JP2582585B2 - Node device of irregular communication network - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to control of a communication network, and particularly to a node device of an irregular communication network.

2. Description of the Related Art Conventionally, communication networks applicable to multimedia communication, such as a local area network (LAN) and a public line network, include a CSMA baseband LAN, a broadband LAN, and a TDMA baseband LAN represented by Ethernet, for example. And a digital PBX. CSMA baseband LAN has a short packet length like data information, text information, etc.
It is suitable for communication of information that occurs suddenly, but when the message length is not limited as in multimedia communication or when data occurs continuously, collisions occur frequently, so high throughput, that is, communication capacity is high. I can't get it.
Lack of applicability as multimedia communication.

The broadband LAN has a shortage of capacity for multimedia communication. In addition, there is a difficulty in the scalability and price of the system. The TDMA baseband LAN is the most suitable for multimedia communication among the conventional methods, but generally has difficulty in system expandability and price. Especially when applied to multimedia communication, the price becomes very high.

In view of such a state of the prior art, the present inventor has already proposed a lattice communication network based on an analogy of a living nerve cell. See, for example, JP-A-58-139543.
In this method, a communication network is formed by connecting a communication control element of a multi-input and one-output signal as a node to a multi-coupling structure, and each node has a communication network form in which digital signals are transferred on a first-come-first-served basis.

This grid communication network is particularly excellent in the following points. One is that the degree of freedom of the network topology is high due to the multi-coupling structure. Therefore, even if fault tolerance (survivability) is high, that is, if there is a failure in a part of the network, communication is adaptively secured through another route. Next, an optimum communication path is selected by the first-come-first-served logic. Therefore, it can be said that the method is suitable for multimedia communication.

By the way, in such a grid communication network, when an outgoing signal spreads over the communication network, a collision occurs between the first outgoing signals transmitted via different routes at each node. Also, a collision between the first outgoing signal and another outgoing signal occurs.

In the lattice communication network proposed so far, when such a collision occurs, each node sets a predetermined period after detecting, for example, an initially inputted outgoing signal, and receives a signal within this period. The signal is ignored as the first outgoing signal or another outgoing signal input via another route, and the signal input after a predetermined period has elapsed is recognized as a return signal, and the path is fixed.

Therefore, a node that detects a collision of an outgoing signal outputs not only the initially inputted outgoing signal but also the subsequently inputted outgoing signal until the path is fixed by recognizing the return signal. continuing. Therefore, for example, when there are a plurality of nodes in which a collision occurs and the first return signal passes through only one of the nodes, the path of the node on which the return signal has passed is fixed first and initialization is performed. on the other hand,
Since the node through which the return signal does not pass is not initialized, the node continues to output the return signal. Therefore, the node initialized by passing the return signal determines that the outgoing signal from the uninitialized node is the occurrence of new communication, that is, the input of a new outgoing signal, and the START sequence is restarted. There is a problem. Such a difference in the initialization of the nodes is a disadvantage particularly when the scale of the communication network is large.

Objective The present invention solves such a problem, and in the event of a collision, prevents the output of the forward signal and the earliest forward signal that are input later from being output, thereby enabling efficient full-duplex communication. It is an object of the present invention to provide a node device of an irregular communication network that can be ensured in a network.

Configuration In order to achieve the above object, the present invention is a node device connected to a transmission path including a transmission line to a terminal or a node device and a reception line corresponding to the transmission line. At least one input means, at least one output means each connected to a transmission line, a connecting means for connecting the input means and the output means, and controlling the connecting means to selectively connect the input means to the output means. And control means for controlling the input means, wherein the control means is connected to the input means, and the first input detection means for identifying the input means from which the signal arrives first among the input means; A first time limit means for starting a time period of a first predetermined time period from identification by the means, and a time period of a second predetermined time period starting from the end of the time period of the first predetermined time period in the first time means A second timed means, and an input detecting means connected to the first timed means and the second timed means for detecting whether a signal has arrived at the input means from the receiving line, and the control means comprises: By controlling the connection means, the connection between the input / output means is disconnected in an idle state for a transmission line not included in the communication already set, and the connection means is controlled in response to the identification in the first-come input detection means, and the identification is performed. The input means is connected to at least all output means other than those corresponding to the identified input means among the output means, and signals are transferred to them.The input detection means controls the signal transfer among the input means. It monitors whether a signal arrives from the receiving line to the input means corresponding to the output means, and receives a signal during the second predetermined period or the first and second predetermined periods among the monitored input means. Identify input means and control means Disconnects the input means which has received the signal first from the output means corresponding to the input means which has received the signal in one of the first predetermined period and the second predetermined period, and the control means Further, if there is an input unit that has not received a signal within the first predetermined period and the second predetermined period, and has received a signal after the lapse of the second predetermined period, the connection unit is controlled. After the elapse of the second predetermined period, the input means receiving the signal is output to the output means corresponding to the input means from which the signal arrives first, and the input means from which the signal arrives is sent to the second predetermined time. It is characterized in that after the passage of the period, the signal is connected to the output means corresponding to the input means, the connection between those input / output means is fixed, and the connection of all other input means to the output means is cut off. It was done.

Hereinafter, the present invention will be specifically described based on examples.

An irregular communication network to which the node device according to the present invention is applied,
As illustrated in FIG. 11, the node device 10 is advantageously implemented as a grid-like communication network connected two-dimensionally or three-dimensionally by a transmission line 12, but the network configuration is essentially irregular. is there. For example, a net configuration having another shape such as a linear shape or a loop shape may be adopted.

The node device 10 is provided with a plurality of, in this example, eight input / output ports, to which other node devices 10 and / or terminals 14 can be connected via a transmission line 12.
There is no limitation on the number of input / output ports, and at least one input / output port may be used. If the node device 10 is within the capacity of the input / output port, the node device 10 or the terminal connected via the transmission path 12
There is no limit on the number of 14. In addition, the entire network is a single node device
Alternatively, a plurality of node devices 10 may be mounted on a single printed wiring board, for example, and the entire device may be treated as one node device to substantially increase the input / output port capacity. Good.

In this embodiment, the terminal 14 is a terminal device capable of transmitting and receiving data asynchronously, and includes a processing system such as a personal computer, a service station such as a file station or a print station, and the like. The data is advantageously transferred in the form of message packets. As will be described later, in the case of a full-duplex terminal, the terminal 14 which advantageously transmits a response signal immediately upon receiving a packet addressed to the terminal 14 is advantageously used.

The transmission line 12 is, for example, an optical transmission line using an optical fiber or an electric transmission line such as a stranded wire or a coaxial cable. In the present embodiment, data is transmitted in analog or digital. It has a full duplex configuration. The transmission path 12 between the node device 10 and the terminal 14 may have a half-duplex configuration.
In addition, the transmission path between the node devices 10 depends on the traffic.
A plurality of 12 may be provided.

Referring to FIG. 1, the node device 10 has an input port 20 to which a reception line from the transmission line 12 is connected, and an output port 30 to which a transmission line to the transmission line 12 is connected. They are interconnected via a section 40. The input port 20 has eight reception or input channels in this embodiment.
i0 to i7, and the output port 30 has eight corresponding transmission or output channels o0 to o7. As a result, up to eight other node devices 10 and terminals 14 can be connected to the node device 10 via the transmission path 12. Of the output channels o0 to o7, the output channels having the same numbers as the input channels i0 to i7, that is, the “corresponding” output channels are connected to the transmission line 12 on the same route.

The switching gate unit 40 is a gate circuit corresponding to the connection means and selectively interconnecting any of the input channels i0 to i7 and any of the output channels o0 to o7. The input port 20 is also connected to a start control unit 60 and an end control unit 70 via a control gate unit 50. The control gate unit 50 is a gate circuit that appropriately controls connection of a signal from the input port 20 to the start control unit 60, and control signals from the start control unit 60 and the end control unit 70 to the switching gate unit 40 and the end control unit 70. It is. Start control unit 60
Identifies the input channel from which the input signal arrived first,
It is a functional unit that 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 input signal in the input channel of the communication path that has already been set, and performs processing for terminating the communication.
The switching gate unit 40, the start control unit 60, and the end control unit 70 are mutually connected by a gate set bus 80.

The control means includes a control gate unit 50, a start control unit 60, an end control unit 70, and a sequence control unit 90 that controls them.

For the sake of simplicity, a specific configuration of the switching gate section 40 has four 4-input NAND gates 42 corresponding to the number of output channels, that is, in this example, as shown in FIG. The switching gate unit 40 further includes 4 × (4-1) three-input NAND gates 44 and 4 × (4
-1) / 2 flip-flops 46, an AND gate 48, and an exclusive OR (EXOR) gate 49 are connected as shown.

More specifically, the output 43 of the 2-input AND gate 41 of each of the input channels i0 to i3 is the 4-input NA of all the output channels except for the corresponding one of the output channels o0 to o3.
One input of the ND gate 42 is commonly connected via a NAND gate 44. The other input of the NAND gate 44 receives an output from a start control unit 60 described later.

The input from the start control unit 60 to the NAND gate 44 detects a collision, that is, a port to which a forward signal other than the earliest forward signal is input in a node device to which a plurality of forward signals are input, as described later. Low level output is output. Of the NAND gates 44, those to which a low-level signal is input from the start control unit 60 have high-level outputs,
Thus, when the output from the termination control unit 70 is at a high level, the output of the NAND gate 42 is at a low level. Accordingly, the output from the output port corresponding to the input port that has detected the collision is thereby stopped.

The output from the output port corresponding to the input port in which the collision has occurred is stopped from the write pulse of the START sequence to the write pulse of the END sequence. That is, as described later, after disconnecting the output port corresponding to the input port that detected the collision, until the return signal is detected and the path is fixed, or the disappearance of the earliest outgoing signal is detected and the initial Until it becomes.

In the preceding stage of the NAND gate 44, four 2-input AND gates 41 corresponding to the number of input channels, that is, in this example, four 2-input AND gates 41 are provided, and when one input 45 is energized from the control gate unit 50, the input port 20 and the internal circuit of the switching gate unit 40 are selectively connected.

When the control line of the specified channel of the gate set bus 80 becomes high level, the switching gate unit 40 specifies the input / output channel to be interconnected in response to the negative clock signal of the WRITE 0 input from the sequence control unit 90. And interconnect the two channels. The connection between the specified channel and the unspecified channel is disconnected. In addition, the connection state at that time is retained for the channel not specified at this time. As a result, a multi-channel connection that allows a communication path for a combination of a plurality of input / output channels at the same time in one node device 10 is performed.

In this manner, the state of all NAND gates 44 can be set by one control. According to this configuration, the NAND
The number of functional units for maintaining the state of the gate 44, that is, the number of flip-flops 46 can be minimized.

For simplicity, the control gate unit 50 includes four OR gates 52, an inverter 54, a three-input NAND gate 56, an EXOR gate 58, and an AND gate 52, as shown in FIG.
The gate 51 is connected and configured as shown. OR
The gate 52 receives the signal 53 from the start control unit 60 and the end control unit 70.
Gate of the switching gate 40
OR gate for outputting to the four AND gates 41. The inverter 54 and the three-input NAND gate 56 are circuits that take the logical product of the signal from the input port 20 and the signal from the end control unit 70 and output the result to the start control unit 60. In this embodiment, one of the three inputs of the NAND gate 56 is not used. EXOR gate 58 and AND gate 51
A signal from the termination control unit 70 is detected when detecting the end of the communication in which the communication path is set, and a start control is performed when detecting the interruption of the first, that is, the first outgoing signal arriving on the first input channel. This is a circuit for selectively outputting the output of the unit 60 to the end control unit 70.

The specific configuration of the start control unit 60 has four inputs and four outputs for simplicity.
As shown in FIG. 4, the case of a channel includes a first-arrival input signal detection unit 60a and an input signal detection unit 60b. The first-arrival input signal detection unit 60a is a functional unit that identifies a channel of an input signal that has arrived first among input channels i0 to i3 according to a first-come-first-served logic. This corresponds to the number of input channels,
That is, four flip-flops 62, a group of NAND gates 66, a four-input NAND gate 68 and an inverter 61,
A three-input NAND gate 63, a bus buffer 65, and a mode switch 67 are connected as shown in the figure.

The flip-flop 62 is a circuit that holds the state of the input channel from which the input signal has arrived. A group of NAND gates 66
Gives priority between the outputs 64 of the flip-flops 62. The four-input NAND gate 68 and the inverter 61 have a holding function of responding to an input signal to any one of the flip-flops 62 and setting the S terminals of all the flip-flops 62 to a low level to fix their states. , First
This is a circuit for notifying the sequence control unit 90 that the second forward signal has arrived.

The three-input NAND gate 63 calculates the logical sum of the output of the group of NAND gates 66 and the output of the input signal detection unit 60b, and outputs the logical sum to the gate set bus 80 via the bus buffer 65.
Output to The mode changeover switch 67 is always open in this embodiment.

The input signal detector 60b is a circuit that detects whether an input signal has arrived at the input port 20. This is configured by connecting flip-flops 69 and 120, four NAND gates 122, and a four-input OR gate 124 as shown. The flip-flop 69 is a two-state circuit for holding the state of the input channel from which the input signal has arrived. The flip-flop 120 is a circuit for storing the output state of the flip-flop 69 and fixing the state by setting their S inputs to low level. That is, the flip-flop 120 stores the port at which the flip-flop 69 has detected a collision. In this case, neither the input port to which the earliest outgoing signal is input nor the port to which other outgoing signals are input, that is, the port in which the collision occurs, is not distinguished. The output from the flip-flop 120 is input to the OR gate 121, and the output from the NAND gate 63 is ORed with the output from the switching gate unit 40.
Output to As for the output from the OR gate 121, only the colliding port, that is, the port to which the non-earliest outgoing signal is input is at the low level. As a result, only the colliding port is output to the switching gate unit 40 as a low level.

The NAND gate 122 is a gate circuit that controls connection of the output of the flip-flop 69 to the first-arrival input detection unit 60a. O
The R gate 124 is a circuit for calculating the logical sum of the output of the flip-flop 69 and notifying the sequence controller 90 that the first return signal has arrived.

The termination control unit 70 includes a communication termination detection unit 70a and a connection storage unit 70b, as shown in FIG. 5 for the case of four channels. The communication end detection unit 70a includes four NOR gates 7
2, a shift register 74, an AND gate 76, and one 4-input OR gate 78 are connected as shown. The NOR gate 72 calculates the logical sum of the signal from the input port 20 and the signal from the output port 30. The shift register 74 is a circuit for detecting the end of communication based on a signal detection time described later. The AND gate 76 is a circuit that calculates the logical product of the output of the shift register 74 and the output of the control gate unit 50. The four-input OR gate 78 is a circuit for notifying the sequence control unit 90 that there is a communication that has been completed among the communication in which the communication path is fixed, or that the first outgoing signal from the first input channel has been interrupted. The selection of which information to report is made by the control gate unit 50. As can be seen, the termination control unit 70 identifies the termination of the communication when both of the two input channels included in the communication with the fixed communication path have no signal.

The end of communication is identified by a signal detection time, that is, a state in which there is no signal for a time determined by a communication end detection time constant, or that a predetermined logical state has continued. The “communication end detection time constant”, that is, the third predetermined period is a time for detecting that the communication has ended without any further signal following the forward signal or the return signal. In the case of full-duplex communication, the length is set to the time required to distinguish the end of true communication from the sequence of "0" or "1" which is the information content. Usually, some extra time is given to this. For example, in the case of Manchester coding, a time length of one bit is used, and in a coding rule of inserting "0" into "1" of 6 consecutive bits in NRZI, a time length of 7 bits or more is taken. Usually, the time length is set to twice those, that is, the time length of 2 bits or 14 bits, respectively.

When half-duplex communication is included together with full-duplex communication, the length of the communication end detection time constant is determined by the propagation delay time for reciprocating the maximum effective network length, and after the terminal 14 has finished receiving the forward signal or the return signal. It is set substantially equal to the sum of the time required to start transmitting the return signal or the forward signal. Usually, some extra time is added to these.

The connection storage unit 70b includes four flip-flops 71 for storing the channel on which the communication path has been fixed, and an AND gate 73 for controlling writing and erasing of the storage.
And a bus buffer 75 for controlling the connection of the output to the gate set bus 80, as shown in the figure.

According to such a configuration, the shift register 74 can always detect the end of communication for all channels.
That is, since the communication end can be detected for the channel not selected by the control gate unit 50, there is no delay corresponding to the communication end detection time constant in the detection of the communication end when the switching is performed.

Further, in the case of full-duplex communication and in the case of including both full-duplex communication and half-duplex communication, the communication end detection time constant may be set according to each. Therefore, there is no need to change the hardware of the device itself.

In addition, instead of these four NOR gates 72, four NAND gates are used.
If a gate is provided, the logical product of the input channel and the output channel can be obtained. In this way, the termination control unit 70 identifies the termination of the communication when there is no signal in one of the two input channels included in the communication with the fixed communication path.

As shown in FIG. 6, the sequence control unit 90 includes five shift registers 91 to 95 and a gate group 96 for generating necessary control signals by appropriately combining their output states.
When occurrence and termination of communication conflict, a flip-flop 97 for giving priority to termination of communication, a mode changeover switch 98, and a boot switch 99 are connected as shown in the figure. The system clock CK1 or CK0 is connected to clock input terminals of the shift registers 91 to 95. In this embodiment, the flip-flop 95 is not used, and the mode changeover switch 98 is always open. The boot switch 99 is an operation switch that is operated only when the node device 10 is started up and initializes all flip-flops in the node device 10. Sequence control unit 90
However, in the case of full-duplex communication and in the case of including both full-duplex communication and half-duplex communication, it is not necessary to change the hardware of the device itself. FIG. 7 shows the operation timing of the sequence control unit 90.

An outline of communication control in the node device 10 will be described. Here, for convenience, the term “transmitting terminal” refers to a terminal that transmits a signal to the transmission path 12, and the “receiving terminal” refers to a signal that transmits a signal to the transmission path 12.
It refers to the terminal receiving from 12. Also, the term "originating terminal" refers to a terminal in which a connection is not set up with another terminal, that is, a terminal that starts transmitting information addressed to a specific terminal from an idle state, and the "receiving terminal" is the terminating terminal. A destination terminal that returns a response to information for the first time.
A signal transmitted from the transmitting terminal is called an "outgoing signal", and a signal transmitted from the receiving terminal, particularly a signal returned in response to the outgoing signal, is called a "return signal".

In a certain node device 10, in an idle state in which a connection is not set between specific input / output channels, the connection gate of the switching gate unit 40 is in a closed state, and the connection between the input / output ports is in a disconnected state.

When an input signal arrives at any of the input channels i0 to i7 in the idle state, it is output to the switching gate unit 40. The first-arriving input signal detection unit 60a
The channel from which the input signal arrives first among 0 to i7, that is, the “first-come-first-served input channel” is detected by first-come-first-served logic. In response to the detection of the first input channel, the switching gate unit 40 connects the first input channel to all output channels other than the output channel corresponding to the first input channel. As a result, broadcasting is performed in which a signal received from the first-arrived input channel is transferred to all output channels other than the corresponding output channel. At this time, there is a time delay in the operation of the first-arrived input signal detection unit 60a and the switching gate unit 40, so that the leading part of the input signal may be partially missing.

The sequence control unit 90 is activated by the detection of the first input channel by the first input signal detection unit 60a, and the sequence control unit 90
Starts timed monitoring for a first fixed time based on a link time constant and a second fixed time as a signal detection time starting from the first fixed time. The input signal detector 60b detects an input signal within the period of the link time constant.

The "link time constant", i.e., the first predetermined period and the signal detection time, i.e., the second period, are the first, i.e., from the same transmission source, from input channels other than the input channel that first detected the input signal. Even if a first outgoing signal is received or another first outgoing signal from another transmission source is received and a collision occurs, the first outgoing signal is eliminated, This is the time to distinguish the first, or first, return signal returned in response to the first outgoing signal associated with the first incoming channel.

The length of the “link time constant” period is
The propagation delay time for reciprocating the maximum allowable distance between the terminals or between the terminals 14 is set to be substantially equal. The propagation delay time includes a delay caused by the node device 10 itself. Usually, some extra time is added to this. The length of the signal detection time is, for example, 1 in the case of Manchester coding.
In the case of a coding rule in which “0” is inserted into “1” of 6 consecutive bits in NRZI, a time length of 7 bits or more is taken. Usually, the time length is set to twice those, that is, the time length of 2 bits or 14 bits, respectively.

The switching gate unit 40 includes a first-arrived input signal detection unit 60a.
When an input signal arrives from any of the input channels within the period of the link time constant stored in the input signal detection unit 60b and the signal detection time after the detection of the first-arrived input channel, the input signal becomes the earliest. This signal is not an outgoing signal but a signal that is colliding with another outgoing signal.This signal is controlled so that it is not output to the output channel. Is controlled so that the earliest outgoing signal is not output to the output channel to be output.

When a signal is input to the input signal detection unit 60b after the period of the link time constant and after the elapse of the signal detection time, since this signal is a return signal, the output corresponding to the input channel of the earliest outgoing signal is output. Connect to a channel.

The termination control unit 70 instructs the sequence control unit 90 when the time defined by the communication termination detection time constant elapses,
The sequence control section 90 resets the first-arrival input signal detection section 60a and the input signal detection section 60b to an initial state.

The end of the communication may be detected by monitoring an input signal from the first-arrived input channel, detecting that the input signal has disappeared, and performing a recovery process. The input signal from both of the other input channels may be monitored to detect that either of them has disappeared, and to perform a recovery process. The detection of the disappearance of the input signal is performed when the logic state of the signal is in a predetermined state for the period of the communication end detection time constant.
For example, it is performed by detecting that it is maintained at "0".

In the above-described embodiment, the input signal detecting section 60b stores the input channel during which the input signal has not arrived during the period according to the link time constant and during the signal detection time.
However, instead of simply storing, after the elapse of the synchronization period, the input channel from which such an input signal has not arrived is connected to the output channel corresponding to the first-arriving input channel, and the output channels of all other input channels are connected. May be configured to be disconnected.

After the time of the link time constant and the signal detection time have passed to the input channel where such an input signal has not arrived,
When the first return signal arrives, the input channel that received the first return signal corresponds to the output channel corresponding to the first input signal, and the first input channel corresponds to the input channel that received the first return signal. To fix the path between the input and output channels. All other input channels are disconnected from the output channels. With this configuration, it is possible to avoid receiving a signal such as noise other than the first return signal to be received from another input channel.

For explanation of the present embodiment, a communication procedure in the system of the present embodiment will be described with reference to FIGS. 10A to 10G for a grid communication network in which four node devices 10 are connected in a grid. In the illustrated communication network, four node devices 10a to 10a
0d are connected in a grid pattern by the transmission path 12. Terminals 14a and 14d are connected to the node devices 10a and 10d, respectively. In the figure, the hatched side indicates the transmitting side, and the thick line indicates the flow of the information signal.

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

First, as shown in FIG. 10A, in a first step, a transmitting terminal that wants to transmit data for the first time from an idle state, for example, 14a sends a first outgoing signal in the form of a packet to the node device 10a through the transmission line 1a. . The first outgoing signal includes a destination address indicating a destination terminal, for example, 14d. The node device 10a detects the first outgoing signal as the first incoming signal. That is, the channel on which the input signal arrived first, that is, the “first-come-first-served input channel” is identified by the first-come-first-served logic. So, the first-come-first-served
All output channels 12ab except the output channel corresponding to 12a
And the first outgoing signal to 12ac or the like. That is, the first outgoing signal is broadcast to all routes of the node device 10a. Further, the node device 10a disconnects the connection of the other input port to the output port.

Next, in the second step, as shown in FIG. 10B, the other node devices 10b, 10c, and 10d also
The first outgoing signal is received from b, 12ac, 12bd, and 12cd, and the same broadcast is performed. These node devices 10b, 10c and 10d also disconnect the input ports other than the input port that received the first outgoing signal from the output ports.

In this example, in the node device 10c, outgoing signals are input from the transmission lines 12ac and 12cd, and a collision occurs. The node device 10c recognizes the input channel 10ca from the transmission path 12ac as the first-arrival input channel, and broadcasts it to another transmission path such as the transmission path 12cd. At this time, since the connection of the input channels other than the input channel 10ca to the output channels is cut off, the input signal from the input channel 10cb is not output to another node device.

Similarly, the node device 10d is connected to the transmission line 12bd
Although the first outgoing signal arrives also from d, the input channel 10da from the transmission line 12bd is recognized as the first-arrival input channel, and the first outgoing signal from the transmission line 12bd is regarded as the transmission lines 12d and 12cd.
Broadcast to other transmission paths. Transmission line 12cd
Since the input channel 10db from is not the first-arrival input channel, the connection of the input channel 10db to the output channel is disconnected, and the input signal from the transmission path 12cd is not broadcast to other transmission paths. In this way, the first outgoing signal transmitted from the terminal 14a and broadcast is transmitted to the network without duplication. Thus, the first outgoing signal via the shortest path reaches the terminal 14d.

In the third step, the node devices 10a to 10d perform the operation in the period of the signal detection time constant starting after the end of the period of the link time constant starting from the detection of the first-arrived input channel, or in the period of the link time constant and the period of the signal detection time constant. If all input channels are monitored and an input signal is received during that period, that is, if a collision occurs as described above, the first input channel to the output channel corresponding to the input channel of the transmission path that received the input signal Is disconnected. As a result, the tenth
As shown in FIG. c, the earliest input outgoing signal is not output to the transmission path to which the outgoing signal has been input after the earliest input outgoing signal. That is, in the node device 10c, the earliest outgoing signal from the node device 10a is not output to the output channel 10cd 'corresponding to the input channel 10cd that has received the non-earliest outgoing signal from the node device 10d. In the node device 10d, the earliest outgoing signal from the node device 10b is not output to the output channel 10dc 'corresponding to the input channel 10dc that has received the non-earliest outgoing signal from the node device 10c.

Also, an input channel that has not received an input signal during the period of the link time constant and the period of the signal detection time constant is identified.

In the fourth step, the terminal 14 connected to the node devices 10a to 10d receives the first outgoing signal. Each terminal 14 checks the destination address contained in the first outgoing signal with its own address. In this example, the terminal 14d sends the first, ie, first, return signal to the transmission line 12d because the destination address matches that of its own station. In the node device 10d, among the input channels corresponding to the output channel that transmitted the first outgoing signal, the input signal does not arrive within the period defined by the link time constant and the period of the signal detection time constant, and After the end of the period, the input channel on which the signal arrives, that is, the input channel 10dd on which the return signal arrives
Identify. This is connected to an output channel 10db 'corresponding to the first-arrived input channel 10db.

In this example, as shown in FIG. 10D, the node device 10d
When a signal is received from the transmission line 12d after a lapse of a period according to the link time constant and the signal detection time constant, the signal, that is, the input channel 10dd that has received the first return signal,
It is connected to the output channel 10db 'corresponding to the first input channel. Therefore, the first return signal received from the transmission line 12d is transmitted from the node device 10d to the transmission line 12bd.

At the same time, the first input channel that has received the first forward signal is output from the output channel corresponding to the input channel that has received the first return signal, and the connection of the other input channels to the output channel is disconnected. That is, in this example, the transmission path 12bd is mutually connected to the transmission path 12d.

In the fifth step, the node devices 10b, 10c and 10a
Performs the same control as the node device 10d. Therefore,
As shown in FIG. 10E, the first return signal arrives at the transmitting terminal 14a by reversing the path on which the first forward signal has been transferred. The first outgoing signal has a certain length,
Also, the terminal device such as the terminal 14d is configured to transmit the first return signal immediately after identifying the destination address of the first outgoing signal, so that the first return signal is the first return signal. It is transmitted while overlapping with the outgoing signal. Therefore, even if terminals other than the terminals 14a and 14d are connected to this network, those terminals cannot participate in this communication. As a result, confidentiality of communication with other terminals, which is important for the communication system, is maintained, and multi-channel communication is enabled.

As shown in FIG. 10E, in the fifth step, the node device 10c does not receive the first return signal from the transmission line 12cd or the like and the first outgoing signal received so far on the transmission line 12ac When it disappears, this is detected and the connection of all input channels to the output channel is cut off. That is, the input signal is not received during the period determined by the link time constant and the signal detection time constant, and the first return signal does not arrive even after the lapse of the input signal, and the first forward signal is not received. Is detected, the connection of all input channels to the output channels is disconnected. This means that the communication is not routed through the node device 10, the route is fixed, or the communication is not established, and the transmitting terminal stops transmitting the first outgoing signal. Therefore, in other cases, the arrival of the first return signal is guaranteed within the terminal response monitoring time starting from the detection of the first-arrived input channel. When the transmitting terminal 14a stops transmitting the first outgoing signal halfway because the first outgoing signal does not reach the receiving terminal 14d for some reason and therefore the first returning signal is not returned. The same is true for

When both the full-duplex communication and the half-duplex communication are included, the node device 10c stops receiving the first outgoing signal, and thereafter, the first return signal is output even after a period according to the network time constant elapses. When it is detected that no incoming signal has arrived, this is detected and the connection of all input channels to the output channels is cut off.
That is, for any input channel that has not received an input signal, if it is detected that the first return signal has not been received within the terminal response monitoring time starting from the end of the first outgoing signal, all input channels Disconnect the channel from the output channel.

By such connection control, the calling terminal 14a and the called terminal 1
One communication path is set and fixed for communication with 4d. Each node device 10 can perform setting control of communication newly occurring for a path that is not fixed.

Therefore, in the sixth step, when another first outgoing signal newly transmitted from another terminal arrives at the node device 10d, as shown in FIG. 10F, the first first outgoing signal from the aforementioned transmitting terminal 14a is transmitted to the node device 10d. Like the outgoing signal, the signal is propagated through the network by an unfixed communication path. At this time, a path already used for another communication is not used.

The receiving terminal for this new first outgoing signal is
In response to this, the first return signal is transmitted. This first
As shown in the seventh step in FIG. 10G, the seventh return signal also has the same route as that of the first return signal for the terminating terminal 14d described above, and has a route opposite to that of the new first forward signal. To reach the calling terminal. As a result, a new communication path is fixed.

Thus, each node device 10 detects the presence or absence of an input signal and performs sequential control regarding the link time constant, the signal detection time, and the network time constant. The same applies to the control for communication termination. For example, 1 for full-duplex communication
When the originating terminal is authorized to continue and terminate communication, it is detected that the first outgoing signal has disappeared for a pair of input channels for which the communication path has been fixed, or the pair of input channels has been detected. Is detected, the connection of the pair of input channels to the output channel is disconnected.

In the case of the half-duplex communication or the case where it is not necessary to set the priority for the transmitting station and the receiving station even in the full-duplex communication, there is no input signal in both of the pair of input channels whose path is fixed. Is detected, the connection of the input channel to the output channel is disconnected.

Alternatively, instead of using the interruption of the input signal on both of the pair of input channels as the condition for detecting the termination, the communication between the input channel and the corresponding output channel is detected by detecting the absence of the input signal. You may comprise so that it may complete | finish. These are exactly the same, because if there is no input signal on one of the fixed input channel pairs of the communication path, there is naturally no output signal on the output channel corresponding to the other input channel.

Alternatively, instead of using the interruption of the input signal on both of the pair of input channels as the condition for detecting the termination, the termination of the communication is detected by detecting the absence of the input signal on both the input channel and the corresponding output channel. May be configured as follows. If there are no input signals in both input channel pairs having a fixed communication path, there is naturally no output signal in the corresponding output channel, which means the same.

In the present embodiment, in principle, the node device 10 does not distinguish between a connection between the node devices 10 or a connection between the node device 10 and the terminal 14. Therefore, the node device 10
The user is not conscious of whether another node device 10 is connected to the transmission line 12 connected thereto or whether the terminal 14 is connected. Therefore, the terminal 14 must behave as if it were the same as the other node devices 10 from the node device 10 to which the terminal 14 is connected.

The basic restrictions on the communication procedure required for the terminal 14 in the present embodiment are as follows. Advantageously, the terminal 14 is capable of transmitting and receiving data in the form of a packet, but is not necessarily limited thereto.

As shown in FIG. 8, the message packet 100 as the first outgoing signal includes at least the preamble P and the destination address D prior to the message M. The preamble P needs to last at least a predetermined length. This is for synchronizing the terminal 14, but the leading part of the preamble P is partially removed every time the node device 10 passes through the node device 10 for the time required for controlling the first-come-first-served logic. Therefore, even if the node device 10 takes into consideration the longest path of the network, it is transmitted in such a length that a portion of the length necessary for the recovery of synchronization remains. There are no other restrictions on packet 100,
Usually, it has the address of the originating terminal 14, that is, the source address S. The message M may be followed by a check code area such as a CRC and a packet end code E, followed by a postamble for maintaining the synchronization of the terminal.

If the first outgoing signal is received and it is determined that the destination address of the packet is addressed to the own station, the terminal 14
Is the response signal immediately after the determination in the case of a terminal of full-duplex communication (FIG. 8) and immediately after the end of the first outgoing signal in the case of a terminal of half-duplex communication (FIG. 12). , The first return signal is transmitted. Although there is no restriction on the first return signal, the response packet 10 as the first return signal
2 usually takes the same format as the first outgoing signal as shown in FIG. 8 or FIG.
It has a destination address D and the address of the terminating terminal 14, ie the source address S, followed by a code indicating an acknowledgment ACK or a negative acknowledgment NACK. This may be followed by message M. If a full-duplex communication function such as voice communication or video communication such as videophone is required,
2, a message M is also added. As mentioned above, the first
The third return signal is guaranteed to be transmitted to the calling terminal preferentially.

The transmitting terminal 14 monitors the reception of the first return signal transmitted from the receiving terminal within a predetermined length of the “terminal response monitoring time”. If the reception of the first return signal is detected within the terminal response monitoring time, it is determined that the receiving terminal is in a state where it can respond normally, and communication can be continued. If the reception of the first return signal is not detected within the terminal response monitoring time, it is determined that the receiving terminal is not in a state where it can normally respond, and the communication is stopped as shown in FIG. Calling terminal
14 can then retransmit the first outgoing signal. This may be the same control as in the case of the CSMA method, for example. With these functions, the path between the calling and called terminals is fixed, and communication can be performed while occupying the communication channel.

For terminals with full-duplex communication, the "terminal response monitoring time"
It is also the time that starts when the calling terminal starts transmitting the first outgoing signal. The length is the propagation delay time for reciprocating the maximum effective network length, the maximum effective accumulated node delay time obtained by accumulating the node delay time in each node device 10 for the maximum number of node devices 10 considered in the network, and the terminating terminal. Is set substantially equal to the sum of the time required to start transmitting the first return signal after starting to receive the first outgoing signal. Usually, some extra time is added to this. The node delay time is a delay time generated because the preamble of the first outgoing signal is deleted by the first-come-first-served logic.

In the case of a system that includes both full-duplex communication and half-duplex communication, the “terminal response monitoring time” is a time that starts from when the transmitting terminal has finished transmitting the first outgoing signal. The length is the sum of the propagation delay time for reciprocating the maximum effective network length and the time required for the called terminal to start transmitting the first return signal after finishing receiving the first outgoing signal. Is set substantially equal to This is the same as the time defined by the communication end detection time constant described above. Usually, some extra time is added to this.

The receiving terminal may notify the transmitting terminal of the completion after receiving the first outgoing signal correctly. That is, this is realized by transmitting the first return signal immediately after the reception of the first forward signal is completed. This includes an acknowledgment ACK or a negative acknowledgment NACK.

In this embodiment, since the transmission line 12 is a full-duplex transmission line,
Even if the terminal is a half-duplex device, if the following functions are added to its network interface, the time required to determine whether the terminal is in a receivable state is equivalent to that of a full-duplex terminal. Can be shorter. That is, the network interface unit receives the first outgoing signal, reads the destination address of the packet, determines whether or not the packet is addressed to its own station, and determines that it is addressed to its own station. Immediately after
It is configured to transmit a receivable signal as the response signal. The "receivable signal" may be in the form of a single pulse without any restrictions, for example. This corresponds to the first return signal, and is transmitted preferentially in each node device 10. Such additional functions are advantageously realized by slightly modifying the network control of the half-duplex terminal.

When transmitting the forward signal or the backward signal following the first forward signal or the first backward signal, that is, when transmitting a plurality of packets continuously, the terminal 14 determines the communication interval between the packets. It is sufficient that the time does not exceed the time specified by the end detection time constant. In other words, when communication is continued, that is, when the set communication path is used in a fixed manner, the next time the packet specified by the communication end detection time constant elapses after the end of the packet being transmitted. May be transmitted.

For example, in the case of full-duplex communication, a dummy signal such as a postamble is inserted between successive packets, that is, between the Nth packet and the (N + 1) th packet so that the communication end detection time constant does not time up. To If half-duplex communication is involved, when the packet being received ends,
The transmission packet is transmitted before the communication end detection time has elapsed. In other words, when the receiving terminal finishes receiving the N-th outgoing signal, within a time period defined by the communication end detection time constant,
Preferably, immediately after transmitting the N-th return signal, the transmitting terminal, upon ending the reception of the N-th return signal, also within the time defined by the communication end detection time constant, preferably immediately immediately after the (N + 1) -th return signal The outgoing signal of is transmitted. For example, if the packet format is not used for audio or video communication,
It is sufficient that the no-signal state is shorter than the communication end detection time constant.

 To end the communication, the terminal may stop transmission.

As long as the restrictions on these communication procedures are followed, the degree of freedom regarding other points is high, and the following effects can be obtained. First, there is no limit on the maximum and minimum packet lengths, and the packet format does not have to be taken. Next, the number of continuous repetitions of the forward information and the return information is not limited, and the communication channel may be occupied. If the data rate is equal to or less than the maximum eta rate determined by the hardware constituting the network, the data rate can be freely determined between the transmitting and receiving terminals. Fourth, full-duplex communication and half-duplex communication can be freely selected and may be mixed.

In summary, in the present embodiment, the node device is configured to transfer the first return signal overlapping with the first outgoing signal.
10 controls. That is, the first incoming channel is detected, and the other input channel from which the signal arrives, that is, the same first outgoing signal transmitted through another path or another first outgoing signal transmitted from another signal source is detected. Is the incoming input channel, i.e., the input channel where the collision occurred. This identification is performed by the time defined by the link time constant and the signal detection time. By cutting off the connection of the input channel having the collision thus detected to the output channel, the output of the non-earlier outgoing signal to other channels is stopped, and the output channel corresponding to the input channel having the collision is output. Is disconnected, the output of the earliest outgoing signal to the transmission line to which the non-earliest outgoing signal has been input can also be stopped. Therefore, it is possible to prevent the malfunction of the node device due to the signal from the port where the collision has occurred, thereby improving the throughput of the entire network.

Further, according to the present embodiment, it is possible to transfer the first return signal overlapping the first forward signal. Therefore, upon detecting the reception of the first outgoing signal addressed to the own station, the terminal 14 may immediately transmit the first return signal, and after a lapse of a predetermined time after the reception of the first outgoing signal. , The control for transmitting the first return signal need not be performed.

As described above, in the present embodiment, multi-channel communication in which one node device 10 allows a plurality of communication at the same time is realized. The high fault tolerance of the grid communication network that forms links by first-come-first-served logic while avoiding failed nodes and failed lines is maintained.

By providing such full duplex communication, the following effects can be obtained. First, since the first return signal can be transferred overlapping with the first outgoing signal, it is possible to detect whether communication has been established by detecting the first return signal, and to control transmission continuation and retransmission at an early stage. Can be performed. Therefore, there is no problem even if the packet of the first outgoing signal is long,
The availability of long packets has the effect of increasing the system throughput, ie, the effective data rate.

Next, active reception control is realized. More specifically, only the terminal that has transmitted the first return signal, that is, only the terminating terminal, can normally receive the first outbound signal. Other terminals cannot intercept those signals.

Further, half-duplex communication is possible with the same node device, that is, the same algorithm, and mixed use with full-duplex communication is allowed.

Furthermore, the communication procedure required for the terminal is simple, and the network interface unit of the terminal can be configured at a small size and at low cost.
Therefore, a highly versatile and efficient communication system is established. Sometimes effectively applied to full-duplex communication.

Effect According to the present invention, by enabling the transfer of the first return signal overlapping with the first outgoing signal, the time until the communication path is fixed is reduced, and the multi-channel Full full-duplex communication is provided. In particular, when a collision occurs in the node device, the output of the non-earliest outgoing signal to another channel is stopped, and the output of the earliest outgoing signal to the transmission path to which the non-earliest outgoing signal is input is also stopped. Therefore, it is possible to prevent the node device from malfunctioning due to a signal from the port where the collision has occurred, and to improve the throughput of the entire network.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a functional block diagram showing an embodiment of a node device of an irregular communication network according to the present invention, and FIG. 2 is a circuit showing a specific circuit configuration example of a switching gate unit in the node device. FIGS. 3 to 6 are circuit diagrams similar to FIG. 2 showing specific circuit configuration examples of a control gate unit, a start control unit, an end control unit, and a sequence control unit, respectively, in the node device. FIG. 7 is a timing chart showing the operation timing of the sequence control unit shown in FIG. 6, and FIG. 8 is the first return signal normally returned in response to the first forward signal in full-duplex communication. FIG. 9 shows a packet flow in the case where the first return signal responding to the first outgoing signal is not returned normally in full-duplex communication. Figures, Figures 10A through 10G FIG. 11 is a state diagram showing states at each stage of communication control in an example in which the present invention is applied to a grid communication network of four nodes. FIG. 11 is a diagram of a communication network configuration applied to the grid communication network of the present invention. FIG. 12 is a diagram showing an example of a relay system. FIG. 12 shows a flow of a packet when a first return signal is normally returned in response to a first forward signal in half-duplex communication. FIG. Description of Signs of Main Parts 10 Node Device 40 Switching Gate Unit 50 Control Gate Unit 60 Start Control Unit 60 a First-arrival Input Signal Detection Unit 60 b Input Signal Detection Unit 70 End Control Unit 80 Gate set bus 90 Sequence control unit i0 to i7 Input channels o0 to o7 Output channels

Claims (7)

(57) [Claims] 1. A node device connected to a transmission line including a transmission line to a terminal or a node device and a reception line corresponding to the transmission line, wherein at least one input means to which the reception line is connected. At least one output unit to which each of the transmission lines is connected; connecting unit for connecting the input unit and the output unit; and controlling the connecting unit to selectively connect the input unit to the output unit. A node device of an irregular communication network having control means for causing the control means to be connected to the input means, and to identify an input means of the input means from which a signal arrives first; A first time limit means for starting a time period of a first predetermined time period from the identification by the first arrival input detection means; and a second predetermined time period from the end of the time period of the first predetermined time period in the first time time means. time A second timed means for initiating a is connected to the first timed means and the second time period means,
Input detection means for detecting whether or not a signal has arrived from the reception line to the input means, wherein the control means controls the connection means so that the transmission path is not included in communication already set In the idle state, the connection between the input / output means is disconnected, the connection means is controlled in response to the identification by the first-come-first-served input detection means, and the identified input means is at least the identification means of the output means. Connected to all output means other than the input means corresponding to the input means, and the signal is transferred to them. The input detection means is an input means corresponding to the output means of the input means which has transferred the signal. Monitoring whether or not a signal arrives from the reception line, identifying input means receiving the signal during the second predetermined period or the first and second predetermined periods among the input means being monitored, The control means Is the second predetermined period or the first and second
Disconnecting the input means from which the signal has arrived first to the output means corresponding to the input means having received the signal during the predetermined time period, the control means further comprises: the first predetermined time period and the second time period. If there is an input unit that has not received a signal within the predetermined period of time, the input unit has received a signal after the lapse of the second predetermined period, the input unit controls the connection unit to perform the second predetermined period. The input means that has received the signal after the elapse of the signal is output to the output means corresponding to the input means that the earliest signal has arrived, and the input means that has received the earliest signal is the signal after the elapse of the second predetermined period. Receiving the received input means, connecting the input means to the output means, fixing the connection between the input and output means, and disconnecting all other input means from the output means. Network node device. 2. The apparatus according to claim 1, wherein said control means sets the first predetermined period and the second predetermined period for a transmission path not included in the communication which has been set. After the lapse of the period, no signal arrives at any of the input means which did not receive the signal within the first predetermined period and the second predetermined period, and the input means which received the signal first received the signal. The node device controls the connection means to disconnect the connection between all of the input / output means when there is no signal being sent. 3. The apparatus according to claim 1, wherein said control means receives said one of a pair of input means included in said fixed connection. A node device, wherein upon detecting an end of a signal, the connection device is controlled to disconnect the pair of input devices from the output device. 4. The apparatus according to claim 2, wherein said control means starts a time period of a third predetermined period when said signal takes a predetermined logic state. When the signal received by the input means included in the fixed connection detects that the predetermined logical state is maintained for a third predetermined time, the connection means is controlled to control the input / output means. A node device for disconnecting a connection between the nodes. 5. The apparatus according to claim 1, wherein the control means detects the end of the first outgoing signal for a transmission path not included in the communication which has already been set, and sets the fourth predetermined signal. A fourth time period means for starting a time period of the period, monitoring the reception of the first return signal for a fourth predetermined time starting from detection of the end of the forward signal, and monitoring the first time during the fourth predetermined period. When the return signal is not received, the connection means is controlled to connect the input means to which the signal has arrived first to all the output means, and all the input means have the signal to arrive first. A node device for disconnecting a connection to an output unit corresponding to an input unit. 6. An apparatus according to claim 1, wherein said control means determines that both signals received by a pair of input means included in said fixed connection are terminated. And a fourth timer means for starting a time period of the fourth predetermined period upon detection, wherein the first recovery means starts the end of the signal received by one of the pair of input means. Monitoring the reception of the signal, and when not receiving the first return signal during the fourth predetermined period, controlling the connection means to control the transmission path not included in the communication already set; A node device, wherein the connection between the pair of input means and the output means is disconnected. 7. The apparatus according to claim 5, wherein said control means starts a time limit of said fourth predetermined period when said signal takes a predetermined logic state. 4, when the signal received by the input means included in the fixed connection detects that the predetermined logic state is maintained for the fourth predetermined time, the connection means is controlled to control the input. A node device for disconnecting a connection between output means.