WO2014167703A1

WO2014167703A1 - Network system, communication method, and network device

Info

Publication number: WO2014167703A1
Application number: PCT/JP2013/061030
Authority: WO
Inventors: 江端　智一; 義則望月; 小泉　稔
Original assignee: 株式会社日立製作所
Priority date: 2013-04-12
Filing date: 2013-04-12
Publication date: 2014-10-16

Abstract

Each of a plurality of network devices is provided with: a clock for providing a first time synchronized with the other network devices; and a plurality of ports. A schedule server determines a route through which the plurality of network devices are to transfer a frame, generates a first schedule indicating a combination of the ports through which the determined route passes, and a time at which the frame is to be transferred, and delivers the first schedule to the plurality of network devices. Each of the plurality of network devices receives the first schedule from the schedule server, refers to the first schedule, and, in cases when the first time is the time indicated by the first schedule, outputs, unchanged, the frame inputted to one port in the combination of ports indicated by the first schedule, from another port in the combination.

Description

Network system, communication method, and network apparatus

The present invention relates to a network system.

When the introduction of the network to the manufacturing site started in the mid-1980s, various automation devices including PLC (Programmable Logic Controller) were used stand-alone at the manufacturing site.

After that, factory automation (FA) was promoted by personal computer of automation equipment using serial communication by RS-232C or RS-422 etc. which was standard equipment of personal computer. However, when using FA for serial communication, there is a problem that the connection between the personal computer and the automation device is one-to-one, and the transmission distance between the personal computer and the automation device is limited to a maximum of 15 m. For this reason, a product that solves these problems in FA by using a converter of RS-422 or RS-485 has appeared.

At this time, network systems dedicated to FA appeared, but PLC vendors mainly used their own methods to connect their PLCs to personal computers. For this reason, PLCs of different vendors and their personal computers cannot be interconnected. In addition, a dedicated network unique to the PLC vendor has been used for the FA network.

When using FA developed by a single vendor, users were unable to use a wide variety of I / O devices. On the other hand, there was an increasing demand for using multi-vendor devices, so standardization began in the late 1980s. Various network proposals for networking have been proposed.

In the first half of 1990s, Ethernet was used as a local area network (LAN) communication medium in an information network, but it was used for FA because there was a problem in the performance of transferring frames in real time. There were few. However, the advent of switching hubs has dramatically improved this problem. In addition, since the Internet has become widespread and switching hubs can be obtained at low cost, Ethernet has been used for industrial use and Ethernet has been introduced into FA.

Ethernet used for industrial use has been applied to networks for connecting controllers (PLCs, etc.) in FA since the mid-1990s. Furthermore, after 2000, Ethernet has also been applied to field networks connecting actuators and sensors.

The control system for controlling the automation equipment includes a network using Ethernet used for industrial use, a master, and a plurality of slaves. The master in the control system transmits the frame storing the command to the slave via the network. Then, the slave executes an instruction read from the received frame, and stores and transfers necessary information in the received frame.

Conventionally, a technique related to EtherCAT has been proposed (see, for example, Non-Patent Document 1), and EtherCAT is a standard for a fieldbus of a network using Ethernet.

An EtherCat master (hereinafter referred to as “master”) in EtherCAT generates one Ethernet frame (hereinafter referred to as “Ethernet frame”) for transmitting control information to a plurality of EtherCat slaves (hereinafter referred to as “slave”). Specifically, the master stores control information for controlling a plurality of slaves in a predetermined area included in one ether frame. Each of the plurality of slaves reads the control information from the received Ethernet frame, and controls an actuator or a sensor connected to the slave according to the control information.

Also, the slave stores information acquired by the sensor in a predetermined area of the ether frame received from the master or another slave, and transfers the ether frame after storing the information to another slave or master. Thereby, all of the plurality of slaves refer to one Ether frame transmitted from the master, and store information in the transmitted Ether frame.

Then, after the reference or information storage is completed by all the slaves, one Ether frame is transmitted to the master again. Thereby, the control process for one cycle in EtherCAT is completed.

The frame length of the Ether frame is 1518 bytes (12144 bits). When an Ethernet frame is transferred through a 100 Mbps network at a speed close to the speed of light (about 300,000 km / s), the total length of the Ethernet frame in the network reaches 36.4 km.

Also, when an Ethernet frame is transferred through a 1 Gbps network at a speed close to high speed, the total length of the Ethernet frame in the network reaches 3.64 km.

The slave in EtherCAT refers to and stores information in an On-the-Fly method and in bit units in an ether frame having a total length of 36.4 km transmitted from the master. The method of referring to and storing the Ether frame in the On-the-Fly method and in bit units is that the slave receives the Ether frame bit by bit, references the received bit without buffering it, and receives it. In this method, information is stored in bits and output to the next transfer destination.

According to this method, since the slave does not buffer or route the Ethernet frame, it can transfer the Ethernet frame at a very high speed. Specifically, according to this method, the time required for the Ethernet frame to pass through one slave is calculated to be 121.44 μsec. According to the EtherCat standard, each slave can transfer an Ether frame in units of 125 μs, and can transfer a frame in real time from the master to the slave.

As described above, control systems implemented by EtherCat and the like have conventionally realized high-precision real-time processing and short-term cyclic communication using a fixed-length ether frame. Then, the master can cyclically control the transmission of the ether frame, so that the entire control system including the control network can be synchronized.

When EtherCAT technology is applied to a control system such as a railway system, an electric power system, or a water supply system that has a wide service area, a developer uses a wide-range control LAN ranging from several kilometers to 100 km as a network of control systems. It was necessary to install. And the developer needed to construct | assemble the control system using the installed wide-area control LAN.

However, in such a wide-area control LAN, all slaves cyclically receive one Ethernet frame. Therefore, each of the masters in two or more different wide area control systems cannot synchronize each of the slaves using one wide area control LAN.

As a result, the developer has to install a plurality of wide-area control LANs for different wide-area control systems, resulting in an increase in development cost. Moreover, even if a plurality of control LANs are installed at the same place, the plurality of control LANs cannot share one installed equipment, resulting in an increase in development cost and an increase in maintenance cost. It was.

Therefore, for example, a method in which wide-area control systems having different properties such as a railway system, a power system, or a water supply system share a general-purpose carrier wide-area network has been conventionally considered.

However, unlike a private wide-area LAN, a general-purpose carrier wide-area network has a large number of communication devices such as routers and switches, and needs to accommodate a plurality of user terminals simultaneously. For this reason, in a general-purpose carrier wide area network, frame transfer delay, fluctuation, loss, or the like that cannot be predicted may occur.

Especially, in a wireless network installed in a wide area, such a transfer delay is likely to occur. For this reason, it is difficult to transfer a frame in real time in a general-purpose carrier wide area network.

In the prior art, a technique has been proposed in which time-synchronized edge nodes transmit and receive frames according to a time table to perform traffic leveling and avoid congestion. Further, a technique for further improving the leveling efficiency by providing a buffer in the edge node has been proposed (see, for example, paragraphs [0029] to [0034] of Patent Document 1).

Furthermore, conventionally, a technique called circuit switching has been proposed (see, for example, Non-Patent Document 2). The circuit switching in Non-Patent Document 2 is a communication switching method in which a physical or virtual transmission path is set and the line is occupied from the start to the end of communication, or data communication using the method. Since circuit switching does not require data to be stored and retransmitted unlike packet communication, it is only necessary to provide switching equipment having a simple function. Further, when line switching is used, the line is occupied between the terminating devices, so the connection speed and QoS are guaranteed, and transmission delay due to congestion or the like does not occur in principle.

JP 2002-314590 A

In the control LAN using the technology of Non-Patent Document 1, data is cyclically transmitted to all slaves. When such a plurality of control LANs are used, two or more different control LANs cannot share one network system. This is because when one network system is shared, a transfer delay or the like occurs in any of a plurality of control LANs, and security cannot be strictly maintained.

For example, when three control LANs of a railway operation management system with a transfer cycle of 200 ms, a power system monitoring and control system with a transfer cycle of 300 ms, and a water supply management system with a transfer cycle of 1000 ms are implemented in one network, Since the transfer cycles in one control LAN are different, data of a plurality of control LANs are transmitted at the same timing in any of the transfer cycles. Then, any of the plurality of data transmitted at the same timing is buffered in the network device, and there is a possibility that data transfer delay occurs in any of the systems.

In addition, when the above-described three control LANs are mounted on a single network, data is shared between different control LANs, resulting in a security problem.

Also, since the frame format of data transferred by each of the plurality of control LANs is different, it is necessary for the network device to identify different frame formats. Furthermore, even though it is necessary for one control LAN, unnecessary data is transmitted to another control LAN, so that network resources are wasted.

On the other hand, when the technique of Patent Document 1 is used, in a network composed of a large number of communication devices such as routers or switches, the time during which the edge nodes transfer frames is controlled between the edge nodes, thereby being effective by congestion Avoid narrowing the bandwidth used. For this reason, when the technique of Patent Document 1 is used, data transfer can be performed accurately and efficiently.

However, in a control system, real-time transfer is an absolute requirement in many cases, and there is a problem that it is not sufficient to guarantee statistical or relative real-time characteristics as in the technique of Patent Document 1. .

Here, the real-time property means that the frame output from the transmission source is guaranteed to arrive at the transmission destination within a predetermined time regardless of the network usage state.

Furthermore, when the circuit switching method of Non-Patent Document 2 is used, a plurality of terminals cannot share a line regardless of the amount of data flowing through the line. For this reason, when the technique of Non-Patent Document 2 is used, the network utilization efficiency is poor, and it is relatively difficult for a plurality of terminals to communicate with each other at different speeds. In addition, since the communication between the terminal devices continues to occupy the line, it is difficult to dynamically switch the route, and it also takes time to switch even when the route is switched.

A typical example of circuit switching described in Non-Patent Document 2 is a subscriber telephone network “Public Switched Telephone Network (PSTN)”. PSTN is SS7: Signaling System No. 7 (common line signal No. 7) is used. SS7 notifies the line control device of the switching control command, but cannot switch the line at the speed of accuracy required by the control network. This is because a switching command is issued to the line control device every time a transmission destination address (telephone number or the like) transmitted by the transmission source is input.

An object of the present invention is to provide a system that effectively uses a network, in which a plurality of different control systems are mounted on one network system, and real-time communication is guaranteed in each of the plurality of control systems. is there.

According to a representative aspect of the present invention, a network system having a plurality of network devices and a schedule server connected to the plurality of network devices, each of the plurality of network devices being another network device. A clock providing a first time synchronized with a plurality of ports, and a plurality of ports, wherein the schedule server determines a route through which the plurality of network devices transfer frames, and the determined route passes A resource allocation unit that generates a first schedule indicating a combination of the ports to be transmitted and a time at which the frame is transferred, and a distribution unit that distributes the first schedule to the plurality of network devices, Each of the plurality of network devices receives the first schedule from the schedule server. With reference to the receiving unit and the first schedule, when the first time is the time indicated by the first schedule, it is input to one port of the port combination indicated by the first schedule And a switch unit that outputs the frame as it is from the other port of the combination.

According to an embodiment of the present invention, it is possible to install a plurality of network systems that guarantee real-time communication in one network system.

Issues, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

It is explanatory drawing which shows the system which mounted EtherCat and VLAN of the present Example 1. It is explanatory drawing which shows the outline | summary of a process of the network system of a present Example 1. FIG. It is explanatory drawing which shows the structure of the real-time network of the present Example 1. It is explanatory drawing which shows the structure of the communication system which has several real time networks of the present Example 1. FIG. It is a block diagram which shows the hardware constitutions of the schedule server of the present Example 1. FIG. 2 is a block diagram illustrating a hardware configuration of a switch according to the first embodiment. FIG. 3 is a block diagram illustrating a hardware configuration of a master according to the first embodiment. 2 is a block diagram illustrating a hardware configuration of a communication interface according to the first embodiment. FIG. It is explanatory drawing which shows the Ethernet frame of this Example 1, and two types of control frames. It is explanatory drawing which shows the construction information of the real-time network of the present Example 1. It is explanatory drawing which shows the structure of the software of the schedule server of a present Example 1, and a process of software. It is a flowchart which shows the process of the resource allocation module of a present Example 1. It is explanatory drawing which shows the whole time table of the present Example 1. FIG. It is explanatory drawing which shows the connection relation of the switch of the present Example 1. FIG. It is explanatory drawing which shows the combination information of the port in the switch of a present Example 1. FIG. It is explanatory drawing which shows the separate time table of the switch of a present Example 1. FIG. It is explanatory drawing which shows the separate timetable of the master of the present Example 1. It is explanatory drawing which shows the structure and process of the software of the switch of a present Example 1. FIG. It is a flowchart which shows the process of the switching module of the switch of the present Example 1. It is explanatory drawing which shows the structure and process of the software of the communication interface of the master of a present Example 1. FIG. It is explanatory drawing which shows the queue which the memory of the master of the present Example 1 has. It is a flowchart which shows the process of the switching module of the master of the present Example 1. It is explanatory drawing which shows the real-time network to which the path | route which does not overlap physically of Example 1 was allocated. It is explanatory drawing which shows the real-time network with which the physical path | route of a present Example 1 overlaps. It is explanatory drawing which shows the whole time table of the present Example 1. FIG. It is explanatory drawing which shows the separate time table of the switch and master after changing the effective period of the real-time network of the present Example 1. It is explanatory drawing which shows the communication system applied to the railway operation management system of the present Example 2. It is explanatory drawing which shows the whole time table of the present Example 2. It is explanatory drawing which shows a part of individual time table of the present Example 2. FIG.

Hereinafter, description will be made with reference to the drawings.

The outline of the first embodiment will be described with reference to FIGS. 1 and 2 by taking a system using EtherCAT and Virtual LANs (VLAN) as an example.

FIG. 1 is an explanatory diagram showing a system in which EtherCat and VLAN according to the first embodiment are mounted.

The system 1001 and the system 1002 shown in FIG. 1 control and synchronize devices using EtherCat. The system 1001 and the system 1002 have a plurality of slaves (a plurality of EC slaves) in EtherCat.

The system 1001 is the control system A, and has a master for the control system A (EC master A). The system 1002 is the control system B, and has a master for the control system B (EC master B). Further, the cycle in which the EC master A transmits an ether frame to the EC slave is different from the cycle in which the EC master B transmits an ether frame to the EC slave.

The EC slave uses the On The Fly method to transfer an Ether frame by outputting 1 input and 1 output for each bit. For this reason, the EC slave can transfer the Ethernet frame in real time.

However, since EC master A and EC master B transmit an ether frame in 125 μs, if the transfer cycle is 100 ms, the network of system 1001 and system 1002 is unused in the period of 799/800. is there. For this reason, network resources may not be used effectively in a control system based on EtherCat.

In addition, the developer cannot integrate two or more control systems, such as the system 1001 and the system 1002, having different periods for cyclically transmitting the Ethernet frame into one network. Specifically, when the EC master A and the EC master B repeat the transmission of the ether frame with different periods, the ether frame transmitted from the EC master A and the ether frame transmitted from the EC master B are equal to each other. One network device may arrive at the same time and either one may be buffered. For this reason, since the control system using EtherCat needs to install a network for each control system, the development cost may increase.

In order to solve such problems as utilization of network resources and increase in development cost, two or more control systems may be implemented by VLAN. A VLAN 1011 illustrated in FIG. 1 is a VLAN for mounting the system 1001, and a VLAN 1012 is a VLAN for mounting the system 1002.

Thus, by using VLAN switches, two or more control systems can coexist in one network.

However, the VLAN switch cannot transfer the Ethernet frame bit by bit using the On the Fly method. This is because the VLAN switch cannot receive an Ethernet frame from the same port at the same time and cannot simultaneously output an Ethernet frame to the same port. The VLAN switch waits for an Ethernet frame. This is because (buffering) is required. If the VLAN switch buffers the ether frame even once, the real-time property in the transfer of the ether frame is impaired.

FIG. 2 is an explanatory diagram illustrating an outline of processing of the communication system according to the first embodiment.

The communication system of the first embodiment is implemented by a single network. The control system according to the first embodiment includes a plurality of different control systems. For example, two control systems of a T-LAN (Time-multiplexed control LAN) 1021 and a T-LAN 1022 are mounted.

In this embodiment, the T-LAN is a real-time network that is effective for a predetermined period. One real-time network implements one control system.

In the network system of the first embodiment, the EC master A, the EC master B, and the plurality of switches are synchronized in time. The T-LAN 1021 is a network configured in a predetermined period. The T-LAN 1022 is a network in which the T-LAN 1022 is not configured and configured in a predetermined period.

The period in which the T-LAN 1021 is valid and the period in which the T-LAN 1022 is valid are alternately switched at a predetermined time.

In the first embodiment, each switch is set in advance so that the Ethernet frame transmitted from the EC master A is transferred only to the device of the control system A while the T-LAN 1021 is valid. This creates a physical (electrical or optical) conductor dedicated to the control system A between the EC master A and the switch.

That is, a physical conductor dedicated to the control system A is mounted on the network system during a predetermined period. For this reason, when an Ethernet frame is transmitted from the EC master A, the Ethernet frame passes through a conductor dedicated to the control system A and is transferred without reading out tag and address information in the switch. For this reason, the time for which the switch buffers the Ethernet frame is extremely short enough to be ignored.

As a result, the T-LAN 1021 can transfer an Ether frame in bit units using the On-the-Fly method, and can transfer an Ethernet frame in real time as much as EtherCat. Then, like the T-LAN 1021, the T-LAN 1022 can transfer an Ethernet frame in bit units using the On the Fly method in a predetermined period.

In order to implement such T-LAN 1021 and T-LAN 1022, it is necessary to synchronize time between the master and the switch with high accuracy. For example, when the T-LAN 1021 and the T-LAN 1022 are installed in a control system (power management system or railway management system, etc.) installed in a range of about 100 square kilometers, the master and the switch are GPS (Global Positioning System). Etc. need to be time synchronized.

In a network including a plurality of routes, such as the T-LAN 1021 and the T-LAN 1022 shown in FIG. By sequentially changing the route, it is possible to configure a plurality of T-LANs in one network. As described above, EtherCat can transfer a real-time Ether frame in units of 125 μs, and in principle, can switch 8000 times per second. For this reason, when all the T-LANs to be mounted communicate cyclically in units of one second, 8000 types of T-LANs are mounted on one network.

For this reason, according to the embodiment described below, since the buffering of the Ethernet frame is not performed or is rarely performed in the network device, delay and fluctuation do not occur, and the scale of the network is affected. This makes it possible to realize a real-time control network that enables a plurality of communications.

FIG. 3 is an explanatory diagram illustrating the configuration of the real-time network according to the first embodiment.

The real-time network according to the first embodiment includes a schedule server 20, one master controller 60 (hereinafter referred to as master 60), at least one control Ethernet 8 (hereinafter referred to as control LAN 8), and at least one slave device 7 (hereinafter referred to as master LAN). , Referred to as slave 7).

The master 60 is an EC master and controls an apparatus controlled in the real-time network by transmitting an instruction or the like. The master 60 is cascade-connected to the slave 7 by the control LAN 8.

The slave 7 is an EC slave and is directly connected to a device controlled by the master 60. The slave 7 is an interface between the device controlled by the master 60 and the master 60.

The devices controlled by the real-time network include, for example, an actuator, a servo drive, an I / O, a temperature / humidity meter, a temperature / humidity regulator, an inverter, and a stepping motor.

The master 60 has a network interface, and transmits and receives a predetermined control frame at a predetermined time. The master 60 outputs an ether frame (control frame) in which control information such as a request is stored toward the slave 7.

Each of the slaves 7 receives the control frames transmitted from the master 60 in the order of connection. Each of the slaves 7 reads out control information from an area assigned to each of the slaves 7 in the control frame transmitted from the master 60.

Then, the slave 7 executes processing on the device connected to the slave 7 according to the read control information. Then, the slave 7 stores the execution result of the process, sensor data, and the like in an area allocated to the slave 7 in the control frame.

Then, the slave 7 transfers the control frame storing the execution result and the like to the other slave 7 connected to itself. After all the slaves 7 store the execution result or the like in the control frame, the control frame is transmitted to the master 60.

The slave 7 performs the process of reading the data included in the control frame and storing the data in the control frame directly on the control frame bit by bit without buffering the control frame. For this reason, the control system of Example 1 is real-time and can perform high-speed control processing.

The schedule server 20 is connected to the master 60 by Ethernet or RS-232C. The schedule server 20 assigns each of the slaves 7 to the area of the control frame, and sets a cycle in which the master 60 cyclically transfers the control frame to the slave 7.

FIG. 4 is an explanatory diagram illustrating a configuration of a communication system having a plurality of real-time networks according to the first embodiment.

The communication system shown in FIG. 4 includes a schedule server 20, a plurality of masters 60 (master 60a to master 60c), a plurality of switches 10 (switch 10a to switch 10d), a plurality of control LANs 8 (control LAN 8a to control LAN 8c), and a plurality of switches. It has a slave 7 (slave 7a to slave 7c).

The switch 10 is a network device for connecting the other switch 10, the master 60 and the slave 7. The communication system according to the present embodiment includes the switch 10 as a network device, but may include a router having a switch function as the network device.

The switches 10a to 10d shown in FIG. 4 constitute a star-shaped network. However, the switches 10a to 10d of the first embodiment may be connected by any other topology.

Further, the devices included in the communication system may be connected by any communication medium as long as they can communicate with each other, and may communicate by wire (optical signal or electric signal or the like) or wirelessly.

The switch 10 has a plurality of ports for connecting to the schedule server 20, the master 60, or the control LAN 8. The switch 10 electrically or optically connects two or more ports in a predetermined period according to a time table held by itself. Accordingly, the switch 10 transfers the input frame according to the transmission source of the input frame and the time table.

SW_A to SW_D are respectively assigned as identifiers to the switches 10a to 10d of the first embodiment. Further, Cont_α to Cont_γ are assigned as identifiers to the masters 60a to 60c of the first embodiment. Also, CLAN_1, CLAN_2, and CLAN_4 are respectively assigned as identifiers to the control LAN 8a to the control LAN 8c of the first embodiment.

In the embodiment described below, the network connecting the control LAN 8, the master 60, and the switch 10 is Ethernet or wide area Ethernet, but may be any network as long as communication is possible. The network of this embodiment may be a WAN, for example.

Further, the network of the present embodiment may be a network having two or more physical or virtual communication lines capable of bidirectional communication. A description of communication lines capable of bidirectional communication is not shown.

The switch 10 and the master 60 of this embodiment are time synchronized with high accuracy. The switch 10 and the master 60 may synchronize the time using, for example, GPS, or may synchronize the time using NTP or IEEE 1588. The switch 10 and the master 60 may synchronize time by any technique as long as the technique ensures the necessary accuracy of time synchronization.

In the embodiment shown below, the switch 10 and the master 60 use a conventional technique for synchronizing time using GPS ("The Role of GPS in Precision Time and Frequency Dissimilation" [online] [January 12, 2013 search. ], See the Internet).

FIG. 5 is a block diagram illustrating a hardware configuration of the schedule server 20 according to the first embodiment.

The schedule server 20 includes devices such as a processor 22, a memory 23, an input controller 26, an Ethernet network interface card (hereinafter referred to as NIC) 21, and the like. Each device of the schedule server 20 is connected by a bus 30.

The processor 22 is an arithmetic unit such as a CPU. The memory 23 holds data and programs.

The NIC 21 is an interface for connecting to the switch 10 and the master 60. The NIC 21 is connected to the switch 10 by Ethernet, but may be connected by any method, and may be connected by RS-232C or the like.

The input controller 26 acquires data input from the user or operator by the keyboard 24 or the mouse 25. The monitor controller 27 has a function of outputting a result of processing by the processor 22 to an output device.

The monitor controller 27 is connected to the monitor 28. The monitor 28 is a display and displays data transmitted from the monitor controller 27. The monitor controller 27 may be connected to a printer, and may output data by any method.

FIG. 6 is a block diagram illustrating a hardware configuration of the switch 10b according to the first embodiment.

FIG. 6 shows the hardware configuration of the switch 10b, but all the switches 10 have the same hardware configuration. Therefore, the hardware configuration of the switch 10 will be described below using the hardware configuration of the switch 10b shown in FIG.

The switch 10 includes devices such as a memory 13, a GPS receiver 15, a synthesizer 17, a microcomputer (microcomputer) 19, and a plurality of NICs 11. Each device of the switch 10 is connected to the microcomputer 19. The NIC 11 is connected to the control LAN 8, the switch 10, and the schedule server 20.

The GPS receiver 15 is connected to a GPS antenna. The GPS receiver 15 also functions as a 1PPS generator that generates a precise 1PPS (Pulse Per Second (pulse at 1 second interval)) signal. The GPS receiver 15 transmits the generated 1PPS signal to the synthesizer 17. The GPS receiver 15 holds the current time received from the GPS antenna and provides it to the microcomputer 19 as necessary.

The synthesizer 17 generates a clock signal having a clock frequency required by the microcomputer 19 based on the 1PPS signal transmitted from the GPS receiver 15. The microcomputer 19 performs processing using the current time provided from the GPS receiver 15 and the clock signal synchronized with the 1PPS signal, so that the processing of each of the switches 10 and the master 60 is synchronized.

The microcomputer 19 has at least one processor and transfers the received frame. The memory 13 holds data, programs, and the like.

The NIC 11 has a plurality of ports and receives or transmits frames. The NIC 11 is an interface for connecting to the switch 10, the schedule server 20, the master 60 and the slave 7.

FIG. 7 is a block diagram illustrating a hardware configuration of the master 60 according to the first embodiment.

The master 60 includes devices such as a communication interface 68, a processor 62, a memory 63, an input controller 65, and a monitor controller 64. Each of the devices of the master 60 is connected by a bus 61.

The processor 62 is an arithmetic unit such as a CPU. The memory 63 holds data and programs. The communication interface 68 is an interface for connecting to the schedule server 20, the switch 10, and the slave 7.

The input controller 65 acquires data input from the user or the operator by the keyboard 66 or the mouse 67. The input controller 65 may input data by any method other than the keyboard 66 or the mouse 67.

The monitor controller 64 has a function of outputting the processing result by the processor 62 to the output device. The monitor controller 64 is connected to the monitor 69. The monitor 69 is a display and displays data transmitted from the monitor controller 64. The monitor controller 64 may be connected to a printer and may output data by any method.

FIG. 8 is a block diagram illustrating a hardware configuration of the communication interface 68 according to the first embodiment.

The communication interface 68 includes a microcomputer 686, a memory 689, a NIC 688, a synthesizer 684, and a GPS receiver 682.

The microcomputer 686 has at least one processor, and transfers a frame received via the NIC 688 to the bus 61. The microcomputer 686 is connected to the memory 689, the NIC 688, and the synthesizer 684.

The microcomputer 686 and the NIC 688 are connected by two paths, a transmission path and a reception path. The functions and effects of the synthesizer 684 and the GPS receiver 682 are the same as the functions and effects of the synthesizer 17 and the GPS receiver 15 of the switch 10b shown in FIG. Specifically, the process of the master 60 and the process of the switch 10 are synchronized by the synthesizer 684 and the GPS receiver 682 generating a clock signal based on the GPS function. The GPS receiver 682 holds the current time.

The communication interface 68 in this embodiment includes the microcomputer 686, but may have any device as long as it can transfer and transmit frames. For example, even if the communication interface 68 transfers a frame using an integrated circuit such as an FPGA (Field-Programmable Gate Array) instead of a microcomputer, the effect and function of the master 60 of this embodiment are not affected. Don't give.

FIG. 9 is an explanatory diagram showing the Ethernet frame and two types of control frames of the first embodiment.

The Ethernet frame shown in FIG. 9 is an example of a frame used in the communication system of the present embodiment. The ether frame shown in FIG. 9 includes a preamble 611, a transmission destination address 612, a transmission source address 613, a type 614, data 615, and a frame check sequence 616.

The preamble 611 stores information for allowing the network interface to recognize the start of transmission of the Ethernet frame. The transmission destination address 612 indicates the destination of the Ether frame, and the transmission source address 613 indicates the transmission source of the Ether frame. A type 614 indicates the type of the data 615, and the data 615 includes data output from an application or the like. The frame check sequence 616 includes information for performing error control.

The frame format of the Ethernet frame is, for example, a frame format of a conventional document (Cisco Systems, Inc. “Ethernet Technologies”, [online], October 16, 2012, [Search January 12, 2013], Internet). May be used.

The control frame is an Ether frame in which control information in the real-time network is included in the data 615. The master 60 in this embodiment transmits two types of control frames. Data 615a and data 615b shown in FIG. 9 are data 615 of the control frame.

The data 615a includes a header area 6151 and a plurality of data areas 6152 (data area 6152-1 to data area 6152-n: n is an arbitrary integer). The header area 6151 stores information that allows the master 60 to recognize the type of the ether frame to the slave 7. Each of the plurality of data areas 6152 is assigned to each of the plurality of slaves 7.

The master 60 stores control information such as requests to each of the slaves 7 in the data area 6152, and each of the slaves 7 stores sensor information and the like in the data area 6152 assigned to itself. Information is stored in the data area 6152 in an array that the slave 7 can read bit by bit.

The data 615b includes control information in a format predetermined between the master 60 and the slave 7. For this reason, an operator of the communication system or the like may set the master 60 to generate a necessary number of control frames according to the type of control.

As the frame format of the control frame data 615, for example, the frame format published in Non-Patent Document 1 may be used.

FIG. 10 is an explanatory diagram showing the construction information 29 of the real-time network according to the first embodiment.

The construction information 29 is information input to the schedule server 20 by the operator of the communication system. The construction information 29 indicates a plurality of real-time networks included in the communication system and a cycle at which a control frame is transmitted in each of the real-time networks. The construction information 29 may include any number of information related to the real-time network.

The construction information 29 includes a T-LAN 291, a master 292, a control LAN 293, a communication period 294, and a communication cycle 295. T-LAN 291 indicates an identifier of a real-time network included in the communication system. The T-LAN 291 shown in FIG. 10 includes T-LAN (# 1), T-LAN (# 2), and T-LAN (# 3) as real-time network identifiers.

The master 292 indicates an identifier of one master 60 in the real-time network indicated by the T-LAN 291. The control LAN 293 indicates an identifier of at least one control LAN 8 to which the slave 7 is connected in the real-time network indicated by the T-LAN 291.

The operator inputs the master 292 and the control LAN 293 in combination. Specifically, the operator combines one identifier of the master 292 with at least one identifier of the control LAN 293 and sets a value in the master 292 and the control LAN 293. Thus, a T-LAN that is valid for a predetermined period is constructed.

The communication period 294 indicates a period during which the real-time network indicated by the T-LAN 291 is valid in the communication system. The communication cycle 295 indicates a cycle in which the real-time network indicated by the T-LAN 291 is valid.

The entries of the construction information 29 are arranged in the order of the real-time network with the highest priority. The priority in the real-time network is arbitrarily determined by the operator of this embodiment. A higher-priority real-time network guarantees control frame transmission.

In the construction information 29 shown in FIG. 10, the T-LAN (# 1) real-time network has the highest priority in the communication system, and the T-LAN (# 3) real-time network has the lowest priority in the communication system. It shows that.

According to the construction information 29 shown in FIG. 10, the T-LAN (# 1) real-time network has the master 60 with the identifier “Cont_α” and the control LAN 8a with the identifier “CLAN_1”. Then, the master 60 (Cont_α) transmits a control frame in a period of 5 ms with a period of 10 ms.

FIG. 11 is an explanatory diagram illustrating the software configuration and software processing of the schedule server 20 according to the first embodiment.

The memory 23 of the schedule server 20 includes an external I / F driver 231, an initial setting module 232, a resource allocation module 233, initial information 234, construction information 29, an entire time table 235, a plurality of individual time tables 236, an application schedule 237, and distribution. A module 238 and a NIC I / F driver 239 are included.

External I / F driver 231, initial setting module 232, resource allocation module 233, distribution module 238, and NIC I / F driver 239 are programs in this embodiment and are executed by processor 22. However, the function of each program may be implemented by a physical device such as an integrated circuit.

Initial information 234 and construction information 29 are input by the operator. The initial information 234 and the construction information 29 are information for generating the entire time table 235 and the individual time table 236.

The overall time table 235 and the individual time table 236 are schedules indicating the time when each of the real-time networks becomes valid, and are generated by the resource allocation module 233.

The application schedule 237 indicates the time when the master 60 and the switch 10 start applying the individual time table 236. The application schedule 237 may be set in advance by an operator.

Note that the memory 23 of the schedule server 20 holds information about the master 60, the switch 10, the control LAN 8, and the slave 7 of the communication system, information about the network topology to which each device of the communication system is connected, and the switch 10 Holds information about the device to which the other port is connected (not shown). For this reason, the schedule server 20 can search for a route in the network of the communication system as necessary.

The operator of the schedule server 20 inputs initial information and construction information to the schedule server 20 using the keyboard 24, mouse 25, or monitor 28. The external I / F driver 231 receives the input initial information and construction information, and inputs the input initial information and construction information to the initial setting module 232.

The initial setting module 232 stores the input initial information and construction information in the memory 23 as initial information 234 and construction information 29. Here, the initial information includes a switching period (for example, 1 ms) and a repetition period (for example, 20 ms).

The switching period is a period from when the switch 10 switches the destination of the port to when the switch 10 can next switch the destination. For this reason, the minimum value of the communication period 294 of the construction information 29 is the same as the value of the switching period. For example, when a 10018-Mbps Ethernet is used for a frame with a length of 1518 bytes, the switching period is about 125 μsec.

The repetitive cycle indicates a cycle in which all the real-time networks included in the communication system transmit a control frame at least once. For this reason, the repetition period is the same as the common multiple of a plurality of values stored in the communication period 295.

The repetition cycle depends on the cycle in which the control frame is transmitted in each of the plurality of real-time networks of the present embodiment. For example, if the communication system includes a real-time network that transmits control frames with a period of 1 second, the repetition period is a common multiple of 1000 ms.

Next, the resource allocation module 233 refers to the initial information 234 and the construction information 29, and allocates a resource (specifically, a port of the switch 10) for the master 60 to transmit a control frame. The resource allocation module 233 notifies the success and failure of the allocation process to the operator of the schedule server 20 via the external I / F driver 231 and the monitor 28.

When the resource allocation is successful, the resource allocation module 233 generates an overall time table 235 and an individual time table 236.

When the individual time table 236 is generated, the distribution module 238 distributes the individual time table 236 and the application schedule 237 to the switch 10 or the master 60 via the NIC I / F driver 239. This distribution method may be any method, and may be a method of transmitting via a network or a method of executing offline.

FIG. 12 is a flowchart showing the processing of the resource allocation module 233 according to the first embodiment.

In the communication system, the resource allocation module 233 allocates a route to each of the real-time networks and distributes a schedule (individual time table 236) to the master 60 and the switch 10 based on the allocated result by the process shown in FIG. The route for transferring the control frame is quickly changed. As a result, the resource allocation module 233 according to the present embodiment can effectively utilize the network resources of the communication system.

The resource allocation module 233 determines whether the initial information 234 or the construction information 29 has been updated (step 100). Specifically, when the initial setting module 232 updates the initial information 234 or the construction information 29, when the initial setting module 232 stores new construction information 29 in the memory 23, or from the initial setting module 232, the initial information 234 is updated. Alternatively, when an event indicating that the construction information 29 has been updated is transmitted, the resource allocation module 233 determines that the initial information 234 or the construction information 29 has been updated.

If the initial information 234 or the construction information 29 has not been updated, the resource allocation module 233 repeats step 100 and waits until the initial information 234 or the construction information 29 is updated.

When the initial information 234 or the construction information 29 is updated, the resource allocation module 233 refers to the construction information 29 and reads the information updated in the construction information 29 as information about a new T-LAN (step 102). When only the initial information 234 is updated, the resource allocation module 233 reads all entries of the construction information 29 as information on the new T-LAN in order to regenerate the individual time table 236.

The information related to the new T-LAN may include information indicating a plurality of T-LANs, or may include information indicating one T-LAN. The process shown in FIG. 12 is an example of the process when the information related to the new T-LAN indicates one T-LAN. When the information on the new T-LAN indicates a plurality of T-LANs, the resource allocation module 233 executes the following

steps

106, 108, 110, and 112 for each new T-LAN.

After step 102, the resource allocation module 233 adds an entry corresponding to the new T-LAN indicated by the read information to the overall time table 235 based on the communication period 294 and communication period 295 of the construction information 29. (Step 104). When the initial information 234 is updated, the resource allocation module 233 newly generates the entire time table 235 based on the updated initial information 234.

After step 104, the resource allocation module 233 determines whether resources can be allocated to a new T-LAN (step 106).

In step 106, the resource allocation module 233 allocates the combination of the ports of the switch 10 to a new T-LAN as a route. When a part of the route assigned to the new T-LAN does not overlap with a part of the route assigned to the existing T-LAN or the T-LAN having a higher priority, the resource assignment module 233 It is determined that the allocation can be performed. Then, the resource allocation module 233 executes Step 108.

In step 108, the resource allocation module 233 determines the entry corresponding to the new T-LAN by storing the route allocated to the new T-LAN in the entire time table 235.

In Step 106, if a part of the route assigned to the existing T-LAN or the T-LAN having a higher priority overlaps with a part of the route assigned to the new T-LAN, the resource assignment module 233 Determines that resources cannot be allocated to the new T-LAN. Then, the resource allocation module 233 determines whether it is possible to allocate a period during which the new T-LAN is valid to a period different from the period during which the existing T-LAN or the T-LAN with high priority is valid. Determine (step 110).

In step 110, the resource allocation module 233 determines that the resource can be allocated if a period different from the existing T-LAN or the like can be allocated to the new T-LAN, and executes step 108.

In step 110, the resource allocation module 233 deletes the entry added in step 104 from the overall time table 235 when a period different from the existing T-LAN or the like cannot be allocated to the new T-LAN. Is notified to the operator via the external I / F driver 231 (step 112).

Hereinafter, specific processing in step 106 will be described.

Various algorithms can be used as the method for determining whether or not resources can be allocated in step 106. Here, the Dijkstra method will be described as an example. For example, the processing in step 106 when the schedule server 20 is connected to a network having each of the switches 10a to 10d shown in FIG. 4 and the construction information 29 shown in FIG. Show.

The resource allocation module 233 allocates a path to the T-LAN in step 106 so that one path between the two switches 10 is not allocated to a plurality of T-LANs.

When using the Dijkstra method, the resource allocation module 233 allocates 1 to the path between the two switches 10 as an initial value of the network cost. The resource allocation module 233 may allocate a value other than 1 as a network cost to each route. The resource allocation module 233 may set the usage frequency of a route to a different usage frequency from other routes by assigning a value other than 1. Further, the resource allocation module 233 allocates infinity (hereinafter referred to as “∞”) as a network cost to a route that has already been allocated to the T-LAN.

In the construction information 29 shown in FIG. 10, T-LAN (# 1) has the highest priority. Therefore, the resource allocation module 233 first determines in step 106 whether or not resources can be allocated to the T-LAN (# 1).

In this case, the resource allocation module 233 needs to calculate a route from the switch 10a connected to the master 60a (Cont_α) to the switch 10b connected to the control LAN 8a (CLAN_1). The resource allocation module 233 calculates the total value of the network costs of each of the plurality of routes from the switch 10a to the switch 10b.

As a result of the calculation, since the total value of the network costs of the path directly connecting the switch 10a to the switch 10b is the smallest, the resource allocation module 233 determines the path directly connecting the switch 10a to the switch 10b as the T-LAN (# 1). Determine the route to be assigned to Since the route assigned to T-LAN (# 1) does not overlap with any T-LAN route, resource assignment module 233 determines that resources can be assigned to T-LAN (# 1). Then, ∞ is assigned to the path directly connecting the switch 10a to the switch 10b.

In step 108, the resource allocation module 233 determines the entry corresponding to the T-LAN (# 1) by storing the determined path in the entry corresponding to the T-LAN (# 1).

Next, the resource allocation module 233 determines in step 106 whether or not resources can be allocated to the T-LAN (# 2). Specifically, the resource allocation module 233 calculates the total value of the network costs of each of the plurality of routes from the switch 10a to the switch 10b. As a result of the calculation, the resource allocation module 233 determines the route having the minimum network cost as the T-LAN ( # 2) is assigned to the route.

Note that since ∞ is already assigned to the route directly connecting the switch 10a and the switch 10b, the resource assignment module 233 cannot assign this route to the T-LAN (# 2). The path that directly connects the switch 10a and the switch 10c, the path that directly connects the switch 10a and the switch 10d, the path that directly connects the switch 10c and the switch 10b, and the path that directly connects the switch 10d and the switch 10c The network costs are all 1.

Therefore, the network cost of the route passing through the switch 10a, the switch 10c, and the switch 10b is 2 in total. The total network cost of the route passing through the switch 10a, the switch 10d, and the switch 10b is 2.

Here, since the network costs of the routes reaching the switch 10b via the switch 10a, the switch 10d, and the switch 10c are all 2 or more, all the routes that pass through the switch 10a, the switch 10d, and the switch 10c are Not adopted.

The resource allocation module 233 includes a T-LAN (T-LAN) among the routes that pass through the switch 10a, the switch 10c, and the switch 10b, and the routes that pass through the switch 10a, the switch 10d, and the switch 10b. Either may be defined as the route assigned to # 2).

Here, the resource allocation module 233 uses, as a route allocated to the T-LAN (# 2), a route that is found first (a route that passes through the switch 10a, the switch 10c, and the switch 10b) among routes having the same network cost. decide. Since the route assigned to the T-LAN (# 2) does not overlap with any T-LAN route, the resource assignment module 233 determines that resources can be assigned to the T-LAN (# 2). Then, ∞ is assigned to the route passing through the switch 10a, the switch 10c, and the switch 10b.

In step 108, the resource allocation module 233 determines the entry for the T-LAN (# 2) by storing the determined path in the entry corresponding to the T-LAN (# 2).

Next, the resource allocation module 233 determines in step 106 whether or not resources can be allocated to the T-LAN (# 3). Since ∞ is assigned as a network cost to the route that directly connects the switch 10a and the switch 10b and the route that passes through the switch 10a, the switch 10c, and the switch 10b, the resource allocation module 233 uses the route that passes through these routes. Cannot be assigned to T-LAN (# 3).

The T-LAN (# 3) requires a path for connecting the master 60c and the switch 10d, and the master 60c and the switch 10d are directly connected. For this reason, the resource allocation module 233 determines a path directly connecting the master 60c and the switch 10d as a path allocated to the T-LAN (# 2). Since the route assigned to T-LAN (# 3) does not overlap with any T-LAN route, resource assignment module 233 determines that resources can be assigned to T-LAN (# 3). And ∞ is assigned to the path connecting the master 60c and the switch 10d.

In step 108, the resource allocation module 233 determines the entry for the T-LAN (# 3) by storing the determined path in the entry corresponding to the T-LAN (# 3).

In the following, a process when the path cannot be assigned to a plurality of T-LANs at the same time because the paths physically overlap (details of the process in step 110) will be described. In step 110, the resource allocation module 233 changes the time at which the route is allocated, and determines whether the allocation is possible.

The resource allocation module 233 allocates a route to a new T-LAN in

steps

106, 108, 110, and 112, and then generates an individual time table 236 based on the entire time table 235 (step 109).

After step 109, the distribution module 238 distributes the individual time table 236 and the application schedule 237 to the master 60 and the switch 10 (step 111). Here, the distribution module 238 transmits the individual time table 236 corresponding to each of the masters 60 and each of the switches 10 to each of the masters 60 and each of the switches 10.

After step 111 or step 112, the resource allocation module 233 returns to step 100 and determines whether the initial information 234 or the construction information 29 has been updated.

FIG. 13 is an explanatory diagram illustrating the entire time table 235 according to the first embodiment.

The entire time table 235 indicates the timing at which the control frame is transmitted in the real-time network (T-LAN) included in the communication system of the first embodiment. The overall time table 235 includes a time 2354 and a T-LAN 2355.

Time 2354 indicates a relative time. Time 2354 indicates the time from 0 seconds to the time indicated by the repetition period for each switching period. For this reason, for example, when the initial information 234 includes the switching period and the repetition period of the example shown in FIG. 11, the resource allocation module 233 determines the 20 pieces of “000” to “019” according to the initial information 234 in step 104. A value indicating the time is stored at time 2354.

Also, time 2354 indicates time synchronized between the switch 10 and the master 60. Specifically, for example, at the time 2354 of “011”, the times at the switch 10 and the master 60 are “12: 34: 56.011”, “12: 34: 56.031”, and “12 "Time 34 minutes 56.051 seconds" (time repeated every 20 ms).

T-LAN 2355 indicates a path through which a control frame is transmitted in the T-LAN included in the communication system. An entry is stored in the T-LAN 2355 so as to correspond to the time 2354 when the T-LAN becomes valid.

13 indicates a route allocated to the T-LAN indicated by the construction information 29 in FIG. Specifically, the T-LAN 2355 shown in FIG. 13 includes an entry 2351 indicating the route of the T-LAN (# 1), an entry 2352 indicating the route of the T-LAN (# 2), and the T-LAN (# 3 ) Includes an entry 2353 indicating the route of).

In step 104, the resource allocation module 233 adds T-LAN 2355 entries in descending order of T-LAN priority according to the communication period 294 and communication period 295 of the construction information 29. In step 104, the resource allocation module 233 adds an entry so that the beginning of the communication cycle is time 2354 “000”.

For example, since the communication period 294 of the T-LAN (# 1) is 5 ms in step 104, the resource allocation module 233 determines that the T-LAN (# 1) at five times 2354 from “000” to “004”. Add an entry indicating. Further, since the communication cycle 295 of the T-LAN (# 1) is 10 ms, the resource allocation module 233 also has an entry indicating T-LAN (# 1) at five times 2354 from “010” to “014”. Add

Further, in step 108, the resource allocation module 233 stores the identifier indicating the path allocated in step 106 in the entry of the entire time table 235 corresponding to the T-LAN to which the path is allocated. In step 108, the resource allocation module 233 changes the T-LAN entry according to the time allocated in step 110.

For example, in the above-described example shown in FIG. 12, since a route for directly connecting the switch 10a to the switch 10b is assigned to the T-LAN (# 1), the entries 2351 “000” to “004” and “010” to In the entry “014”, an identifier “Cont_α-SW_A-SW_B-CLAN_1” indicating the route is stored.

Hereinafter, the process in which the resource allocation module 233 generates the individual time table 236 from the entire time table 235 in step 109 will be described.

FIG. 14 is an explanatory diagram showing a connection relationship of the switch 10a according to the first embodiment.

FIG. 14 shows an image of information that the schedule server 20 has, and the information shown in FIG. 14 shows the connection relationship of the ports that the switch 10a has. According to FIG. 14, the switch 10a has six ports, and identifiers “0” to “5” are assigned to the respective ports.

The master 60a is connected to the port (port “1”) whose identifier is “1”, and the switch 10b is connected to the port “2”. Further, the switch “10c” is connected to the port “3”, and the switch “10d” is connected to the port “4”. The master 60b is connected to the port “5”.

FIG. 15 is an explanatory diagram illustrating port combination information 210 in the switch 10a according to the first embodiment.

The combination information 210 indicates a combination of devices for which the switch 10 transfers frames to each other and a port identifier for connecting the combination of the devices. The schedule server 20 holds the combination information 210 of the ports of all the switches 10 included in the communication system in the memory 23.

15 is the combination information 210 of the switch 10a, and indicates the connection relationship shown in FIG. For example, in the combination information 210 shown in FIG. 15, when an Ethernet frame is transferred from the master 60a to the switch 10b via the switch 10a, the Ethernet frame is transferred between the port "1" and the port "2" of the switch 10a. Indicates that

The combination information 210 is used by the resource allocation module 233 to generate the individual time table 236 from the entire time table 235.

When generating the individual time table 236 from the entire time table 235, the resource allocation module 233 first reads an entry including each identifier of the switch 10 from the entire time table 235. Then, the resource allocation module 233 extracts devices connected to each of the switches 10 from the path indicated by the read entry. Then, the resource allocation module 233 identifies a combination of ports in the switch 10 based on the extracted device and the combination information 210.

Furthermore, the resource allocation module 233 stores an identifier indicating the combination of the specified ports in the entry of each individual time table 236 of the switch 10. Here, the resource allocation module 233 stores an identifier indicating the identified combination of ports in the entry of the individual time table 236 of the switch 10 corresponding to the time 2354 of the read entry.

Also, the resource allocation module 233 reads an entry including each identifier of the master 60. Then, the resource allocation module 233 stores a value indicating that the port is opened in the entry of the individual time table 236 of the master 60 corresponding to the time 2354 of the read entry.

For example, in step 109 shown in FIG. 12, the resource allocation module 233 reads the entry 2351 and the entry 2352 as entries including the identifier of the switch 10a (SW_A). The entry 2351 and the entry 2352 read out here indicate a path of “Cont_α-SW_A-SW_B-CLAN_1” and a path of “Cont_β-SW_A-SW_C-SW_B-CLAN2”.

Therefore, the resource allocation module 233 connects the master 60a (Cont_α) and the switch 10b (SW_B) via the switch 10a, the master 60b (Cont_β) and the switch 10c (SW_C) to the switch 10a. The combination of the ports to be connected through the combination information 210 is specified. The combination specified here is a combination of ports “1-2” and “3-5”.

In step 109, the resource allocation module 233 adds the entries of the individual time table 236 of the switch 10a corresponding to the times 2354 “000” to “004” and “010” to “014” of the read entry. “1-2” is stored. In step 109, the resource allocation module 233 stores “3-5” in the entry of the individual time table 236 of the switch 10a corresponding to the times 2354 “000” to “003” of the read entry. .

The resource allocation module 233 generates the individual time table 236 for each of the switch 10 and the master 60 by performing such processing for all the switches 10.

FIG. 16 is an explanatory diagram illustrating the individual time table 236 of the switch 10 according to the first embodiment.

The individual time table 236 of the switch 10 indicates the port to which the switch 10 is connected and the time at which the port is connected. Each individual time table 236 of the switch 10 corresponds to each of the switches 10.

16 includes an individual time table 236a to an individual time table 236d, and each of the individual time tables 236 corresponds to each of the switch 10a to the switch 10d.

The individual time table 236 includes a time 2361 and a port 2362. Time 2361 corresponds to time 2354 in the overall time table 235. A port 2362 in the individual time table 236 of the switch 10 indicates a combination of connected ports.

For example, the individual time table 236a shown in FIG. 16 indicates that the port “1” and the port “2” of the switch 10a are between the time 2361 “000” and the time 2361 “004” and the time from the time 2361 “010”. In order to transfer a control frame between 2361 and "014", it indicates that they are physically connected. In addition, the individual time table 236a shown in FIG. 16 transfers the control frame between the time 2361 “000” and the time 2361 “003” between the port “3” and the port “5” of the switch 10a. Indicates that they are physically connected.

Note that the resource allocation module 233 holds information regarding the ports of the switches 10a to 10d as combination information 210 as shown in the lower part of FIG.

FIG. 17 is an explanatory diagram illustrating the individual time table 236 of the master 60 according to the first embodiment.

The individual time table 236 of the master 60 indicates the time when the port is opened in order to transmit the control frame. Each of the individual time tables 236 (individual time table 236e to individual time table 236g) shown in FIG. 17 corresponds to each of master 60a to master 60c.

The port 2362 in the individual time table 236 of the master 60 indicates whether or not to open the port for transmitting the control frame. When the port 2362 shown in FIG. 17 includes “◯”, it indicates that the port for transmitting the control frame is opened.

Note that the resource allocation module 233 holds information regarding the ports of the masters 60a to 60c as shown in the lower part of FIG.

FIG. 18 is an explanatory diagram illustrating the software configuration and processing of the switch 10 according to the first embodiment.

The memory 13 of the switch 10 includes a receiving module 192, a switching module 193, and a NIC I / F driver 191. The receiving module 192, the switching module 193, and the NIC I / F driver 191 shown in FIG. 18 are software implemented by a program. Any mounting method may be used as long as the function 191 can be mounted. For example, each function may be physically implemented by an integrated circuit or the like.

In addition, the memory 13 of the switch 10 includes an application schedule 237, an individual time table 236, and an individual time table 194. The application schedule 237 and the individual time table 236 in the memory 13 are transmitted from the schedule server 20.

The NIC I / F driver 191 transmits the data received by the NIC 11 to the reception module 192 or the switching module 193.

The reception module 192 receives the individual time table 236 and the application schedule 237 from the schedule server 20 via the network or offline, and stores the received individual time table 236 and the application schedule 237 in the memory 13. Further, the reception module 192 may receive the individual time table 236 that is set to be activated at the timing indicated by the application schedule 237.

The individual time table 194 indicates a combination of a time at which a T-LAN control frame is transferred and a port at which the control frame is transferred at the time. The individual time table 194 is updated by the individual time table 236.

The switching module 193 transfers the received control frame using the time synchronized between the master 60 and the switch 10 (that is, the time synchronized with the GPS function) and the individual time table 194. In addition, the switching module 193 transfers the received Ethernet frame using a switching table (not shown) held in advance.

The switching module 193 refers to the application schedule 237 when the application schedule 237 is stored in the memory 13. Then, the switching module 193 updates the individual time table 194 with the individual time table 236 when the current time held in the switch 10 is the time to apply the individual time table 236 indicated by the application schedule 237. Specifically, the switching module 193 changes the individual time table 236 to the individual time table 194 and deletes the old individual time table 194.

The application schedule 237 is distributed to all of the master 60 and the switch 10. Therefore, by updating the individual time table of each device according to the application schedule 237, all of the master 60 and the switch 10 start referring to a new individual time table at the same time (or at the same time as close as possible). be able to. Then, the master 60 and the switch 10 can start transferring the control frame according to the schedule indicated by the overall time table 235.

FIG. 19 is a flowchart illustrating the processing of the switching module 193 of the switch 10 according to the first embodiment.

When the switching module 193 detects a timer interrupt of the microcomputer 19 activated by the time-synchronized clock (step 1931), the switching module 193 acquires the current time held in the switch 10 (step 1932), and further stores the individual time table 194. Reference is made (step 1933). As described above, the switch 10 holds the current time by the GPS function of the GPS receiver 15. Further, the timer interruption of the microcomputer 19 occurs at least every switching period described above.

After step 1933, the switching module 193 determines whether or not a port combination is allocated at time 2361 (included in the individual time table 194) corresponding to the acquired current time (step 1935).

Note that any method may be used as a method of associating the acquired current time with the time 2361. For example, when the switching module 193 starts referring to the individual time table 194 updated according to the application schedule 237, the switching module 193 generates a timer whose upper limit is the repetition period indicated by the individual time table 194, and the current time and timer You may hold correspondence with. Then, when a timer interrupt is detected, a time 2361 corresponding to the current time may be acquired based on the current time and the timer.

When a combination of ports is assigned at time 2361 corresponding to the current time, the switching module 193 transfers the Ethernet frame received by any of the assigned ports as a T-LAN control frame (step 1934). ). Specifically, the switching module 193 transfers the bit of the Ether frame received by one port of the port combination at the time 2361 corresponding to the current time to the other port of the port combination, A bit is output from the other port.

At this time, the switching module 193 does not read out header information such as address information (destination address 612 and transmission source address 613 shown in FIG. 9 and transmission source address 613) and data 615 from the bits of the Ethernet frame received by the port, and includes them in the Ethernet frame. The transmitted bits are transferred as they are. Therefore, if the control frame is received by the switch 10 at the time 2361 when the individual time table 236 indicates the combination of ports, the control frame is promptly transferred to the switch 10 without being buffered for routing or the like. Is done.

As a result, the switch 10 can transfer the control frame in real time, and can transfer the control frame in bit units using the On the Fly method.

Moreover, since the switching module 193 transfers the control frame according to the combination of ports indicated by the individual time table 194, a plurality of different real-time network control frames are not transferred through the same route. For this reason, the network configured by the switch 10 can include a plurality of real-time networks.

The switching module 193 repeats Step 1934 until the next timer interrupt occurs. When the timer interrupt occurs (Step 1931), the switching module 193 executes Step 1932.

If no port combination is assigned at time 2361 corresponding to the current time, the switching module 193 reads address information and the like stored in the received ether frame. Then, the switching module 193 transfers the received ether frame based on the switching table held in advance, which is different from the individual time table 194 of the present embodiment, and the read address information (step 1936).

The transfer processing in step 1936 includes, for example, conventional documents (Cisco Systems, Inc., “How LAN Switches Work”, [online], August 1, 2007, [Search January 12, 2013], Internet reference. The switch processing described in (1) may be used.

In step 1934, when an Ethernet frame is received by a port that is not assigned at time 2361 corresponding to the current time, the switching module 193 stores in advance different from the individual time table 194 of the present embodiment, as in step 1936. The Ethernet frame is transferred according to the switching table and the address information.

In step 1936, the switching module 193 transfers the Ethernet frame based on the address information or the like, so that the switch 10 of this embodiment can transfer a frame other than the control frame of the real-time network.

The switching module 193 repeats Step 1936 until the next timer interrupt occurs. When the timer interrupt occurs (Step 1931), the switching module 193 executes Step 1932.

FIG. 20 is an explanatory diagram illustrating the software configuration and processing of the communication interface 68 of the master 60 according to the first embodiment.

The memory 689 of the master 60 includes a NIC I / F driver 6861, a reception module 6862, and a transmission module 6863. The NIC I / F driver 6861, the reception module 6862, and the transmission module 6863 shown in FIG. 20 are software implemented by a program. The master 60 of the first embodiment is the NIC I / F driver 6861, the reception module 6862, and the transmission. Any mounting method may be used as long as the function of the module 6863 can be mounted. For example, each function may be physically mounted by an integrated circuit or the like.

The memory 689 of the master 60 includes an application schedule 237, an individual time table 236, an individual time table 6864, and a queue 6865. The application schedule 237 and the individual time table 236 are transmitted from the schedule server 20.

The NIC I / F driver 6861 transmits the data received by the NIC 688 to the reception module 6862 or the transmission module 6863.

The reception module 6862 receives the individual time table 236 and the application schedule 237 from the schedule server 20 via the network or offline, and stores the received individual time table 236 and application schedule 237 in the memory 689. The reception module 6862 may receive the individual time table 236 set to be activated at the timing indicated by the application schedule 237.

The individual time table 6864 indicates the time when the control frame is transmitted. The individual time table 6864 is updated by the individual time table 236.

The transmission module 6863 transmits the control frame stored in the queue 6865 using the time synchronized with the master 60 and the switch 10 (that is, the time synchronized with the GPS function) and the individual time table 6864. To do. Further, the transmission module 6863 transmits the Ethernet frame stored in the queue 6865 by a switching table (not shown) held in advance.

The transmission module 6863 reads out the control frame held in the queue 6865 and outputs the control frame to the NIC 688 bit by bit. The NIC 688 outputs the input control frame to the network bit by bit.

Further, when the application schedule 237 is stored in the memory 689, the transmission module 6863 refers to the application schedule 237 as with the switching module 193 of the switch 10. Then, the transmission module 6863 updates the individual time table 6864 with the individual time table 236 at the time when the individual time table 236 indicated by the application schedule 237 is applied. Specifically, the transmission module 6863 changes the individual time table 236 to the individual time table 6864 and deletes the old individual time table 6864.

FIG. 21 is an explanatory diagram illustrating the queue 6865 included in the memory 689 of the master 60 according to the first embodiment.

The queue 6865 holds control frames and ether frames transmitted by the master 60 and control frames and ether frames received by the master 60. The queue 6865 includes at least one CLAN transmission queue 601, at least one CLAN reception queue 602, a transmission queue 603, and a reception queue 604.

The CLAN transmission queue 601 is a queue for storing control frames to be transmitted, and the CLAN reception queue 602 is a queue for storing received control frames. The master 60 has CLAN transmission queues 601 and CLAN reception queues 602 corresponding to the number of control LANs 8 that transmit control frames.

When the transmission module 6863 opens one port for transmitting a control frame, each control frame in the CLAN transmission queue 601 is output bit by bit in the order specified in advance from the opened port. When the master 60 receives the control frame, the transmission module 6863 inputs each data included in the control frame to the CLAN reception queue 602 corresponding to the transmission source control LAN 8 for each bit.

The transmission queue 603 and the reception queue 604 are queues for storing ether frames, which are different from the control frames of this embodiment. The transmission queue 603 and the reception queue 604 are queues for transmitting and receiving Ethernet frames other than the control frame at a timing when the port is not used for transmitting the control frame.

FIG. 22 is a flowchart illustrating processing of the transmission module 6863 of the master 60 according to the first embodiment.

When the transmission module 6863 detects a timer interrupt of the microcomputer 686 activated by the time-synchronized clock (step 68631), the transmission module 6863 acquires the current time held by the master 60 (step 6632), and further refers to the individual time table 6864. (Step 68633).

The master 60 holds the current time by the GPS function of the GPS receiver 682. Further, the timer interrupt of the microcomputer 686 occurs at least for each switching period described above.

After step 68633, the transmission module 6863 determines whether or not a value indicating that the port is opened is included in the entry of time 2361 (included in the individual time table 6864) corresponding to the acquired current time. (Step 68635). The same method as the switching module 193 of the switch 10 may be used as a method of acquiring the time 2361 corresponding to the current time.

When the entry indicating that the port is opened is included in the entry at time 2361 corresponding to the current time, the transmission module 6863 transmits the control frame stored in each of the CLAN transmission queue 601 to the NIC I / F driver 6861. And sequentially transmitted to the slave 7 via the NIC 21. Also, the transmission module 6863 stores the bits received by the NIC 21 and the NIC I / F driver 6861 in the CLAN reception queue 602 (step 68634).

The transmission module 6863 repeats Step 68634 until a timer interrupt occurs. Then, the transmission module 6863 executes Step 68632 when the timer interruption occurs.

If it is determined in step 68635 that the entry indicating that the port is opened is not included in the entry at time 2361 corresponding to the current time, the transmission module 6863 transmits the Ethernet frame stored in the transmission queue 603; The received ether frame is stored in the reception queue 604 (step 68636). The transmission module 6863 repeats step 68636 until a timer interrupt occurs.

FIG. 23 is an explanatory diagram showing a real-time network to which a physically non-overlapping route is assigned according to the first embodiment.

Through the above-described processing in the first embodiment, routes are assigned to T-LAN (# 1), T-LAN (# 2), and T-LAN (# 3) as shown in FIG. The paths of T-LAN (# 1), T-LAN (# 2), and T-LAN (# 3) shown in FIG. 23 do not physically overlap.

According to the above-described processing example, in a single communication system including a plurality of real-time networks, routes that do not physically overlap each other are assigned to each real-time network. As a result, the schedule server 20 can cause the switch 10 to transfer the control frame through a unique route, and the switch 10 can transfer the control frame in real time.

In the following, a process in a case where it is determined that a route assigned to each real-time network physically overlaps and a route cannot be assigned to the T-LAN in step 106 shown in FIG.

FIG. 24 is an explanatory diagram showing a real-time network in which physical paths overlap in the first embodiment.

The T-LAN (# 3) shown in FIG. 24 includes a control LAN 8d (CLAN_5) and a control LAN 8e (CLAN_6). One of the routes that the T-LAN (# 3) can take is the route indicated by “Cont_γ-CLAN4-CLAN5-CLAN6”.

When this route is allocated to the T-LAN (# 3), the route of the T-LAN (# 2) and the route of the T-LAN (# 3) between the switch 10c (SW_C) and the switch 10b (SW_B) And overlap. Therefore, the T-LAN (# 2) master 60b and the T-LAN (# 3) master 60c cannot transmit control frames at the same time.

Further, when the T-LAN (# 3) is assigned the above-described route, the control frame is transferred by the control LAN 8c (CLAN_4) and the control LAN 8d (CLAN_5), and further reaches the control LAN 8e (CLAN_6). The switch 10c and the switch 10d use four ports.

FIG. 25 is an explanatory diagram illustrating the entire time table 235 according to the first embodiment.

The overall time table 235 shown in FIG. 25 is the overall time table 235 generated after executing step 110 shown in FIG. 12 when the assigned routes are physically duplicated. The time 2354 of the entry 2353 shown in FIG. 25 is different from the time 2354 of the entry 2352 by 4 ms. Thus, the schedule server 20 can avoid the problem that the route between the switch 10c and the switch 10b is shared by two T-LANs at the same time.

The processing of Step 106 and Step 110 when physical paths overlap is shown below. As the method used in step 106 shown below, the Dijkstra method is used as in the above example. However, any method may be used as the method for assigning the route in step 106.

In the processing described below, the entry of the construction information 29 indicating T-LAN (# 1) and T-LAN (# 2) is the same as the entry of the construction information 29 shown in FIG. ) Indicating the plurality of control LANs 8 (CLAN_4, CLAN_5, and CLAN_6) in the control LAN 293. For this reason, the process of assigning a route to T-LAN (# 1) and T-LAN (# 2) in step 106 is the same as that in step 106 described above.

The order in which control frames are transferred to a plurality of control LANs 8 of T-LAN # 3 is determined in advance. For example, the order of the control LAN 8 to which the control frame is transferred is specified by the order of the values stored in the control LAN 293 of the construction information 29. This is because the slave 7 connected to the control LAN 8 sequentially transfers the control frames. In the T-LAN (# 3) shown in FIG. 24, the control frame is transferred in the order of the switch 10d, the switch 10c, and the switch 10b.

The resource allocation module 233 sets the order of the control LAN 8 in the T-LAN, for example, in the construction information 29 by an operator or the like. The resource allocation module 233 determines a path to be allocated to the T-LAN so that the control frames are transferred in a preset order.

Specifically, the resource allocation module 233 in the above example does not cause the switch 10d and the switch 10b to pass through the switch 10c from the paths that may be allocated to the T-LAN (# 3) in step 106. Exclude routes connected to.

Further, since the route is allocated to the T-LAN (# 3), the resource allocation module 233 can connect the master 60c and the slave 7 of the T-LAN (# 3) at a network cost lower than ∞ in step 106. Search for a route.

The master 60c and the switch 10d are connected by a route having a network cost of 1. The switch 10d and the switch 10c are directly connected by a route having a network cost of 1. All the paths between the switch 10c and the switch 10b have a network cost of ∞.

Therefore, the resource allocation module 233 determines the path between the master 60c and the switch 10d and the path between the switch 10d and the switch 10c. Since the path between the switch 10c and the switch 10b cannot be determined, the resource allocation module 233 determines in step 106 that resources cannot be allocated to the T-LAN (# 3).

If the resource allocation module 233 determines in step 106 that the resource cannot be allocated to the T-LAN, in step 110, the T- that has not been able to determine a route in a period different from the period in which other T-LANs are valid. It is determined whether resources can be allocated to the LAN.

In step 110, the resource allocation module 233 does not change the period during which the T-LAN whose route has already been determined (in this case, T-LAN (# 1) and T-LAN (# 2)) is valid. In addition, a period in which the T-LAN (# 3) can be validated is searched. Specifically, the resource allocation module 233 refers to the entire time table 235 and sets the entry 2353 of the T-LAN (# 3) to the time when the T-LAN (# 1) or T-LAN (# 2) is invalid. Move to 2354.

The resource allocation module 233 moves the entry 2353 of the T-LAN (# 3) at every time 2354, and specifies the time 2354 “004” at which the entry 2352 is not stored. In addition, when the T-LAN (# 3) is temporarily enabled at the time 2354 “004” (temporary communication period), the resource allocation module 233 at the time 2354 in the other communication period of the T-LAN (# 3). , It is determined whether or not the entry 2352 is stored.

As a result of the determination, when the entry 2352 is stored in any of the temporary communication periods of the T-LAN (# 3), the resource allocation module 233 determines that the T-LAN whose route has already been determined is invalid. The time 2354 is searched. If there is no other time 2354 as a result of the search, the resource allocation module 233 determines in step 110 that the resource cannot be allocated to the T-LAN (# 3), and executes step 112.

As a result of the determination, if the entry 2352 is not stored in the entire temporary communication period of the T-LAN (# 3), the resource allocation module 233 sets the T-LAN (# 3) in the communication system in the temporary communication period. Search for routes that can be assigned.

In the route where the switch 10c and the switch 10b are directly connected, the T-LAN (# 2) is invalid during the temporary communication period (time 2354 is “004”), so the network cost is 1. In addition, the network cost of the path connecting the switch 10c to the switch 10b via the switch a is ∞ during the provisional communication period.

For this reason, the resource allocation module 233 determines a route for directly connecting the switch 10c and the switch 10b as a route for connecting the switch 10c and the switch 10b. Then, the resource allocation module 233 allocates a route that directly connects the master 60c, the switch 10d, the switch 10c, and the switch 10b to the route of the T-LAN (# 3) at time 2354 “004”. Then, the resource allocation module 233 determines in step 110 that resources can be allocated to the T-LAN (# 3), and executes step 108.

In step 108, the resource allocation module 233 determines that the path allocated in step 110 to the temporary communication period entry 2353 (in the above example, “Cont_γ-SW_D-CLAN_4-SW_D-SW_C-CLAN_5-SW_C-SW_D-CLAN_6 ") Is stored and the entry 2353 is determined.

If the paths allocated to the T-LAN physically overlap, the resource allocation module 233 can effectively utilize the network resources of the communication system by allocating different times to the effective times in Step 110.

FIG. 26 is an explanatory diagram illustrating the individual time table 236 with the switch 10 and the master 60 after changing the effective period of the real-time network according to the first embodiment.

The individual time table 236 shown in FIG. 26 is a table generated in step 108 based on the entire time table 235 shown in FIG. 25, and is a table when a route is assigned to the T-LAN as shown in FIG. . 26 is an individual time table 236 for the switch 10b, the switch 10c, the switch 10d, and the master 60c.

The difference between the individual time table 236b shown in FIG. 26 and the individual time table 236b shown in FIG. 16 is that a combination of port “3” and port “6” is added to the individual time table 236b shown in FIG. It is.

The difference between the individual time table 236c shown in FIG. 26 and the individual time table 236c shown in FIG. 16 is that the combination of the port “1” and the port “4” and the port “ The combination of 3 "and port" 6 "is added.

The difference between the individual time table 236d shown in FIG. 26 and the individual time table 236d shown in FIG. 16 is that a combination of port “3” and port “6” is added to the individual time table 236d shown in FIG. It is.

The difference between the individual time table 236g shown in FIG. 26 and the individual time table 236g shown in FIG. 17 is that, in the individual time table 236g shown in FIG. 26, the time 2364 at which the value indicating that the port is opened is stored as “004”. “0”, “014”, and “009”.

In the processing shown in FIG. 12, the resource allocation module 233 adjusts so that the effective period of the real-time network does not overlap when the paths allocated to the real-time network physically overlap. As a result, the schedule server 20 can transfer the control frame by a unique route. As a result, the switch 10 does not need to perform buffering processing for routing the control frame. Can be transferred in real time. The schedule server 20 can implement a plurality of real-time networks in one communication system.

According to the first embodiment, the master 60 transmits and receives the control frame in the route and period assigned by the schedule server 20, and the switch 10 transfers the control frame according to the combination of the assigned ports. Further, the transfer function of the switch 10 and the transfer function of the master 60 are synchronized in time. Then, the switch 10 transfers the control frame without performing processing that requires buffering such as routing. For this reason, the transfer delay and loss that occur when transferring the Ethernet frame do not occur when the control frame is transferred. As a result, the communication system according to the first embodiment can transfer the control frame in real time.

The switch 10 and the master 60 according to the first embodiment transfer a control frame according to the individual time table 236 distributed from the schedule server 20 to construct a predetermined physical line for transmitting the control frame at a predetermined time. To do. Since each of the switch 10 and the slave 7 in the constructed line transfers a control frame in bit units by the On the Fly method, a real-time network can be installed even in a communication system installed in a wide area.

In addition, since the schedule server 20 determines each route of the real-time network and the switch 10 transfers the control frame according to the determined route, the schedule server 20 according to the first embodiment quickly changes the route of the real-time network. It is possible to effectively use network resources.

Further, when the schedule server 20 determines the route of the real-time network as a physically different route, the communication system of the first embodiment can implement a plurality of real-time networks in one control system. In addition, even when each route of the real-time network physically overlaps, the schedule server 20 determines the time when the real-time network is valid as a different time. It can be implemented in one control system. As a result, the cost for installing a plurality of real-time networks can be reduced, and security can be appropriately protected even in a single network on which a plurality of real-time networks are mounted.

In the first embodiment described above, a method using GPS is applied as a method of time synchronization between the switch 10 and the master 60. However, any method may be used, for example, a method of time synchronization in a network. NTP or IEEE 1588 may be used.

Also, the time of the switch 10 and the master 60 may be synchronized using the time table of the first embodiment (the entire time table 235 and the individual time table 236). For example, the resource allocation module 233 may receive a trigger signal (bit) for time synchronization from the schedule server 20 between time 2354 “000” and time 2354 “001” so that all the switches 10 and the master 60 can receive the time synchronization trigger signal (bit). The entire time table 235 is generated. Then, by receiving the trigger signal, the switch 10 or the master 60 can correct the time and synchronize the time.

The switch 10 is set so that the switch 10 has one input and multiple outputs for a predetermined period (for example, a control frame input from one master 60 is transferred to a plurality of switches 10 or a plurality of devices). By doing so, it is possible to construct a network system of an arbitrary shape directly connected by a physical route. The schedule server 20 according to the second embodiment generates the entire time table 235 and the individual time table 236 so that the switch 10 has one input and multiple outputs.

The schedule server 20, the switch 10, and the master 60 of the second embodiment have the same functions and configurations as the schedule server 20, the switch 10, and the master 60 of the first embodiment.

FIG. 27 is an explanatory diagram showing a communication system applied to the railway operation management system of the second embodiment.

The communication system according to the second embodiment includes T-LAN (# 1) and T-LAN (# 2). The EC master of T-LAN (# 1) is master 60a (Cont_α), and the EC master of T-LAN (# 2) is master 60b (Cont_β). Each T-LAN includes a signal device 71 and a pointer 72 as the slave 7.

Furthermore, the communication system according to the second embodiment includes switches 10a to 10h (identifiers: SW_A to SW_H). The schedule server 20 is connected to the switch 10d.

The T-LAN (# 1) of the second embodiment is a network for managing the operation on the ring-shaped railway line, and the T-LAN (# 2) of the second example is operated on the straight railway line. It is a network for management. Each of the railway lines is provided with a slave 7 (signal 71 and pointer 72). By controlling the slave 7, the master 60 manages the operation of the railway line.

The railway lines managed by the T-LAN (# 1) and the railway lines managed by the T-LAN (# 2) partially overlap, and the T-LAN (# 1) and the T-LAN (# A part of the switch 10 and the slave 7 installed in 2) overlap. In the second embodiment, the T-LAN (# 1) and T-LAN (# 2) networks are laid regardless of the shape of the railway line. However, the network of the second embodiment is laid so that each master 60 can transmit a control frame to each slave 7 via the switch 10.

In the T-LAN (# 1) and T-LAN (# 2) of the second embodiment, the route from the master 60 to each of the slaves 7 is determined in advance by an operator or the like. Therefore, the resource allocation module 233 of the schedule server 20 determines a predetermined route as a route to be assigned to the T-LAN in step 106 shown in FIG. However, if it is determined in step 106 that the respective routes of the T-LAN are physically duplicated, the resource allocation module 233 determines that the time period in which the T-LAN is valid in step 110 is not duplicated in the entire time table 235. Adjust the entry.

In the T-LAN (# 1) and T-LAN (# 2) of the second embodiment, a part of the route overlaps, so that the resource allocation module 233 determines that the T-LAN (# 1) is effective in Step 110. A different period is assigned to the period during which T-LAN (# 2) becomes valid.

The master 60 and the switch 10 use the individual time table 236 generated by the schedule server 20 to transfer the control frame of each T-LAN during the period assigned to each railway line (T-LAN). For this reason, the administrator of the railway operation management system according to the second embodiment does not need to install a real-time network for each railway line, and can effectively use network resources.

Further, the administrator of the second embodiment does not need to lay a network dedicated to a new railway line (specifically, the switch 10 and the slave 7) even when a new railway line is created. By setting 29 and initial information 234, a real-time network for operation management can be freely generated.

Furthermore, the railway operation management system of the second embodiment is different from the communication system of the first embodiment in that the railway operation management system of the second embodiment differs from the EtherCat of the communication system of the first embodiment in that the control frame is set to the master 60. There is no need to return it. The control frame transmitted from each master 60 is not transferred to the switch 10 and the master 50 after reaching each of the traffic light 71 and the pointer 72.

Therefore, the schedule server 20 can generate a time table so that the switch 10 duplicates the control frame and transfers the control frame to a plurality of devices. As a result, the schedule server 20 can cause the switch 10 to multicast the control frame.

FIG. 28A is an explanatory diagram illustrating the entire time table 235 according to the second embodiment.

The entire time table 235 shown in FIG. 28A is generated when 1 ms is set as the switching period of the

initial information

234 and 20 ms is set as the repetition period of the initial information 234. The overall time table 235 shown in FIG. 28A is generated when the communication period 294 and the communication cycle 295 of T-LAN (# 1) and T-LAN (# 2) are all set to 10 ms.

Therefore, according to the entire time table 235 shown in FIG. 28A, the T-LAN (# 1) is valid during the time 2354 from “000” to “009”, and the time 2354 is changed from “010” to “019”. Until then, T-LAN (# 2) is effective.

FIG. 28B is an explanatory diagram illustrating a part of the individual time table 236 according to the second embodiment.

The individual time table 236 shown in FIG. 28B is a part of the table generated by the resource allocation module 233 in the second embodiment. The individual time table 236d for the switch 10d, the individual time table 236h for the switch 10h, and the individual time table for the switch a. 236a.

For example, the schedule server 20 shows a case where the switch 10d is set in advance to transfer the control frame to the switch 10c, the switch 10e, the switch 10f, and the switch 10h on the railway line of T-LAN (# 1). .

In this case, the resource allocation module 233 uses the port “2” for the switch 10d to communicate with the master 60a and the ports “4” and “5” for the switch 10c, the switch 10e and the switch 10f to communicate with the switch 10d. And “6” are combined, and this port combination is stored in the individual time table 236d in step 109.

In step 1934, the switching module 193 of the switch 10d duplicates three control frames received from the port “2” at the time “000”-“009” indicated by the time 2361 of the individual time table 236d. Each of the duplicated control frames is output from the ports “4”, “5”, and “6” in the individual time table 236d.

According to the second embodiment, the control frame can be transferred in real time in a wide range of networks by multicasting the control frame to the switch 10. In addition, by multicasting the control frame, a plurality of slaves can share a route, so that the network utilization efficiency is improved.

In the second embodiment, the route assigned to the T-LAN is determined in advance. However, the schedule server 20 is based on the network topology indicating the connection relationship between the master 60, the switch 10, and the slave 7 provided in the communication system. The route may be searched.

Further, the present invention is not limited to the above-described embodiments, and includes various modifications. The above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described.

Further, a part of the configuration of one embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of one embodiment. Further, it is possible to add, delete, and replace other configurations for a part of the configuration of each embodiment.

In addition, each of the above-described configurations, functions, processing units, processing procedures, and the like may be realized in hardware by designing a part or all of them, for example, with an integrated circuit. Each of the above-described configurations, functions, and the like may be realized by software by the processor interpreting and executing a program that realizes each function. Information such as programs, tables, and files that realize each function is stored in a memory, a hard disk, a recording device such as SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD. be able to.

Also, the control lines and information lines indicate what is considered necessary for the explanation, and not all the control lines and information lines on the product are necessarily shown. Actually, it may be considered that almost all the components are connected to each other.

∙ A control system having multiple networks with guaranteed real-time communication can be applied to a wide range of social infrastructure.

Claims

A network system having a plurality of network devices and a schedule server connected to the plurality of network devices,
Each of the plurality of network devices includes a clock that provides a first time synchronized with the other network devices, and a plurality of ports.
The schedule server
A resource allocation unit that determines a path through which the plurality of network devices transfer frames, and generates a first schedule indicating a combination of the ports through which the determined path passes and a time at which the frames are transferred;
A distribution unit that distributes the first schedule to the plurality of network devices;
Each of the plurality of network devices is
A receiving unit for receiving the first schedule from the schedule server;
With reference to the first schedule, when the first time is the time indicated by the first schedule, the frame input to one port of the port combination indicated by the first schedule is And a switch unit that outputs the data as it is from the other port of the combination.
The network system according to claim 1,
The plurality of switch units are:
When the first time is a time indicated by the first schedule, a frame input to one port of the combination of ports indicated by the first schedule is read without reading address information from the input frame. , Output as is from the other port of the combination,
When the first time is a time other than the time indicated by the first schedule, the address information of the received frame is read, and the received frame is transferred according to the read address information. Network system.
The network system according to claim 1 or 2,
The resource allocation unit generates the first schedule including a combination of one first port and a plurality of second ports;
The plurality of switch units are:
When the first time is the time indicated by the first schedule, the frame input to the first port is duplicated by the number of the plurality of second ports,
A network system, wherein the plurality of duplicated frames are output as they are from the plurality of second ports.
The network system according to claim 3,
The network system further includes a master device connected to the plurality of network devices,
The resource allocation unit generates a second schedule indicating a time at which the master device outputs a frame,
The delivery unit delivers the second schedule to the master device;
The master device is
A clock providing a second time synchronized with the first time;
A network system comprising: a transmission unit that refers to the second schedule and outputs a frame when the second time is a time indicated by the second schedule.
The network system according to claim 4, wherein
The resource allocation unit
A plurality of network devices determining a first path for transferring a first frame and a second path for transferring a second frame;
When at least a part of the first path and at least a part of the second path overlap, a time at which the second frame is transferred is different from a time at which the first frame is transferred. Decided on
A combination of ports through which the first path passes, a time at which the determined first frame is transferred, a combination of ports through which the second path passes, and a transfer of the determined second frame And generating a first schedule indicating a time to perform.
The network system according to claim 5,
The schedule server holds an application schedule indicating a time to apply the first schedule and the second schedule;
The distribution unit distributes the application schedule to the plurality of network devices and the master device,
The plurality of switch units refer to the received first schedule when the first time is a time indicated by the application schedule,
The transmission system refers to the received second schedule when the second time is a time indicated by the application schedule.
A communication method in a network system having a plurality of network devices and a schedule server connected to the plurality of network devices,
Each of the plurality of network devices includes a receiving unit, a switch unit, a clock that provides a first time synchronized with another network device, and a plurality of ports.
The schedule server has a first processor;
The method
The first processor determines a route through which the plurality of network devices transfer frames, and generates a first schedule indicating a combination of the ports through which the determined route passes and a time to transfer the frames Resource allocation procedure
A delivery procedure in which the first processor delivers the first schedule to the plurality of network devices;
A first procedure in which the plurality of receiving units receive the first schedule from the schedule server;
When the plurality of switch units refer to the first schedule and the first time is a time indicated by the first schedule, one of the port combinations indicated by the first schedule is set to one port. And a second procedure for outputting the input frame as it is from the other port of the combination.
The communication method according to claim 7, wherein
The second procedure is:
When the first time is a time indicated by the first schedule, the plurality of switch units change the frame input to one port of the port combination indicated by the first schedule to the input frame. Without reading out the address information from the other port of the combination,
When the first time is a time other than the time indicated by the first schedule, the plurality of switch units read address information of the received frame, and the received frame is read according to the read address information. And a procedure for transferring the communication method.
The communication method according to claim 7 or 8,
The resource allocation procedure includes a procedure in which the first processor generates the first schedule including a combination of one first port and a plurality of second ports;
The second procedure is:
When the first time is a time indicated by the first schedule, the plurality of switch units replicates the frame input to the first port by the number of the plurality of second ports. Procedure and
A communication method comprising: a step in which the plurality of switch units output the plurality of duplicated frames as they are from the plurality of second ports.
The communication method according to claim 9, comprising:
The network system further includes a master device connected to the plurality of network devices,
The master device includes a second processor and a clock for providing a second time synchronized with the first time;
The resource allocation procedure includes a procedure in which the first processor generates a second schedule indicating a time at which the master device outputs a frame;
The delivery procedure includes a procedure in which the first processor delivers the second schedule to the master device,
The method includes a transmission procedure in which the second processor refers to the second schedule and outputs a frame when the second time is a time indicated by the second schedule. Communication method.
The communication method according to claim 10, comprising:
The resource allocation procedure includes:
A procedure in which the first processor determines a first path through which the plurality of network devices transfer a first frame and a second path through which a second frame is transferred;
When at least one procedure of the first route and at least one procedure of the second route overlap, the first processor transfers the second frame and the first frame. A procedure for determining a different time to be
The first processor determines the combination of ports through which the first path passes, the time to transfer the determined first frame, and the combination of ports through which the second path passes. And a procedure for generating a first schedule indicating the time for transferring the second frame.
The communication method according to claim 11, comprising:
The schedule server holds an application schedule indicating a time to apply the first schedule and the second schedule;
The distribution procedure includes a procedure in which the first processor distributes the application schedule to the plurality of network devices and the master device,
The second procedure includes a procedure in which the plurality of switch units refer to the received first schedule when the first time is a time indicated by the application schedule,
The transmission procedure includes a procedure in which the second processor refers to the received second schedule when the second time is a time indicated by the application schedule.
A network device for forwarding frames,
A clock that provides time, and
Multiple ports,
A receiving unit that receives, from a schedule server, a first schedule indicating a combination of the ports through which a path for transferring a frame passes and a time for transferring the frame;
With reference to the first schedule, and when the time provided by the clock is the time indicated by the first schedule, a frame input to one port of the port combination indicated by the first schedule, And a switch unit that outputs the data as it is from the other port of the combination.
14. The network device according to claim 13, wherein
The switch part is
When the time provided by the clock is the time indicated by the first schedule, the address information of the frame input to one port of the port combination indicated by the first schedule is read from the input frame. And output it directly from the other port of the combination,
When the time provided by the clock is a time other than the time indicated by the first schedule, the address information of the received frame is read, and the received frame is transferred according to the read address information. Network device.
The network device according to claim 13 or 14,
The receiving unit receives the first schedule including a combination of one first port and a plurality of second ports;
The switch part is
When the time provided by the clock is the time indicated by the first schedule, the frame input to the first port is duplicated by the number of the plurality of second ports,
The network apparatus, wherein the plurality of duplicated frames are output as they are from the plurality of second ports.