signal has arrived during the period determined by the active detection time constant are stored in the input signal detection unit 60b even after the period has elapsed, and only the input signals of those input channels can be detected. After the same period has elapsed, the input channel on which such an incoming J signal has not arrived is connected to the output channel corresponding to the first input channel, and all other input channels are disconnected from the output channel. It may be configured as follows.

When the first demodulated signal arrives at the input channel to which such an input signal did not arrive after the period of the input signal detection time constant has elapsed, the input channel that received the first demodulated signal is assigned to the first input channel. to the corresponding output channel, and
Connect the first input channel to the output channel corresponding to the input channel from which the first received signal arrived, fix the path between the input and output channels, and connect all other input channels to the output channel corresponding to the input channel. Connect to all output channels except

For understanding this embodiment, FIGS. 7(a) to (e) show a grid-like communication network in which four node devices 10 are connected in a grid-like manner.
The communication procedure in the system of this embodiment will be explained with reference to . In this illustrative communication network, four node devices 10
A to 10d are connected in a grid pattern by transmission lines 12. The node devices 10a and 10d have terminals 14a and 14d.
are connected to each other. In the figure, the hatched side indicates the transmitting side, and the thick line indicates the flow of information signals.

For full-duplex communication, detection of input signals and connection control between input and output channels based on the detection are performed in the following seven basic steps.

First, as shown in FIG. 7(a), in the first step, the originating terminal that wants to transmit data for the first time from an idle state, for example 14a, sends the first outgoing signal in the form of a packet to the node device through the transmission path 12a. Send to lOa. The first outgoing signal includes a destination address indicating the destination terminal, for example 14d. The node device 10a detects the first outgoing signal as a first-arrival input signal. That is, the channel to which the input signal arrived first, ie, the "first-come-first-served input channel" is identified by first-come-first-served logic. Therefore, the first outgoing signal is transferred to all output channels 12ab, 12ac, etc. except for the output channel corresponding to the first input channel 12a. That is, the first outgoing signal is broadcast to all routes of the node device 10a.

Further, when the node device 10a receives the first outgoing signal on the first-come-first-served input channel, it disconnects all other input channels whose communication paths are not fixed to the output channel, and uses the active signal output unit 200 to Active signal 2 from the output channel corresponding to the first input channel
30a is output.

Next, in the second step, as shown in FIG. 7(b),
The other node devices 10b, 10c, and 10d also connect the first
The second outgoing signal is received and a similar broadcast is performed. In this example, the node device 10c recognizes the transmission path 12ac as a first-come, first-served input channel, and broadcasts to other transmission paths such as the transmission path 12cd. Similarly, the node device 10d
In this case, the first outgoing signal arrives from the same transmission path 12cd in addition to the transmission path 12bd, but it recognizes the transmission path 12bd as the first input channel and transmits only the first outgoing signal from the transmission path 12bd. The signal is broadcast to other transmission paths such as paths 12bd and 12cd, and the signal from transmission path 12cd is not output. Node devices 10c, 10. In case d, if the arrival time difference between the first input signal and another input signal that arrives later than the first input signal is shorter than the time required for connection control, an instantaneous overlap occurs. However, since this occurs in the preamble portion of the message packet, there is no problem.

In this way, the first outgoing signal transmitted from terminal 14a and broadcast from notebook 12 is transmitted throughout the network without duplication. In this way, the first outgoing signal via the shortest route reaches the terminal 14d.

The node devices 10a-10d monitor all input channels during the active detection time constant period starting from the detection of the first input channel, and identify the manual channels that have not received input (No. 8) within that period.

They are stored in the human power signal detection section 60b. In each node device 10, if the input channel, output channel, and other node devices and terminals connected to the node device are functioning normally, the active signal or the first signal is detected within the period of the active detection time constant. The outbound signal should arrive. For example, in the node device 10b, the input port 23
4bx does not receive the signal for some reason, this is stored in the node device 10b.

The node devices 10a to 10d detect input channels that have no input signal within a period determined by an input signal detection time constant that starts after the elapse of the active detection time constant. At this time, the active signal has already ended. Further, at this time, the node devices 10a to 10lod may be configured to connect the thus detected input channel to all output channels other than the corresponding output channel. Further, the node devices 10a to 10d select the input channels that are not stored in the input signal detection unit 60b among the input channels detected in this way, that is, the input channels from which a signal has arrived within the period of the active detection time constant, and select them as corresponding input channels. It may be configured to connect to all output channels other than the output channel.

In the third step, the terminal 14 connected to the node devices 10a to 10d receives the first outgoing signal. At this time, each terminal 14 returns the active signal 232 and checks the destination address included in the first outgoing signal with its own address. In this example, the terminal 14d sends out an active signal 232d, and since the destination address matches that of its own station, it sends out the first return signal to the transmission line 12d. As shown in FIG. 7(C), the node device 10d selects one of the input channels corresponding to the output channel that sent the first outgoing signal.
An input channel in which an input signal does not arrive within a period defined by the input signal detection time constant and a signal arrives after the end of the period defined by the input signal detection time constant is identified. Connect this to the output channel corresponding to the first input channel.

In this example, as shown in FIG. 7(C), the node device 1
0d means that when a signal is received from the transmission line 12d after the period determined by the input signal calibration time constant, that signal, that is, the first
The input channel that received the th demodulated signal is connected to the output channel 12bd corresponding to the first input channel. Therefore, the first demodulated signal received from the transmission path 12d is sent from the node device 10d to the transmission path 12bd.

After that, normally, after a time corresponding to the terminal response monitoring time or the communication end detection time constant when full-duplex communication or half-duplex communication has elapsed, all other input channels are outputted corresponding to that input channel. Connect to all output channels except channel. This can prevent the first outgoing signal of the transmission path 12cd in FIG. 10(c) from being detected by the node device 10d. That is, in this example, the transmission line 12bd is thereby interconnected with the transmission line 12cd.

In the fourth step, the node devices 10a and 10b also perform the same control as the node device 10d.

Therefore, as shown in FIG. 7(d), the first incoming signal reaches the originating terminal 14a by retracing the route along which the first outgoing signal was transferred. The first outgoing signal has a certain length, and the terminal device such as the terminal 14d
Since the configuration is such that the first return signal is transmitted immediately after identifying the destination address of the first outbound signal, the first return signal is transmitted.
The second incoming signal is transmitted while overlapping with the first outgoing signal.

Therefore, even if terminals other than terminals 14a and 14d are connected to this network, these terminals cannot participate in this communication. This maintains the confidentiality of communications between other terminals, which is important for a communication system, and also enables multichannel communications.

As shown in FIG. 7(e), in the fifth step, the node device 10c does not receive the first received signal from the transmission path 12cd, etc., and the first received signal that has been received up to that point is sent to the transmission path 12ac. When the forward signal disappears, this is detected and all input channels are connected to all output channels except the output channel corresponding to that input channel. In other words, if it is detected that no input signal is received during the period of the input signal detection time constant, the first inbound signal does not arrive even after that period, and the first outbound signal is no longer received. , connect all input channels to all output channels except the output channel corresponding to that input channel. This means that the route of the communication has been fixed without passing through the node device 1o, or that the communication has not been established and the originating terminal has stopped transmitting the first outgoing signal. Therefore, otherwise,
The arrival of the first returned signal is guaranteed within the terminal response monitoring time starting from the detection of the first input channel.

For some reason, the first outgoing signal is not received by the receiving terminal 14d.
The same is true when the transmitting terminal 14a cancels transmission of the first outgoing signal midway because the first incoming signal is not returned.

If both full-duplex communication and half-duplex communication are included, the node device 10c will no longer receive the first outgoing signal, and will continue to receive the first outgoing signal even after the period determined by the communication end detection time constant has elapsed. When detecting that no signal arrives, all input channels are connected to all output channels except the output channel corresponding to that input channel. In other words, for any input channel that did not receive an input signal, if it is detected that the first incoming signal has not been received within the terminal response monitoring time starting from the end of the first outgoing signal, all input channels Connects a channel to all output channels except the output channel corresponding to its input channel.

Through such connection control, one communication path is set and fixed for communication between the originating terminal 14a and the receiving terminal 14d. Each node device 10 can control the settings of newly occurring communications regarding unfixed routes.

In this way, each node device 10 detects the presence or absence of an input signal and performs sequential control regarding the active detection time constant, input signal detection time constant, terminal response monitoring time, and communication end detection time constant. The same applies to control regarding termination of communication. For example, in full-duplex communication, when one calling terminal is given the authority to continue and end communication, normally the
The first outgoing signal continues at an interval shorter than the signal end detection time, and if the incoming signal is transmitted intermittently, the first outgoing signal disappears for a pair of input channels for which the communication path is fixed. or upon detecting that there is no input signal on any of the pair of input channels, all input channels are connected to all output channels except the output channel corresponding to that input channel. Naturally, at this time, the demodulated signal is not being transmitted.

In the case of half-duplex communication, or when there is no need to set priorities for the transmitting station and receiving station even in full-duplex communication, it is necessary to confirm that there is no input signal on both of a pair of input channels for which the route has been fixed. is detected and all input channels are connected to all output channels except the output channel corresponding to that input channel.

After an input signal arrives at a certain input channel of a certain node device 10 in which the monitoring circuit 400 is installed in the start control unit 60, a response signal from the receiving terminal is transmitted to another node device 10 before the aforementioned predetermined period elapses. If the signal does not arrive at the input channel, the monitoring circuit 400 generates a pseudo ACK signal to the sequence control section 90.

More specifically, when an input signal arrives first from a specific input channel, its outgoing signal is output to all output channels other than the first input channel. The monitoring circuit 400 monitors the signal 5TART output from the AND gate 68 of the first-arrival input signal detection section 60a. For example, when a broadcast state continues due to a call from a terminal or a transmission to a non-existent destination, the monitoring circuit 400 monitors the signal reception state signal 5TART of this input channel, and if this continues for a predetermined period, The monitoring circuit 400 outputs a pseudo ACK signal to the sequence control section 90.

In response to this ACK signal, the sequence control unit 90
A K sequence is entered, and a signal WRITEO is generated to cause the switching gate section 40 to establish a link between the input and output channels. Its operation is similar to the normal operation described above. However, since only one of the signals O to 3 from the path buffer 65 of the start control unit 60 to the bus 80 is at a high level, which corresponds to the first input channel, the switching gate unit 40 cannot connect the link even if it attempts to establish a link. , which would isolate that input channel from other channels. Therefore, an abnormal state in which the completed link is not set up within the communication network and the entire network is occupied due to continued broadcasting is avoided.

The monitoring circuit 400 also includes an active signal output section 200
Outputs the N0ACK signal to. Active signal output section 2
00 stores the first input port as an oscillation port in the flip-flop 216 according to this signal, and returns the signal of this input port to the corresponding oscillation port.

The procedure for handling the above port failure is shown in Figure 8 (a) to (d).
This is explained by: In the broadcast shown in FIG. 8(a), node A broadcasts the first outgoing signal from first-arrival input port P to nodes B and C. Node B detects this at its first input port and sends an active signal to node A. When departures and landings occur simultaneously, the first outgoing signal (collision signal) is sent out. Node A detects the active signal and the collision signal from the signal from node B within the time limit of the above-mentioned active detection time constant and input signal detection time constant, but when no signal arrives, the fault storage unit (DMU)
210 DF with the first arrival Kaponide Q as a down port
Stored in the reset state of F242. After the broadcast in FIG. 8(b), down port Q of node A
sends a specific signal CLKD to node C, for example, a signal with a constant period. When the fault between nodes A and C is removed in FIG. 8(c), port R of node C receives the signal for a predetermined period of time and becomes an oscillation port. Node C disconnects port R from other ports, returns signal CLKD to the output port, and sends it to node B.

Node A receives the same input signal as signal CLKD at port Q, detects a match between the two, and stops sending out signal CLKD. Node C also returns to normal as the signal from oscillation port Q disappears. The above was a case of a connected port or a down port and an oscillation port, but when both ports are down ports in FIG. 8(d), they send down signals to each other.

At each node, the ports are disconnected, a match between the transmitted and received signals is detected, and each port returns to normal.

Next, an embodiment of the invention according to claims 4 to 6 will be described with reference to FIGS. 11 to 14.

The same parts as those shown in the previous embodiment are indicated using the same reference numerals (the same applies to the following embodiments).

In the case of the above-mentioned embodiment, it is possible for the down port to automatically recover, but at this time, the down port sends out a down signal,
The oscillator port must fold back the input signal. However, when a port of a normal node device receives a down signal, it continues to broadcast the down signal, and the oscillation port detection timer operates, causing the port to become an oscillation port. That is, in the above-described embodiment system, since a down signal and a signal used for communication are not distinguished, broadcasting must be continued for a certain period of time or more, resulting in poor network efficiency. Furthermore, if some kind of signal such as noise comes in during this process, a malfunction may occur in which the signal is regarded as an ACK and the path is fixed.

This embodiment further improves the previous embodiment and solves these problems.

First, FIG. 11 shows a control gate section 50 that receives connection control.
The configuration is shown below. This corresponds to the number of channels, i.e.
A group of four OR gates 51 and an inverter 5
2, exclusive ○Y history gate 53, AND gate 54, NAND gate 55, and AND gate 56,
One OR gate 57 is connected as shown in the figure. - the group OR gate 51 receives the signals 58°5 of the corresponding channels from the start controller 60 and the end controller 70;
9, and if there is a signal 58.59, it is output to the switching gate section 40. 5TA
The RT signal and the signal 59 from the termination control section 70 are exclusive O.
After being exclusive ORed by the R gate 53, each A
The signal is input to the ND gate 54, and the logical AND is performed with the output from the AND gate 51, and the result is sent to the termination control section 70. Further, the signal from the input port 20 and the signal stored in the fault storage section 210 are sent to the start control section 60 via the NAND gate 55. Furthermore, the start control section 6
The signal 58 from 0 and the signal stored in the fault storage unit 210 are logically ANDed by an AND gate 56 and then ORed.
It is sent out from gate 57 as signal FAO. This signal FA○ is the input signal that enters the first-arrival input port.

Further, in the sequence control section 90 configuration of FIG. 12 shown in correspondence with FIG. 6, the signal E is taken out from the shift register 92.

In the configuration of the start control section 60 shown in FIG. 13, which corresponds to FIG.
It is entered as 0.

A monitoring circuit 400 that receives such a signal and detects an oscillation port is configured as shown in FIG.

First, as in the case of FIG. 4, a timer 406 is provided which starts when the 5TART signal becomes H level. When the output QH of this timer 406 becomes H, the signal 404
ACK sequence control is forcibly started.

Therefore, in the timer 406 of this embodiment, one of the two types of clocks, 02 and C3, is selected. 1st
Figure 4 shows the clock frequency switching section 4 for this clock selection.
Has 07. First, two timers 4 for down signal detection.
08,409 is provided.

One timer 408 starts when the 5TART signal becomes H, and uses CI as a clock. the other timer 40
9 also starts when the 5TART signal becomes H, and uses the signal FAO obtained from the control gate section 50 as the clock. Both clocks CI and FAO are configured to stop when their respective outputs QD and QB become H. Here, before the output QD of the timer 408 becomes H, the signal F
The value of the output QB of the timer 409 changes depending on the speed of AO. For example, when using a signal with a frequency of 10 MHz in communication, if the frequency of clock C1 is set to lOMHz, the output QB of timer 409 when FAO is lOMHz
becomes H level. On the other hand, if the down signal is IMHZ, the output QB of timer 409 remains at L level even if the down signal comes to FAO. Due to the difference in the value of the output QB of the timer 409, one of the clocks C2 and C3 is selected after the signal E reaches F (level) by the gate circuit 410 as a clock for the timer 406. In this embodiment, Since the CKI shown in FIG. 12 is set to IMHz, if the frequency of the clock C3 is set to lOMHz, the output QH of the timer 406 can be set to the H level before the signal E from the shift register 92 reaches the ■(level). The frequency of the other clock C2 can be set as appropriate depending on the system configuration, but the point is that it should be set sufficiently slower (lower) than the clock C3.

According to this configuration, when the signal FA○ is a signal for normal communication, the clock of the timer 406 for oscillation port detection is the slower clock C2, and when it is a down signal, the faster clock C3 is selected. Oscillation detection can be completed before ACK reception becomes possible.

Further, an embodiment of the invention according to claims F to 9 is described as a first embodiment of the invention.
This will be explained with reference to FIG. 5 and FIG. 16.

First, since the node devices of the above-mentioned amorphous communication network do not depend on the protocol, they broadcast all the signals they receive. In this case, abnormal signals such as noise are handled in the same way, which may cause incorrect path fixation.

Therefore, in this example, based on the above example,
It was configured as follows. First, in the configuration of the start control section 60 in FIG. 15 shown in correspondence with FIGS. 3 and 13, the signal E
, FAO are input to the monitoring circuit 400, and reset RO, R1 and write clock Wl are input to the monitoring circuit 400.
has been entered.

A monitoring circuit 400 that receives such a signal and detects an oscillation port is configured as shown in FIG.

The monitoring circuit 400 of this embodiment includes two D-type flip-flops 411, in addition to the configuration of the down signal detection section shown in FIG.
A collision signal detection unit 413 based on a collision signal detection unit 412 is provided, and is configured to be input to an AND gate 414 together with the output from the timer 409 to control a gate circuit 410.

Here, when the flip-flop 411 of the above type is reset by the reset signal RO and until the next write clock W1 is input (collision detection time), when at least one or more pulses are input from the FAO, The output QB of the D-type flip-flop 412 becomes H level, and the same operation as in the previous embodiment is performed.

That is, when FAO is a normal communication signal, the clock of the oscillation detection timer is C2, and when it is a down signal or when no pulse is received from FAO within the collision detection time, clock C3 is selected. , oscillation detection ends before it becomes possible to accept ACK. Therefore, even when an abnormal signal pattern that causes an erroneous path comes in, it can be detected before the erroneous path is fixed, and it can be separated from the normal port as an oscillation port.

Effects of the Invention Since the present invention is configured as described above, according to the invention according to claims 1 to 3, when the node device broadcasts, a specific input signal is not inputted from the faulty input signal. If a signal can be sent and this signal is received, it can be determined that the transmission path and input/output means between the node devices are normal, and they will automatically return to normal without affecting other communications, making maintenance and operation of the node device easier. is improved, and a decrease in line efficiency can be prevented.

In addition, according to the invention described in claims 4 to 6, when a normal port receives a down signal and returns it as an oscillation port when performing automatic recovery of a down port, it is possible to prevent a long broadcast. Network efficiency can be improved, and since oscillation detection is finished before ACK reception is possible, malfunctions such as fixing a wrong path as a down port due to noise or the like can also be prevented.

Furthermore, according to the invention described in claims 7 to 9, even if an abnormal signal due to noise or the like is input, the abnormality is detected before the incorrect path is fixed to prevent the incorrect path from being fixed, and , it also prevents long broadcasts,
Network efficiency can be further improved.

[Brief explanation of drawings]

1 to 10 show an embodiment of the invention according to claims 1 to 3, FIG. 1 is a functional block diagram showing an embodiment of a node device of an amorphous communication network, and FIG. FIG. 3 is a circuit diagram showing a specific circuit configuration example of the active signal sending unit in the node device; FIG. 3 is a circuit diagram showing a specific circuit configuration example of the start control unit in the node device; FIG. FIG. 5(a) is a circuit diagram showing a specific circuit configuration of a monitoring circuit in a node device; FIG. 5(b) is a circuit diagram showing a specific circuit configuration of a fault storage unit in a node device; FIG. 6 is a diagram showing a specific circuit configuration of the sequence control unit in the node device, and FIG. 7(a)
- (e) are state diagrams showing the states at each stage of communication control for an example in which the node device shown in Fig. 1 is applied to a grid communication network of four nodes, and Figs. 8 (a) to (d) are , FIG. 9 is a relay system diagram showing an example of a communication network configuration in which the same node device is applied to a grid-like communication network, and FIG. 10 is a control sequence of the sequence control unit. , and FIGS. 11 to 14 show embodiments of the invention according to claims 4 to 6,
FIG. 11 is a circuit diagram showing a specific circuit configuration example of a control gate unit in a node device, FIG. 12 is a diagram showing a specific circuit configuration of a sequence control unit in a node device, and FIG. 13 is a start control unit in a node device. FIG. 14 is a circuit diagram showing a specific example of the circuit configuration of the monitoring circuit in the start control section of the node device, and FIGS. 15 is a circuit diagram showing an example of a specific circuit configuration of a start control section in a node device, and FIG. 16 is a circuit diagram showing an example of a specific circuit configuration of a monitoring circuit in a start control section of a node device. FIG. DESCRIPTION OF SYMBOLS 10... Node device, 12... Transmission path, 14... Terminal, 20... Input means, 30... Output means, 40...
- Connection means, 60a... First-come-first-served input detection means, 60b.
- Monitoring means, 90... Control means, 200... Signal output means, 210... Fault storage means, 250... Verification means, 400... Timing means, 406... Timer for oscillation port detection, 407... Clock frequency switching means,
408, 409...Oscillation port detection timer, 413
...Abnormality detection means applicant Ricoh Atsushi Co., Ltd. AFtlZOO ◇ (b) F7 [F] (Koma3)

Claims (1)

[Claims] 1. At least one output means to which each transmission line in a transmission line to a terminal or other node device is connected; and each reception line in the transmission line corresponding to the transmission line is connected. at least one input means connected to the input means, a connection means for connecting these input means and the output means, and identifying a first-come-first-served input means connected to the input means and to which an outgoing signal arrived first among these input means; a first-come-first-served input detection means for detecting a first-come, first-served input; a time-limiting means for starting a time limit for a predetermined period from identification of the first-come-first-served input detecting means; The apparatus comprises a monitoring means for monitoring, a signal output means for outputting a predetermined active signal from an output means corresponding to the first input means, and a fault storage means for storing a fault between the input means and the output means. , a control means for controlling the connection means and selectively connecting the input means to the output means; the control means controls the connection means to connect the communication already set in the input means; The idle input means for the transmission paths not included are connected to at least all the output means except the output means corresponding to the input means, and the control means connects the connection means in response to the identification of the first input detection means. control, disconnect all input means except the first input means from the corresponding output means, and send the outgoing signal from the first input means to all output means except the output means corresponding to the first input means. and the monitoring means monitors whether a return signal arrives from the receiving line to the input means corresponding to the output means that transferred the outgoing signal among the input means, and the input means receives the return signal. is connected to the output means corresponding to the first-come-first-served input means, and the first-come-first-served input means is connected to the output means corresponding to the input means that received the received signal to fix the connection between these input and output means, while the other All input means are connected to all output means except at least the output means corresponding to this input means, and the active signal is transmitted to the input means corresponding to the output means to which the outgoing signal was transferred for a period exceeding a predetermined period by the time limit means. and if a collision signal is not input, the control means stores the input/output means as a fault in the fault storage means, and controls the connection means to disconnect the fault input means from other input means; When a specific signal is input to the device, the control means releases the memory of the fault storage means and controls the connection means to automatically restore the connection of the faulty input/output means to normal. A node device for an amorphous communication network. 2. The control means has a collation means for collating a specific signal with a signal inputted by the fault input means, and causes the fault output means to send out the specific signal, and when the collation means detects a match, the fault is detected. 2. The node device for an amorphous communication network according to claim 1, wherein the connection of the input/output means of the input/output means is restored to normal. 3. When the incoming signal is not returned to the other input means and the outgoing signal is received beyond the predetermined time limit set by the time limit means, the signal of the input means separated by the other input means is made to correspond to this input means. 2. The node device for an amorphous communication network according to claim 1, further comprising input/output return means for outputting to the output means in the control means. 4. The control means includes an oscillation port detection means having an oscillation port detection timer and a clock frequency switching means, and determines that a port receiving a specific signal is an oscillation port;
Claim 1 characterized in that the clock frequency of the oscillation port detection timer is immediately changed to a higher one.
A node device of the amorphous communication network described above. 5. The node device for an amorphous communication network according to claim 4, wherein the specific signal is a down signal sent from a down port. 6. The clock frequency switching means changes the clock frequency of the oscillation port detection timer so as to detect the oscillation port as an oscillation port and return the input signal before the received signal can be accepted. 5. The node device of the amorphous communication network according to 5. 7. The control means includes an abnormality detection means having an oscillation port detection timer, and when a collision signal cannot be detected within the collision signal detection time at the first input port, this signal is determined to be abnormal, and the oscillation port is immediately detected. 2. The node device for an amorphous communication network according to claim 1, wherein the node device handles the node device as a node device. 8. Has a clock frequency switching means, and when a collision signal cannot be detected within the collision signal detection time at the first input port, this signal is determined to be abnormal, and the clock frequency of the oscillation port detection timer is immediately set to a higher one. 8. The node device for an amorphous communication network according to claim 7, wherein the node device is changed to be treated as an oscillation port. 9. According to claim 8, the clock frequency switching means changes the clock frequency of the oscillation port detection timer so as to detect the oscillation port as an oscillation port and return the input signal before receiving the received signal. A node device of an amorphous communication network.