JPH05252170A - Optical communication network and communication system - Google Patents

Abstract

PURPOSE:To heighten the working efficiency of a broadcast bus by identifying the kind of collision when it occurs on the broadcast bus, and deciding priority for the use of the broadcast bus when the collision between two bodies occurs. CONSTITUTION:A carrier detector 12 which detects a carrier on the broadcast bus and a code violation detector 13 which detects the interference state of a received signal by a code violation rule are provided at every node in a network in which plural nodes are connected to the broadcast bus with bidirectionability and communication between the nodes is performed, and the collision between two bodies is detected by the carrier detector 12, and the collision among multiple bodies over three is detected by the code violation detector 13. A node main body 5 changes the transmission mode of a packet based on a detection result.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical LAN (local area network) and a communication protocol (communication control procedure) for the optical LAN. In particular, the present invention relates to an optical communication network and a communication method for detecting a collision on a network and controlling communication.

[0002]

2. Description of the Related Art Generally, in a LAN, a plurality of nodes are connected to a bus, and communication between each node is performed via the bus. A broadcast bus is one of the buses used in such a LAN. The broadcast bus is a bus in which a signal transmitted from a certain node can be simultaneously received by all the nodes. Ethernet (registered trademark) is well known as a LAN using this broadcast bus, and its communication protocol is CSMA / CD.
(Carrier Sense Multiple A
ccess / Collision Detectio
n) method, IEEE 802.3
Is standardized as. In Ethernet (registered trademark), a coaxial cable is used as a transmission medium. Each node is connected to the coaxial cable, and the node to be transmitted first checks whether there is a signal from another node on the coaxial cable, and if there is no transmission from another node, starts transmission. However, in such a case,
There is a chance that more than one node will accidentally start transmitting at the same time. Such a state is called a collision, and Ethernet (registered trademark) detects this collision state by the voltage level of the coaxial cable.

A node that has detected a collision sends a jam signal for a certain period of time, and then waits for a random time. This jam signal should be longer than the maximum round trip propagation delay of the network. This is necessary to inform all nodes connected to the network of the collision state. Also, waiting for a random time is
This is because if a plurality of nodes that could not be transmitted try to retransmit at the same time when the line becomes free, a collision will occur again.

On the other hand, optical communication is also being adopted in LANs. However, in a LAN using an optical fiber as a transmission medium, when a node is added, it is not possible to simply add a tap to add a node, unlike a LAN using a coaxial cable.

In order to solve this problem, therefore, a network has been proposed in which the transmission and reception of each node are divided into separate terminals and all the nodes are distributed by a star coupler. (E.G. Rawson et a
l. , "Fibernet: Multimode Op
mechanical Fibers for Local Co
mputer Networks ”, IEEE Tra
nsaction onCommunication
s, Vol. COM-26, No. 7, p395 (19
78)).

FIG. 13 shows an example of the above proposal of an optical communication network using a star coupler. In FIG. 13, 26
Reference numerals a and 26b are optical fibers, 27 is a node, 25 is a mixing rod type star coupler, and 24 is a terminal. The signal from each node 27 is converted into an optical signal by each light emitting element 22 and supplied to the star coupler 25 via each optical fiber 26a. These optical signals are mixed once by the star coupler 25, then distributed to the respective light receiving elements 23 via the respective optical fibers 26b and converted into electric signals again, and these electric signals are converted to the respective nodes 2
7 is supplied. As a result, the signal transmitted from one node can be transmitted to all the nodes (multicast property), and a communication network similar to Ethernet (registered trademark) can be constructed. ..

In this proposed example, if the number of nodes connected to the star coupler increases, the reception level per node will decrease. Therefore, it is conceivable to add a star coupler and a relay amplifier to expand the network, but in this case, a new problem occurs. That is, the star coupler distributes the signal transmitted from a certain node to the reception port of that node, so that a feedback loop is formed between the star couplers connected to each other, and there is a relay amplifier between the star couplers. If so, oscillation will occur. Therefore, when the star coupler is used, the nodes can be connected to the network by the number of terminals provided in advance in one star coupler.

Further, the property that the own transmission signal is distributed to the own reception port makes collision detection difficult. This is because the actual passive star coupler does not have very good distribution ratio uniformity, and it is difficult to apply level difference detection to collision detection.

Therefore, as a collision detection method applied to a network using a passive star coupler as shown in FIG. 13, a code foul law (CRV: Code Rule) is used.
(Viation) has been proposed. (Small edge:
"Study on Collision Detection Circuit Construction for Optical Star Network", Showa 57
Annual Conference of the Institute of Electronics and Communication Engineers, Optical and Radio Wave Division 341 (198)
See 2)).

The code foul law is Ethernet (registered trademark)
Manchester code used for 2 bits of 1-bit information
It uses the fact that it is represented by a code consisting of bits, that is, it has redundancy.

The Manchester code is such that the rising edge of the central portion of one cycle of the reference clock shown in FIG. 14A corresponds to "1" of data and the falling edge corresponds to "0" of data. is there. FIG. 14B shows
As an example, an example of the Manchester code representing the data “110111” is shown. As can be seen from the figure, in the normal state, the Manchester code is at the H (high) level or more than one cycle of the reference clock in the figure (a). L (low) level does not continue. However, considering the case where the collision signal as shown in FIG. 7C collides with the transmission signal as shown in FIG. 7B, the light intensity of the received signal is as shown in FIG. When this is demodulated, a bit pattern as shown in FIG. That is,
In some cases, the H level continues for one cycle or more of the reference clock. When such an unexpected code (foul code) is detected, a CRV (foul code) signal as shown in FIG. Here, the collision signal shown in FIG. 6C is a Manchester code with a phase shift. Since the nodes are not synchronized, it is uncertain in what phase each Manchester code is added.

The code foul law is that the detection of such a code (foul code) which should not be possible is regarded as a collision. As is clear from the above principle, this method will only detect the collision detection stochastically, but it has already been reported that it can be put to practical use if appropriate hardware is used.

In order to solve the problem that the nodes connectable to such a network are limited, the applicant of the present invention is connected to the same node among a plurality of transfer coefficients expressing the transfer characteristics of the star coupler. Japanese Laid-Open Patent Application No. Hei 2 (1999) -264-96, a feedback loop is not formed even when a star coupler is used in combination by setting the transfer characteristic of a signal between an input terminal and an output terminal forming a pair to 0.
98370 specification. Further, in the optical communication network configured by connecting the star couplers as shown in the specification of the application, the own transmission signal does not return to the transmitting node. Therefore,
By monitoring the receiving port during transmission, it can be determined that a collision has occurred if any signal is detected, and a collision detection mechanism can be easily realized.

Further, in such a network, even when a certain node is transmitting, that node can receive a signal from another node, and transmission and reception can be performed simultaneously. That is, the optical communication network described in the above-mentioned application specification is a bidirectional bus.

Further, in a network constructed by combining star couplers, bidirectional communication using a single optical fiber is also filed by the applicant as Japanese Patent Application No. 3-158614. Furthermore, by performing wavelength division multiplexing in such a network, a so-called multi-channel LAN in which a plurality of broadcast buses using one optical fiber as a transmission medium are arranged in parallel
According to the present applicant, it is also possible to construct a multimedia LAN capable of handling both data communication and transmission of signals such as voice and video signals that require real-time performance by using the plurality of broadcast buses. , Japanese Patent Application No. 3
It has been filed as No. 158580.

Regarding the operation method of the multi-channel LAN, the CSMA / CD method is mainly studied theoretically (for example, Ikebata, Okada, "Load balancing /
Regionally distributed complex multi-channel CSMA / CD system ”, IEICE Transactions (B), Vol. J70-B, N
o. 12, p1466-p1474, (1987)).

By the way, in the case of performing bidirectional communication, the above-described collision detection method of monitoring its own receiving port and determining that a collision has occurred if any signal is detected cannot be used. .. In such a bidirectional bus, when bidirectional communication is performed using a contention-based protocol, when a communication is started between the first node and the second node, the third node happens to
That is, there is a probability that nodes other than the parties will start transmission. This collision of the third node is similar to the collision with other transmitting nodes in one-way communication.

In a network using a broadcast bus such as Ethernet (registered trademark) described above, a signal transmitted by a certain node can be received by all other nodes. It is not preferable from the viewpoint.

Therefore, in order to solve such a problem, the applicant of the present invention monitors the broadcast bus for a time τ ₁ longer than the round trip propagation delay time τ _{0 of} the broadcast bus even after the transmission node starts transmission. However, the reply node starts a reply after a waiting time τ ₂ longer than τ ₁ after receiving the packet addressed to the own node, thereby detecting a collision of the third node in Japanese Patent Application No. 3-97405. In the issue.

[0020]

When communicating using the contention-based protocol in a network composed of broadcast buses as described above, a plurality of nodes simultaneously start transmission on the same broadcast bus. It is inevitable that it will happen, that is, a collision will occur. Even if transmission is started after monitoring the status of the broadcast bus, multiple nodes will start transmission without knowing that each other has started transmission at a time interval less than the round trip propagation delay time of the broadcast bus. This is because there is a probability that

In the case of a conventional contention-based protocol, for example, the above-mentioned CSMA / CD method, when a collision occurs, the nodes involved in the collision have a fixed time (approximately equal to the round-trip propagation delay time of the broadcast bus). After sending the jamming signal during (selected time), wait a random time and try again. The jamming signal is sent to ensure collision detection. Therefore, when a collision occurs, no effective communication is performed, and a jamming signal simply flows on the broadcast bus. Moreover, none of the transmission requests of the colliding nodes are canceled, and a potential accumulation of transmission requests occurs in the form of waiting for a random time.

Here, the case where only two nodes transmit at the same time within the round-trip propagation delay time of the broadcast bus is called two-body collision, and similarly, when only three nodes transmit at the same time, three-body collision occurs. When only four nodes transmit at the same time, we call it four-body collision. The case of three or more collisions is called a multi-body collision. Therefore, a two-body collision is not included herein as a multi-body collision. When defined in this way, it is known that the collisions that occur on the broadcast bus are almost two-body collisions, and that many-body collisions of three or more collisions rarely occur. This will be described below.

Suppose a transmission request (hereinafter called a call) is randomly generated at a plurality of nodes connected to one broadcast bus. Let Λ be the average frequency of calls per unit time. [Lambda] is expressed as, for example, 1000 calls / second if a transmission request of 1000 calls is generated in one second. It is known that such a phenomenon is represented by a stochastic process called Poisson distribution. This Poisson distribution is a function that gives the probability of n calls occurring when the broadcast bus is monitored for a certain time τ. However, n is generally a positive integer. The Poisson distribution is shown in the following equation.

[0024]

[Equation 1] However, when n = 0, it is given by the following equation.

[0025]

[Equation 2] Here, Λ = 1000 calls / second (= 10 ³ calls /
Sec), the probability that only one call is observed during τ = 50 μsec is calculated as Pτ (1) = 4.76 × 10 ^-2 . Similarly, the probability that only two calls are observed during τ = 50 μsec is Pτ (2) = 1.19 × 10 ⁻³ , 3
The probability that only the call is observed is Pτ (3) = 1.98 ×
Calculated as 10 ^-5 . The probability that no call is observed during τ = 50 μsec is Pτ (0) = 0.951.
Is required.

If τ is considered as a round trip propagation delay time, τ =
The fact that only one call is observed within 50 μsec means that no collision occurs. Similarly 2
If only Cole is observed, a two-body collision will occur,
If only 3 calls are observed, this corresponds to a collision of 3 bodies. Also, Pτ (0) is a broadcast bus with τ = 50
Monitoring for μsec corresponds to the case where no call is made, that is, the line remains idle.

Comparing the probability of two-body collision and the probability of three-body collision from the above results, if Λ = 1000 calls / sec and τ = 50 μsec, Pτ (3) / Pτ
(2) = 1.60 × 10 ⁻² = 1.6%. That is, most of the collisions are two-body collisions, and three-body collisions rarely occur. It is known that the rate of three-body collisions increases as the average frequency Λ of calls per unit time increases, but in the case where Λ is an order of magnitude larger, Λ = 1
Even if 0 ⁴ calls / second, Pτ (3) / Pτ (2) = 2
It is about 0%. FIG. 15 shows a graph in which the horizontal axis represents Λ (calls / second) and the vertical axis represents the value of Pτ (3) / Pτ (2).

It is known from the actual measurement results that the average call occurrence frequency Λ in Ethernet (registered trademark) is, for example, about 30 calls / second on average in one day. However, as a peak value, the daily average is 50
It is also known that 60 to 60 times more calls are generated (for example, JF Snoch and JA Hu.
pp, "Measured performance"
of an Ethernet Local Comp
uter Network ”, Communicati
ons of A. C. M. , Vol. 23, no. 1
2, p711-p729 (1980)). Therefore, the value of λ = 1000 calls / second is a value almost close to the instantaneous maximum value of the call occurrence frequency.

As can be understood from the above description, if the collision of two bodies can be avoided, it is possible to prevent the reduction of the line efficiency due to the collision to a considerable extent. Also, if 2
It will be understood that, in the case of a body collision, if it is possible to determine the priority of using the broadcast bus among the nodes involved in the collision, there is an effect similar to effectively avoiding the two body collision. That is, even if a two-body collision occurs, if either node yields the broadcast bus, the other transmission request will be canceled, and the accumulation of potential transmission requests will be eased. Is.

In view of the above circumstances, the present invention identifies the type of collision when a collision occurs on the broadcast bus and determines the priority of using the broadcast bus when two vehicles collide. The purpose of this is to improve the efficiency of using the information bus.

[0031]

In order to achieve the above-mentioned object, an optical communication network of the present invention has a structure in which a plurality of nodes are connected to a bidirectional broadcast bus and packets are used for communication between the nodes. In each network, each node has means for detecting a carrier on the broadcast bus,
Means for detecting the interference state of the received signal.

In the communication system of the present invention, in the optical communication network, the means for detecting the carrier detects two collisions, and the means for detecting the interference state of the received signal includes three or more bodies. It is characterized by detecting a multi-body collision.

The packet issued by each node is provided with a code string indicating the priority of the packet, and when a two-body collision occurs, the right to use the broadcast bus is determined by comparing the code strings indicating the priority. It is characterized by doing.

[0034]

According to the present invention, when a node detects a carrier on the broadcast bus when it tries to transmit,
When transmission is started, collision of two bodies will occur. Further, when the received signal is in the interference state, two or more collisions have already occurred on the broadcast bus, and therefore, when transmission is started, three or more collisions will occur. In this way, by performing both the detection of the carrier and the detection of the interference state of the received signal, when each node has a collision on the network, whether the collision is a two-body collision or not.
It is possible to discriminate whether the collision is more than one body.

Further, according to the communication system of the present invention, the two-body collision and the many-body collision are discriminated from each other, and in the two-body collision, the use priority order of the lines is determined based on the priority given to the packet in advance. Therefore, the use efficiency of the line is improved as compared with the case where the lines are randomly scrambled.

[0036]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The features of the present invention will be specifically described below with reference to the accompanying drawings. FIG. 1 is a schematic diagram of an embodiment of an optical communication network of the present invention.

The plurality of nodes 1 are connected to each other via a star coupler 2. In the example of FIG. 1, 4
Two mutually connectable 8-terminal star couplers 2 are shown, and the star couplers 2 are connected to each other via a bidirectional optical amplifier 3 to form a network.
Each node 1 is connected to the star coupler 2 via the optical fiber 4, and communication between the nodes 1 is performed by packets. In addition, as the bidirectional optical amplifier 3,
For example, a semiconductor laser amplifier is used.

In the first embodiment, the node 1 shown in FIG. 1 is composed of a node body 5 and a communication interface 6 as shown in FIG. The communication interface 6 is provided with a light-receiving device 9 such as a photodiode and a light-emitting source 8 such as a laser diode one by one, and an optical demultiplexer / multiplexer 7 performs demultiplexing and multiplexing of optical signals. The optical fiber 4 led out from the optical demultiplexer / multiplexer 7 is connected to one terminal 14 of the star coupler 2 shown in FIG. The transmission port 10 is hardware that controls transmission, and the reception port 11 is hardware that controls reception. The receiving port 11 has a carrier detection device 12 and a code violation detection device 13
Are connected to detect the occurrence of a carrier and a code foul, and send a signal to the node body 5. Therefore, the node body 5 and the communication interface 6 are connected by four signal systems.

FIG. 3 shows the interconnectable 8 shown in FIG.
The structural example of the terminal star coupler 2 is shown. An optical demultiplexer / multiplexer 15 having a branching ratio of 7: 1, an optical demultiplexer / multiplexer 16 having a branching ratio of 2: 1, and an optical demultiplexer / multiplexer having a branching ratio of 1: 1 on a substrate 18 made of glass or polycarbonate. A required distribution ratio is obtained by connecting 17 by an optical waveguide to form an optical integrated circuit. The optical demultiplexer / multiplexer 15 having a branching ratio of 7: 1 has seven branches to the optical demultiplexer / multiplexer 16 and one branch to the other optical fiber 4. The optical demultiplexer / multiplexer 16 having a branching ratio of 2: 1 has two branches to the optical demultiplexer / multiplexer 17 side and one branch to the optical demultiplexer / multiplexer 16 side. Eight optical fibers 4 are connected to the substrate 18. Reference numeral R _{C in} FIG. 3 is the radius of curvature of the optical waveguide. This interconnectable star coupler is described in detail in Japanese Patent Application No. 3-158614 filed by the present applicant.

In the above-mentioned network, when only one node transmits, all the other nodes connected to the network can receive the content. Also, if two nodes send at the same time,
Although the two transmitting nodes can correctly receive the transmission contents of the other party, there is a characteristic that only the interference signal can be received by the nodes other than the two nodes, that is, confidentiality. This is because the network is bidirectional. When three or more nodes transmit at the same time, any node can receive only the interference signal.
Therefore, the confidentiality of the communication content can be maintained when the two nodes are transmitting to each other. This is due to Japanese Patent Application No. 2-6999 filed by the applicant.
No. 9 is described in detail.

FIG. 4 is a schematic diagram showing the structure of a second embodiment of a node used in an optical communication network to which the communication system of the present invention is applied. By replacing the node 1 shown in the schematic diagram of the optical communication network of FIG. 1 with the structure shown in FIG. 4, a multi-channel optical communication network by wavelength multiplexing can be configured.

In the structure of the node shown in FIG. 4, four systems of communication interfaces each having the same hardware as the communication interface 6 shown in FIG. 2 are provided. In the communication interfaces 6a to 6d of FIG. 4, the emission wavelengths of the laser diodes for transmission are changed, respectively, and λ _a to
It emits light with a wavelength of λ _d . The wavelength multiplexing multiplexer 19 multiplexes the optical signals of wavelengths λ _{a to} λ _d .
Bidirectional optical amplifier 3 of optical communication network shown in FIG.
Is a semiconductor laser amplifier, the wavelength band that can be amplified is as wide as 50 to 70 nm, and the wavelengths λ _{a to} λ _d are, for example, 1
If the wavelength is changed in steps of 0 nm, it is possible to collectively amplify the optical signals of wavelengths λ _{a to} λ _d by one semiconductor laser amplifier.

Next, an embodiment of a communication system, that is, a protocol according to the present invention will be described. FIG. 5 is a state transition diagram of a transmitting node in the communication system according to the first embodiment of the present invention. The protocol shown in FIG. 5 is a protocol applied when a single transmission channel (broadcast bus) is used, and is a protocol for avoiding a two-body collision. The meaning of each variable in FIG. 5 is as follows. CS is a variable indicating the presence or absence of carrier sense, and indicates that there is a carrier when it is "1" and no carrier when it is "0". CRV is a variable indicating the occurrence of a code foul, and indicates that the code foul occurs when it is "1" and does not occur when it is "0". P _i represents the priority of the packet sent from the own node as a variable,
Similarly, P _o represents the priority of the packet sent by the partner node as a variable. ">", "<", "="
Each symbol of represents the superiority, inferiority, and peers of the priority relationship. For example, if P _i > P _o , it means that the packet sent by the own node has a higher priority.

When there is a transmission request from the upper protocol,
First, carrier sensing of the transmission channel is performed. If a carrier is detected (CS = 1), it means that another node is using the transmission channel. Therefore, after waiting a random time, carrier sensing of the transmission channel is performed again. If no carrier is detected (CS =
0), start transmission. When the transmission of the packet header is completed, carrier sensing is performed again while continuing the transmission. This carrier sense continues until the transmission ends normally. If no carrier is detected until the end of transmission (CS = 0), the process ends normally. If any carrier is detected before the transmission ends normally (CS = 1),
This means the occurrence of a collision.

At the same time when the carrier is detected (CS = 1), the transmitting node switches the transmission content to a jamming signal and takes in the detected signal. The captured signal includes the header part of the packet transmitted by another node. The packet structure is shown in FIG. The header section refers to a preamble, a code string 20 indicating a priority, a destination address, and a source address as shown in FIG. The meaning of the code string 20 indicating the priority will be described later.

The transmitting node compares the code string indicating the priority of the fetched packet with the code string indicating the priority of the packet transmitted by its own node. If the priority of the packet transmitted by its own node is high (P _i >)
P _o ), while continuing to transmit the jamming signal, continue carrier sensing for a certain period of time, confirm that the other carrier has disappeared from the transmission channel (CS = 0), and then retransmit the packet again from the beginning. .. If the opponent's carrier does not disappear within a certain period of time (CS = 1), a standby for a random time is entered. On the other hand, if the other party's packet has a high priority (P _i <P _o ), the transmission is immediately stopped and a random time is waited. When the priorities are exactly the same (P _i = P _o ), a jamming signal is transmitted for a fixed time and then a standby for a random time is started. In the transition diagram of FIG. 5, in the process of "carrier sense (jamming signal) for a fixed time", the fixed time is limited to
This is to prevent a situation in which two nodes that have collided with each other due to some error determine that their own node is prioritized, and contention continues indefinitely on the transmission channel.

A code for cyclically determining the priority is used for the code string indicating the priority. This is easy to understand when considering rock paper scissors in familiar examples. There are four types of 2-bit code, "00", "01", "10", and "11", but using only three of them, "00" is goo, "01" is par, If "10" is the cutoff, "0"
1 is prioritized over "00", "10" is prioritized over "01", and "00" is prioritized over "10". In other words, the priority is cyclically determined. The priority of is relatively determined.

In the present embodiment, 24 pieces of 2-bit codes as described above (that is, 48 bits in total) are arranged, and the "win / loss" is compared in order from the beginning, and the "win" is given priority. It controls as the right is given. The probability that all of these 24 code strings are equal is (1/3) ²⁴ =
It is 3.5 × 10 ⁻¹² , and it is practically rare that all the code strings are the same. This priority determination method can evenly distribute the channel usage right to each node.

There are only three types of rock-paper-scissors, goo, choki, and par, but with "11" as the fourth hand, it has priority over all other codes "00", "01", and "10". Can also be defined as If "11" is consecutively arranged from the head of the code string indicating the priority, the number of times "11" continues indicates the absolute priority of the packet. This absolute priority is given under a certain rule according to the nature of the packet or the contents of the packet.

In the above description, only the case of two-body collision was considered. If three or more multi-body collisions occur, there is no guarantee that the contents of the packet of the other party can be received correctly. Therefore, when a multi-body collision occurs, it is necessary to take a control procedure to wait for a random time after sending a jamming signal for a fixed time, as in the normal CSMA / CD protocol. Multi-body collision can be detected by the occurrence of a code foul, as described above.

In the state transition diagram of FIG. 5, a code violation occurs in each state of "carrier sense while continuing transmission", "comparison of priority while transmitting jamming signal", and "carrier sense (jamming signal) for a certain period of time". If you do (CR
Such a control procedure is adopted for V = 1).

Further, the node to be received constantly monitors the transmission channel, and when the packet addressed to the own node is recognized, it receives the packet and transmits it to the upper protocol. In this embodiment, the protocol on the receiving side is simple, so the state transition diagram is not shown.

Next, the protocol in the second embodiment of the communication system of the present invention will be described. That is, it is a protocol when a plurality of transmission channels are configured in parallel by wavelength multiplexing using a node as shown in FIG. FIG. 7 is a state transition diagram on the transmitting side in the second embodiment. The main difference between the state transition diagram shown in FIG. 7 and the state transition diagram shown in FIG. 5 is that “random selection of channels” is added in the state transition diagram shown in FIG.

When there is a transmission request from the upper layer protocol, the node to be transmitted simultaneously senses carriers of a plurality of channels and starts transmission on a channel randomly selected from the vacant channels. The following control procedure is almost the same as the state transition shown in FIG. 5, but the processing when the priority is lost is different. In the state transition shown in FIG. 5, the transmission is immediately stopped and then a standby for a random time is started. However, in the state transition shown in FIG. Immediately restart transmission on that channel (randomly select multiple channels). With the addition of this control procedure, even a node that has lost the contention for a circuit can perform transmission without waiting without waiting if there is another vacant channel. Therefore, the delay time required for communication can be reduced. In FIG. 7, the carrier sense is C
It is represented by the variable S _n . n represents the number of the transmission channel. Further, when CS _n = 1 in the state of “carrier sense (jamming signal) for a fixed time”, the transmission is stopped and then the carrier sense is resumed. This is a process for avoiding infinite conflicts due to some kind of error. In this embodiment, since there are a plurality of channels, instead of shifting to the standby state for the random time as shown in FIG. 5, a search for a vacant channel is performed. The waiting time can be reduced by adopting this processing.

In this protocol, the receiving side can use the same protocol as in the first embodiment. The difference from the first embodiment is that since a plurality of packets may be received at the same time, the throughput is simply increased to match it. There is no fundamental difference in the control procedure.

As a method of constructing a transmission channel having multi-channel bidirectionality, other than wavelength multiplexing, a method described in Japanese Patent Application No. 3-238328 filed by the present applicant. Can also be used. The above protocol is based on the above-mentioned Japanese Patent Application No. 3-23832.
It goes without saying that the present invention is also applicable to the network described in No. 8 specification.

The third embodiment of the communication system of the present invention is one in which a control protocol for confidential communication is added to the first embodiment of the communication system. FIG. 8 shows a state transition diagram on the transmitting side in the third embodiment of this communication system, and FIG. 9 shows a state transition diagram on the receiving side. In this embodiment, the packet structure is shown in FIG. 10 with a slight modification from the packet shown in FIG. 6 used in the previous two embodiments. That is, the header part of the packet includes the code 21 indicating the packet type. The code 21 indicating the packet type is 2 bits. Here, the code string 20 indicating the priority is 46 bits in total (23 in 2 bit pairs). Packet type PT
"00" represents a normal packet, "01" represents a packet on the transmitting side in confidential communication, and "10" represents a reply packet on the responding side in confidential communication. Note that "11" is undefined. PT in FIGS. 8 and 9 is a variable indicating this packet type, and R is a variable that becomes “1” when the source address of the received packet matches the address of the communication partner, and “0” otherwise. Yes, A is a variable that becomes “1” when the destination address of the received packet matches the address of the own node, and “0” otherwise.

The state transition diagram shown in FIG. 8 is based on the state transition diagram shown in FIG. 5, but the control procedure after the transmission of the header is different between the normal packet and the packet for confidential communication. ing. Normal packet (PT = 0
For 0), the same processing as the processing shown in the state transition diagram of FIG. 5 is performed. In the case of confidential communication (PT = 01), immediately after transmission of the header is completed, the jamming signal is transmitted. Transmission is started only when the source address of the packet received in this state matches the address of the communication partner (R = 1) and the packet type is PT = 10. When the source address of the received packet does not match the address of the communication partner (R = 0), a jamming signal is transmitted for a fixed time and then a standby for a random time is started. Further, as shown in FIG. 8, “carrier sense while transmitting jamming signal”,
Even when CRV = 1 in the "transmission" process, a jamming signal is transmitted for a fixed time and then a standby for a random time is started.

On the receiving side, as shown in the state transition diagram of FIG. 9, the transmission channel is monitored, the carrier is received (CS = 1), and the destination address of the received packet is detected. If the destination address does not match the address of the own node (A = 0), the process returns to channel monitoring. When the destination address of the received packet matches the address of its own node (A = 1), the type of the reply request, that is, the packet type is detected, and when it is a normal packet (PT = 00), the carrier is detected. As long as it is (CS = 1), the packet is fetched and transmitted to the upper layer protocol, and then the processing is ended.

When the addresses match (A = 1) and the packet type is PT = 01, that is, confidential communication, a jamming signal is started to be transmitted. Then, while transmitting the jamming signal, the packet is taken in and transmitted to the upper layer protocol. If a code violation occurs (CRV = 1) while the jamming signal is being transmitted, the transmission of the jamming signal is immediately stopped.

According to the above protocol, the node of the communication partner sends a jamming signal, so that two-body collision is intentionally caused on the transmission channel, and as a result, the interception of the communication content by other nodes can be prevented. It is possible.

The fourth embodiment of the communication system of the present invention is a modification of the above-mentioned third embodiment. In the above-described third embodiment, since the two-body collision is intentionally generated, there is a problem that the probability of the three-body collision increases, and the usefulness of the mechanism for avoiding the two-body collision is reduced. .. On the contrary,
In the fourth embodiment, the node that should respond does not reply immediately after receiving the packet, but waits for the round trip propagation delay time of the broadcast bus before starting the reply. By adding this procedure, it is possible to prevent an increase in the probability of occurrence of a three-body collision and not to reduce the usefulness of the two-body collision avoidance mechanism. FIG. 11 shows a state transition diagram on the response node side according to the fourth embodiment. The state transition diagram on the transmitting side is the same as in FIG.

The state transition diagram shown in FIG. 11 is based on the state transition diagram shown in FIG. 9. However, in the case of PT = “01”, the “reply request detection” process and the “jamming signal transmission” are performed. While receiving, while processing, "wait for a certain period of time"
The difference is that the process of is added.

The fifth embodiment of the communication system of the present invention is obtained by adding the control procedure of confidential communication shown in the third embodiment to the multi-channel communication system shown in the second embodiment. This is a communication protocol including all the functions described so far. FIG. 12 is a state transition diagram of a transmission side of a communication protocol having both functions of avoiding collision of two bodies and secret communication in a network having multi-channel transmission channels. The state transition diagram shown in FIG. 12 is the state transition diagram shown in FIG. 7 in the second embodiment and FIG. 8 in the third embodiment.
12 is combined with the state transition diagram shown in FIG. 12, and can be understood more easily than the description of each embodiment given above.
The description of the state transition diagram in FIG. The processing on the response side can be handled by the processing of the state transition diagram shown in FIG. 9 or 11.

[0065]

As described above, according to the present invention,
By distinguishing between a two-body collision and a multiple-body collision and performing different control between the two-body collision and the multiple-body collision, it is possible to effectively avoid the two-body collision and improve the line utilization efficiency. The delay time required for packet transmission can be reduced on average.

[Brief description of drawings]

FIG. 1 is a schematic diagram of an optical communication network configured by using an interconnectable 8-terminal star coupler applied to the present invention.

FIG. 2 is a schematic diagram showing the configuration of a node applied to the first embodiment of the optical communication network of the present invention.

3 is a plan view showing a circuit configuration of an 8-terminal star coupler used in the optical communication network shown in FIG. 1. FIG.

FIG. 4 is a schematic diagram showing a configuration of a node applied to a second embodiment of the optical communication network of the present invention which constitutes a plurality of bidirectional transmission channels by wavelength division multiplexing.

FIG. 5 is a state transition diagram of a transmission side of a communication protocol for avoiding a collision between two bodies on a network having a single bidirectional transmission channel.

FIG. 6 is a schematic diagram showing the format of a packet used by the communication protocol of the present invention.

FIG. 7 is a state transition diagram of a transmission side of a communication protocol for avoiding collision between two bodies on a network having a multi-channel bidirectional transmission channel.

FIG. 8 is a state transition diagram on the transmission side of a communication protocol that performs both two-body collision avoidance and confidential communication on a network having a single bidirectional transmission channel.

FIG. 9 is a state transition diagram on the response side of a communication protocol that performs both two-body collision avoidance and confidential communication.

10 is a schematic diagram showing a packet format in which a code indicating a packet type is added to the packet format shown in FIG.

FIG. 11 is a state transition diagram on the response side of a communication protocol that performs both collision avoidance and confidential communication when a certain waiting time is provided when responding.

FIG. 12 is a state transition diagram of a transmission side of a communication protocol for performing both two-body collision avoiding confidential communication on a network having a multi-channel bidirectional transmission channel.

FIG. 13 is a configuration diagram showing a configuration example of a conventional optical LAN using a star coupler.

FIG. 14 is a timing chart showing the principle of the code foul play method.

FIG. 15 is a graph showing a ratio of a two-body collision probability and a three-body collision probability with respect to the number of calls per unit time.

[Explanation of symbols]

1 node, 2 star coupler, 3 bidirectional optical amplifier, 4 optical fiber, 5 node body, 6, 6a-6
d communication interface, 7 optical demultiplexer / multiplexer, 8 light emission source, 9 light receiver, 10 transmission port, 11 reception port, 12 carrier detection device, 13 code foul detection device, 14 terminals, 15 light with branching ratio 7: 1 Minute / multiplexer,
16 optical demultiplexer / multiplexer with 2: 1 branching ratio, 17 1: 1 branching ratio
1 optical demultiplexer / multiplexer, 18 substrate, 19 wavelength multiplexing multiplexer, 20 code string indicating priority, 21 packet type code, 22 light emitting element, 23 light receiving element,
24 terminals, 25 mixing rod type star coupler, 26a, 26b optical fiber, 27 node, λ
a to λd Wavelength of signal light corresponding to the communication interfaces 6a to 6d, CS Variable indicating carrier sense, CS
n A variable indicating carrier sense in a multi-channel system, a variable indicating the priority of the packet issued by the P _i own node, a variable indicating the priority of the packet issued by the P _o partner node, and a CRV code foul occurrence Variable, CRV
_n Variable indicating occurrence of code violation in multi-channel system, R Variable indicating whether destination address of received packet matches address of own node, Variable indicating PT packet type

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Office reference number FI technical display location 7341-5K H04L 11/00 320

Claims (8)

[Claims] 1. In a network in which a plurality of nodes are connected to a bidirectional broadcast bus and communication between the nodes is performed by packets, each node includes means for detecting a carrier on the broadcast bus. , An optical communication network comprising means for detecting the interference state of a received signal. 2. In a network comprising a plurality of bidirectional broadcast buses and a plurality of nodes connected to all or part of the broadcast buses, each node is connected to its own node. An optical communication network comprising means for detecting a carrier on a news bus and means for detecting an interference state of a received signal. 3. The optical communication network according to claim 1, wherein the means for detecting the interference state of the received signal is a detection means based on the code violation method. 4. The communication system in the optical communication network according to claim 1, wherein the means for detecting the carrier detects a two-body collision, and the means for detecting an interference state of the received signal. A communication method characterized by detecting a collision of three or more bodies. 5. A packet issued by each node is provided with a code string indicating the priority of the packet, and when a two-body collision occurs, the right to use the broadcast bus is determined by comparing the code strings indicating the priority. The communication method according to claim 4, wherein the communication method is determined. 6. The communication system according to claim 5, wherein all or a part of the code string indicating the priority is a code string for cyclically determining the priority. 7. The method according to claim 4, wherein the node to be returned in response to the signal from the transmitting node intentionally causes two-body collision to give confidentiality to the communication. The communication system according to claim 6. 8. The communication system according to claim 7, wherein the node to be returned emits a signal after a delay time of a constant time to reduce the probability of occurrence of multi-body collision.