DE102020117759A1 - Method and device for fault-tolerant operation of a network arrangement, network switching unit and vehicle - Google Patents

Abstract

The invention relates to a method for fault-tolerant operation of a network arrangement, the network arrangement consisting of a number of network subscriber stations (10, 20, 30, 40) which are connected to one another by communication lines (ET1, ET2, ET3, ET4) in a ring. The method has the steps of determining a preferred circulation direction for the transmission of data packets from an originating network subscriber station (10) to a destination network subscriber station (30), sending a data packet from the originating network subscriber station (10) to the destination network subscriber station (30) along the preferred circulation direction, and checking whether the transmitted data packet arrives at the destination network subscriber station (30). When a connection failure between two network subscriber stations (10, 20, 30, 40) is detected, the steps of sending the data packet back to the source network subscriber station (10), reversing the direction of circulation in the ring network and sending the data packet from the source network subscriber station (10 ) to the destination network subscriber station (30) in the reverse direction of circulation.

Description

The invention relates to the technical field of data communication, in particular for a vehicle. There are high requirements for transmission security, while at the same time there is a high demand for increased data rates. The invention relates to a method and a device for fault-tolerant operation of a network arrangement, a network switching unit and a vehicle.

A large number of control units are installed in modern vehicles. A number of control units are used for the drive train alone, but there are also other control units that are installed in the area of the vehicle body and control units that belong to the infotainment area. Recently, more and more environmental sensors have been installed. In particular, autonomous/automatic driving makes environmental sensors such as lidar sensors, corresponding to light detection and ranging, radar sensors, corresponding to radio detection and ranging, ultrasonic sensors and cameras necessary. These are devices for which the need for high data rates for in-vehicle communication is constantly growing. These environmental sensors relate to high-resolution imaging sensors and currently require data rates in the gigabit/s range. The communication bus systems typically used in vehicles such as CAN, corresponding to the Controller Area Network, or Flexray do not offer the necessary data rates. Therefore, the use of (automotive) Ethernet technology is recommended as a new communication system, as it is inexpensive and offers high data rates.

One problem, however, is the fail-safety of classic Ethernet topologies (bus topology, star topology or tree topology). The failure of a central element (be it a network node or a central bus connection) results in a total failure of the entire communication system. That would be fatal for vehicle communication networks.

For this reason, the use of ring topologies is being considered as a way of increasing the fail-safety of vehicle communication systems, since there are always two possible communication paths available there. Alternatively, there are replication approaches such as the bus standard IEEE P802.1 CB, High availability seamless redundancy (HSR) and Parallel Redundancy Protocol (PRP) according to the IEC 62439-3 standard, which duplicate or replicate the traffic, ie the same Send communication flow in several ways and remove the duplicates at the recipient.

In addition, in the area of software-defined networking (SDN) there are failover mechanisms such as OpenFlow failover groups, which monitor the status of a connection and switch to backup routes in the event of an error. This can also be combined with the Bidirectional Forwarding Detection (BFD) protocol to speed up error detection. The BFD protocol is described in detail in the IETF standard document RFC 5880. Additionally, on the document "Fast Recovery in Software-Defined Networks" by Niels LM van Adrichem, Benjamin J. van Asten and Fernando A. Kuipers of Delft University of Technology, published in 2014 Third European Workshop on Software-Defined Networks, IEEE.

From the document DE 10 2014 225 802 A1 an Ethernet-based communication system for vehicles is known. Services in the form of Ethernet multicast messages are sent to an Ethernet switch (SW) by a sending participant, which forwards the messages of the service selectively only to those other participants who were previously on the Ethernet switch (SW) for the corresponding registered for service.

From the document U.S. 2012/0 109 446 A1 it is known to enable electronic components to communicate with one another using different bus protocols, to use gateways. A gateway performs some degree of protocol conversion to facilitate communication between controllers that communicate using different communication protocols.

From the document U.S. 2013/0 322 434 A1 a network device is known which can work as robustly as possible in error situations. The proposed network setup enables crossed data transmission. The formation of a ring-shaped network arrangement with a plurality of network devices is made possible by the crossover linkage, which can take place, for example, within the network device via an interconnection or wiring. This results in communication paths equipped with consistent data that are practically independent of one another in a single ring-shaped physical network. Overall, this results in a particularly robust network arrangement in which errors are also Liable control devices are manageable in a network and yet consistent data can be present in the network.

The disadvantage of ring topologies, however, is that they require special protocols such as STP (IEEE 802.1 D) and RSTP (IEEE 802.1w), which block individual ports to prevent loops from forming. These are then no longer available for communication. In the case of replication approaches, less than half of the possible bandwidth is available, since part of it is occupied by redundancy. With the SDN-based solutions, a central controller is required, which must be connected to all switches. This means an additional number of connections. Alternatively, there are also solutions without a central controller, but in which a so-called "crankback" path is created, i.e. lines are unnecessarily loaded twice (see source above). For these reasons, ring topologies have so far been avoided in the automotive sector.

From the patent DE 101 61 186 B4 a ring topology is known for an Ethernet network that uses fast response to ring breaks to reroute traffic. In order to avoid loops, all data is normally sent via a defined port. The second port is deactivated. In addition, the Ethernet standard was modified to detect connection errors more quickly. In the event of an interruption, the forwarding port is changed, bypassing the interruption. The one from the patent DE 101 61 186 B4 known solution can only be used with a modified Ethernet standard. In addition, the detection of the interruption and the switching of the routing cannot be separated. This is a disadvantage, for example, if you want to modify the routing but still need fast error detection. In addition, routing becomes inefficient since only one port is used for forwarding. For example, if you want to communicate with the direct neighbor on the blocked port, in this case all traffic is routed via the other two nodes. The blocked port also prevents dedicated replication of critical traffic types.

The aim of the invention is to find a practicable implementation for the introduction of fault-tolerant ring topology based on Ethernet technology for use in vehicles. This is the object of the invention.

This object is achieved by a method and a device for fault-tolerant operation of a network arrangement according to claim 1 and claim 9, and a network switching unit according to claim 11 and a vehicle according to claim 14.

The dependent claims contain advantageous developments and improvements of the invention according to the following description of these measures.

In a general embodiment of the invention, the invention relates to a method for fault-tolerant operation of a network arrangement, the network arrangement consisting of a number of network subscriber stations which are connected to one another by communication lines in the form of a ring, so that the data packets in the ring network can travel from an originating network subscriber station to a destination -Network subscriber station can circulate. The method is characterized by the steps: determining a preferred circulation direction for the transmission of data packets from the source network subscriber station to the destination network subscriber station, sending a data packet from the source network subscriber station to the destination network subscriber station along the preferred circulation direction, checking whether the sent Data packet arrives at the target network subscriber station. The following steps are also carried out: If a connection abort between two network subscriber stations is detected, the data packet is sent back to the source network subscriber station, the direction of circulation is reversed and the data packet is sent from the source network subscriber station to the destination network subscriber station in the reverse circulation direction. Because there are two directions of circulation in a ring bus topology, if a bus connection fails on a segment of the ring bus topology, the direction of circulation can simply be reversed and the destination network subscriber station can still be reached via a detour.

In order to implement this, it is advantageous if the network subscriber stations have a programmable network switching unit or are connected to a programmable network switching unit. Such network switching units are known as network switches and are used in many ways in the LAN area. These are also available in the SDN area as software-based network switching units. In order to achieve the desired scalability and configurability, it is advantageous if such an SDN-based programmable network switching unit is used.

It is then particularly advantageous if the preferred circulation directions for the communication to the individual network subscriber stations are entered in assignment tables in the respective network switching units. So-called source address tables are typically created in commercially available network switching units, in which the MAC addresses of the originating network subscriber station that sent the packet are listed with the port number. According to the invention, another table is used, in which it is recorded for the respective network switching unit via which port the data packet is to be forwarded for each destination address. This can be done separately for each input port of the network switching unit. Particularly in the case of network configurations that are not planned to change during operation, it is possible to design the assignment tables based on the target address specification.

With the table in this form, it is then possible to also enter the reversal of the circulation direction in a ring bus segment in the allocation table if the connection is broken. This can advantageously be entered in the assignment table from the outset, in the manner of a backup rule. The network switching unit only needs to be informed that the connection has been broken and that the direction of transmission must therefore be reversed. This can preferably be done with a signaling message that a control unit sends to the network switching unit.

The detection of a connection failure in a ring bus segment is already integrated in the Ethernet protocol at the bit transmission layer level, corresponding to layer 1 of the OSI/ISO layer model of data communication. According to the Ethernet 10BaseT standard, the bit transmission layer transmits so-called "heartbeat" messages with a period of 16 ±8 ms, which must be answered. If no response is received, after a set time in the range of 50-150 ms it is assumed that the network connection has failed, e.g. because a network connector has been unplugged.

In order to achieve a faster reaction to a connection termination, it is advantageous to use the BFD protocol in one embodiment of the invention. This is done in such a way that signaling packets are transmitted cyclically between the two network switching units between which the ring bus segment is located, and if the signaling packets are absent, the connection abort in this ring bus segment is recognized by the two network switching units.

The network switching unit, which recognizes the connection failure itself, can then send the incoming packets back in the opposite direction of circulation. One problem, however, is that the other network subscriber stations do not know anything about the connection termination and therefore continue to forward the data packets in the preferred circulation direction (“crankback” path). In order to solve this problem, an advantageous development of the invention is that the network switching unit that detects the connection abort sends the respective other neighboring network switching units an error message in which the connection abort in the relevant ring bus segment is reported. Typically, the two network switches will detect the loss of connectivity between them. Therefore, the configuration can also be such that both network switching units inform their neighboring network switching units about this. If there are only four network subscriber stations in the ring bus network, it is sufficient if each network switching unit that detects the loss of connection informs its other neighboring network switching unit.

After the message about the disconnection has arrived, it is advantageous if the respective network switching unit is configured in such a way that its allocation table states that the data traffic is routed in the opposite direction to the preferred circulation direction.

In order to meet the required data rate at acceptable costs, it is advantageous if a variant of the Ethernet standard family is used as the network technology for connecting the network subscriber stations.

The 100BASE-T1 or 1000BASE-T1 variant of the Ethernet standard family can be used here in an advantageous manner. This variant was specially developed for use in vehicles. The automotive industry has specified the bit transmission layer as the communication standard for Automotive Ethernet. This standard is published by the IEEE in two variants as 100BASE-T1 and 1000BASE-T1 as part of the IEEE P802.3bp specification.

The disconnection can have various causes. When using the invention in a vehicle, an accident would be a typical event that can trigger the disconnection. However, it would also be conceivable that the existing vibrations could cause a network connector to come loose and the connection to be lost as a result. With such causes only the repair in a workshop will fix the connection loss. In other areas of application, however, it would be quite possible for the user to fix the connection failure himself. In any case, an advantageous development of the invention is that the network switching unit, which recognizes that the fault in the relevant ring bus segment has been rectified, informs its neighboring network switching unit that the fault in the relevant ring bus segment has been rectified. The network switching unit that has been informed is then configured after the arrival of the notification that the fault has been rectified in such a way that it is again entered in its assignment table that the data traffic is to be routed again along the preferred path.

For a device for performing the method with a number of network subscriber stations, which are connected to one another by communication lines in a ring, it is advantageous for the network subscriber stations to have a programmable network switching unit or to be connected to a programmable network switching unit, the programmable network switching unit has an assignment table in which a preferred circulation direction for the transmission of data packets from an originating network subscriber station to a destination network subscriber station is entered. It is also advantageous that the network switching unit also has a test unit for testing the connection to an adjacent network subscriber station, with the programmable network switching unit being configured in such a way that a data packet received in the preferred circulation direction is detected when a connection is broken between the two adjacent network subscriber stations sends back to the originating network subscriber station in the opposite direction of circulation. This can then forward the reflected data packets. In this way, data traffic can be maintained despite the connection being lost.

It is advantageous for the test unit if the test unit is designed to send an error message to the other neighboring network subscriber station to report the connection abort in the relevant ring bus segment. The neighboring network subscriber station can then adjust to the fact that the connection in the relevant ring bus segment is interrupted.

A further embodiment of the invention consists in a network switching unit for a device for carrying out the method according to the invention. This has a test unit for testing the connection to a first neighboring network subscriber station in a relevant ring bus segment, the test unit being designed to send an error message to another neighboring network subscriber station, in which the connection abort in the relevant ring bus segment is related to the first neighboring network subscriber station is communicated. In this case, it is advantageous if the network switching unit is also configured in such a way that it sends back a data packet received in a preferred circulation direction to an originating network subscriber station in the opposite circulation direction. As described above, there are two communication routes in a ring bus system to reach a network subscriber station. This is exploited here to maintain the flow of communication.

For this purpose, it is advantageous if the network switching unit has an allocation table in which the preferred circulation direction for the transmission of data packets from the source network subscriber station to a destination network subscriber station is entered, with the network switching unit also being configured in such a way that, after recognizing of the connection abort in the relevant ring bus segment or after receipt of the error message in the allocation table an entry that indicates that the data packets are to be forwarded in the opposite direction to the preferred circulation direction.

It is just as advantageous if the network switching unit has an allocation table in which the preferred circulation direction for the transmission of data packets from the source network subscriber station to a destination network subscriber station is entered, with the network switching unit also being configured in such a way that, after recognizing of the connection abort in the relevant ring bus segment or after receipt of the error message in the allocation table an entry that indicates that the data packets are to be forwarded in the opposite direction to the preferred circulation direction. This can be easily implemented with a programmable network switching unit.

In a further embodiment, the invention also relates to a vehicle in which a device for carrying out the method according to the invention is integrated. The invention is particularly advantageous for use in vehicles. Modern vehicles are increasingly equipped with environmental sensors permitted, which are required for the automatic driving function. These generate data streams that significantly increase the data transport volume in the vehicle.

So it is interesting to use the invention in the vehicle if the vehicle is equipped with the following network subscriber stations, a camera control unit and/or ultrasonic sensor control unit and/or RADAR sensor control unit, corresponding to radio detection and ranging and/or LIDAR -Sensor control unit, corresponding to light detection and ranging.

An embodiment of the invention is shown in the drawings and is explained in more detail below with reference to the figures.

• 1 ein Fahrzeug, welches mit Umgebungssensoren ausgestattet ist;
• 2 ein Blockdiagramm für eine Ringbus-Netzwerkanordnung gemäß der Erfindung;
• 3 den Datenfluss vor und nach Auftreten eines Verbindungsabbruches auf einem Ringbus-Segment; und
• 4 das Format einer Signalisierungsnachricht, mit der einer benachbarten Netzwerkteilnehmerstation mitgeteilt wird, dass ein Verbindungsabbruch erkannt wurde.

Show it:

• 1 a vehicle equipped with environmental sensors;
• 2 a block diagram for a ring bus network arrangement according to the invention;
• 3 the data flow before and after a connection abort occurs on a ring bus segment; and
• 4 the format of a signaling message with which a neighboring network subscriber station is informed that a connection failure has been detected.

This description illustrates the principles of the inventive disclosure. It is thus understood that those skilled in the art will be able to devise various arrangements which, while not explicitly described herein, embody principles of the inventive disclosure and are intended to be protected within their scope.

The method for fault-tolerant operation of a network arrangement is described below using the example of an arrangement of environmental sensors used in a vehicle. The vehicle is in 1 shown and is provided with the reference number 100.

A passenger car is shown. However, any other vehicle could also be considered as the vehicle. Examples of other vehicles are: buses, commercial vehicles, in particular trucks, agricultural machinery, construction machinery, motorcycles, rail vehicles, etc. The use of the invention would generally be applicable to land vehicles, rail vehicles, water vehicles and aircraft, including airplanes. In addition, there are also possible uses outside of the vehicle sector, such as in home automation or industrial automation.

A camera that is intended to fulfill its function as a reversing camera is designated by the reference symbol K1. A camera that is used as a front camera for monitoring the surroundings is designated by the reference symbol K2. Then another camera K3 is installed in the left exterior mirror, which observes the left side of the vehicle environment. Accordingly, another camera K4 is installed in the right exterior mirror (not shown). All cameras have the same structure and are equipped with the same hardware and software. In order to enable the Surround View function, the four cameras are typically connected to a central camera control unit, which receives the video data streams from all four cameras K1-K4 and forwards them to a computer, where the data are processed in order to analyze the images accordingly perform the autonomous driving function. In addition, other environmental sensors are installed in the vehicle. Two radar sensors are denoted by R1 and R2. The radar sensor R1 records the area around the front of the vehicle and the radar sensor R2 records the area around the rear of the vehicle. Reference character L1 designates a lidar sensor. This is used to measure the front area of the vehicle environment.

the 2 shows the components of the environment monitoring, which is used, for example, for the automatic driving function of the vehicle 100. The same reference numbers designate the same components as in FIG 1 . Four network subscriber stations 10 to 40 are shown, which are connected to one another with communication lines ET1 to ET4 such that a ring-shaped communication system is formed. Reference number 10 designates a lidar device with lidar sensor L1, network switching unit S1 and microcontroller μC1. Reference number 20 designates a computing device with network switching unit S2 and microcontroller μC2. A control unit CU1 to CU4 is located in each of the microcontrollers μC1 to μC4. The arithmetic unit 20 can be a powerful computer that performs a wide range of arithmetic tasks for the vehicle, for example evaluating the images from the various environmental sensors. Reference number 30 designates a camera device with camera sensors K1 to K4, network switching unit S3 and microcontroller μC3. Reference number 40 designates a radar device with radar sensors R1 and R2, network switching unit S4 and microcontroller μC4. The network subscriber stations 10 to 40 are connected to one another by Ethernet network lines ET1 to ET4. For this, the network subscriber stations must be equipped with the necessary Ethernet ports. For use in vehicles, the variant IEEE P802.3bp is used, which is designed for use in vehicles. It is also known as 100BASE-T1. Connectors other than the RJ51 connectors known from LAN technology are also used there.

In the 2 the various Ethernet ports per network switching unit are numbered consecutively with the reference symbols P1 to P3. Port P1 is used to connect the neighboring network subscriber station clockwise. Port P2 is used to connect the neighboring network subscriber station counterclockwise. The representation in the 2 chosen so that the network switching units S1 to S4 are each represented as a separate hardware component. The reference symbols ZT1 to ZT4 designate assignment tables whose function serves to configure the network switching units S1 to S4, which will be explained in more detail below. In addition, a microprocessor μP1 to μP4 is shown for each network switching unit S1 to S4. However, it is noted that, according to the principle of software-defined networking SDN, the network switching units S1 to S4 can also be implemented in software and are even preferably implemented. Then only the microcontrollers μC1 to μC4, the ports P1, P2 and the environmental sensors would have to be realized as hardware components.

As mentioned, a variant of the Ethernet standard family is used as the network technology for networking the network subscriber stations 10 to 40 . This is particularly advantageous, not only because of the available data transport capacity, but also because it involves networking components from the infotainment area. In this area there are ready-to-use transmission protocols such as IP (Internet Protocol), UDP (User Datagram Protocol), RTP (Real-Time Transport Protocol), RTSP (Real-Time Streaming Protocol) and TCP (Transmission Control Protocol). can.

Allocation tables ZT1 to ZT4 are used to organize the communication flow in the ring bus network consisting of the four network subscriber stations 10 to 40. The following describes in detail how the assignment tables ZT1 to ZT4 are structured and which entries are set therein for each network switching unit S1 to S4.

Each network subscriber station 10 to 40 is configured in such a way that it forwards the data packets in a preferred circulation direction if it is not the target addressee itself. As the preferred direction of circulation, clockwise forwarding is used as an example. If it is the target addressee itself, it will forward the data to the local port to which a processing unit of the local network subscriber station is connected. This is in the example shown 2 respectively the port P1.

In the exemplary embodiment, it is assumed that IPv6 addresses are used. However, the invention could also be implemented using IPv4 addresses. The length of IPv6 addresses has been extended from 32 bits to 128 bits compared to IPv4 addresses. Therefore, a different spelling was agreed here. The addresses are no longer specified with three-digit decimal numbers separated by a period, but with four-digit hexadecimal numbers separated by a colon.

An IPv6 address is written in hexadecimal notation as eight 16-bit blocks, each separated by a colon. For simplification, the leading zeros in each block can be omitted and a long block of zeros can be combined into a double colon "::".

The entries in the assignment tables ZT1 to ZT4 have the following form:
<Priority, Inbound Port, Destination IP Address, Outbound Port>.
The priority information is used to classify different entries.

The ports are as in 2 shown numbered, e.g. port P1 of switch 1 goes to port P2 of switch 2 and port P1 of switch 2 goes to port P2 of switch 3 etc. Port P3 stands for the local network subscriber station. The data packets are forwarded there if they specify the address of the local network subscriber station as the destination address. The arithmetic unit μC1 to μC4 of the local network subscriber station, which processes the supplied data, is connected to the port P3. The destination addresses of the respective network subscriber stations 10 to 40 are in 2 specified. The network subscriber station 10 is assigned the destination address 2001:db8::1. The network subscriber station 20 is the target address mapped to resse 2001:db8::2. The network subscriber station 30 is assigned the destination address 2001:db8::3. The network subscriber station 40 is assigned the destination address 2001:db8::1.

As mentioned, the data packets should normally be forwarded in a clockwise direction. That is, if the network subscriber station 10 would like to communicate with the network subscriber station 30, the data packets are routed via the network switching unit S2.

If the network subscriber station 40 wants to communicate with the network subscriber station 30, the data packets are routed to S3 via the network switching unit S1 and S2.

The following entries are made for this in the assignment table ZT1: Assignment table ZT1 normal case priority input port destination address exit port goal rules 100 P1 ipv6. dst=2001 :db8::1 P3 100 p2 ipv6_dst=2001 :db8::1 P3 Forwarding Rules 100 p2 ipv6_dst=2001 :db8::2 P1 100 p2 ipv6_dst=2001 :db8::3 P1 Source Rules 100 P3 ipv6_dst=2001 :db8::1 P3 100 P3 ipv6_dst=2001 :db8::2 P1 100 P3 ipv6_dst=2001 :db8::4 p2 100 P3 ipv6_dst=2001 :db8::3 P1 additional rules 99 p2 ipv6. dst=2001 :db8::2 p2 99 p2 ipv6_dst=2001 :db8::3 p2 99 P1 ipv6_dst=2001 :db8::3 p2 99 P1 ipv6_dst=2001 :db8::4 p2

Various rules are provided in the table. A distinction is made between rules that apply to the normal case and additional rules (so-called back-up rules) that apply to an error. The entry in the Priority field is used to distinguish between them. The rules for the normal case (preferred case) are marked with the entry "100" for the priority specification. The back-up rules are provided with the priority indication "99". A distinction is made between target rules, forwarding rules and source rules, for the normal case and additional rules for the error case.

The destination rules relate to the case that the data packets have the address of their own network subscriber station as the destination address. The port P3 is entered as the output port for the network subscriber station 10 if the data arrive via one of the input ports P1, P2 and have the entry 2001:db8::1 as the destination address.

The forwarding rules determine that the data packets are output via port P1 if they arrive via port P2. In that case they come from the network switching unit S4.

The source rules determine how forwarding takes place when the data packets come from the local processing unit μC1 and are therefore delivered via port P3. They are then returned to port P3 if the address 2001:db8::1 is specified as the destination address. They are output via port P1 if the address 2001:db8::2 is entered as the target address (direct neighbor). They are output via port P2 if the address 2001:db8::4 is entered as the target address (direct neighbor). you will over Port P1 is output if the address 2001:db8::3 is entered as the destination address (1-hop remote network subscriber station 30, therefore clockwise forwarding via network switching unit S2).

The additional rules become valid when an error occurs. They therefore have the lower priority specification so that they cannot normally become active. At the same time, it is prevented in such a way that an additional rule can overwrite a normal case rule. The first additional rule states that the data packet is reflected back if the address of the network subscriber station 20 is entered as the destination address. This rule is used when the forwarding rule for the destination address 2001 :db8::2 returns an error message. This can be caused, for example, by a fault in the connecting line between network subscriber stations 10 and 20. The second additional rule states that the data packet is reflected back if the address of the network subscriber station 30 is entered as the destination address. This rule is used when the forwarding rule for the destination address 2001 :db8::3 returns an error message. This can be caused, for example, by a fault in the connecting line between network subscriber stations 20 and 30. The third additional rule states that the data packet is delivered to port P2 if it was received via port P1 and the address of network subscriber station 30 is entered as the destination address. This corresponds to a forwarding of packets that originate from the network switching unit S2, but which no longer have a direct connection to the neighboring network switching unit S3. The fourth additional rule states that the data packet is delivered to port P2 if it was received via port P1 and the address of network subscriber station 40 is entered as the destination address. This corresponds to a forwarding of data packets that originate from the network switching unit S2, but which no longer have a direct connection to the neighboring network switching unit S3 and therefore could not forward them in a clockwise direction. This rule would also come into play if the connection between network subscriber stations 40 and 30 is disrupted.

Instead of the precautionary entry of additional rules, it would also be possible to modify the rules for the normal case if the error occurs. However, this would have the disadvantage that the network switching units could not react immediately and data packets would therefore have to be buffered.

The entries in the following allocation tables are listed for the other network subscriber stations 20, 30 and 40: Allocation table ZT2, normal case priority input port destination address exit port goal rules 100 P1 ipv6. dst=2001 :db8::2 P3 100 p2 ipv6_dst=2001 :db8::2 P3 Forwarding Rules 100 p2 ipv6_dst=2001 :db8::3 P1 100 p2 ipv6_dst=2001 :db8::4 P1 Source Rules 100 P3 ipv6_dst=2001 :db8::2 P3 100 P3 ipv6_dst=2001 :db8::3 P1 100 P3 ipv6_dst=2001 :db8::1 p2 100 P3 ipv6_dst=2001 :db8::4 P1 additional rules 99 p2 ipv6_dst=2001 :db8::2 p2 99 p2 ipv6_dst=2001 :db8::4 p2 99 P1 ipv6_dst=2001 :db8::4 p2 Allocation table ZT3, normal case priority input port destination address exit port goal rules 100 P1 ipv6_dst=2001 :db8::3 P3 100 p2 ipv6_dst=2001 :db8::3 P3 Forwarding Rules 100 p2 ipv6_dst=2001 :db8::4 P1 100 p2 ipv6. dst=2001 :db8::1 P1 Source Rules 100 P3 ipv6_dst=2001 :db8::3 P3 100 P3 ipv6_dst=2001 :db8::4 P1 100 P3 ipv6_dst=2001 :db8::2 p2 100 P3 ipv6_dst=2001 :db8::1 P1 additional rules 99 p2 ipv6_dst=2001 :db8::4 p2 99 p2 ipv6_dst=2001 :db8::1 p2 99 P1 ipv6_dst=2001 :db8::1 P1 Allocation table ZT4. normal case priority input port destination address exit port goal rules 100 P1 ipv6_dst=2001 :db8::4 P3 100 p2 ipv6_dst=2001 :db8::4 P3 Forwarding Rules 100 p2 ipv6_dst=2001 :db8::1 P1 100 p2 ipv6. dst=2001 :db8::2 P1 Source Rules 100 P3 ipv6_dst=2001 :db8::4 P3 100 P3 ipv6_dst=2001 :db8::1 P1 100 P3 ipv6_dst=2001 :db8::3 p2 100 P3 ipv6_dst=2001 :db8::2 P1 additional rules 99 p2 ipv6_dst=2001 :db8::1 p2 99 p2 ipv6_dst=2001 :db8::2 p2 99 P1 ipv6_dst=2001 :db8::2 p2

3 shows the typical error case of a connection abort on a ring bus segment. It is the segment between network switching unit S2 and network switching unit S3. The control units CU1 to CU4 can react to different situations and change the entries in the assignment tables. Especially in the area of software-defined networking SDN, it is advantageous that the control units CU1 to CU4 are also implemented as software modules, which are processed, for example, on the microcontrollers μC1 to μC4.

Normally, a communication flow is routed from network switching unit S1 (top left) to network switching unit S3 (bottom right) via network switching unit S2 (top right), see communication route KR1.

If there is an error in the ring bus network, as shown, on the ring bus segment ET2, an alternative route must be used. For this, the network switching unit S2 and the network switching unit S3 first detect the error. This can be done either through the bit transmission layer (Layer 1) or with the BFD protocol, as described above. The network switching units S2 and S3 will both recognize the loss of connection and report it to their directly neighboring network switching units S1 and S4. This is done with the signaling messages SN1 and SN2.

After that, network switching unit S2 mirrors the communication flow back to network switching unit S1, see communication route KR2. Network switch S1 has implemented an overhead path that forwards the communication flow to network switch S4. This then finally forwards to the actual destination, network subscriber station 30. There is still a problem in that network switching unit S1 from network subscriber station 10 still sends the data via network switching unit S2 ("crankback" path), since it has nothing about the connection abort between network switching unit S2 and S3 white. This results in an unnecessary load on the network switching unit S2 or the ring bus segment ET1 between network switching units S1 and S2.

So that the neighboring nodes, network switching units S1 and S4, can also recognize this error status, there are now 2 options: either they react to the alternative route used, which is not normally used, or the control units of network switching units S2 and S3 signal their Neighbors abort the connection via the signaling messages SN1 and SN2. In both cases the network switches have to modify their allocation tables. When this is done, the communication is now routed directly via the network switching unit S4, communication route KR3.

Should the defective connection be reestablished, this is communicated by renewed signaling messages from the control units CU1 and CU2, whereupon the routing tables are reset to their original state.

If a connection breaks down, for example between network switching units S2 and S3, the allocation table ZT1 is changed as follows: Allocation table ZT1, connection break between S2 and S3 priority input port destination address exit port goal rules 100 P1 ipv6_dst=2001 :db8::1 P3 100 p2 ipv6_dst=2001 :db8::1 P3 Forwarding Rules 100 p2 ipv6. dst=2001 :db8::2 P1 100 p2 ipv6_dst=2001 :db8::3 P1 Source Rules 100 P3 ipv6_dst=2001 :db8::1 P3 100 P3 ipv6_dst=2001 :db8::2 P1 100 P3 ipv6_dst=2001 :db8::4 p2 100 P3 ipv6_dst=2001 :db8::3 p2 additional rules 99 p2 ipv6_dst=2001 :db8::2 p2 99 p2 ipv6_dst=2001 :db8::3 p2 99 P1 ipv6_dst=2001 :db8::3 p2 99 P1 ipv6_dst=2001 :db8::4 p2

After the arrival of the signaling message from network switching unit S2, the entry for the output port of the source rule for network subscriber station 30 is switched to P2 in assignment table ZT1 for the normal case. The data packets in the ring bus network are thus routed counterclockwise from network subscriber station 10 to network subscriber station 30 . This applies to the assignment tables ZT1.

The following change applies to table ZT2: assignment table ZT2, connection abort between S2 and S3 priority input port destination address exit port goal rules 100 P1 ipv6_dst=2001 :db8::2 P3 100 p2 ipv6_dst=2001 :db8::2 P3 Forwarding Rules p2 p2 Source Rules 100 P3 ipv6_dst=2001 :db8::2 P3 100 P3 ipv6_dst=2001 :db8::3 P4 P2 100 P3 ipv6_dst=2001 :db8::1 p2 100 P3 ipv6_dst=2001 :db8::4 P1 additional rules 99 p2 ipv6_dst=2001 :db8::2 p2 99 p2 ipv6_dst=2001 :db8::4 p2 99 P1 ipv6_dst=2001 :db8::4 p2

The forwarding rules are deleted, which means that the additional rules become valid. In addition, the source rule for destination address 2001 :db8::3 is changed. Port P2 is entered as the output port, since the direct route to neighboring station 30 is not available. In addition, it is mentioned that the additional rule for input port P1 with destination address 2001:db8::4 is not used because no data can come via input port P1 due to the connection abort.

The following change applies to table ZT3: assignment table ZT3, connection abort between S2 and S3 priority input port destination address exit port goal rules 100 P1 ipv6_dst=2001 :db8::3 P3 100 p2 ipv6_dst=2001 :db8::3 P3 Forwarding Rules 100 p2 ipv6_dst=2001 :db8::4 P1 100 p2 ipv6_dst=2001 :db8::1 P1 Source Rules 100 P3 ipv6_dst=2001 :db8::3 P3 100 P3 ipv6_dst=2001 :db8::4 P1 100 P3 ipv6_dst=2001 :db8::2 P1 100 P3 ipv6_dst=2001 :db8::1 P1 additional rules 99 p2 ipv6_dst=2001 :db8::4 p2 99 p2 ipv6_dst=2001 :db8::1 p2 99

First, the forwarding rules are no longer used because port P2 is no longer available for forwarding the data packets as long as the connection is broken. In addition, the source rule for destination address 2001 :db8::2 is changed. Port P1 is entered as the output port, since the direct route to neighboring station 20 is not available. It is also mentioned that the additional rules for input port P2 with destination addresses 2001 :db8::1 and 2001 :db8::4 are not used because the rules are only used in the event of an error between network switching unit S1 and network switching unit S2. In addition, the additional rule for input port P1 with destination address 2001:db8::1 is deleted, since the rule is only used in the event of an error between switch 1 and switch 2, which is not present. For this reason, this rule does not necessarily have to be deleted.

The following applies to table ZT4: Allocation table ZT4, disconnection between S2 and S3 priority input port destination address exit port goal rules 100 P1 ipv6_dst=2001 :db8::4 P3 100 p2 ipv6_dst=2001 :db8::4 P3 Forwarding Rules 100 p2 ipv6_dst=2001 :db8::1 P1 100 p2 ipv6_dst=2001 :db8::2 P1 Source Rules 100 P3 ipv6_dst=2001 :db8::4 P3 100 P3 ipv6_dst=2001 :db8::1 P1 100 P3 ipv6_dst=2001 :db8::3 p2 100 P3 ipv6. dst=2001 :db8::2 P1 additional rules 99 p2 ipv6. dst=2001 :db8::1 p2 99 p2 ipv6. dst=2001 :db8::2 p2 99 P1 ipv6_dst=2001 :db8::2 p2

No changes are necessary in this assignment table. The forwarding rule for destination address 2001:db8::2 is not required since network switching unit S3 can no longer reach network switching unit S2 directly.

This principle presented works regardless of where the disconnection occurs. The reason why the rule changes at network switches S2 and S3 are asymmetric/unequal is that the disconnect for network switch S2 is "in front of it" and for network switch S3 it is "behind", ie network switch S2 can no longer forward (routing in clockwise direction) and must use the alternative routes.

4 now shows the structure of a signaling message. The signaling message is transmitted as a UDP packet according to the UDP protocol user datagram protocol. The UDP packet is embedded in an IP packet, which in turn is embedded in an Ethernet packet 40, transmitted. At the beginning of the Ethernet data frame is the Ethernet MAC header, followed by the IP header and the UDP header in field 41. Two fields can also be provided for this. Feld42 follows in the user data field of the UDP packet. Field 42 contains the signaling message number SNN. In the exemplary embodiment, eight different signaling messages are distinguished for each event, depending on which ring bus segment has failed or has been repaired. It is programmed in the control units CU1 to CU4 which changes must be made in the assignment tables for which signaling message.

Since the utilization of the still active ring bus segments increases in the event of an error, it is advantageous if critical communication flows are prioritized using quality of service mechanisms such as "bandwidth shaping". The control units can also be used for this purpose.

In addition, with the solution presented, it is possible to set which traffic flows are to be replicated on both circulation paths or which are to be secured only by the fail-over mechanism described above.

The disclosure is not limited to the exemplary embodiments described here. There is room for various adaptations and modifications that those skilled in the art would contemplate based on their skill in the art as well as belonging to the disclosure. All examples mentioned herein, as well as conditional language, are intended to be understood as not being limited to such specifically cited examples. For example, it will be appreciated by those skilled in the art that the block diagram presented herein represents a conceptual view of exemplary circuitry. Similarly, it will be appreciated that an illustrated flowchart, state transition diagram, pseudo-code, and the like are various variations representing processes that may be stored substantially on computer-readable media and thus executable by a computer or processor.

It should be understood that the proposed method and associated devices can be implemented in various forms of hardware, software, firmware, special purpose processors or a combination thereof. Specialty processors can include Application Specific Integrated Circuits (ASICs), Reduced Instruction Set Computers (RISC), and/or Field Programmable Gate Arrays (FPGAs). Preferably, the proposed method and device is implemented as a combination of hardware and software. The software is preferably installed as an application program on a program storage device. Typically, it is a computer platform based machine that includes hardware such as one or more central processing units (CPU), random access memory (RAM), and one or more input/output (I/O) interfaces. An operating system is typically also installed on the computer platform. The various processes and functions described here can be part of the application program or a part that is executed via the operating system.

Reference List

10

1. Network subscriber station

20

2. Network subscriber station

30

3. Network subscriber station

40

4. Network subscriber station

50

signaling message

51

IP & UDP headers

52

payload field

100

vehicle

CU1

1. Control unit

CU2

2. Control unit

CU3

3. Control unit

CU4

4. Control unit

P1

1st port

p2

2nd Port

P3

3rd port

ET1

1. Ring bus segment

ET2

2. Ring bus segment

ET3

3. Ring bus segment

ET4

4. Ring bus segment

K1

1. Camera sensor

K2

2. Camera sensor

K3

3. Camera sensor

K4

4. Camera sensor

KR1

1. Communication route

KR2

2. Communication route

KR3

3. Communication route

L1

lidar sensor

µP1

1. Processor unit

µP2

2. Processor unit

µP3

3. Processor unit

µP4

4. Processor unit

R1

1. Radar sensor

R2

2. Radar sensor

S1

1. Network switching unit

S2

2. Network switching unit

S3

3. Network switching unit

S4

4. Network switching unit

SN1

1. Signaling message

SN2

2. Signaling message

ZT1

1. Mapping table

ZT2

2. Mapping table

ZT3

3. mapping table

ZT4

4. Mapping table

QUOTES INCLUDED IN DESCRIPTION

This list of the documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• DE 102014225802 A1 [0006]
• US 2012/0109446 A1 [0007]
• US 2013/0322434 A1 [0008]
• DE 10161186 B4 [0010]

Non-patent Literature Cited

• "Fast Recovery in Software-Defined Networks" by Niels L. M. van Adrichem, Benjamin J. van Asten and Fernando A. Kuipers of Delft University of Technology, published in 2014 [0005]

Claims (15)

Method for fault-tolerant operation of a network arrangement, the network arrangement consisting of a number of network subscriber stations (10, 20, 30, 40) which are connected to one another by communication lines (ET1, ET2, ET3, ET4) in a ring, so that the data packets in the ring network can circulate from one network subscriber station (10, 20, 30, 40) to another network subscriber station (10, 20, 30, 40), characterized by the steps of specifying a preferred circulation direction for the transmission of data packets from an originating network subscriber station (10) to a Target network subscriber station (30), sending a data packet from the source network subscriber station (10) to the target network subscriber station (30) along the preferred circulation direction, checking whether the sent data packet arrives at the target network subscriber station (30), when a connection break between two network subscriber stations is detected (10, 20, 30, 4th 0), returning the data packet to the source network subscriber station (10), reversing the direction of circulation and sending the data packet from the source network subscriber station (10) to the destination network subscriber station (30) in the reverse circulation direction. procedure after claim 1 , wherein the network subscriber stations (10, 20, 30, 40) have a programmable network switching unit (S1, S2, S3, S4) or are connected to a programmable network switching unit (S1, S2, S3, S4) and the preferred circulation directions for the communication to the individual network subscriber stations (10, 20, 30, 40) can be entered in the respective network subscriber stations (10, 20, 30, 40) in allocation tables (ZT1, ZT2, ZT3, ZT4). procedure after claim 2 , whereby the reversal of the circulation directions is entered in the assignment tables (ZT1, ZT2, ZT3, ZT4). procedure after claim 2 or 3 , the two network switching units (S2, S3) sending (SN1, SN2) an error message (SN1, SN2) to their respective other neighboring network switching units (S1, S4), in which the connection abort in the relevant ring bus segment ( ET2) is communicated. procedure after claim 4 , the respective other network switching unit (S1, S4) being configured after the error message (SN1, SN2) has arrived so that it is entered in its respective assignment table (ZT1, ZT4) that the data traffic is to be forwarded in the opposite direction of circulation. procedure after claim 4 or 5 , the two network switching units (S2, S3) determining in one step whether the connection abort in the relevant ring bus segment (ET2) has been rectified, and if so, that they carry out a reconfiguration of their allocation table (ZT2, ZT3), to return to normal operation. procedure after claim 6 , wherein the other network switching unit (S1, S4) is informed by the neighboring network switching unit (S2, S3) that the connection abort in the relevant ring bus segment (ET2) has been rectified, and after receipt of the notification that the fault has been rectified is configured in such a way that its assignment table (ZT1, ZT4) is entered so that the data traffic is to be forwarded along the preferred circulation direction. Method according to one of the preceding claims, a variant of the Ethernet standard family being used as the network technology for connecting the network subscriber stations (10, 20, 30, 40). Device for carrying out the method according to one of the preceding claims, having a number of network subscriber stations (10, 20, 30, 40) which are connected to one another in a ring by communication lines (ET1, ET2, ET3, ET4), characterized in that the network subscriber stations ( 10, 20, 30, 40) have a programmable network switching unit (S1, S2, S3, S4) or are connected to a programmable network switching unit (S1, S2, S3, S4), the programmable network switching unit (S1 , S2, S3, S4) has an assignment table (ZT1, ZT2, ZT3, ZT4) in which a preferred circulation direction for the transmission of data packets from an originating network subscriber station (10) to a destination network subscriber station (30) is entered, the Network switching unit (S1, S2, S3, S4) further comprises a processor unit (μP1, μP2, μP3, μP4) with an installed network protocol stack, which is used for testing the connection to a b Neighboring network subscriber station (30) is designed, wherein the processor unit (uP2) is designed so that it at Detection of a broken connection to a neighboring network subscriber station (30), the programmable network switching unit (S2) configured so that it sends back a data packet received in the preferred circulation direction in the opposite circulation direction to the original network subscriber station (10). device after claim 9 , wherein the processor unit (.mu.P2, .mu.P3) is designed to send an error message (SN1, SN2) to the other neighboring network subscriber station (10, 40) to notify the connection abort in the relevant ring bus segment (ET2). Network switching unit, for a device according to claim 9 or 10 , characterized in that the network switching unit (S2, S3) has a processor unit (µP2) with an installed network protocol stack which is designed for testing the connection to an adjacent network subscriber station (10, 40), the processor unit (µP2, µP3) is designed to send an error message (SN1, SN2) to another neighboring network subscriber station (10, 40), in which the connection abort in the relevant ring bus segment (ET2) is communicated to the first neighboring network subscriber station (20, 30). and the programmable network switching unit (S2, S3) is configured in such a way that, if the connection is terminated, it sends back a data packet received in the preferred circulation direction to the originating network subscriber station (10, 40) in the opposite circulation direction. network switching unit, after claim 11 , wherein the network switching unit (S2, S3) has an assignment table (ZT2, ZT3) in which the preferred circulation direction for the transmission of data packets from the source network subscriber station (10, 40) to a destination network subscriber station (30) is entered, the Network switching unit (S2, S3) is further configured in such a way that, after detecting the loss of connection in the relevant ring bus segment (ET2) or after receiving the error message (SN1, SN2), it changes the assignment table (ZT2) in such a way that it changes the specification means that the data packets should be forwarded in the opposite direction to the preferred circulation direction. network switching unit claim 12 , wherein the network switching unit (S2, S3) is designed such that, after determining that the fault has been rectified in the relevant ring bus segment (ET2), it is configured such that its assignment table (ZT2, ZT3) is entered that data traffic should be forwarded again along the preferential circulation direction. Vehicle, characterized in that in the vehicle according to a device claim 9 or 10 is integrated. vehicle after Claim 14 , wherein the network subscriber stations (10, 20, 30, 40) a camera control unit and/or ultrasonic sensor control unit and/or RADAR sensor control unit, corresponding to Radio Detection and Ranging and/or LIDAR sensor control unit, corresponding to Light detection and ranging.

Images