DATA PROCESSING PROCEDURES

Description

[0001] The invention relates to a data processing method, a system for carrying out the data processing method, and a computer program product.

[0002] Automated industrial processes require computer-aided data processing methods that must meet high security standards. Even a single error in such a data processing method can interrupt the industrial process or even cause damage to property or personal injury. Errors in computer-aided data processing methods can arise from hardware, software, or data transmission. Various methods are already known for detecting and preventing such errors. For example, to detect and prevent hardware-related errors, data processing takes place on redundant hardware components. For instance, a data processing program is executed multiple times on one or more cores of a processor within a hardware component.This has the disadvantage that the processor's computing resources are occupied solely for the redundant execution of the data processing program. Furthermore, data processing procedures become longer due to the need for error detection and prevention methods, as these methods need to be more effective and reliable in detecting and preventing errors. However, such an extension of the data processing procedure hinders the fastest possible execution. In particular, control systems that rely on data processing procedures are subject to high demands for the fastest possible data processing in order to achieve high operational reliability.

[0003] In the article "Combined task- and network-level scheduling for distributed time-triggered systems", published by Springer Science+Business Media New York in 2015, the authors Silviu S. Craciunas and Ramon Serna Oliver describe so-called Ethernet-based time-triggered networks (TTEthernet).

[0004] In the slide set from NASA Johnson Space Center (NASA: National Aeronautics and Space Administration) entitled "On Time-Triggered Ethernet in NASA's Lunar Gateway", author Andrew Loveless reports on Time-Triggered Ethernet in NASA's Lunar Gateway.

[0005] In the article "Time Triggered Architecture: A Consistent Computing Platform", published in IEEE MICRO, IEEE SERVICE CENTER, LOS ALAMOS, CA, US, in Volume 2, No. 4, the authors Reinhard Maier, Günther Bauer, Georg Stöger and Stefan Poledna describe a time-triggered architecture.

[0006] The object of the invention is to provide an alternative data processing method with which a high level of operational reliability can be achieved in a time-efficient manner.

[0007] This task is solved by a data processing method with the features of claim 1.

[0008] Furthermore, the invention is based on the objective of providing a system for carrying out the data processing method according to the invention.

[0009] This task is solved by a system with the features of the subordinate system claim.

[0010] Furthermore, the invention is based on the objective of providing a computer program product for carrying out the data processing method according to the invention with the system according to the invention.

[0011] This task is solved by a computer program product with the features of the subordinate computer program product claim.

[0012] Advantageous embodiments of the present invention can be found in the dependent subclaims.

[0013] In the data processing method according to the invention, data is transmitted between different network participants in a network according to a communication cycle consisting of successive time intervals defined in a communication scheme. In particular, the data is transmitted in a time-controlled manner. Preferably, the time settings of the network participants are synchronized according to a common clock signal. This can be achieved, for example, using a so-called "Precision Time Protocol (PTP)," which is known to those skilled in the art. During at least one time interval of the communication cycle, predetermined network participants are authorized to either send or receive the data. The communication cycle can, for example, be implemented based on a time-division multiplexing method.Time-division multiple access (TDMA) is also known by its English name. TDMA is preferably a synchronous method. In synchronous TDMA, time intervals of a fixed length are used. During these intervals, predetermined network participants are authorized to send data, and other predetermined network participants are authorized to receive data. This allows communication within a network to be easily and physically divided. Furthermore, it enables cost-effective convergence with other networks.

[0014] Furthermore, it is conceivable that the data is transmitted according to a secure transmission protocol. This could, for example, be a transmission protocol according to the IEC 61784-3 standard. In this way, the integrity of the transmitted data can be ensured. Moreover, security requirements for data transmission in a network can be easily met.

[0015] In another advantageous implementation, data is transferred between different network participants according to a "publish / subscribe" protocol. For this purpose, the data can be transferred, for example, according to the "OPC UA PubSub" standard. This eliminates the need for a sender to address the data. The data can thus be transmitted cost-effectively from one or more senders to one or more receivers unknown to the senders.

[0016] Preferably, at least two predetermined network participants are authorized to transmit data during a time interval of the communication cycle. Particularly preferably, during this time interval, at least one of the at least two predetermined network participants is authorized to send data, and at least one other of the at least two network participants is authorized to receive data. This enables a reliable data stream from a sender to a receiver.

[0017] Furthermore, the data processing method according to the invention provides several data processing devices as network participants. A data processing device can, for example, be one or more microprocessors, a computer, or another device known to those skilled in the art for executing machine-readable instructions. Each of the several data processing devices undergoes a process cycle. In each process cycle, successive process steps are executed, in which at least a portion of the data is either read in, processed, or output. In particular, the process cycle corresponds to a basic scheme based on the input-processing-output principle. The arrangement and sequence of the individual process steps can be arbitrary.

[0018] Furthermore, in the data processing method according to the invention, the communication cycle is synchronized with the process cycle of at least one of the several data processing devices in such a way that, during a temporal cycle of the communication cycle, a predetermined process step of the process cycle of the at least one of the several data processing devices is executed during a time interval of the communication cycle assigned to this predetermined process step. This makes it possible to keep the duration of the data processing method short. The duration of the data processing method can be kept largely independent of the number of network participants involved in the data processing method, and in particular, independent of the number of data processing devices.Furthermore, in the data processing method according to the invention, the communication cycle and the process cycle of the aforementioned at least one of the several data processing devices are repeated multiple times.

[0019] Using the data processing method according to the invention, a data flow between sender and receiver can be reliably assigned. Furthermore, it enables data from multiple network participants to be received, processed, and / or output simultaneously. Several data processing devices can run process cycles in parallel and execute process steps concurrently. Depending on the process steps, a data flow between predetermined senders and predetermined receivers can be uniquely assigned. This effectively prevents the detrimental lengthening of the data processing method with an increasing number of network participants, particularly data processing devices.

[0020] Preferably, the aforementioned communication cycle, as well as all process cycles of the multiple data processing devices synchronized with the aforementioned communication cycle, are executed multiple times. This enables the use of the data processing method in cyclically repeating processes, such as control processes.

[0021] In an advantageous refinement of the data processing procedure, it is stipulated that the predetermined process step of the process cycle is executed completely within the time interval of the communication cycle assigned to that predetermined process step. Each process step can thus be uniquely assigned to a time interval. This makes it possible to complete a process step before a change in the permissions of network participants occurs in a subsequent time interval. This prevents errors during data transmission, such as the loss of data being transmitted.

[0022] In an advantageous embodiment of the further training, the duration of the predetermined time interval is adjusted to the duration of the predetermined process step in such a way that the duration of the predetermined time interval is essentially the same as the duration of the predetermined process step. In this way, periods of the communication cycle during which no data transmission takes place can be prevented.

[0023] In a further advantageous embodiment, it is provided that during a first-order time interval of the communication cycle, at least one of the several data processing devices is authorized to receive data. During this first-order time interval, input data is read by this at least one data processing device. This makes it possible to easily predetermine a data flow from a group of network participants to at least one data processing device. Input data can, for example, be data acquired by sensors. In particular, input data can be the current state data of a process to be controlled.

[0024] In an advantageous embodiment of the aforementioned further development, it is provided that during the first type of time interval of the communication cycle, the multiple data processing devices are authorized to receive data. Furthermore, during this first type of time interval, identical input data is read by each of the multiple data processing devices. In this way, the duration for reading input data is largely independent of the total number of data processing devices.

[0025] Furthermore, an advantageous embodiment provides that during a second type of time interval of the communication cycle, input data is processed into an output data set by at least one of the several data processing devices. In this way, data can be transmitted between any group of network participants during the second type of time interval. This allows for the utilization of available transmission capacity. In particular, within the second type of time interval, any number of network participants can be authorized to send data, and other network participants can be authorized to receive data.

[0026] In an advantageous embodiment of the aforementioned further development, it is provided that during the second time interval, a checksum is calculated from the output data set by means of at least one of the several data processing devices. Based on a checksum, the content of the output data set can be provided in summarized form. In particular, output data sets with the same content have the same checksum.

[0027] In a further advantageous embodiment of the aforementioned refinement, it is provided that during the second type of time interval, each of the several data processing devices processes the input data into an output data set using the same deterministic data processing program. In this context, a deterministic data processing program is understood to be a program that, upon repeated execution, calculates the same output data set each time, starting from the same input data. Deviations between the output data calculated by the deterministic data processing program from the same input data can only differ in terms of hardware-related errors. In this way, an arbitrarily scalable number of independently calculated output data sets can be generated in a time-efficient manner, and in particular, simultaneously.It is conceivable that the aforementioned data processing program is stateless. In the case of a stateless data processing program, the output data records depend solely on the input data. Preferably, during the second type of time interval, a checksum is calculated from the output data record by each of the several data processing devices. The checksum enables a simple and quick comparison of the output data records.

[0028] Furthermore, it is conceivable that at least some of the multiple data processing devices process the input data into an output data set using a data processing program that is identical to the aforementioned data processing program, instead of the same data processing program. In this context, "two identical data processing programs" means that both programs, starting from the same input data, are programmed differently to calculate the same output data set. In particular, identical data processing programs are programmed diversely. A comparison of the output data sets of the differently programmed data processing programs makes it possible to identify systematic errors based on programming errors, in addition to the previously mentioned random errors.

[0029] Another advantageous refinement provides that during a third type of communication cycle, data transmission is restricted exclusively to transmission between the multiple data processing devices. This limits the output of potentially erroneous data records to output to the multiple data processing devices. The potential impact of an erroneous output data record on other network participants can thus be avoided.

[0030] In an advantageous embodiment, it is provided that during the third type of time interval, only the first data processing device of the multiple data processing devices is authorized to receive data, while other data processing devices of the multiple data processing devices are authorized to send data. During the third type of time interval, each of the other data processing devices outputs a data record, which is then read in by the first data processing device. This enables centralized and time-efficient data processing with exactly one data processing device.

[0031] In a further advantageous embodiment, it is provided that during the third type of time interval, the additional data processing devices each output the checksum calculated in the second type of time interval. Here, this checksum is output by some of the additional data processing devices instead of the output data record. In this way, the amount of data to be transmitted in the network can be reduced, and in particular, the third type of time interval can be shortened.

[0032] In a further advantageous embodiment, it is provided that during a fourth-order time interval of the communication cycle, only the first data processing device is authorized to send data. During this fourth-order time interval, the first data processing device compares its own output data record with at least one input output data record. In an advantageous alternative embodiment, during the fourth-order time interval, the first data processing device compares a checksum of its own output data record with the input checksums. This comparison enables simple and cost-effective detection of random and / or systematic errors in the output data records. By comparing checksums instead of output data records, the fourth-order time interval can be shortened.

[0033] Furthermore, an advantageous refinement provides that, if the output data record matches at least one input output data record during the fourth time interval, the output data record is output by the first data processing device. Alternatively, it is provided that, if the checksum of the output data record matches the input checksums during the fourth time interval, the output data record is output by the first data processing device. This allows for a simple and cost-effective verification of the output data record using an input output data record or the input checksums. It is also conceivable that a separate time interval is provided for the output data record to be output by the first data processing device.In an advantageous embodiment, during the fourth type of time interval, at least one network participant designated to receive the output data set is authorized to do so. For example, a network participant configured as an actuator is authorized to receive the output data set. This enables a clear assignment of the data flow between the network participants.

[0034] Furthermore, an advantageous refinement provides that, in the event of a discrepancy between the output data record and a read-in output data record during the fourth time interval, the first data processing device outputs such a data record that matches at least one other read-in output data record. This, however, requires that at least two output data records are read in by the first data processing device during the third time interval. If, on the other hand, only one output data record is read in by the first data processing device during the third time interval, output by the first data processing device is prevented.Alternatively, it is provided that if a discrepancy arises between the checksum of the first data processing device's own output data record and at least one of the read checksums during the fourth time interval, the first data processing device outputs a data record whose checksum matches at least one other checksum. When comparing checksums, it may be sufficient for the first data processing device to read only one output data record from at least one of the other data processing devices during the third time interval, in addition to the checksums. This allows for simple and reliable verification of output data records. By outputting a verified output data record from another data processing device, an interruption of the data processing process can be prevented.

[0035] The data processing method according to the invention can be carried out using the system according to the invention.

[0036] The system according to the invention comprises a network configured to transmit data between different network participants according to a communication cycle consisting of successive time intervals defined in a communication scheme. This allows predetermined network participants to be authorized to send or receive data during one or more predetermined time intervals. The communication scheme, communication cycle, and time intervals are, in particular, those already described previously in connection with the data processing method. Furthermore, the system according to the invention comprises several data processing devices, each configured as a network participant.The data processing devices correspond in particular to those described in connection with the data processing procedure. Each of the multiple data processing devices is configured to read, process, and / or output at least a portion of the data in a process cycle consisting of successive process steps. The process cycles and the associated process steps correspond in particular to those previously described in connection with the data processing procedure.Furthermore, in the system according to the invention, the communication cycle is synchronized with the process cycle of at least one of the several data processing devices in such a way that, during a time interval of the communication cycle, a predetermined process step of the process cycle of the at least one of the several data processing devices can be executed during a time interval of the communication cycle assigned to this predetermined process step. In addition, the system according to the invention provides for multiple iterations of the communication cycle and the process cycle of the aforementioned at least one of the several data processing devices. In this way, a system can be provided that offers arbitrary scalability with regard to a number of network participants without adversely extending the duration of the data processing procedure.

[0037] In an advantageous advanced training system, at least one sensor is included as a network participant. This enables the simple provision of input data.

[0038] Another advantageous feature is the inclusion of at least one actuator as a network participant. This allows a system to be quickly and reliably transformed into a target state based on an output data set.

[0039] The computer program product according to the invention can be used to cause the system according to the invention to execute the data processing method according to the invention. For this purpose, the computer program product includes machine-readable instructions. This enables a cost-effective implementation of the data processing method.

[0040] The properties, features, and advantages of the invention described above, as well as the manner in which they are achieved, are explained in more detail in the following description of an embodiment of the invention in conjunction with the figures. Where appropriate, the same reference numerals are used in the figures for the same or corresponding elements of the invention. The embodiment serves to illustrate the invention and does not limit the invention to the combinations of features specified therein, including functional features. Furthermore, all specified features can be considered in isolation and combined appropriately with the features of any claim.

[0041] They show: FIG 1 an illustration of an embodiment of a data processing method according to the invention and a schematic representation of an embodiment of a system according to the invention for carrying out the data processing method; FIG 2 an illustration of a time-lapse of the embodiment of the data processing method.

[0042] FIG 1 Figure 1 shows an embodiment of a network 10. The embodiment is schematically represented here as a logical topology. A physical topology of the network 10 can be derived from the one shown in Figure 1. FIG 1 The logical topology shown may differ. Furthermore, it illustrates FIG 1 An exemplary embodiment of a data processing method that processes actual state data of a controlled process into target state data of the controlled process. The network 10 can, in principle, have any number of network participants. In FIG 1 For the sake of clarity, only those network participants relevant to the following exemplary implementation will be described and illustrated in the following description.

[0043] In the present embodiment, the network participants of network 10 are a first data processing device 26, three further data processing devices 28, two sensors 30, and two actuators 32. The two sensors 30 are each configured to acquire actual state data of the process to be controlled. The two actuators 32 are each configured to receive target state data of the process to be controlled. Based on this target state data, the process to be controlled is transformed from an actual state to a target state in the present embodiment by means of the two actuators 32. Furthermore, in the present embodiment, the target state data is calculated by means of a deterministic data processing program based on the actual state data.The same deterministic data processing program is executed by both the first data processing device 26 and the three other data processing devices 28.

[0044] In network 10, data can be transferred between the various network participants. In this embodiment, the data is transferred according to a communication cycle 14 defined in a communication scheme 12. In this embodiment, the communication cycle 14 is implemented using a synchronous time-division multiplexing (TDM) method. This can be implemented, for example, using the IEEE 802.1Qbv standard. For this purpose, the time settings of the network participants are synchronized according to a common clock signal. The time settings of the network participants can be synchronized, for example, using a Precision Time Protocol (PTP). Furthermore, the data is transferred between the network participants based on a publish / subscribe protocol. For example, the OPC UA PubSub standard can be used as the publish / subscribe protocol.Furthermore, in the present embodiment, the data is transmitted between the network participants according to a secure transmission protocol.

[0045] FIG 2 Figure 1 illustrates a temporal progression of the exemplary implementation of the data processing method. In this exemplary implementation, the communication cycle 14 consists of five consecutive time intervals 16, 18, 20, 22, and 24, each with a fixed duration according to the synchronous time-division multiplexing method. The durations of the time intervals 16, 18, 20, 22, and 24 can vary. The communication scheme 12 defines which network participants are authorized to send or receive data in which time interval. Furthermore, Figure 12 illustrates... FIG 2 a first process cycle 34, which is traversed by the first data processing device 26. The first process cycle 34 has five consecutive process steps 38, 40, 42, 44, 46. Furthermore, it illustrates FIG 2 a further process cycle 36. The further process cycle 36 is traversed by each of the three further data processing devices 28. Furthermore, in the present embodiment, the further process cycle 36 has three successive process steps 48, 50, 52.

[0046] Communication cycle 14 is synchronized with the first process cycle 34 and the subsequent process cycle 36 such that, during a complete cycle of communication cycle 14, each process step 38, 40, 42, 44, 46 of the first process cycle 34 and each process step 48, 50, 52 of the subsequent process cycle 36 is fully executed during exactly one of the time intervals 16, 18, 20, 22 of communication cycle 14 assigned to the respective process steps 38, 40, 42, 44, 46, 48, 50, 52. Furthermore, in this embodiment, communication cycle 14, the first process cycle 34, and the subsequent process cycle 36 all have the same start time.

[0047] In the present embodiment, during a first time interval 16 of the communication cycle 14, both the first data processing device 26 and the three further data processing devices 28 are authorized to receive data. Furthermore, during the first time interval 16, the two sensors 30 are authorized to send data. This enables the first data processing device 26, in a first process step 38 of the first process cycle 34, and the three further data processing devices 28, in a first process step 48 of the subsequent process cycle 36, to read the actual state data of the process to be controlled, provided by the two sensors 30, as input data. This input data is read simultaneously by the first data processing device 26 and the three further data processing devices 28.During the first time interval 16, the input data is completely read in by the aforementioned data processing devices 26, 28. Regardless of the present embodiment, any number of data processing devices can be considered as network participants, through which the same input data can be processed without extending the time of the data processing procedure.

[0048] During a second time interval 18, a second process step 40 of the first process cycle 34 and a second process step 50 of the subsequent process cycle 36 are executed simultaneously. In this process, the first data processing device 26 and the three other data processing devices 28 each use the same deterministic data processing program to completely process the input data into one output data record. Alternatively or additionally, it is conceivable that at least one of the three other data processing devices 28 and / or at least one additional data processing device (not shown in detail) executes a different data processing program completely during the second time interval 18 instead of the same data processing program.The other data processing program would differ from the previously mentioned programs only in that it is programmed differently to calculate the same output data set based on the same input data. A discrepancy between the output data set calculated by one of the same data processing programs and the output data set calculated by the other program can therefore be due to programming errors as well as random, hardware-related errors. In this embodiment, the output data set contains the target state data. Furthermore, 18 checksums are calculated from the respective output data sets during the second time interval. The same checksums result for output data sets with identical content.If the checksums differ, the content of the output data records will also differ. During the second time interval 18, any network participant can be authorized to send data and any other network participant can be authorized to receive data. These may include, in particular, network participants of network 10 that are not shown or described in detail in connection with this embodiment.

[0049] The second time interval 18 of the communication cycle 14 is followed by a third time interval 20 of the communication cycle 14. During the third time interval 20, data transmission is limited exclusively to transmission between the first data processing device 26 and the three other data processing devices 28. The first data processing device 26 is authorized to receive data, and the three other data processing devices 28 are each authorized to send data. During the third time interval 20, each of the three other data processing devices 28 fully executes a third process step 52 of the further process cycle 36. In this exemplary embodiment, one of the three other data processing devices 28 outputs both the calculated output data set and a checksum thereof.Of the remaining two of the three additional data processing devices 28, only a checksum of the respective calculated output data records is output in this third process step 52. Furthermore, during the third time interval 20, a third process step 42 of the first process cycle 34 is executed by means of the first data processing device 26. In this third process step 42 of the first process cycle 34, the data output by means of the three additional data processing devices 28 is read in by means of the first data processing device 26.By restricting the transmission of data during the third time interval 20 exclusively to a transmission between the first data processing device 26 and the three further data processing devices 28, it can be ruled out that the output data set issued by means of said one of the three further data processing devices 28 may be mistakenly issued to the actuators 32.

[0050] During a fourth time interval 22 of the communication cycle 14, only the first data processing device 26 is authorized to send data. During this fourth time interval 22, the two actuators 32 are authorized to receive data. During this fourth time interval 22, both a fourth process step 44 of the first process cycle 34 and a fifth process step 46 of the first process cycle 34 are fully executed. Alternatively, it would be conceivable that the fourth process step 44 and the fifth process step 46 could each be assigned their own time interval in which the respective process steps can be fully executed. In the fourth process step 44, the first data processing device 26 compares a checksum of its own output data record with the checksums read in during the preceding third process step 42 of the first process cycle 34.

[0051] If the checksum of the own output data set matches the respective read checksums, the own output data set is output by the first data processing device 26 in the fifth process step 46 of the first process cycle 34 during the fourth time interval 22.

[0052] In the event of a discrepancy between the checksum of its own output data record and at least one of the read checksums, the first data processing device 26 outputs a verified output data record. To do this, the first data processing device 26 determines which of the checksums differs from the others. If its own checksum is the only one that differs from the read checksums, the output data record read in the third process step 42 is output in the fifth process step 46. However, if one of the read checksums differs from its own checksum and the two other read checksums, the first data processing device 26 again outputs its own output data record in the fifth process step 46.

[0053] This approach allows for easy verification of the output data records. Alternatively, in the event of a discrepancy between the checksums, either the output of a data record can be omitted or the data processing procedure can be aborted. For example, the process being controlled can be placed in a safe state, in which the process is interrupted.

[0054] Finally, in the present embodiment, the communication cycle 14 has a fifth time interval 24. No process step of the aforementioned process cycles 34, 36 is assigned to this fifth time interval 24. During the fifth time interval 24, any network participant may be authorized to send data and any other network participant may be authorized to receive data.

[0055] Furthermore, it illustrates FIG 2The communication cycle 14, the first process cycle 34, and the subsequent process cycle 36 are executed multiple times. Each time, the communication cycle 14, the first process cycle 34, and the subsequent process cycle 36 are executed synchronously. In this embodiment, this is achieved by ensuring that the start times of the communication cycle 14, the first process cycle 34, and the subsequent process cycle 36 coincide at the beginning of each subsequent iteration.

[0056] Although the invention has been further illustrated and described in detail by the preferred embodiments, the invention is not limited by the disclosed examples and other variations can be derived from them by a person skilled in the art without leaving the scope of protection of the invention.

Claims

1. Data processing method, in which - data is transmitted between various network subscribers of a network (10) in accordance with a communication cycle (14) consisting of consecutive time intervals (16, 18, 20, 22, 24) defined in a communication scheme (12); - during at least one time interval (16, 18, 20, 22, 24) of the communication cycle (14) predefined network subscribers are authorised either to send or to receive the data in each case; - a plurality of data processing apparatuses (26, 28) are provided as network subscribers; - a process cycle (34, 36) is executed by each of the plurality of data processing apparatuses (26, 28); - consecutive process steps (38, 40, 42, 44, 46, 48, 50, 52) are carried out in the respective process cycle (34, 36), in which at least a part of the data is either read in, processed or output in each case; - the communication cycle (14) is synchronised with the process cycle (34, 36) of at least one of the plurality of data processing apparatuses (26, 28) in such a way that, during an execution time of the communication cycle (14), a predefined process step (38, 40, 42, 44, 46, 48, 50, 52) of the process cycle (34, 36) of the at least one of the plurality of data processing apparatuses (26, 28) is carried out during a time interval (16, 18, 20, 22) of the communication cycle (14) assigned to this predefined process step (38, 40, 42, 44, 46, 48, 50, 52); and - the communication cycle (14) and also the process cycle (34, 36) of the said at least one of the plurality of data processing apparatuses (26, 28) are executed multiple times.

2. Data processing method according to claim 1, characterised in that the predefined process step (38, 40, 42, 44, 46, 48, 50, 52) of the process cycle (34, 36) is carried out completely within the time interval (16, 18, 20, 22) of the communication cycle (14) assigned to this predefined process step (38, 40, 42, 44, 46, 48, 50, 52).

3. Data processing method according to claim 1 or 2, characterised in that - during a first type of time interval (16) of the communication cycle (14) at least one of the plurality of data processing apparatuses (26, 28) is authorised to receive data; and - during this first type of time interval (16) the input data is read in by this at least one data processing apparatus (26, 28).

4. Data processing method according to claim 3, characterised in that - during the first type of time interval (16) of the communication cycle (14) the plurality of data processing apparatuses (26, 28) are authorised to receive data; and - during this first type of time interval (16) the identical input data is read in by the plurality of data processing apparatuses (26, 28) in each case.

5. Data processing method according to one of the preceding claims, characterised in that during a second type of time interval (18) of the communication cycle (14), by means of at least one of the plurality of data processing apparatuses (26, 28) input data is processed to an output dataset.

6. Data processing method according to claim 5, characterised in that during this second type of time interval (18), the input data is processed to an output dataset by each of the plurality of data processing apparatuses (26, 28) by means of the same deterministic data processing program.

7. Data processing method according to one of the preceding claims, characterised in that during a third type of time interval (20) of the communication cycle (14), a transmission of data is exclusively restricted to a transmission between the plurality of data processing apparatuses (26, 28).

8. Data processing method according to claims 6 and 7, characterised in that - during the third type of time interval (20) only a first data processing apparatus (26) of the plurality of data processing apparatuses (26, 28) is authorised to receive data and further data processing apparatuses (28) of the plurality of data processing apparatuses (26, 28) are authorised to send data; and - during the third type of time interval (20) one output dataset is output in each case by the further data processing apparatuses (26, 28), which is read in by means of the first data processing apparatus (26).

9. Data processing method according to claims 6 and 8, characterised in that - during the second type of time interval (18) a checksum of their output dataset is computed by means of the further data processing apparatuses (28); and - during the third type of time interval (20) this checksum is output in each case by means of the further data processing apparatuses (28), wherein this checksum is output by a part of the further data processing apparatuses (28) instead of the output dataset.

10. Data processing method according to claim 9, characterised in that - during a fourth type of time interval (22) of the communication cycle (14), exclusively the first data processing apparatus (26) is authorised to send data; and - during the fourth type of time interval (22) by means of the first data processing apparatus (26), its own output dataset is compared with at least one output dataset read in or a checksum of its own output dataset is compared with the checksums read in.

11. Data processing method according to claim 10, characterised in that in the case of a match between its own output dataset and the at least one output dataset read in or in the case of a match between the checksum of its own output dataset and the checksums read in during the fourth type of time interval (22), its own output dataset is output by the first data processing apparatus (26).

12. Data processing method according to claim 10, characterised in that in the case of a deviation between its own output dataset and an output dataset read in or in the case of a deviation between the checksum of its own output dataset and a checksum of the checksums read in, such an output dataset that matches the at least one further output dataset read in or of which the checksum matches at least a further checksum is output by the first data processing apparatus (26).

13. System for carrying out the data processing method according to one of the preceding claims, having - a network (10), which is configured to transmit data between various network subscribers of the network in accordance with a communication cycle (14) consisting of consecutive time intervals (16, 18, 20, 22, 24) defined in a communication scheme (12); - a plurality of data processing apparatuses (26, 28), which are each embodied as a network subscriber and which are each configured to read in, to process and / or to output at least a part of the data in a process cycle (34, 36) consisting of consecutive process steps (38, 40, 42, 44, 46, 48, 50, 52); in which - predefined network subscribers of the network (10) are each able to be authorised to send or to receive the data during one or more of predefined time intervals (16, 18, 20, 22, 24); - the communication cycle (14) is synchronised with the process cycle (34, 36) of at least one of the plurality of data processing apparatuses (26, 28) in such a way that, during a time in which the communication cycle (14) is executing, a predefined process step (38, 40, 42, 44, 46, 48, 50, 52) of the process cycle (34, 36) of the at least one of the plurality of data processing apparatuses (26, 28) is able to be carried out during a time interval (16, 18, 20, 22) of the communication cycle (14) assigned to this predefined process steps (38, 40, 42, 44, 46, 48, 50, 52); and - there is provision for the communication cycle (14) and the process cycle (34, 36) of the said at least one of the plurality of data processing apparatuses (26, 28) to be executed multiple times.

14. Computer program product, containing machine-readable instructions, which cause the system according to claim 13 to carry out a data processing method according to one of claims 1 to 12.