WO2001058092A1

WO2001058092A1 - Network user, in particular field device with telegram path guidance between ports and a microprocessor interface and a network therewith

Info

Publication number: WO2001058092A1
Application number: PCT/DE2001/000414
Authority: WO
Inventors: Karl Glas; Dieter Klotz; Christoph MÜNCH
Original assignee: Siemens Aktiengesellschaft
Priority date: 2000-02-02
Filing date: 2001-02-02
Publication date: 2001-08-09
Also published as: DE10004426A1

Abstract

The invention relates to a network user, in particular a field device, comprising several ports for connection of further network components. An interface (25), a so-called microprocessor-interface, is provided to link the ports (28...31) with a user-internal processor bus. A control unit (46), a so-called switch control, provides message route management between the ports (28...31) and the microprocessor interface (25). The network user can, thus, be connected in a linear structure without using further switches. Furthermore, a redundant layout of the transmission channels and a clock-time synchronisation are possible.

Description

description

NETWORK SUBSCRIBER, ESPECIALLY FIELD DEVICE, WITH TELEGRAM DEFLECTION BETWEEN PORTS AND A MICROPROCESSOR INTERFACE, AND NETWORK WITH SUCH A NETWORK SUBSCRIBER

The invention relates to a network participant, in particular a field device, and a network in which a plurality of such network participants are interconnected.

From the Siemens publication "Industrial Ethernet with Switchmg and Fast Ethernet Technology *, Basic Information 10/98, order number 6ZB5530-1AH02-0BA0, a network subscriber, a so-called optical switch module (OSM) with several ports for connecting further network components, is already known. To set up an Ethernet or Fast Ethernet network, for example, end devices are connected to the ports and are therefore connected to each other via the OSM. A star-shaped network topology is created. The OSM is a so-called Layer 2 component, which analyzes the destination and source address of the circulating telegrams and forwards the telegrams via their ports according to their addresses, that is, carries out telegram control. A decoupling of the individual communication channels is achieved and the full transmission capacity is available for each communication channel.

Networks based on star-shaped structures are widespread in the field of office communication. In automation systems, Ethernet is often used for communication at the control level and the cell level. However, Ethernet has found practically no input at the field level. At the field level, fieldbuses, such as PROFIBUS, are used which have a bus structure, ie in which the network participants are interconnected in a line. Communication takes place at the field level with a different physics and a different protocol than in the line or control level. So-called gateway components must be used for vertical data traffic across the boundaries between the individual levels. This is associated with disadvantages such as low speed in the vertical direction, a lack of transparency in the data, and high software and maintenance costs for the automation system. In addition, increased performance is desired at the field level compared to current field buses. These disadvantages are exacerbated by the fact that vertical data traffic in automation systems increases steadily, since operational data acquisition, alarm management, visualization and remote maintenance require this to an increased extent.

From the essay "The network is the controller * by Martin

Buchwitz and Martin Jetter, published in Electronics 8/1999, pages 38 to 56, a controller for use in automation technology is known which has a switch with multiple ports. The individual components of the controller are connected to the switch via a Fast Ethernet interface and can communicate with each other via the switch. However, field devices continue to be connected in a star shape to the external ports of the controller or to an external switch.

The invention has for its object to provide a network participant and a network that are easily configurable for communication and require little effort.

To solve this problem, the new network participant of the type mentioned has the features specified in the characterizing part of claim 1. A network with several such network components as well as advantageous embodiments of the invention are described in claim 6 or the subclaims. The invention has the advantage that network participants, in particular field devices, can be interconnected in a line structure in the manner customary for users of field buses. A separate switch, as would be required with a star-shaped structure, is not required. In particular in the case of a network according to Ethernet, Fast Ethernet or Gigabit Ethernet specification, the invention enables the construction of a large network, since only the distance between two network components may not exceed certain limits, but the length of the line structure non ^¬ is limited. In addition, the integration of switch functions in the network participants has the advantage that, particularly in the case of Ethernet, the CSMA / CD access control can be deactivated and the network receives a deterministic behavior. The area of application of the network participants and the network is thus expanded to include application cases in which real-time behavior is required.

If four ports are provided on the network subscribers for connecting further network components, subscribers can be interconnected to form a two- or three-dimensional network structure, since with these two ports m one line is used to integrate the network subscriber, two more are used to connect the line to another line can be.

An embodiment of the interfaces implemented by the ports in accordance with the Ethernet or Fast Ethernet specification has the advantage that technology knowledge already available from other areas can be used when implementing, for example, a fieldbus with such network components. In this way, a continuous network is obtained for the office, control, line and field levels, which enables transparent access to any data. A gateway for coupling network areas with different physics and different protocols is advantageously not required. Also stand out Networks based on the Ethernet specification due to the high performance of data transmission. They offer cost advantages thanks to a widely available technology and components that are available in high numbers of pieces. It is possible to connect a large number of network participants to a network. The advantages of the line or bus structure of the network favored in the field area, for example the advantage of simple connection of the subscribers, are combined by the invention with the above-mentioned advantages of networks based on the Ethernet specification according to IEEE 802.3. The switch functions integrated into the network participants assume the function of a previously separately installed network component, for example a switch, which is no longer required.

A further improvement in applications in automation technology, in particular in the use of time-critical applications, is achieved if the control unit of the switch function is designed in such a way that a transmission priority of the telegrams to be sent is evaluated and telegrams with high priority over telegrams with a lower one Priority will be sent.

A microprocessor for correcting an internal clock on the basis of time information received using m telegrams has the advantage that different network participants can carry out synchronized actions.

With such network participants, as already explained above, a network topology with a linear arrangement of the participants can be produced in a simple manner. If the two line ends are connected to each other by a network participant, this creates a ring-shaped network topology. The network participant at one line end has the function of a redundancy manager, which is designed to close the two line ends in the fault-free case, ie when there is no interruption within the rows disconnect and connect the two line ends in the event of an error. Within the series, a rapid elimination of an occurring error is advantageously made possible and thus the availability of the network is increased. Such a redundant ring, with respect to each operation timing of the communication comprises a linienfor strength Netzwerktopo ^¬ logy, is hereinafter referred to also as a series since it only with respect to the redundancy differs from this.

A communication channel between two rows formed from several network participants can be designed redundantly. M by a simple manner, a first communication channel ^¬ by a connection of one port of a first network subscriber, which is arranged m the first row, of a second network subscriber which is arranged m the second row is formed with a port. Furthermore, a second communication channel is realized by connecting a port of a third network subscriber, which in turn is arranged in the first row, to a port of a fourth network subscriber m in the second row. The third network participant can monitor the first communication channel for freedom from interruption and, when the first communication channel is interrupted, can set the second communication channel m to the active state. In this way, you can obtain a redundant design of the communication channels between rows with little effort. This type of redundancy can be implemented inexpensively and is characterized by a short reconfiguration time.

For the redundant design of the communication channel between two rows, no additional hardware expenditure, in particular for cables, is required if the third network participant evaluates message telegrams, so-called port select telegrams, to monitor the first communication channel, which the first network participant cyclically uses with an uninterrupted communication channel the network participant sends which m are arranged in the first row.

Advantageously, a small maximum number of devices to be traversed for cross-row telegram traffic can be achieved if of two operational communication channels that communication channel is activated, which is arranged by connecting a port of a network participant, the m of the first row at the smallest distance from the center of the row is formed with a port of a network subscriber which is arranged in the second row. A conflict that arises when two network participants are equidistant can be resolved in a simple manner by prioritizing device addresses, e.g. by giving preference to the network device with the smaller device address.

The network participants and a network constructed with such network participants are characterized by great freedom with regard to the interconnection of participants. Three-dimensional structures can also be realized. For this purpose, for example, the first row is connected to a third row via at least one third communication channel. The second row is connected directly or indirectly to the third row via a communication path that does not lead via network participants of the first row, so that a loop would exist when all communication channels of the network were active. At the same time, the network participants in the network are designed such that, by exchanging configuration telegrams, they ensure that the network is loop-free with regard to communication and that each network participant communicates with every other network participant via exactly one active communication path, i.e. a communication path that only includes communication channels that are in the active state. Since the redundant transmission paths present in loops are managed by configuration telegrams, no additional management hardware is required.

The established algorithm of the tensioning tree method can advantageously be carried out to establish the loop-free nature of the network, with each row as a virtual network subscriber and each operational and potentially active port of a network subscriber of a row, by connecting it to an operational port of a network subscriber another row, an operational communication channel is formed as an operational port of the virtual network subscriber.

In a network, time synchronization of the network participants can also advantageously be carried out. For this purpose, a first network participant sends a first telegram to a second network participant, which contains the time of the first network participant corrected for a transmission time delay. The telegram runtime is stored in the second subscriber over the physical transmission link between the first network subscriber and the second network subscriber. It can, for example, be entered manually or previously measured by another telegram traffic. The second network participant measures the time delay since receipt of the first telegram and corrects the time received in the first telegram by the running time and a measured reception time delay. The second network subscriber is thus advantageously able at any time to determine a synchronized time from the sum of the time received in the first telegram and the delay in reception time supplemented by the transit time. Variable transmission and reception delays have no effect on the result of the synchronization. If the second network participant is also designed to send a second telegram for time synchronization to a third network participant, which is based on the runtime and the delay between receipt of the first If the received time contains the corrected telegram and the second telegram, the iterative forwarding of the respectively corrected time from network participants to network participants becomes possible. Identical correction mechanisms can be applied in the receiving network participants. The runtime stored in each network subscriber corresponds to the runtime over the physical transmission link between the last sending and receiving network subscriber.

As an alternative to sending a telegram for time synchronization from a first network participant to a second, it is also possible to synchronize the time with two telegrams. For this purpose, the first network subscriber sends a first telegram for time synchronization to a second network subscriber and at the same time stores a time of the first network subscriber corrected for the transmission time delay. The runtime of the telegram is stored in the second network participant over the physical transmission link between the first network participant and the second network participant. The second network participant measures the time delay since receipt of the first telegram without, however, stopping the timer used for this. The first network participant now sends a second telegram, which contains the time of the first network participant corrected for the transmission time delay, to the second network participant. The second network subscriber corrects the time received in the second telegram by the runtime and the reception delay, which is measured in relation to the reception of the first telegram. This procedure with two telegrams has the same advantages that are also associated with the procedure for time synchronization with only one telegram.

For a network with more than two network participants, the method with two telegrams can also be tive process to be continued. For this purpose, the second network participant forwards the first telegram to a third network participant and measures its delay time for the telegram forwarding. In a third telegram, the second network participant sends a received time to the third network participant, corrected for the running time and the delay time of the telegram forwarding of the first telegram.

The actual transmission time delay in the time synchronization is advantageously taken into account if the first network subscriber has a first timer for determining the transmission time delay, which he starts a list of the transmission orders when a telegram m is entered and after the telegram is made available for physical transmission as a value reads the transmission time delay by which the time of the telegram entry is to be corrected.

Likewise, the actual reception time delay during transmission is advantageously taken into account if the second network subscriber has a second timer for determining the reception time delay, which he starts when a first telegram is received from a physical transmission link.

The runtime of the telegram over the physical transmission link between the first network participant and the second network participant can be stored as the start value m the second timer before it starts. This has the advantage that the sum of the transit time and the delay in reception time can be determined by only one timer.

In a further development, the start and end of the duration of a telegram can each be determined as the point in time at which a characteristic field of a telegram with a fixed distance from the telegra starts a medium independent Interface of the first network participant leaves or m enters a medium independent interface of the second network participant. Measurements of the transmission time delay, transit time and reception time delay are advantageous regardless of the length of the respective telegram.

If the network components meet the Ethernet, Fast Ethernet or Gigabit Ethernet specification, then the Type field can be used with advantage as the characteristic field of the telegram. In the second network participant, upon receipt of the type field, it is known that it is a telegram that is used for time synchronization, and the necessary mechanisms can be initiated. The data field is located behind the type field and can be changed in the transmitter until the type field is output to the Media Independent Interface. This makes it possible to transmit the time corrected for the transmission time delay in the same telegram.

The runtime of a telegram over the physical transmission link can be exactly determined by the first network participant sending a first telegram to a second network participant for determining the runtime and starting a response time timer after the telegram has been provided for the physical transmission. After receiving the first telegram for determining the runtime, the second network participant starts a timer for measuring the time spent in the vehicle and stops the timer after providing a second telegram for physical transmission to the first network participant. The measured residence tent is transmitted to the first network participant in the second telegram to determine the running time. After receiving the second telegram from the physical transmission, the first network participant stops the response time timer and calculates the duration of a telegram over the physical transmission path as half the difference between the measured response time and the residence tent in the second network participant. With reference to the drawings, in which exemplary embodiments of the invention are shown, the invention and embodiments and advantages are explained in more detail below.

Show it:

1 shows a communication structure of an automation system, FIG. 2 shows a block diagram of an interface of a network participant, FIG. 3 shows a series connection of network participants with line-shaped network topology, FIG. 4 shows a series connection of network participants with two communication channels, FIG. 5 shows a two-dimensional connection of network participants,

Figure 6 shows a two-dimensional interconnection with redundant

Figure 7 shows a two-dimensional network after redundancy management, Figure 8 shows a three-dimensional network after redundancy management, Figure 9 shows the basic structure of a telegram and Figure 10 shows the structure of an order list.

FIG. 1 shows the structure of a communication structure in an automation system. Communication takes place continuously at the control, line and field level through a network, the data transmission of which complies with the Fast Ethernet standard according to IEEE 802.3u. At the field level there are a sensor 1, for example a pressure transducer, a sensor 2, here an ultrasonic flow transducer, an actuator 3, a control valve for setting a flow, and a programmable logic controller 4 with twisted-pair lines 5, 6 and 7 interconnected in bus form. The programmable logic controller 4 forms together with the two

Sensors 1 and 2 and the actuator 3 a control circuit in which the position of the control valve m depending on the Measured values of the pressure transducer and the flow transducer is specified. The programmable logic controller 4 is connected to a switch 9 via a twisted pair line 8. A cell computer 10, a master computer 11 and a firewall 12, with which a secure Ethernet access is implemented, are also connected to the switch 9 in a star-shaped topology. With a line 13, further cell computers, not shown in the figure for the sake of clarity, from adjacent cells of the automation system are connected to the switch 9.

The example shown in FIG. 1 clearly shows the advantage of high data transparency across all levels. The same transmission standard is used for control, line and field level. The field devices 1, 2 and 3 are connected to the programmable logic controller 4 m in the manner customary in the user m of a linear topology. Another advantage is the consistent use of uniform addresses for the individual network participants. An address conversion, as would be required if different standards were used at the different levels, can be omitted in the new network and the new network participants in an automation system.

Figure 2 shows the basic structure of the communication interface of a network participant, for example the sensor 2 m Figure 1. Application-specific circuit parts, such as a mechanical-electrical converter element, a device for signal preprocessing and a voltage supply are not shown for reasons of clarity. The parts combined in a rectangle 20 can be integrated in an ASIC (Application Specific Integrated Circuit). Communication with the application-specific switching points of the network subscriber takes place via a microprocessor bus 21, to which a RAM 22, a microprocessor 23 and a microprocessor interface 25 are connected. With broken lines, m is the representation ment of the microprocessor 23 indicated that the integration m the ASIC is optional. Its functions could be taken over by an external processor. The task of the microprocessor 23 is the execution of user programs and communication functions, for example the handling of TCP / IP. Another task can be the management of send and receive lists of telegrams of different priority in the external RAM 22. The microprocessor 23 selects an order from a transmission list in the external RAM 22 and starts a DMA controller 26, which is referred to as DMA 1 control, via a microprocessor interface 25, after having previously m the number of the DMA controller 26 sending data bytes and a pointer that points to the data byte to be sent. If the send job has been completely transmitted by the DMÄ controller 26 to a transmit buffer 27, the microprocessor 23 removes this send job from the send list in the RAM 22 and processes the next send job if the send list is not empty and still in the transmit buffer 27 free space is available.

Four Ethernet controllers 28, 29, 30 and 31 are also integrated in the ASIC 20. Each of these Ethernet controllers enters the data bytes of a completely received telegram via a multiplexer 32, a DMA controller 33, which is also referred to as DMA 2 control, and the microprocessor interface 25 m a receive list in RAM 22. The microprocessor 23 accesses the reception list and evaluates the received data in accordance with an application program.

The microprocessor interface 25 forms the essential interface between the Ethernet controllers 28... 31 and the microprocessor bus 21. It controls or operates the write and read accesses via the DMA controller 33 or the DMA controller 26 done on the RAM 22. Are both DMA controllers 26 and 33 simultaneously DMA- Requirements, so the microprocessor interface 25 decides on the access rights of the two DMA channels. Via the microprocessor interface 25, the microprocessor 23 can also write parameter registers 34 which are required for operating the communication interface of the network subscriber. Examples include a pointer to the start of the high-pore memory area in a transmit buffer of an Ethernet controller, a pointer to the start of the high-priority memory area in each of the receive buffers of the Ethernet controller

Operating mode register for general control bits, an address of the series to which the network participant belongs, a cycle time for so-called port select telegrams and settings for various monitoring time intervals.

The transmit buffer 27 has a size of more than three kilobytes and is divided into a memory area for high-priority and a memory area for low-priority telegrams. The ratio of the two memory areas can be parameterized. The memory areas of the transmit buffer for high and low priority data are each implemented as a ring buffer. Sending the data from the transmit buffer 27 via one or more Ethernet controllers 28 ... 31 begins when the number of bytes entered in a telegram is greater than the parameterizable fill level, or when a complete telegram from the RAM 22 was copied into the transmit buffer 27 and at least one ethernet controller 28 ... 31 is free for transmission.

The Ethernet controllers 28 ... 31 are each constructed identically. Their structure is explained in more detail using the example of the Ethernet controller. A device 40, which is referred to as Transmit Control, contains a control unit which is responsible for the transmission of telegrams, for repetitions, transmission abort etc. It forms the interface between the internal controller clock and the send clock. For storing transmit status information for low-priority users and high-speed telegrams, a transmit status register m of the device 40 is provided. If a telegram was sent without errors via the port, a corresponding interrupt is generated.

A Media Independent Interface 41 (MII) integrates the MAC sublayer of Layer 2 according to the seven-layer model, i.e. of the data link layer. This forms an interface to a module 36 for physical data transmission. The MII 41 also contains a transmit function block 42 and a receive function block 43. In addition, a MAC control block (not shown in FIG. 2), an address filter, a statistic payer and a host interface are integrated. Control and configuration data can be transmitted to module 36 and via the media independent interface

Status information can be read from this. The individual functions of the transmit function block 42 are: mapping of the bytes to be sent, detection of collisions in half-duplex operation and execution of a back-off algorithm, provision of transmit status information to the device 40 after termination of one Transmission process, compliance with the idle time inter-packet gap (IPG) between two telegrams, supplementing the transmission data with a preamble, a start-off-frame delimeter (SFD) and a parameterizable cyclic redundancy check word (CRC) , Filling a telegram with pad bytes if the telegram length was <60 bytes, and aborting a transmission process on request.

The receive function block 43 makes the received bytes available to a device 44, which is referred to as receive control. The receive function block 43 recognizes the start-of-frame delimeter and a VLAN frame. He checks the Langenfeld and the CRC word m telegrams. After the reception process has ended, receive status information is made available to the device 44. Block 43 recognizes and removes preambles and start-of-frame delimeters in the case of telegrams. If the free storage space falls below one Receive buffer 45 of the Ethernet controller 28 in full duplex mode a predetermined threshold value, the MAC control block sends a pause control telegram for flow control via the block 36. This telegram causes the connected network subscriber, as long as none Send data telegrams to the Ethernet controller 28 until the time interval sent with the pause control telegram has expired. The address filter carries out telegram filtering in accordance with unicast, multicast and broadcast addresses. To do this, the destination address (DA) received in a telegram is compared with filter addresses. The statistic payers store statistical information about send and receive operations. The host interface allows access to parameter registers and statistics payer of the Ethernet controller 28 by the neighboring network participants.

The device 44 contains a control unit which is responsible for receiving telegrams. It forms the interface between the internal clock of the Ethernet controller 28 and the receive clock.

The receive buffer 45 has a size of more than 3 kilobytes. It is divided into a memory area for high-pore and m a memory area for low-priority telegrams. The ratio of the two memory areas can be parameterized.

The memory areas are each implemented as a ring buffer.

The DMA controller 33 controls the DMA transfer from one of the receive buffers m the Ethernet controllers 28 ... 31 to the RAM 22. The DMA transfer begins when m of the receive buffers, for example in the receive Buffer 45, the number of data bytes received has reached a parameterizable minimum fill level or a telegram has been completely received. At the same time, this receive buffer must be selected for the DMA transfer by a module 46, which is referred to as switch control. A multiplexer 47 is connected upstream of the device 40, which is controlled by a control unit 46, which is referred to as switch control.

Switch-Control 46 controls the forwarding of data between the Ethernet controllers 28 ... 31 and the storage of received data if they are intended for the respective network participant. Since the application of the invention is not limited to networks according to the Ethernet specification, the Ethernet controllers 28 ... 31 are also generally referred to below as port 1, port 2, port 3 or port 4. Which ports are released for the forwarding of received data depends on the network structure m the participant is involved in. Switch-Control 46 controls the following actions as a function of the operating status, the network structure, the received destination address and the telegram priority:

If the received destination address is the same as the subscriber's own address, the received telegram is transmitted via the microprocessor interface 25 m to the RAM 22 without being forwarded to other ports.

If a broadcast telegram is received at a port, the telegram m is transferred to RAM 22 and made available to the other released ports for transmission.

If a telegram with a multicast address that corresponds to one of the multicast addresses stored in a filter table 48 is received at a port, the telegram m is transferred to the RAM 22 and made available to the other released ports for transmission. If the received destination address is different from your own subscriber address and the multicast addresses, the telegram is made available to the other released ports for sending without being saved for further processing.

For telegrams with so-called VLAN bytes, for example, eight priority levels are available. Stand If several telegrams are to be sent simultaneously, the order in which the telegrams are sent is determined according to their transmission priority.

In a monitor mode operating state, all telegrams which meet the parameterized filter conditions are transmitted to the RAM 22.

The forwarding of telegrams taking into account a modified tensioning tree algorithm.

Switch control 46 contains further parameters, the meanings of which will be explained in more detail later: a row address R _E3 , which corresponds to the row connected to port 3, a row address R ₁₄ , which represents the address of the row connected to port 4, a number N _{R1 of} the transmission links up to port 1, a number N * - _^ der

Transmission routes up to port 3, a value | N _Ri - N _R | - of the respective network participant, a value | N _K ι - N _R | _{^ α} from the sender of a received port select telegram, an amount | N _Ri -

as the smallest value received so far, a source address A _s _. _n that of a received telegram, one

Source address A _st ored of the telegram with the smallest value of amount | N _R ι - N _R | _{m; L} n, a best received combination (Root_ID.Cost .Transmιtter_ID.Port_ID) for port 3, a best received combination (Root_ID.Cost. Transmιtter_ID. Port_ID) _F for port 4, a best received combination (Root_ID. Cost. Transmιtter_ID . Post_ID) of the series, a message interval counter for a cycle time of port select telegrams, a timeout counter for a timeout interval on port 1, a timeout counter for a timeout interval on port 2, a timeout -Payers for one

Timeout interval on port 3, an active time payer for a time interval that begins with the last receipt of a port select telegram on port 4, a combination aging payer for a maximum time interval within which a configuration telegram is received must, otherwise the saved combination Root_ID. Cost. Transmιtter_IC. Port_ID is deleted, a payer for the time interval Δt _r - ^ _e i _ü. after which Por: 3 switches a series from inactive to active, which corresponds to twice the worst case throughput time of a port select telegram through the series, and a payer for the time interval Δt _ne t-deiay. after which em port is switched from potentially active to active and which corresponds to twice the worst case throughput time of a configuration telegram through the network.

The user can enter multicast and virtual LAN identification addresses, so-called VLAN addresses, in the filter table 48. A multicast or VLAN telegram is only accepted if the address received matches one of the addresses m in the filter table 48.

A device 50 for redundancy control is intended to

Ensure the network that detected physical errors do not impair communication between the network components. Firstly, there is redundancy within each row formed with network participants. To do this, the network participants connected in series must form an Rmg, which is open at one point in the event of a fault, but can be closed in the event of an error. Secondly, redundancy with several communication channels between the rows is possible. To do this, a number of network participants must be connected to an adjacent row via at least two ports. Only one communication channel between two rows is ever active, ie data telegrams are only exchanged via this communication path. If an active communication channel between two rows is identified as faulty, it is deactivated and switched to another communication channel. In order to implement the redundancy, the device has 50 cycle time registers with a parameterized cycle time for test telegrams, a cycle time number for generating a cycle time interval, a control unit for switching over to a redundant communication channel and for initiating a so-called Lmk-Up or lmk-down telegrams, em Series run time register with a parameterized worst-case cycle time of a message used by a number and em Rei ^¬ henlaufzeitzahler for generating a series term.

Certain events are communicated to the microprocessor 23 via a device 51 for interrupt control, which is also referred to as interrupt control. Essentially, these are messages from sent or received telegrams and error messages. The device 51 contains an interrupt request register, an interrupt mask register, an interrupt register and an interrupt acknowledge register. Every event is stored in the interrupt request register. Individual events can be suppressed via the interrupt mask register. Only events that are not masked by the interrupt mask register appear in the interrupt register. The entry m the interrupt request register, however, is independent of the interrupt mask in the interrupt mask register. Bits in the interrupt request register can be reset with write access to the interrupt acknowledge register. Em module 52 contains special user functions which have to be integrated in the communication interface of the network participant. A partial function is implemented with a module 53 for time synchronization, another with a module 54 for aquidistance, which will be explained in more detail later. For the ports 1 to 4 there is a delay timer 1 to 4 with the reference numerals 57, 58, 59 and 60, which determines the transmission time between the respective network subscriber and the network subscriber connected via the respective port. The respective delay timer is also used as a lead time timer (DLZ timer) for the respective port. Furthermore, for each port there is an aquidistance timer, an auxiliary timer for a transmission time Δti via the respective port and a parameter Δt _DZ , which is the sum of the throughput times in the send and receive direction and the line runtime between corresponds to the communication interface and the network participant connected via the respective port. There is also a local clock 37 m from the network subscriber, the time of which can be read and set via the microprocessor bus 21.

An integrated serial peripheral interface (SPI) 55 is a simple but powerful serial bus system for connecting peripheral components, e.g. EEPROMs. Em mertegπeses I / O interface 56 is a parallel interface with 12 configurable Em- and outputs. LEDs for status display can be controlled via this interface, for example.

Each port of the communication interface can be parameterized operated in half-duplex or in full-duplex mode. While half-duplex mode is set on one port, full-duplex mode can be parameterized on another port at the same time. In full duplex mode, telegrams can be sent and received at the same time. This is not possible in half-duplex mode.

An application-specific application program, which can be stored on the RAM 22, for example, carries data to be sent m an order list in the RAM 22 em. The DMA controller 26 copies data from this job list into the transmit buffer 27. Compiled telegrams are passed on to the released Ethernet controllers 28 ... 31. If a transmission conflict occurs because other telegrams are currently being routed through the communication interface controlled by switch control 46, the transmit buffer 27 should be able to store two complete Ethernet telegrams. The sending of the data from the transmit buffer 27 begins when a number of data bytes to be parameterized or a complete telegram has been transferred from the RAM 22 into the transmit buffer 27 and at least em

Ethernet controller is free. The telegram remains stored in the transmit buffer 27 until it is clear of all given Ethernet controller 28 ... 31 was sent. The number of data bytes of a telegram, which must be stored in the transmit buffer 27 at least before being sent, must be parameterized so that the telegram is sent without gaps. Otherwise the telegram will be received incorrectly by other network participants. If telegrams of different priority are stored in the transmit buffer, the telegrams are sent according to their transmission priority.

FIG. 3 shows an example of an interconnection of three network participants 61, 62 and 63 m in a line-shaped structure. For a better clarity of the representation, the ports 1 to 4 of the communication interface of the network participants 61, 62 and 63 m are subdivided into switching points T1 to T4 for the sending direction and switching points R1 to R4 for the receiving direction. For example, port 2 can also be referred to as port T2 / R2. For a linear structure, in the manner shown, port T2 / R2 of network participant 61 is connected to port T1 / R1 of network participant 62 and port T2 / R2 of network participant 62 is connected to port T1 / R1 of network participant 63. The data transmission takes place with a twisted pair cable for each transmission. The ports involved can thus be operated in full duplex mode. End devices can optionally be connected to the open ports T3 / R3 and T4 / R4 in FIG. 3 and to the port T1 / R1 of the network subscriber 61 or the port T2 / R2 of the network subscriber 63 and thus coupled to the network.

FIG. 4 shows a number of network participants 70, 71, 72 and 73, each of which can exchange data with one another via two communication channels. The communication channels are each connected by a connection of the port T2 / R2 to the port Tl / Rl of the neighboring network participant and the port

T4 / R4 realized with the port T3 / R3 of the neighboring network subscriber m in the manner shown. For example, as a high-pore and a low-priority communication channel and the data throughput can be doubled. There is no data exchange between the communication channels, ie a telegram received at the circuit part R1 can - if necessary - only be forwarded by the circuit part T2. Both communication channels are operated in full duplex mode.

An example of a two-dimensional interconnection of the network subscribers is shown in FIG. 5. Network subscribers 80, 81 and 82 are interconnected in a row in the manner already described with reference to FIG. 3. In addition, network participants 83, 84 and 85 form a series and network participants 86, 87 and 88 form a series. Terminals 89, 90 and 91 are connected to ports T4 / R4 of network participants 80, 81 and 84, respectively, and terminals 92, 93 and 94 are connected to ports T3 / R3 of network participants 83, 84 and 87 Connecting the port T4 / R4 of the network subscriber 82 to the port T3 / R3 of the network subscriber 85 is a communication channel between the respective rows. Correspondingly, two communication channels are formed between the network participants 83 and 86 and between the network participants 85 and 88. To ensure that there is no loop in the network, only one of the communication channels may be activated at a time.

The rows formed from the subscribers 80 to 82, 83 to 85 and 86 to 88 are each assigned a unique row address R _k , which is stored in a parameter register “row address”.

To clarify the redundancy control, another two-dimensional network is shown in FIG. 6 e. 8 network participants 100 ... 107, 110 ... 117, 120 ... 127 and 130 ... 137 are connected in a row. Both the broken lines and the solid lines between the network participants represent communication channels. However, it must be ensured that only one communication path is used between any two network participants in the entire network. Loops would occur with several possible communication paths, ie telegrams would multiply and circulate. To such situations vermei ^¬ the one who Spannmg tree algorithm has been developed. Data telegrams are only received by ports, forwarded to ports and sent by ports that are contained in the voltage tree. The remaining ports are deactivated. Deak ^¬ tivierte communication channels are illustrated m FIG 6 by broken lines, activated by solid lines. For redundancy within a row, the two line ends are connected to one another, for example at network participants 100 and 107. In the case of an error, the communication channel created in this way is deactivated, and in the case of an error m the active state is set. This redundancy requires an uninterrupted line structure. Since the Spannmg-Tree algorithm may also have interrupted a connection via the ports T1 / R1 and T2 / R2, it cannot be used unchanged. In the following, a method is presented which ensures loop freedom for a network of interconnected rows of network participants without having to interrupt a row. If necessary, only ports T3 / R3 are deactivated. Only one communication channel may be active between two rows at a time, ie data telegrams are exchanged via this communication path. There is no data exchange via the other communication paths between two rows. The selection of the only active communication channel between two rows is made using port select telegrams. These are telegrams that are only forwarded within a row. There is no exchange between the rows.

The task of these port select telegrams is to find a network participant that can communicate with a port T3 / R3 adjacent row is connected and based on the number of network participants is as far as possible from both ends of the line structure, ie the smallest distance from the center of the row. These properties define the only network participant in the series that is actively connected to an adjacent series via port T3 / R3. All other connections via ports T3 / R3 of network participants in the same row are adjacent to this fourth deakti ^¬. No data telegrams are exchanged via deactivated communication channels.

Port select telegrams are clearly identified by an identifier in the type field.

FIG. 7 shows the result of the redundancy control in a two-dimensional network in a different type of representation. The number entered in the individual boxes corresponds to the respective address of the network participant. Port Tl / Rl is on the left, port T2 / R2 on the right, port T3 / R3 on the top and

Port T4 / R4 on the lower side of the network participants. An active communication channel operated in full duplex mode is shown by two solid parallel lines between two participants. The communication channels drawn in with two broken lines are deactivated.

The data area of port select telegrams that are received by a network participant via port T2 / R2 contains information about the number N _{κ of} the transmission links between port Tl / Rl of the network participant at the “right * edge of the row and port T2 / R2 of the respective network subscriber and contains the number N _{Ri of} the subscriber who was the last to forward or send the received port select telegram. The data area of port select telegrams that are received via port T1 / R1 contains the number N _{R1 of} the transmission links between port T2 / R2 of the network subscriber at the “left * edge of the row and the port T1 / R1 of the respective network subscriber as well the number N _{R ^} which is valid for the network subscriber who was the last to forward or send the received port select telegram.

Regardless of the port via which a port select telegram was received, it contains the value | N _Ri - N _R _ li _fιat0 r at

Initiator of the received port select telegram, the 16 bit address R (0 <k <p, p number of rows) of the row to which the initiator of the port select telegram belongs, and a valid bit V for the received value of IN _RI - N _{R ^} linii-iat i and for the row address R. V = 0 means that the received values are invalid. Such a port select telegram was fed in by the network subscriber m, which is not connected to its neighboring row via an operational port T3 / R3. If V = 1, the received values are valid. The port select telegram was fed into the row by a network participant who is connected to an adjacent row via an operational port T3 / R3. A connection to the adjacent row is ready for operation if Δt within a parameterizable time interval. _ιrreo r (timeout interval) a port select telegram is received at port P3 / R3. This is only possible if the port T3 / R3 of the network subscriber in its own row and the port T4 / R4 of the network subscriber in the adjacent row are each in an “lmk-pass *” state. In this state, telegrams can be transmitted in both directions.

In the steady state, only the network participant sends every row cyclically, ie m every reporting interval Δt _M , port select telegrams, which is the only one actively connected to the next row via port T3 / R3. This shows whether this connection between two rows is still active. The following describes a procedure according to which this network subscriber can be tuned:

1. Network participants also send port select telegrams to be forwarded or compiled themselves via their port T3 / R3.

2. Upon receipt of a port select telegram via port T3 / R3, each network participant recognizes whether it is connected to another row via this port.

Network participants set the valid bit V to one if a port select telegram has been received via port T3 / R3.

3. Port select telegrams received on port T3 / R3 are sent back to the sender unchanged. The network subscriber previously saves the address R _n received with the port select telegram of the row connected via port T3 / R3.

4. A timeout counter is assigned to port T3 / R3, which is incremented with an adjustable clock. Each port select telegram received on port T3 / R3 resets this payer. The network participant sets the valid bit V to zero if within a parameterizable

Timeout interval Δttimeo no port select telegram is received.

5. Network participants also send forwardable or self-made port select telegrams

Port T4 / R4.

6. When a port select telegram is received via port T4 / R4, the network participant recognizes that it is connected to another row via this port. 7. Port select telegrams received in port T4 / R4 are returned to the sender unchanged. The network subscriber ^¬ previously stores the program with the port select Tele ^¬ received address R _n of the through port T4 / R4 the connected number.

8.Network participants send their own, i.e. self-made, port select telegrams

8.1. via port T1 / R1 with N _R2 = 1 and port T2 / R2 with N _R1 = 1 when initializing with amount IN _Ri - N _{R Inι} t- _la t-_ =

FFH and valid bit V = 1 if a port select telegram has already been received via port T3 / R3, or valid bit V = 0 if no port select telegram has been received via port T3 / R3. 8.2 via Tl / Rl with N _R2 = 1 and port T2 / R2 with (N _R ι + l) if no port select telegram or data telegram was received at port T2 / R2 within a parameterizable time interval Δttimecj. A timeout counter is assigned to port T2 / R2, which is incremented with an adjustable clock. Each on port

T2 / R2 received port select telegram or data telegram resets this payer.

8.3 via port Tl / Rl with (N _R ^^ + 1) and port T2 / R2 with

(N _R ι + 1) if a port select telegram is received at port T2 / R2 with the received value N _K. ^ _' Not equal to the stored value of N _R _. In addition, N _RZe mpf is saved.

8.4 via port T2 / R2 with (N _R1 + 1), if a port select telegram is received at port T2 / R2 with N _R1Ld3 , _{^ end} -ι not equal (N _R1 +1).

8.5 via port T2 / R2 with N _Ri = 1 and port Tl / Rl with (N _r +1), if within a parameterizable time interval Δt, _ιme jt no port select telegram or data telegram was received at port Tl / Rl , The port T1 / R1 is assigned a timeout counter, which is incremented with an adjustable clock. Each on port Tl / Rl received port select or data telegram resets this counter.

8.6 via port T2 / R2 with (N _R ι _emp f + l) and port Tl / Rl with

(N _R2 +1) if a port select telegram is received at port Tl / Rl with a received value N _R ι _emp f not equal to the stored value of N _Ri . In addition, N _R _{ιempf is} stored.

8.7 via port Tl / Rl with (N _R2 +1) if a port select telegram is received at port Tl / Rl with N _R2 iassender not equal (N _R2 +1).

9. Network participants that are ready for operation

Communication channel are connected to another row via port T3 / R3 and receive a port select telegram with a valid bit V set to one, compare | N _RI - N _R2 limtiator of the received port select telegram with amount | N _R1 - N _R2 | _s of your own station:

9.1 If | NRI - N _R2 limtiator <IN _Ri - N _RS , then

| N _RI - N _R2 limtiator in the data field of the port select telegram not changed during forwarding. The

Network subscriber saves

IN _RI - N _R2 | _mιn = INRI - N _R2 limtiator and sets the address A _s tored = Amitiator. Aι _nιtιator is the source address of the received port select telegram. The value | N _RI - N _R2 is the distance of the receiving station from the center of the row.

9. 2 IS t | NRI - N _R2 = llmtiatcπ I NRI - N _R J _S, SO wi rd Aimtiator wi th its own station address A _s compared: In Aimtiator <A _s the port select message is no change in the data field | NRI - N _R2 limtiator forwarded. The network participant saves

I NRI - N _R2 Lin = I NRI - NR ₃ And Astored = Aι _{nl t} ιator.

If Aimtiator = A ₃ , the received port select telegram is filtered out, since this case is only possible in the event of an error. If A _n i _t i _ator > A _s , the received port select telegram is filtered out and a separate port select telegram with IN _R ι - N _R2 limtiaor is set to I _RI - N _R2 | _{s sent} via the ports T1 / R1 and T2 / R2. The network participant saves

I RI - N _R2 Ln = I NRI - N _R2 ls, and A _Ξ t ₀ red = A _s .

9. 3 I st I NRI - N _R2 itiator> IN _Ri - N _R2 L, the received port select telegram is filtered and a separate port select telegram with IN _Ri - N _R2 limited is set to IN _Ri - N _R | _{s sent} via the ports T1 / R1 and T2 / R2. The network participant saves

I NRI - N _R2 Ln = I NRI ~ N _R2 Is And Asto red = A _s .

10. Network participants that are not connected or connected to another row via a non-operational communication channel via port T3 / R3 and receive a port select telegram with a valid bit V = 1 route the telegram with the received valid bit V and the value I _R ι - N _R2 lim iator within the series. The network user saves

I NRI - NR2 Iπiin ⁼ I NRI - NR ₂ limti ator und A _{Ξ t} ored ⁼ Ai mi at- •

11. Network participants that receive a port select telegram with a valid bit V = 0 forward the telegram with V = 0 and the value I NRI - N _R2 limtiator within the series. The stored values of | N _R1 - N _R2 | _mιn and A _Ξ -ι._ .. remain unchanged.

12. An operational port T3 / R3 is activated when IN _Ri - N _R2 Ln = IN _R I - N _R2 | _G and A _3t ored = A _{Ξ is} stored. This applies to a period of time Δt _rθ wdeia _y , - which corresponds to at least twice the worst case throughput time of a port select telegram through a series.

13. An active port T3 / R3 of a network participant is deactivated if the network participant does not receive a port select telegram from within the timeout interval Δttimeout another row. In addition, the communication channel via port T3 / R3 to the other row is marked as not ready for operation.

An active port T3 / R3 is also deactivated if the network participant receives a port select telegram with t | NRI - N _R2 limt iator <IN _R1 - NR_ L θdβr

IN _R I - N _R2 limtiator = I _Ri - N _R2 and Aι _nιt ι _{a 0} r <A _s . This network participant stores the received values IN _RI - N _R2 limtiaor and A _nl tι _a or and forwards the received telegram. The communication channel via port T3 / R3 to the other row remains operational.

14. Network participants that are ready for operation

Communication channels connected to another row via port T3 / R3 send cyclically m each message interval Δt _M their own port select telegrams.

14.1 via the port T1 / R1 with (N _R2 +1) and the port T2 / R2

(N _R ι + l), if an active port T3 / R3 is deactivated, with | N _R1 - N _R _limtι _α - _r = FFH and valid bit V = 1 until another port select telegram

Network subscriber is received;

14.2 via the port Tl / Rl with (N _R2 +1) and the port T2 / R2 with (N _RI +1), if within a parameterizable time interval Δtι- _ιme _ t neither at port Tl / Rl nor at port T2 / R2 em port select telegram or data telegram was received by the network participant with the only active communication channel to the adjacent row, ie em telegram with the source address A _ΞTO red 14.3 via port Tl / Rl with (N _R2 +1) and port T2 / R2 with (N _RI + 1) if IN _Ri - N _R _ L _r = | N _Ri - N _F I and _Ξ t- _ore d = A _{s is} stored.

In the relaxed state of the method, each network work participant in the row knows the network subscriber who is connected via its port T3 / R3 to an adjacent row and has the smallest distance from the center of the row. Only over this active communication channel, data telegrams are exchanged between the two rows. The connections of the other network participants to the next row are deactivated. FIG. 7 shows a network in this swung state.

If, in deviation from this exemplary embodiment, the only active communication channel of a row to the adjacent row is at the edge of the row, then | N _RI - N _R _ L _n by | N _R1 - N _{R mα} iNπ - N _{R Irι} , _{ιat r} <IN _RX - N _R _ I through IN _R ι - NR. _, Iι _rιl tι _a mt> | N _RI - N _R _ L to replace.

FIG. 8 shows a network with a three-dimensional structure. The arrangement of the ports on the individual boxes, which each represent a network participant with the entered address, is the same as that shown in FIG. 7. The communication channels are also shown in the same way. In the case of a three-dimensional structure, a series m of linear structure is built up with several network participants by connecting the ports T1 / R1 and T2 / R2. Several of these rows are interconnected to form a three-dimensional structure, as shown in FIG. 8. To do this, there must be at least one communication channel between ports T3 / R3 and T4 / R4 between two rows. Multiple communication paths between two rows are permitted. Terminals can be connected to ports T3 / R3 and T4 / R4 of the network nodes that are not used for communication channels between rows. Each row is a unique row address R *. assigned with 0 <k <p, where p is the number of

Rows of the selected network structure. The respective row addresses are indicated on the left side of FIG. 8 next to the respective row. The associated address of the series is stored in each network participant in the "Row address *" parameter register.

However, it must be ensured that between any two network participants in the entire network only one Communication path is used. Loops were tion paths at several communica ^¬ occur, ie messages could multiply and circulate. To avoid loops, port select telegrams are used in combination with a modified voltage tree algorithm.

For this purpose, the data area of the port select telegrams is expanded by the address R _{n of} the adjacent row to which the port T3 / R3 or the port T4 / R4 of the network subscriber who compiled the telegram is connected.

The task of the port select telegrams expanded by the row address R _{n of} the adjacent row is to find a network subscriber which is connected to a row with the address R _r via port T3 / R3 and in relation to the number of

Network subscriber is as far as possible from both ends of its row, ie has the smallest distance from the middle of the row. This defines the only network participant in the row that is potentially actively connected to the adjacent row with the address R _r via port T3 / R3. Port T3 / R3 is switched from potentially active to active if the modified voltage tree algorithm also switches this port active. Data telegrams are only exchanged via active communication channels between the rows. With the method previously described for the two-dimensional network structure, the network participant can be found who has the smallest distance from the center of the row and whose port T3 / R3 is connected to the row with the address R _N. The port select telegrams ensure that between two rows that are directly connected to each other via communication channels, only one communication channel via port T3 / R3 is potentially active at any time. In order to reliably prevent loops that are closed over more than two rows, a modified tension tree method ensures that no loops occur across the entire three-dimensional network. Characteristics of the modified voltage tree method are that each row is regarded as a virtual switch, with the potentially active communication channels via ports T3 / R3 or via ports T4 / R4 to other rows as ports of the virtual switch and that a communication channel is viewed via a port T4 / R4 is potentially active if it is connected to a potentially active port T3 / R3 of another row, which is also regarded as a virtual switch. Furthermore, the following entries are provided in the data field for configuration telegrams:

1. Root_ID: A 64 bit address R _{R of} the virtual switch, which is assumed to be “root”.

2. Transmιtter_ID: A 64 bit address R _{τ of} the virtual switch to which the sending network participant belongs. The addresses R _R and R _τ each correspond to the address of the row that is regarded as a virtual switch.

3. Cost: Smallest number of rows that a telegram has to go through from a sender to the Root_ID.

4. Port_ID: A 16 bit address P _{H. of} the port via which the sending virtual switch sends the configuration telegram. RP _ID is equal to the address R _{n of} the row that is connected to the port via which the virtual switch sends with the Transmιtter_ID.

With these definitions, the Spannmg-Tree algorithm can be applied to a network of virtual switches. It is based on the described configuration telegrams that are sent and received by virtual switches. Only the potentially active or active ports T3 / R3 or T4 / R4 of a row, ie a virtual switch, evaluate received configuration telegrams. The deactivated ports T3 / R3 or T4 / R4 evaluate the configuration telegrams and then filter them. The voltage tree method switches the ports T3 / R3 or T4 / R4 from potentially active to active, which ensure that there is only one communication path between any two network participants in the network and therefore no loops occur. The remaining ports T3 / R3 or T4 / R4 remain potentially active or deactivated. Data telegrams are only exchanged via active communication channels between the rows.

A virtual switch is always ready to receive configuration telegrams at its ports and stores the configuration message for each port with the "best * combination of Root_ID". Cost. Transmitter_ID. Port_ID. Not only are the combinations received compared for each port, but also the combination that the virtual switch would send to this port is compared. A combination Kl is "better *" than another combination K2 if 1. Root_ID of Kl <Root_ID of K2,

2. Root_ID of Kl = Root_ID of K2 and Cost of Kl <Cost of K2,

3. Root_ID of Kl = Root_ID of K2 and Cost of Kl = Cost of K2 and Transmitter_ID of Kl <Transmitter_ID of K2 or

4. Root_ID of Kl = Root_ID of K2 and Cost of Kl = Cost of K2 and

Transmitter_ID of Kl = Transmitter_ID of K2 and Port__ID of Kl <Port_ID of K2.

The root port of a virtual switch is the port with the

"Best * received combination

K _R = K _E = Root_ID.Cost .Transmitter_ID.Port_ID.

The root port is the port of a virtual switch with the shortest distance to the Root_ID.

The combination of the root port is communicated to all network participants of the virtual switch via port select telegrams. This means that every network participant has the necessary information to decide whether a port of is potentially to be switched actively to active. A potentially active port is switched to active if Root_ID. (Cost + 1) .Transmιtter_ID.Port_ID from the root port is "better * than Root__ID. Cost. Transmιtter_ID. Port_ID of the considered port.

The condition that leads to the activation of a port of a virtual switch must be valid for a period of time Δt _ne tdeiay which corresponds to at least twice the worst case throughput time of a configuration telegram through the network before the relevant port is actually activated. Only configuration telegrams received from the root port are forwarded to the active ports of the virtual switch. Configuration telegrams are only sent or forwarded via the active ports T3 / R3. The only ones

The potentially active ports T4 / R4 of the connected virtual switches are therefore receivers of these telegrams. In addition, configuration telegrams are only sent or forwarded via the active ports T4 / R4. The only receivers of such telegrams are the potentially active ports T3 / R3 of the connected virtual switches. A so-called combination aging counter is assigned to each port of a virtual switch. This counter is reset with each received or forwarded configuration telegram and restarted. The combination aging counter is therefore only activated for the potentially active or active ports in a row and is incremented with a parameterizable time cycle. If the combination aging payer reaches the parameterizable threshold "maximum age *" for a potentially active or an active port, the combination Root_ID.Cost .Transmιtter_ID saved for this port. Port_ID deleted and recalculated.

The exchange of information within a virtual

Switches, ie within a series of network participants with a lien-shaped structure, are implemented with port select Telegrams that are similar to the port select telegrams of a network with a two-dimensional structure described above. The data area of the port select telegrams of a network with a three-dimensional structure is expanded by a 16 bit address R _{n of} the port T3 / independent of the receiving port compared to the data area of the port select telegrams for a network with a two-dimensional structure. R3 connected virtual switches, ie the neighboring row, the validity of which is indicated by the valid bit V already described. If V = 0, the received value of the row address R _{n is also} invalid. The port select telegram was fed into the virtual switch by a network participant, which is connected to the virtual switch via an operational port T4 / R4. If V = 1, the address R _{n is} valid. In addition, a potentially active bit P _{pA is} inserted in the data area of the port select telegram. Depending on the receiving port, this bit has two meanings:

_{1.P pA} m port select telegrams received on port T4 / R4 informs the network subscriber whether port T3 / R3 of the sending network subscriber of the connected virtual switch is potentially active or active (P _pA = 1) or deactivated ( P _pA = 0).

2. P _pA m port select telegrams received on port T1 / Rl or port T2 / R2 informs the network subscriber whether port T4 / R4 of the initiator of the port select telegram is potentially active or active (P _pA = 1) or deactivated (P _pA = 0).

Furthermore, the data area of the port select telegram is expanded by a 16 bit address Ri of the virtual switch connected via port T4 / R4. The value of the address is only necessary if P _pA = 1 is set.

In addition, the port select telegrams become a value

Transfer active timers at the time of transmission. The active timer measures the time since the last reception of a port Select telegram via port T4 / R4 with the potential active bit set, ie with P _pA = 1, has passed. The value of the active timer is only valid if P _pA = 1.

Furthermore, the data area of port select messages for three-dimensional network structure contains the best receive ^¬ ne combination for this port K _E = Root_ID.Cost .Transmιtter_ID. Port_ID that was sent or forwarded in the data field of a configuration telegram from a row connected to a potentially active or active port T3 / R3 or T4 / R4 with the address R _n or R.

In addition, the best combination known to date at the potentially active or active ports T3 / R3 and T4 / R4

Transfer row, ie the virtual switch, in the data field of port select telegrams: K _R = Root_ID.Cost .Transmιtter_ID.Port_ID.

A network node that receives a port select telegram from a potentially active or active port T3 / R3, ie a port select telegram with P _pA = 1, from another row on a deactivated port T4 / R4 switches port T4 / R4 to potentially active or active and sends a configurable number of port select telegrams with P _p ; = 1.

The ports T4 / R4 are each assigned an active timer, which is incremented with an adjustable clock. The active timer measures the time since the last receipt of a port select telegram with the potential active bit set P _p "= 1. Each port select telegram received via port T4 / R4 with the potential active bit set P _pA - 1 resets the active timer and restarts it. Received port select telegrams with P _pA = 0 reset the active timer without starting it. A network participant with a potentially active or active port T4 / R4 deactivates this port if 1. a port select telegram with P _pA = 0 is received via port T4 / R4, 2. via port T1 / R1 or T2 / R2 A port select telegram is received with P _pA = 1 and a received value of the active timer, which is less than the own active time, whereby the own active timer is reset in this case without being restarted, or 3. the time measured by the active timer reaches a parameterizable maximum value.

In emgeschwungenen state all network participants via the ports T3 / R3 connected series em send m each virtual switch, that is, each row, with a potentially active T3 / R3 Verbmdung to a Benach ^¬ disclosed cyclically every reporting interval .DELTA.t _M each port select telegram via the ports T1 / R1 and T2 / R2.

In addition, the following network participants send m each virtual switch in a port select telegram via the ports Tl / Rl and T2 / R2:

1. Each network participant during initialization with K _R = K _E = source address .0. Source address. Port_ID, 2. every network participant with a potentially active communication channel via port T3 / R3 to a row with address R _r when it receives an configuration telegram via port T3 / R3,

3. each network participant with a potentially active communication channel via port T4 / R4 to a row with the address R _x when it receives a configuration telegram via port T4 / R4,

4. Each network participant in which the combination aging payer of a potentially active or active port reaches the threshold value “maximum age *, with the combinations K _E and K _R. Stored for this port deleted and recalculated and firstly using e port select telegram

K _R = K _E = source address. 0. Source address .Port_ID is sent, 5. each network participant with a potentially active or active port whose stored combination K _{R is} "better * than the combination K _R in the received port select telegram, whereby the combination K _R stored for this port is deleted and recalculated and wherein a port select telegram with the best combination received so far for this port

K _E = Root_ID.Cost .Transmιtter_ID. Port_ID and with K _R = mm {K _E , received value of K ,.} is sent and 6. Each network participant, whose port T4 / R4 is switched from deactivated to potentially active or active, sends a parameterisable number of port Select telegrams with P _p = 1.

All receivers of a port select telegram with a deactivated communication channel to another row store the values K _E and K _R sent by the associated potentially active or active port of the virtual switch.

FIG. 8 shows the result of the application of the port select telegrams in combination with the modified voltage tree method to each row of the three-dimensional network shown.

Redundancy in the network is intended to ensure that physical errors, electromagnetic interference, network expansions or component replacement do not impair communication between the network components. The prerequisite for this is not only a quick detection of errors or network modifications and a quick network reconfiguration, but also a network area that is as small as possible, which during the reconfiguration time differs from the Effects of the error or the network modification is affected.

Redundancy management enables redundancy within each row of a network, with respect to the communication channels between two rows connected to one another, and redundancy m with respect to the entire network. The modified Spannmg-Tree algorithm guarantees freedom from loops.

This type of redundancy advantageously enables short reconfiguration times with minimal hardware expenditure and can therefore be implemented with little effort. In addition, the network area is limited which is affected by the effects of an error or a network configuration during the reconfiguration time.

In order to ensure redundancy in a series of network subscribers connected in the form of a line, e Rmg is formed, as is the case in FIG. 6, for example, with network subscribers 100 ... 107. A network subscriber, for example network subscriber 100, which is located at one end of the row, must be operated in redundancy mode. It acts as a redundancy manager.

By setting a redundancy bit in the parameter register, this network participant is switched to the redundancy mode. To check the series, it cyclically sends a Testl telegram to port 1 with the MAC address of port 1 as the source address. The cycle time is, for example, 10 ms. Test 2 is sent cyclically on port 2 with the MAC address of port 2 as the source address. The cycle time is also 10 ms, for example. The Test2 telegram is sent offset by half the cycle time, 5 ms after the Testl telegram. If the row is not interrupted, the Testl telegrams sent on port 1 are received again on port 2, and the Test2 telegrams m are also reversed Direction. In this case, the communication channel between the two ports 1 and 2 within the network participant, which is operated as a redundancy manager, is disconnected, so that all data telegrams received at port 2 are filtered out and from the telegrams received at port 1 only those to the own station address directed to be accepted.

The network participant operated as redundancy manager closes the Rmg, i.e. it forwards received telegrams between ports 1 and 2 if a test telegram is not received at one of the two ports within a parameterizable time interval of, for example, 100 ms or if an "lmk-down *" telegram is received by a network participant of the respective row that receives one Interruption of the communication channel to the next network participant. The port of the network participant affected by the interruption is deactivated. A prerequisite for the reactivation of this port is the reestablishment of the connection to the other network participant for a certain minimum time of, for example, 1.6 s or the receipt of one

"L k-up" telegram from the redundancy manager. When the ring is closed, an "lmk-down *" telegram is sent to ports 1 and 2 of the redundancy manager in order to inform all other network participants in the series about the new series structure. Test telegrams continue to be sent cyclically.

The network subscriber operated as a redundancy manager opens the Rmg when a test telegram is received again over the previously interrupted route or when a telegram “lmk-up *” is received by the network subscriber in the series whose communication channel to the neighboring network subscriber has been for a certain minimum time, for example 1 , 6 s is no longer interrupted. When the ring is opened, an "lmk-up *" telegram is sent to ports 1 and 2 of the redundancy manager in order to send all other network participants in the series via the new ring structure inform. Test telegrams continue to be sent cyclically. Each network participant in the series resets the registers necessary for the forwarding of telegrams by receiving an "lmk-up" or "lmk-down *" telegram.

A redundant implementation of the communication channels between two rows requires at least two separate communication paths. However, a maximum of one path may be used for data exchange between the rows. This potentially active communication channel between two rows is selected with the help of port select telegrams. If a potentially active communication channel is identified as faulty, it is deactivated and another communication path is switched to potentially active. The following applies to the switchover time from deactivated to potentially active:

Switching time> Δt _tιmeθ ut + Δt _rθ wdeiay. where Δtr _lme out is the timeout interval and Δt _r owdeiay corresponds to twice the worst case throughput time of a port select telegram through the series. The switchover time therefore depends on the number of network participants that form a row. It is, for example, for a series of 50 network participants in the order of 200 ms, if a timeout interval of 150 ms is assumed.

Redundancy in a three-dimensional network is also possible. If the loop structure already provides freedom from loops, ie there is no network redundancy, every potentially active communication path between two rows is also active. In this case, it is not necessary to use the modified tension tree algorithm described. With network redundancy, the modified Spannmg-Tree algorithm ensures loop-free between the rows. A reconfiguration of a network with the modified tension-tree algorithm is only necessary for errors or network modifications that are not processed by the redundancy within a row or the redundancy of the communication channels between two rows. In the case of a network with such network participants, the transmission time from the sender to the receiver depends on the number of network participants over which the telegram is forwarded and cannot be neglected. The transmission time of a telegram increases for everyone

Network subscriber who forwards the telegram to a subscriber- _specific delay Ti e Δt _1Λ which is composed of the following times:

1. Throughput time through the physical layer structure, for example module 36 FIG. 2, in the reception direction from the reception of the first bit of a nibble until the nibble on the MII is output as valid with m “RX__DV = 1 *. This time is, for example, 21 T _B ιt for DP83843 PHYTER from NSC, where T _B .t corresponds to 100 ns at 10 MBaud transmission rate and 10 ns at 100 MBaud transmission rate.

2. Residence tent in the network participant from receiving a nibble to sending the same nibble. If the subscriber is currently sending its own telegram and a telegram has already been entered in the receive buffer, data forwarding can be delayed by up to (3k x 8) x T _B1 compared to undisturbed operation.

3. Throughput time through the physical layer structure, via which the telegram is forwarded in the transmission direction, from the first rising edge of a transmit

Clock signal after providing a nibble on the MII up to the first transmitted bit of this nibble. This pass-through tent is for example for DP83843 PHYTER

4. Runtime over the lines between two neighboring network participants. The sum of the times specified under 1, 3 and 4 is a fixed quantity and is referred to as the throughput time Δt _DZ . It can either be parameterized or measured by the network participants. A change in this throughput time Δt _D L: is only possible if a network participant is out of the network removed or added to the network or when the cabling is changed.

The throughput time Δt _D z can be determined with the following sequence of telegrams, which network participants execute, for example after initialization or on request:

1. Each network participant who is newly added to the network sends its four neighboring network participants a so-called DLZ telegram, i.e. em first telegram for running time determination. This telegram is clearly identified in the 16 bit type address.

2. The new network participant starts a DLZ timer 1 after the last nibble of the type field of the DLZ telegram has been made available to the media independent interface (MII) of port 1 for sending. Accordingly, he starts a DLZ timer 2, 3 and 4 for sending via ports 2, 3 and 4. 3. Each of the maximum of 4 neighboring network participants starts after receiving the last nibble of the type field of the DLZ telegram on his MII its DLZ timer of the respective port. The received DLZ telegram is not forwarded, but is sent back to the sender, supplemented by the residence time, the respective Ethernet controller of the network participant. If this neighboring network participant has passed the last nibble of the type field of the DLZ telegram modified in this way to its MII directed to the newly connected network participant, it stops the DLZ timer and sends the dwell time stored in the DLZ timer with the data field of the telegram to the newly connected network participant. 4. The network subscriber newly added to the network stops the assigned DLZ timer 1, 2, 3 or 4 when he receives the last nibble of the type field at his MII of the respective port. 5. For example, for port 1 of the newly connected network participant, the throughput time Δt _D zι can be calculated according to the formula: Δt _D Lzι = (T _DR -T _D ): 2, with T _DR - response time measured with the DLZ timer 1 and T _DA - measured stay tent in the neighboring network participants.

Correspondingly, the throughput tents .DELTA.t _D LZ2 .DELTA.t _D LZ3 and At _D z. calculated for the remaining ports of the new network participant. The flow tents determined in this way are stored as parameters in module 52 (FIG. 2).

6. With a further telegram, the newly connected network participant sends the measured pass tents to the neighboring network participants via the respectively assigned port.

The described determination of the pass-through tents is only necessary for every second network participant when initializing a network.

Time synchronization has the task of synchronizing the clocks of several or all network participants. The communication channels between network participants are advantageously operated in full duplex mode so that the transmission of telegrams exhibits deterministic behavior.

The transmission time from a sender to a receiver in a network depends on the number of network participants over which the telegram is routed and cannot be neglected.

Time synchronization can be carried out with two special telegrams, for example. Figure 9 shows the general structure of a telegram. The first field 140 contains a destination address, ie an address of the subscribers to whom the telegram is directed, for example wise 48 bits long. E second field 141 contains a source address, the address of the respective transmitting Netzwerkteilneh ^¬ mers whose length ebenj alls for example, 48 bit amounts. An identifier of the telegram is transmitted in a type field 142 with, for example, 16 bits. The user data of the telegram are sent in a data field 143 of variable length. Completed ^¬ det is the telegram by a check sequence 144 that is used essentially for checking the transmission quality. Telegrams for time synchronization can be identified by a special multicast address as destination address 140 and / or by a new type address to be defined in type field 142.

FIG. 10 shows an order list which can be stored, for example, in the RAM 22 in FIG. 2. Telegrams that are to be transmitted via the network are entered in such an order list. If the transmission is not prioritized, the telegram below is transmitted next. It can therefore happen that, for example, a completed telegram 151 is only transmitted when two telegrams 152 and 153 previously entered in the job list have been transmitted. Depending on the number of pending orders, the delay in sending a telegram in a network participant can therefore be varied after it has been entered in the order list. The following describes one way in which the influence of the transmission time delay in time synchronization can be avoided:

1. With a time master, i.e. A network participant, on whose clock the clocks of the other network participants are synchronized, carries an SM-TimeO telegram, an first telegram for time synchronization, m an order list e, starts for each port P, P = 1, 2, 3, 4, a delay timer and remembers the start time of these timers.

2. The delay timer of each port P, P = 1, 2, 3, 4, is stopped at the time master after the last nibble of the type field of the SM-TimeO telegram was made available to the MII of the respective port for sending. The transmission delay time is thus determined. Em nibble is defined as e half a byte, that is, it is a sequence of 4 bits.

3. After receiving the last nibble of the type field of the SM-TimeO telegram at the MII of its port P, P = 1, 2, 3 or 4, each neighboring network participant starts its delay timer with the value of the respective throughput time Δt _D Lz _P. previously measured by the method described above or entered by an operator. The address of the time master that was received as the source address in the SM-TimeO telegram is also saved. 4. At the point in time at which the neighboring network subscriber creates the last nibble of the type field of the SM-TimeO telegram for forwarding to the MII of another port, the value of the delay timer assigned to port P, P = 1, 2, 3 or 4, is stored. However, the delay timers continue to run. The stored delay times of the ports each correspond to the transmission time Δt _{2 of} this network subscriber defined above. By the start value .DELTA.t _D z _P of the delay timer, the duration of the tele ^¬ program on the physical transmission is required to draw added.

5. The time master then carries an SM timel

Telegram, which contains the start time of the delay timers in the data field, m the order list e. Before sending this SM-Ti telegram via a port P, P = 1, 2, 3, 4, it replaces the start time of the delay timers with the time at which the last nibble of the type field of the SM-Ti eO telegram was made available to the MII of this port for sending, ie by the sum of the start time of the delay timers and the measured delay time of the respective port P. The time corrected by the transmission time delay is thus entered in the SM timel telegram. 6. Each neighboring network participant that receives an SM-Timel telegram adds the stored delay time Δt _{x of} the SM-TimeO telegram previously forwarded via port P, P = 1, 2, 3 or 4 to the one in the SM- Timel telegram received time and sends the time received in this way with an updated SM-Timel telegram via another port to the next neighbor. SM-Timel telegrams are only accepted by the network participant that previously sent an SM-TimeO telegram.

When the SM-Timel telegram is received, the time slave knows, i.e. the neighboring network participant, the start time of his delay timer. The synchronized time is the sum of the time received in the SM-Timel telegram and the delay time of the time slave for the respective receiving port. The neighboring network subscriber thus corrects the time received in the second telegram by the runtime and the reception time delay.

As an alternative to the option described above, the following procedure enables time synchronization with just one telegram:

1. The time master starts the delay timers assigned to the ports and carries the time telegram with the start time of these timers m the job list em.

2. The delay timer of each port P, P = 1, 2, 3, 4, is stopped at the time master after the last nibble of the type field of the time telegram is available for transmission to the media-independent interface of port P was made, ie after the last nibble of the type field was made available for physical transmission. The network participant then adds the start time of the delay timers specified in the time telegram to the value of the delay timer of the respective port P and sends this sum as a time corrected by the transmission time delay with a first telegram for time synchronization via the respective port P. 3. Each neighboring network participant starts after receiving the last nibble of the type field of the first telegram for time synchronization on a port P,

P = 1, 2, 3 or 4, its assigned delay timer with the value of the respective lead time Δt _DLZP . It therefore measures the time delay since receipt of the first telegram.

4. At the point in time at which the neighboring network subscriber creates the last nibble of the type field of the first telegram for time synchronization for forwarding to the MII of a port, the value of the delay timer which is assigned to this port is stored. However, the delay timers continue to run. The stored delay times assigned to the individual ports each correspond to the transmission time Δt _{x of} this network subscriber. It is added to the received start time of the delay timer and forwarded with a second telegram for time synchronization via another port to the next, ie a third, network participant.

With the receipt of a first or second telegram for

The time slave knows the start time of its delay timer in time synchronization. The synchronized time is the sum of the time received in a first or second telegram and the delay time of the time slave for the receiving port P.

The described time synchronization options can be used in a corresponding manner for the synchronization of aquidistance timers m for the network participants. The task of aquidistance timers is to enable several or all network participants to perform specified actions aquidistantly. In control systems, this function is often referred to as an “electronic wave *. For all network participants that are connected to one another via the network, a clock beat should be generated, with the clock of which respective target values are transferred and actual values are queried. An application example is the measurement of the electrical fresh performance, we m the required current and voltage measured values are recorded by separate transmitters and queried over the network.

It is assumed that an aquidistant cycle is controlled by only one aquidistance master. The network participant that takes over the function of an aquidistance master has a timer that is loaded at the start with the parameterizable value of the aquidistance interval. The timer is free running and is decremented with every bit clock. When the timer has expired, it is loaded again with the parameterized value of the aquidistance interval and a new cycle begins. The difference between an aquidistance timer and a clock is therefore the direction of travel. To correct a timer status transmitted with telegrams, the time delays do not have to be added as with the time, but have to be subtracted. The term “time synchronization *” used above should therefore be understood to include the synchronization of aquidistance timers.

The following options are available for synchronizing active actions:

1. The Aquidistanz master carries the Aquidistanz telegram m the order list em. It stores the value of the

Aquidistanz-Timers from when the last nibble of the type field of the Aquidistanz telegram is transferred to the MII of the four ports P, P = 1, 2, 3, 4, ie provided for the physical transmission. This value Δt _Aqjl stored for each port P, which corresponds to the time remaining until the aquidistance interval has elapsed, is sent with the aquidistance telegram via port P to the neighboring network subscriber.

2. Each network participant starts after receiving the last nibble of the type field of the aquidistance telegram at the MII of a port, ie when receiving the aquidistance Telegram from the physical transmission link, an auxiliary timer with the value of the throughput time Δt _D _.zρ-

3. The value of the auxiliary timer is saved at the point in time at which the neighboring network subscriber creates the last nibble of the type field from the equidistance telegram for forwarding to the MII of another port. The stored value of the auxiliary timer corresponds to the transmission time Δti of this network participant for port P. This stored time Δti becomes from the received remaining time Δt _AqU i to the next

Subtracted cycle start. The neighboring network participant forwards the corrected remaining time (Δt _Aqui - Δti) with the aquidistance telegram via the other port to the next neighboring network participant. In addition, he loads the corrected remaining time into his equidistance timer, which is decremented with every cycle.

4. If the equidistance timer of an equidistance slave has expired, it is first loaded with the parameterized value of the aquidistance interval and decremented with each bit cycle. As soon as a new aquidistance telegram is received from the equidistance master, the equidistance slave loads the remaining time determined in the manner described (Δt _AqU i ^_ Δt _; ) into the equidistance timer until the next cycle _begins .

For the described synchronization of aquidistance timers, the maximum transmission time between a transmitter and a receiver in the network should be less than the length of the aquidistance interval.

In the exemplary embodiment, a network according to the Ethernet specification was described. However, the invention is also readily applicable to Fast Ethernet, Gigabit Ethernet or other types of networks.

Claims

claims

1. Network subscriber, in particular field device, with several ports for connecting further network components, since - characterized in that an interface (25), a so-called microprocessor interface, is provided for connecting the ports to an internal processor bus (21) and that a control unit ( 46), a so-called switch control, is provided for routing telegrams between the ports and the microprocessor interface.

2. Network subscriber according to claim 1, so that four ports (28 ... 31) are provided for connecting further network components.

3. Network participant according to claim 1 or 2, d a d u r c h g e k e n n e e c h n e t that the ports (28 ... 31) meet the Ethernet, Fast Ethernet or Gigabit Ethernet specification.

4. Network subscriber according to claim 3, which also means that the control unit (46) is designed such that a transmission priority of the telegrams to be sent is evaluated and that telegrams with high priority are sent before telegrams with low priority.

5. Network subscriber according to one of the preceding claims, that the microprocessor (23) is present for correcting an internal clock (37) on the basis of time information received in telegrams.

6. Network with several network participants according to one of the preceding claims, characterized in that the network participants (1 ... 4) m one lm-shaped network topology are interconnected.

7. Network according to / claim 6, d a d u r c h g e k e n n - z e i c h n e t that the two line ends at one

Redundancy manager is connected, which is designed to separate the two line ends in the event of a fault, i.e. if there is no interruption within the row, and to connect the two line ends to one another in the event of a fault.

8. Network according to claim 6 or 7, characterized in that the network contains at least two rows formed from a plurality of network participants (100 ... 107; 110 ... 117), which are connected to one another via at least two communication channels, of which only the first one active and the second is deactivated, the first communication channel being formed by connecting a port of a first network participant (103), which is arranged in the first row, to a port of a second network participant (110), which is arranged in the second row and wherein the second communication channel is formed by a connection of a port of a third network subscriber (104) arranged in the first row to a port of a fourth network subscriber (117) arranged in the second row, and wherein the third network subscriber (104) monitors the first communication channel for uninterrupted freedom and if the first communication channel the second communication channel m the active state.

9. Network according to claim 8, characterized in that the third network participant (104) for monitoring the first communication channel evaluates message telegrams, so-called port select telegrams, which the first network participant (103) cyclically to the network participants with an uninterrupted first communication channel (100 ... 107) sends which are arranged in the first row.

10. Network according to claim 8 or 9, d a d u r c h g e - k e n n z e i c h n e t that of two ready

Kom umkationskanalen that communication channel is actively switched, which is formed by a connection of a port of a network participant, which is arranged m of the first row at the smallest distance from the center of the row, with a port of a network participant, which is arranged m of the second row.

11. Network according to one of claims 8 to 10, characterized in that the first row is connected to at least one third communication channel with a third row, that the second row is connected directly or indirectly via a communication path to the third row, which is not leads over network participants of the first row, so that a loop was created when all communication channels of the network were active, and that the network participants of the network are designed in such a way that, by exchanging configuration telegrams, they ensure that the network is loop-free and that everyone Network subscriber is connected to every other network subscriber via exactly one active communication path.

12. Network according to claim 11, characterized in that the network participants are designed such that the tension-tree method is carried out to produce the loop-free network, each row as a virtual network participant, a so-called virtual switch, and each operational port one Network subscriber of a row, by the connection of which to an operational port of a network subscriber of another row, an operational communication channel is formed as an operational port of the virtual switch.

13. Network according to one of claims 6 to 12, characterized in that there is a first network participant, which is designed to send e first telegram for time synchronization to a second network participant, which contains a time corrected by a transmission time delay of the first network participant that a second network participant is present, with which the transit time of the telegram over the physical transmission path between the first network participant and the second network participant is stored and which is designed to measure the time delay since receipt of the first telegram and the time received in the first telegram to correct the turnaround time and the reception time delay.

14. Network according to claim 13, characterized in that the second network participant is further configured to send e second telegram for time synchronization to a third network participant, the one by the throughput time and the delay time between receipt of the first telegram and transmission of the second telegram corrected received time contains.

15. Network according to one of claims 6 to 12, characterized in that a first network participant is present, which is designed to send em first telegram for time synchronization to a second network participant and to store a time corrected by a transmission delay of the first telegram of the first network participant that there is a second network participant, with which the throughput time of the telegram over the physical transmission path between the first network participant and the second network participant is stored and which is designed to measure the time delay since receipt of the first telegram, the first network participant furthermore is formed, a second telegram, which is about the send time-delay corrected time of the first network participant contains to send to the second network participant and wherein the second network participant is further configured to correct the time received in the second telegram by the throughput time and the reception time delay.

16. The network of claim 15, ^¬ dadurchgekenn characterized in that the second network station is further configured to, the first message to a third

To forward the network subscriber, to measure its delay time for the forwarding of telegrams and to send a third telegram to the third network subscriber which contains a received time corrected for the throughput time and the delay time for forwarding the telegram in the first telegram.

17. Network according to one of / claims 13 to 16, characterized in that the first network participant has a first timer (57) for determining the transmission time delay, which he starts a list of send requests when entering a telegram and after providing the telegram for physical transmission as Reads the value of the transmission time delay at which the time at which the telegram m the list of transmission orders was entered is to be corrected.

18. Network according to claim 17, so that the second network subscriber has a second timer for determining the reception time delay, which he starts upon receipt of a first telegram from a physical transmission link.

19. Network according to claim 18, α a characterized in that the throughput time of the telegram over the physical transmission link between the first network participant and the second Network participants as the starting value m the second timer is stored before its start.

20. Network according to one of claims 13 to 19, since - you rchgek characterized that the beginning and end of the throughput time of a telegram are each determined as the time at which em characteristic field of a telegram with a fixed distance from the beginning of the telegram e Media Independent Interface of the first participant left or arrives at the media independent interface of the second network participant.

21. Network according to claim 20, so that the network complies with the Ethernet, Fast Ethernet or Gigabit Ethernet specification and that the characteristic field of the telegram is the type field.

22. Network according to one of claims 13 to 21, so that the first network subscriber is furthermore designed according to the

Recording m to send the network a first telegram for the determination of the runtime to an adjacent subscriber and to start a response time timer after the telegram has been provided for the physical transmission, and that the second network subscriber is further trained to receive a telegram for the determination of runtime from the physical transmission to start a timer for determining the residence tent and, after a second telegram for physical transmission has been provided, to stop the timer and to transmit the measured residence tent t _m m to the first network participant for determining the running time, and wherein the first network participant is further designed to after receiving the second telegram to determine the runtime from the physical transmission, to stop the response time timer, to evaluate the measured response time t _DR and the measured stay time toA of the second network subscriber and one Lead time t _DL z to determine the physical transmission

(* DR ^{~ l} DA) ^l DLZ