Synchronized data network system and method for initializing and synchronizing the same

Description

[0001] The invention relates to a synchronized data network system, in particular a data network system used as an avionics system data network for deterministic and potentially flight safety-relevant communication between the data network participants. Furthermore, the invention relates to a method for initializing and synchronizing the data network system.

[0002] When data network components are used in an avionics system, there is often a requirement to enable or facilitate deterministic communication between participants in the avionics system data network, as exemplified in US 2006 / 168612 A1. This is generally a stringent requirement, for example, when communication is a critical part of the system's function with direct or indirect implications for flight safety. Determinism is an abstract concept and is described in document DO-297 "Integrated Modular Avionics (IMA) Development Guidance and Certification Considerations," November 8, 2005, RTCA SC-200, EUROCAE WG-6, as "the ability to produce a predictable outcome based on previous operations, with the outcome occurring repeatably within a specified time window."

[0003] One approach to achieving deterministic behavior in a data network implementation is through absolute control of network traffic in time and space. This allows for a bandwidth guarantee for communication between any two participants in the network during a precisely defined time period. This is described in US 2018 / 062764 A1.

[0004] A key component in achieving this is ensuring that every data network node in a data network system, connected to other data network nodes via direct data transmission links, has a common understanding of the current system time with a defined minimum acceptable deviation in precision. For a deterministic data network system to function, it is therefore essential that all components or data network nodes within the network are synchronized and share a common time base. This is also demonstrated in WO 2021 / 119675 A2.

[0005] This common time can be achieved through various time synchronization tools, such as industry-standard methods like the Precision Time Protocol (PTP) and its corresponding developments for modern time-sensitive networks.

[0006] Industry-standard time synchronization (hereafter simply synchronization) methods are optimized for use in complex, dynamically changing, open networks and require sophisticated procedures to establish a common time base among all participants. In contrast, avionics network systems are often pre-built and, moreover, are closed systems to limit complexity. Under these circumstances, it is therefore necessary to find efficient approaches to synchronization and simpler means of verifying deterministic behavior.

[0007] In contrast to established avionics system data network standards (such as ARINC-664, Part 7, AFDX) or recently developed standards (such as CAIN on TSN), it is necessary to achieve significantly lower complexity for the network components in order to reduce costs, simplify the network computing capacity used for synchronization, reduce the time required to achieve security certification, and increase efficiency.

[0008] The invention is therefore based on the objective of creating a synchronized data network system and a method for initializing and synchronizing it, which can achieve a lower complexity of network components, a reduction in costs, a significant simplification of the computing capacity used for synchronization, a reduction in the time required for security certification, and an increase in efficiency.

[0009] This problem is solved by the synchronized data network system according to claims 1 and 15 and by the methods according to claims 13 and 14. Advantageous embodiments and further developments of the invention are set forth in the dependent claims.

[0010] According to the invention, a synchronized data network system is provided, which comprises a data network that connects data network nodes to each other via direct data transmission links between adjacent data network nodes for data exchange. Among these data network nodes, a chain of nodes consisting of a successive first, second, and third data network node is provided according to the invention, wherein the second data network node is adapted to receive data sent by the first data network node on a first direct data transmission link between the first and second data network nodes and to forward this data to the third data network node on a second direct data transmission link between the second and third data network nodes.The second data network node is adapted to delay the forwarding of data on a one-hop data forwarding path between the first and third data network nodes in such a way that the hop latency is equal to a predetermined fixed common hop latency.

[0011] A synchronized data network system is therefore envisaged, in which the data network nodes participating in the network are designed in such a way that when data is forwarded from a sending data network node to a receiving data network node, an intermediate forwarding data network node delays the forwarding of the data so that the duration of the hop or the hop latency is no longer dependent on network-specific or node-specific latencies, but is adjusted or leveled for each forwarding node to a common, fixed hop latency that is the same for all.

[0012] Although this delay slightly increases response times or transmission latencies, the data throughput remains the same, and at the same time the synchronization of the entire network is massively simplified, since synchronization data packets, which are usually normal data packets with appropriate header information, allow the slave clocks of the slave data network nodes to be synchronized to the master clock in the simplest way by means of synchronization data packets, which are usually normal data packets with appropriate header information, sent from a master data network node, by multiplying the timestamp of the sent synchronization packet upon arrival at a slave data network node by the number of hops multiplied by the fixed common hop latency.

[0013] The synchronization of a data network system described below can also be used in a decentralized data network if several network components have reliable master clocks. Thus, the invention provides a highly redundant method for synchronizing a data network system for deterministic communication among the network participants.

[0014] In the case of implied synchronization of the data network system by means of a master data network node which has a master clock, it is preferred if the master data network node sends synchronization data packets with a master clock timestamp t tx to slave data network nodes and the slave clocks of the slave data network nodes are synchronized with the master clock time according to the number of hops h of the synchronization data packet completed between the master data network node and the respective slave data network node, and by means of the common hop latency d CH according to the formula t master = t tx + d CH * h.

[0015] To level out latency across a one-hop data link, there are two ways to define a one-hop data link. For example, one possibility is that the one-hop data link corresponds to a data link that starts at the sending end of the first data network node, runs along the direct data link between the first and second data network nodes, passes through the second data network node, and ends at the sending end of the second data network node.On the other hand, it is also possible to define the one-hop data forwarding path in such a way that the one-hop data forwarding path corresponds to a data forwarding path that starts at the receiver side of the second data network node, runs through the second data network node and via the direct data transmission path between the second and third data network nodes, ending at the receiver side of the third data network node.

[0016] For a practical implementation of the synchronized data network system, it is advantageous if the second data network node has a data transmission unit connected to a direct data transmission link to send data to other data network nodes; a data reception unit connected to a direct data transmission link to receive data from other data network nodes; and a latency adjustment unit adapted to calculate a latency adjustment period specific to the current one-hop data transmission link using a stored specific hop latency of the current one-hop data transmission link and the predetermined fixed common hop latency period.and includes a delay buffer that is interposed between the data receiving unit and the data sending unit to delay the forwarding of data packets on the current one-hop data forwarding path by the specific latency matching duration received from the latency matching unit.

[0017] It is advantageous if the latency matching unit includes a timer for measuring a loop latency on a one-hop data link to be measured; a memory for storing the intrinsic node latency, the loop latencies of the one-hop data links of the second data network node, and the predetermined fixed common hop latency; and a computing unit adapted to calculate the latency matching duration using the formula d ARX = d CH - d RT / 2- d SRX_B.

[0018] For the actual implementation of an initialization and synchronization of the data network system of the invention, it is advantageous if the second data network node comprises a communication control unit for controlling an initialization and synchronization, which includes an initialization control unit that controls the initialization process of the data network nodes; and a synchronization control unit that controls the synchronization of the data network nodes.

[0019] It is advantageous if the initialization control unit causes the second data network node, acting as a slave data network node, to perform the following steps: Sending, from the communication control unit, a loop latency measurement command to the data transmission unit; sending, by the data transmission unit, a loop signal or ping signal via a suitable direct data transmission link of a one-hop data forwarding link to be measured to a data reception unit of a suitable receiving data network node; starting a timer upon sending the loop signal by the latency adjustment unit; receiving, by the data reception unit, the loop signal sent back by the suitable data network node via the suitable direct data transmission link; stopping the timer by the latency adjustment unit upon receiving the returned loop signal; reading the loop latency from the timer;Calculate the specific hop latency of the one-hop data link to be measured by adding the intrinsic node latency to half the round-trip latency; repeat the steps for all one-hop data links to be measured; and store the determined hop latencies of all one-hop data links to be measured.

[0020] Optionally, if during initialization the common hop latency is determined by a master data network node from all determined hop latencies of all one-hop data forwarding paths to be measured, a further step of sending the determined hop latencies of all one-hop data forwarding paths to be measured to the master data network node can take place.

[0021] If the second data network node is to function as the master data network node, it is advantageous for the initialization control unit to cause the second data network node, as the master data network node, to perform the following steps: receiving, from the slave data network nodes, the respective determined hop latencies of the one-hop data forwarding paths to be measured for all slave data network nodes; and setting a common hop latency from the transmitted hop latencies, whereby the predetermined fixed common hop latency is greater than all transmitted hop latencies.

[0022] To ensure that the fixed common hop latency is not chosen to be unnecessarily large, it is advisable for the fixed common hop latency to correspond to a time duration equal to a hop latency with the longest duration scaled with a safety factor greater than 1 or with a safety offset.

[0023] When synchronizing the data network system by the second data network node, which acts as a slave data network node, it is advantageous if the synchronization control unit causes the second data network node, as a slave data network node, to perform the following steps: receiving synchronization data packets with a master clock timestamp t tx from a master data network node with a master clock; and synchronizing the slave clock of the second data network node with the master clock time according to the number of hops h completed by the synchronization data packet between the master data network node and the second data network node, and using the common deterministic hop latency d CH according to the formula t master = t tx + d CH * h.

[0024] If the second data network node is to function as the master data network node, it is advantageous if the synchronization control unit instructs the second data network node, as the master data network node, to send synchronization data packets with a master clock timestamp t tx to the slave data network nodes at a time t tx.

[0025] To easily provide the slave data network nodes with the number of hops along with the master clock timestamp, and to ensure that synchronization occurs due to a legitimate synchronization of a master's data packet, it is highly advantageous if the synchronization data packet contains, in addition to the master clock timestamp, the number of hops and an identifier of the master data network node.

[0026] The synchronized data network system according to the invention offers significant advantages over the prior art when used as an avionics system data network for deterministic and / or flight safety-relevant communication between the data network nodes. It is advantageous for the data network to have a ring, star, or net structure with bidirectional data exchange.

[0027] Furthermore, according to the invention, a method for initializing the synchronized data network system of the invention is provided, comprising the following steps: sending a loop latency measurement command from a communication control unit of a slave data network node to the data transmission unit of the slave data network node; sending a loop signal or ping signal by the data transmission unit of the slave data network node via a corresponding direct data transmission link of a one-hop data forwarding link to be measured to a data reception unit of a corresponding further slave data network node; starting a timer of the slave data network node when the loop signal is sent by a latency adjustment unit of the slave data network node;Receiving, by the data receiving unit of the slave data network node, the loop signal sent back by the corresponding other slave data network node via the corresponding direct data transmission path; stopping the timer by the latency matching unit of the slave data network node upon receipt of the returned loop signal; reading the loop latency from the timer of the slave data network node; calculating the specific hop latency of the one-hop data transmission path to be measured by adding the intrinsic node latency to half the loop latency; repeating the steps for all one-hop data transmission paths to be measured; and storing the determined hop latencies of all one-hop data transmission paths to be measured.

[0028] Optionally, if the common hop latency is determined during initialization by a master data network node from all determined hop latencies of all single-hop data forwarding links to be measured, the following further steps can be performed: Sending the determined hop latencies of all single-hop data forwarding links to be measured to the master data network node; Receiving, by the master data network node, the respective determined hop latencies of the single-hop data forwarding links to be measured from all slave data network nodes; Determining, by the master data network node, a common hop latency from the transmitted hop latencies, whereby the predetermined fixed common hop latency is greater than all transmitted hop latencies.

[0029] Furthermore, according to the invention, a method for synchronizing a synchronized data network system of the invention with at least one master data network node and slave data network node is provided, comprising the following steps: sending, by the master data network node, at a time t tx, synchronization data packets with a master clock timestamp t tx to the slave data network nodes; receiving, by the slave data network nodes, from the master data network node with a master clock timestamp t tx; and synchronizing the slave clock of the slave data network node with the master clock time according to the number of hops h completed by the synchronization data packet between the master data network node and the slave data network node, and by means of the common deterministic hop latency d CH according to the formula t master = t tx + d CH * h.

[0030] According to the invention, a synchronized data network system, a master data network node and a slave data network node are also provided, each adapted to carry out the above methods according to the invention.

[0031] The invention will be explained in more detail below, for example, with reference to the drawing. The drawing shows: Fig. 1 a highly simplified schematic view of a synchronized data network system in a ring structure according to the invention, Fig. 2 a chain of nodes consisting of a first, second and third data network node, which are connected to each other by a first and second direct data transmission path, Fig. 3A a time-lapse diagram of the delay in forwarding the data through the second data network node according to a first embodiment of the invention, Fig. 3B a time-lapse diagram of the delay of the data by the second data network node according to a second embodiment of the invention, Fig. 4A a chain of nodes consisting of a first, second, third and fourth data network node, illustrating the delay of data forwarding for the purpose of synchronization according to the first embodiment of the invention, Fig. 4B a chain of nodes consisting of a second, third, fourth and fifth data network node, which illustrates the timing of the data forwarding delay for the purpose of synchronization according to the second embodiment of the invention, Fig. 5 a highly simplified schematic block diagram of a data network node according to the invention, Fig. 6A and 6B Components of the data network nodes and their role in the initialization of the synchronized data network system according to the invention, Fig. 7 a process flow diagram of the method for initializing the synchronized data network system according to the invention, and Fig. 8 a process flow diagram of a method for synchronizing a synchronized data network system according to the invention.

[0032] In the various figures of the drawing, corresponding building elements are provided with the same reference symbols.

[0033] In Fig. 1 Figure 1 shows an embodiment of a synchronized data network system 10 according to the invention. The synchronized data network system 10 comprises a data network 100 that connects data network nodes 110 to each other via direct data transmission links x between adjacent data network nodes 110 for the purpose of data exchange. In the embodiment of the invention shown according to Figure 10, the data network nodes 110 are connected via direct data transmission links x between adjacent data network nodes 110 for the purpose of data exchange. Fig. 1 The data network nodes 110A, 110B, 110C, and 110D are therefore interconnected in a ring structure via the direct data transmission paths xAB, xBC, xCD, and xAD. Although the data network system 10 in Fig. 1 Although the data network system 10 has a ring structure, it is also possible to provide a data network 100 with a star or net structure. The invention can be implemented on any network topology and is therefore not limited to a specific network structure. The topology of the network structure is thus free; the only requirements are the bidirectionality of the connections between the nodes and that at least one node must have more than one interface to neighboring nodes in order to create a topology with more than two participants. The number of data network nodes 110 is also arbitrary. Fig. 1 The four data network nodes shown, 110A, 110B, 110C and 110D, are to be understood as purely an example.

[0034] In the data network system 10, 110 prefabricated avionics components are preferably provided as data network nodes to establish an avionics system data network. This network enables deterministic and potentially flight safety-relevant communication between the avionics components acting as data network nodes. Such deterministic communication is essential, particularly in flight networks or when using autonomous flight or driving applications, as synchronized communication and the traceability of causal chains within an avionics system are required, especially for flight safety verification and, of course, for flight safety itself. These systems are used when it comes to saving human lives, fulfilling military missions, or transporting valuable assets.Avionics components that can be used as data network nodes 110 according to the invention include a wide variety of components, such as computer systems for flight control or mission management, human-machine interfaces like display or input devices, attitude sensors, engine sensors and actuators, control surface sensors and actuators, or weapon systems in which weapons can be triggered or decoys deployed via the synchronized data network system. Although use in avionics is particularly preferred, the use of the synchronized data network system according to the invention is not limited to this application, but can also be used, for example, in the fields of navigation, space travel (such as satellite systems), or in large, security-relevant facilities.

[0035] In the data network system 10 according to the invention, the direct data transmission path x between two data network nodes 110 can enable bidirectional or dual-simplex communication. The temporal behavior of data sent through the network 100 is mainly characterized by a unidirectional hop latency dH and a combined round-trip latency dRT, which for each link depends on external factors such as the physical medium or environmental conditions and internal factors such as the design or implementation of the data network node 110.

[0036] Due to the variation in hop latencies d H across the data network system 10, the runtime for transmitting data packets from a source data network node 110 to a destination data network node 110 is not a direct function of the data routing path chosen by the data network 100 and the number of hops on that data routing path. Rather, it is necessary to consider all specific latencies arising from external or internal network factors, which entails an enormous computational effort if the data network system 10 is to be implicitly synchronized. Here, implicit synchronization means synchronizing all network node secondary clocks to a master clock of a master data network node 110M ( Fig. 5 ) by sending synchronization data packets from the master data network node 110M to the data network nodes 110.

[0037] Furthermore, such synchronization is not only computationally intensive, but also dependent on runtime initialization and thus also affects the determinism of the system, as it requires tolerances across the entire runtime-dependent range of values.

[0038] In Fig. 2 The synchronized data network system 10 according to the invention is shown in further detail, with a more detailed discussion of the various latencies that are generated either by the data network nodes 110 or by the direct data transmission links x. Fig. 2 Figure 1 shows a chain of data network nodes 110 according to the invention, consisting of a successive first, second, and third data network node 110A, 110B, and 110C, which are interconnected via direct data transmission paths xAB and xBC when data is transmitted from data network node 110A to data network node 110C. Bidirectional communication is also possible via the direct transmission paths xCB and xBA; however, for the sake of simplicity, only the forwarding direction from node 110A to 110C will be considered.

[0039] According to the invention, the second data network node 110B is adapted to receive data sent on a data forwarding path x ABBC from the first data network node 110A on the first direct data transmission path x AB between the first and the second data network node 110A, 110B and to forward it on a second direct data transmission path x BC between the second and the third data network node 110B, 110C to the third data network node 110C.

[0040] According to the invention, a distinction is made between the direct data transmission path x, which is part of the data network 100, and a data forwarding path x ABBC, which also includes the transmission path through a data network node 110 with a corresponding intrinsic node latency.

[0041] In particular, the one-hop data forwarding paths x AAB, x ABB, x BBC, x BCC, or hereinafter also more generally x XZZ or x ZZY, are of interest for the invention, since they are the smallest unit of a data forwarding hop from one data network node 110 to another data network node 110. There are two possible approaches to defining a one-hop data forwarding path: (a) The one-hop data forwarding path can be defined as corresponding to the data forwarding paths x ABB and x BCC. Here, the direct data transmission path x AB from the sending data network node 110A to the forwarding data network node 110B, as well as the data forwarding path through data network node 110B, are combined as the one-hop data forwarding path. If data network node 110C is the forwarding data network node 110, the direct data transmission path x BC and the data forwarding path through data network node 110C are combined as the one-hop data forwarding path x BCC.Therefore, if the second data network node 110B is the forwarding data network node 110, the one-hop data forwarding path corresponds to a data forwarding path x ABB, which starts at the sender side of the first data network node 110A, runs via the direct data transmission path x AB between the first and the second data network node 110A, 110B, through the second data network node 110B, and ends at the sender side of the second data network node 110B. (b) According to a second embodiment, the one-hop data forwarding link x AAB , x BBC consists of the forwarding link with corresponding latency of the forwarding node 110 and the subsequent direct data transmission link x AB , x BC, over which the data is sent from the forwarding data network node 110A, 110B to a corresponding further data network node 110B, 110C.This means that if the second data network node 110B is the data-forwarding data network node, the one-hop data forwarding path corresponds to the data forwarding path x BBC, which starts at the receiver side of the second data network node 110B, passes through the second data network node 110B, and then via the direct data transmission path x BC between the second and third data network nodes 110B and 110C, ending at the receiver side of the third data network node 110C.

[0042] As from Fig. 2 As can be seen further, the corresponding forwarding paths through the data network node 110 and through the direct data transmission paths x are assigned corresponding latencies, which either relate to the runtime or latency of data forwarding through the data network node 110 d SRX_A , d SRX_B , d SRX_C or to the runtime or latency of forwarding via the direct data transmission paths x AB and x BC with the latencies d AB and d BC.

[0043] Regarding the one-hop data forwarding paths, the sum of the intrinsic node latency d SRX and the transmission latency over the direct data forwarding paths x yields the hop latencies d H_AAB, d H_ABB, d H_BBC, and d H_BCC. Although the hop latencies in Fig. 2 Since all are represented as being of the same size, the problem arises in a previously known data network system 10 precisely here, that due to the variance of the intrinsic node latencies and the transmission latencies, an implicit synchronization of the data network system 10 becomes very complex.

[0044] However, this problem is overcome by the synchronized data network system 10 according to the invention, as for example in the Fig. 3A und 3B is illustrated. Fig. 3A This shows a detailed view from Fig. 2 , where a delay in data transmission takes place according to a first embodiment, wherein the in Fig. 3B shown excerpt of the Fig. 2 represents a delay in data transmission according to the second embodiment of the invention.

[0045] As in Fig. 3A As shown, the second data network node 110B is adapted to delay the forwarding of data on a one-hop data forwarding path x ABB between the first and third data network nodes 110A, 110C such that the hop latency d H_ABB is equal to a predetermined fixed common hop latency d CH. Here, the one-hop data forwarding path corresponds to a data forwarding path x ABB that starts at the sender side of the first data network node 110A, runs via the direct data transmission path x AB between the first and second data network nodes 110A, 110B, through the second data network node 110B, and ends at the sender side of the second data network node 110B.

[0046] In Fig. 3B The second data network node 110B is configured to delay the forwarding of data on a one-hop data forwarding path x BBC between the first and third data network nodes 110A, 110C such that the hop latency d H_BBC is equal to a predetermined fixed common hop latency d CH. The one-hop data forwarding path corresponds to a data forwarding path x BBC that starts at the receiver side of the second data network node 110B, passes through the second data network node 110B, and then continues via the direct data transmission path x BC between the second and third data network nodes 110B, 110C, ending at the receiver side of the third data network node 110C.

[0047] Even though, for the sake of completeness, both embodiments (a) ( Fig. 3A ) and (b) ( Fig. 3B Since the second embodiment (b) is to be treated as follows, the first embodiment (a) is preferred, as the practical implementation of the second embodiment (b) is somewhat more difficult. In particular, the second embodiment requires in Fig. 3B Outgoing locally generated packets in the RX path of the second data network node 110B are created, time-stamped, and delayed towards the TX send link of the second data network node 110B. This also delays data transfers between neighboring node 110 that would not otherwise need to be delayed (see also the paths IDP -> RDP_loc and IDP -> RDP_fwd_AC in [documentation missing]). Fig. 5 ).

[0048] According to the invention, the intrinsic node latency d SRX_B is extended by a variable node latency d AX_ABB or d AX_BBC by means of a delay circuit in a forwarding data network node 110 such that, regardless of the data transmission path latencies d AB or d BC and regardless of the intrinsic node latency d SRX_B, the forwarding time or forwarding latency or one-hop data forwarding latency d H_ABB or d H_BBC during a forwarding hop through the second data network node 110B is always equal to a predetermined common hop latency d CH, no matter which data network node 110 or which associated direct data transmission path x of a selected one-hop data forwarding path is chosen in the synchronized data network system 10 according to the invention.

[0049] The delay circuit in the second network node 110B and all other forwarding data network nodes 110 does delay data transmission in the data network system 10 according to the invention; however, this approach greatly simplifies the synchronization of the entire data network system 10. This is evident in the Fig. 4A for the first embodiment of the invention and in Fig. 4B illustrated for the second embodiment of the invention.

[0050] Thus, synchronization of the data network system 10 according to the invention can, for example, be achieved by means of a master data network node 110M ( Fig. 5 ) corresponding slave data network node 110S of data network system 10 to a master clock CLK_M ( Fig. 5 ) are synchronized. For this purpose, the master data network node 110M is configured to periodically send synchronization data packets SDP with a master clock timestamp t tx to slave data network node 110S, or upon instruction from a user, and to synchronize the slave clocks CLK_S of the slave data network node 110S according to the number of hops h completed by the synchronization data packet SDP between the master data network node 110M and the respective slave data network node 110S, and using the common hop latency d CH according to the formula t master = t tx + d CH * h with the master clock of the master data network node 110M.

[0051] Synchronization using this simple formula above is made possible by the fact that, as in the Fig. 4A and 4B shown, a forwarding of a synchronization data packet SDP from the data network nodes 110A to 110D ( Fig. 4A ) or from data network nodes 110B to 110E ( Fig. 4B The transmission time from the master data network node 110M to a slave data network node 110S to be synchronized is a clearly defined time duration, namely the number of hops from the master data network node 110M to the respective slave data network node 110S multiplied by the predetermined fixed common hop latency dCH. The data transmission paths xAB, xBC, xCD, xDE can be of various types, including, for example, a wired or cable connection, a fiber optic connection, or a wireless radio connection for a direct data transmission path x. Furthermore, the intrinsic node latencies dSRX can also differ, since during data forwarding, the hop latency is always adjusted or leveled to a common predetermined fixed hop latency dCH, regardless of the data network node 110 being chosen.

[0052] It is emphasized that although the synchronization of the data network system 10 according to the invention using a master data network node 110M is preferred, synchronization can also take place in a decentralized data network system if several network nodes are equipped with clocks suitable as master clocks. It is also conceivable, for example, that not only one master data network node 110M is provided, but that a plurality of the data network nodes 110 or even all data network nodes 110 can operate as master data network nodes 110M, thereby maximizing the reliability and redundancy of the synchronized data network system 10 according to the invention. In this case, only appropriate coordination between the data network nodes 110 is necessary to ensure that no two data network nodes 110 perform the master role at the same time.

[0053] Using the synchronization method according to the invention, it is therefore possible to achieve a simple synchronization of all data network nodes by means of a common hop latency for all connections in a closed network, which is the same for all data network nodes and corresponding forwarding paths.

[0054] In Fig. 5 Figure 1 shows a block diagram of a data network node 110 according to the invention. Here, for example, the second data network node 110B comprises a data transmission unit TX, which is connected to a direct data transmission link x to send ODP data to other data network nodes 110, which are either locally generated data packets to be sent TDP_loc or forwarded data packets to be sent RDP_fwd. The second data network node 110B further comprises a data reception unit RX, which is connected to a direct data transmission link x to receive IDP data from other data network nodes 110. As shown in Figure 1, the data transmission unit TX is connected to a direct data transmission link x to receive IDP data from other data network nodes 110. Fig. 5 As shown, the second data network node 110B also has a latency matching unit (LMU) that is adapted to calculate a latency matching duration d ARX specific to the current one-hop data forwarding path x ABB, x BBC using a stored specific hop latency d H_ABB, d H_BBC of the current one-hop data forwarding path x ABB, x BBC and using the predetermined fixed common hop latency d CH. In the case of forwarding data packets IDP from the data receiving unit RX via the data transmitting unit TX to further data network nodes 110, the data packet RDP_fwd (e.g.RDP_fwd_AC or RDP_fwd_CA) by means of a delay buffer CDB interposed between the data receiving unit RX and a data sending unit TX intended for forwarding, in order to delay the forwarding of data packets RDP_fwd_AC on the current one-hop data forwarding path x ABB , x BBC by the specific latency matching duration d ARX received from the data latency matching unit LMU.

[0055] The structure of the LMU latency adjustment unit is described in Fig. 6A The device shown includes a timer for measuring a loop latency dRT on a one-hop data link xABB, xBBC to be measured, as well as a memory for storing the intrinsic node latency dSRX_B, the loop latencies dRT of the one-hop data links xABB, xBBC of the second data network node 110B, and the predetermined fixed common hop latency dCH. Furthermore, the latency adjustment unit LMU includes a processing unit adapted to calculate the latency adjustment period dARX_ABB specific to a one-hop data link using the formula dARX = dCH - dRT / 2 - dSRX_B.

[0056] As in Fig. 5 As shown, a delay of the data packets only occurs when they are intended for forwarding, whereas received RDP_loc data packets that are not synchronization data packets and are intended for local processing in the corresponding second data network node 110B are not delayed. Thus, the configurable delay of data forwarding, preferably according to the first embodiment, is provided on the receiver side of each node, exclusively for packets that have not reached their destination address and which are forwarded to the data transmitting unit TX for forwarding from the data receiving unit to the data transmitting unit.In this way, packets that are processed locally can be sent to the end user promptly and directly without further delay, in contrast to an implementation according to the second embodiment of the invention, in which a delay occurs on the sender side for the forwarding path to be overcome until the next data network node 110.

[0057] As in Fig. 5 As further shown, the second data network node 110B has either a master clock CLK_M or a slave clock CLK_S, depending on whether the second data network node 110B operates as a master data network node 110M or as a slave data network node 110S. If the second data network node 110B operates as a slave data network node 110S, the slave clock CLK_S is synchronized with the master clock CLK_M of another master data network node 110M by a synchronization data packet SDP, either without delay after receipt according to the second embodiment, or after a delay due to the delay buffers CDB according to the first embodiment.The second data network node 110B also has a communication control unit LCU / CONFIG for controlling initialization and synchronization, wherein the communication control unit LCU / CONFIG includes an initialization control unit LCU, which controls the initialization process of the data network node 110, and a synchronization control unit CONFIG, which controls the synchronization of the data network node 110.

[0058] In the following, an initialization process of the synchronized data network system according to the invention and then a synchronization process of the data network 10 according to the invention will be described.

[0059] During network initialization, all nodes in the network are prompted to independently determine the latencies on all data transmission paths to their neighboring nodes. The data network-node interface implementation must be symmetrical or with a known asymmetry that is taken into account by the neighboring nodes. This allows the hop latency to be correctly derived from a loop latency measurement.Before network initialization is complete, each node is instructed to insert an additional delay or specific latency adjustment period d ARX into the data receive path in conjunction with all data transmission paths and connections to the neighboring node, so that the resulting hop latency on all these connection paths corresponds to a configurable common value for the entire network. This value is determined from all single-hop data forwarding latencies before the network is operational. This value is the predetermined fixed common hop latency d CH, or common deterministic hop latency d CH.

[0060] It should be emphasized that the network initialization process is entirely decentralized, and the distinction between slave data network node 110S and master data network node 110M is initially irrelevant, as the future master data network node 110M also initially undergoes the same initialization process as all other data network nodes 110, functioning as slave data network node 110S. During this process, all nodes determine the round-trip latency to their neighbors as part of the network initialization (generating requests, receiving responses, and receiving requests and generating responses). There are no roles involved, and all nodes are equal. Once each node has determined the necessary information, the target latency is set, and the initialization phase ends.The uniform hop latency then makes implicit time synchronization possible simply by the defined master communicating regularly with all other nodes (unidirectionally, no response is needed).

[0061] In detail, the initialization from the perspective of the second data network node 110B in its role as slave data network node 110S, for example during a probe of the one-hop data forwarding path to its neighboring node 110A, proceeds as follows, as shown in the Fig. 6A and 6B is illustrated.

[0062] During the initialization process of the synchronized data network system 10, the initialization control unit LCU instructs the second data network node 110B, acting as a slave data network node 110S, to perform the following steps. First, the communication control unit LCU sends a round-trip measurement request (RTMR) to the data transmitter unit TX. The data transmitter unit TX then sends a round-trip signal or a ping signal via a corresponding direct data transmission path x, in this case x AB, the one-hop data link x ABB to be measured (for example, in the case of an implementation of the first embodiment of the invention), to a data receiver unit RX of a corresponding receiving data network node 110A. Simultaneously, the timer of the latency adjustment unit LMU is started by the LMU when the round-trip signal is sent.The loop signal or ping signal transmitted to data network node 110A is received by the data receiving unit RX of data network node 110A and processed using a measurement request response signal (MRR). Fig. 6B The data signal is sent back directly to the receiver unit RX of the second data network node 110B via the data transmission unit TX of data network node 110A. As soon as the data receiver unit RX has received the loop signal sent back from the corresponding first data network node 110A via the corresponding direct data transmission path x AB, the timer is stopped by the latency adjustment unit LMU upon receipt of this returned loop signal.

[0063] After reading the loop latency d RT from the timer of the latency adjustment unit LMU, the latency adjustment unit LMU calculates the specific hop latency d H_ABB of the example of Fig. 6A and Fig. 6B The one-hop data transmission path x ABB to be measured is determined by adding the intrinsic node latency d SRX_B to half the round-trip latency d RT. The intrinsic node latency d SRX_B can be determined by the second data network node 110B itself by internally measuring the forwarding latency of data packets from the data receiving unit RX to the data transmitting unit TX. This intrinsic node latency value can also be measured during the manufacturing of the participating data network node and stored in the memory of the latency matching unit LMU.

[0064] Since the second data network node 110B must measure all one-hop data forwarding paths x XBB , x BBY according to the first or second embodiment for the complete initialization of the initialized data network system 10, the initialization process described above is repeated until the second data network node 110B has measured all one-hop data forwarding paths to data network nodes 110 directly connected to it.

[0065] In the event that a "worst case" hop latency is determined during the verification phase of the platform, this will then be made known to each data network node 110 as a predetermined common hop latency d CH as a configuration.

[0066] However, it is also possible for the network to dynamically determine the common hop latency d CH itself in order to optimize the overall latency. In this case, this can be done, for example, via a master data network node 110M, but an external reader can also be used, which is connected to the network once during initialization to read all determined hop latencies d H_XZZ , d H_ZZY of all single-hop data forwarding paths x XZZ , x ZZY of all data network nodes 110 to be measured.

[0067] However, there is generally no exchange of the determined hop latencies between the nodes, and there are no master / slave roles in this regard. Each node is solely responsible for determining the round-trip latency to its neighbor and calculating the necessary delay to adjust the routing paths through the local node to a target latency established in advance within the system.

[0068] After the complete measurement of all one-hop data forwarding links and the associated hop latencies d H_XBB , d H_BBY of all one-hop data forwarding links to be measured x XBB , x BBY, the determined hop latencies can be sent to the master data network node 110M, in case the master data network node 110M is intended for determining the common hop latency d CH.

[0069] Therefore, if the second data network node 110B operates as such a master data network node 110M, the initialization unit LCU instructs the second data network node 110B, as the master data network node 110M, to perform the following steps. The second data network node 110B receives from the slave data network nodes 110S the respective determined hop latencies d H_XZZ, d H_ZZY of the one-hop data transmission paths x XZZ, x ZZY of all slave data network nodes 110S to be measured. Once the master data network node 110M has all hop latencies of all one-hop data forwarding links to be measured of all slave data network nodes, the master data network node 110M determines a common hop latency d CH from the transmitted hop latencies.The master data network node 110M is relatively free in the way it is defined; the only condition is that the predetermined fixed common hop latency d CH is equal to or greater than the largest of all transmitted hop latencies d H_XZZ , d H_ZZY, since otherwise at least one transmitted hop latency would be greater than the fixed common hop latency d CH and thus synchronization in the sense of the invention would not be possible.

[0070] Besides the fact that the common hop latency dCH must be greater than all hop latencies of all single-hop data forwarding links in the network 10, it is also advantageous if the time difference between the defined common hop latency dCH and the longest hop latency transmitted by the slave data network node 110S is chosen to be as small as possible, in order to avoid unnecessarily delaying data traffic within the synchronized data network system 10 due to an unnecessarily large common hop latency dCH. Thus, a typical hop latency in an avionics data network system 10 according to the invention is in the range of 350 ns, where, for example, a defined common hop latency dCH of 500 ns would be an acceptable value for the common hop latency. It should be emphasized that the figures mentioned only represent orders of magnitude, as only a small part of the latency is caused by cable lengths.The majority of the latency is introduced by the transceivers and by the implementation of the network interface of the nodes.

[0071] In a practical implementation of determining the fixed common hop latency d CH, the fixed common hop latency d CH can correspond to a time duration equal to the longest hop latency d H_XZZ , d H_ZZY scaled with a safety factor greater than 1 or with a safety offset. For example, the safety factor can be a value greater than 1 and less than 1.5, less than 1.4, less than 1.3, less than 1.2, less than 1.1, or less than 1.05. The safety offset can be in a range between 100 ns and 200 ns. It should be noted here that, as described above, the cable lengths on the data transmission paths x account for only a small proportion of the total latency during data transmission in network 10. The required area depends on the runtime variations (minimum) and the resources used for delay buffer (maximum).In any case, the additive latency is typically low compared to the latencies already introduced by the transceivers and cable lengths. It is important to note that for short transmission distances (e.g., under 10m), the cable length has a much smaller impact than the interface architecture and the transceivers within the nodes themselves. For example, in a prototype implementation, total latencies of approximately 300ns were measured for a transmission link with 3m connecting cables, of which approximately 5ns / m was attributable to the fiber optic cable.

[0072] In any case, the offset should be set so that if the system is modified or an additional cable run of 10 cm to 50 cm is added, the system does not need to be completely reconfigured and initialized. The larger the offset, the more robust the data network system is when the components' latencies change. It should also be emphasized here that discrete values ​​are not very useful. The larger the safety offset, the greater the permissible changes in cable length, provided the runtime variation is known.

[0073] The safety offset therefore only defines the maximum compensation potential, which doesn't mean that this maximum must always be fully utilized. It depends on the target network. Even in a system that can variably compensate for node delays of up to 1 second, it's possible to work with a maximum offset of 5 ns, since all nodes exhibit very low delays that are determined before commissioning and then set as the target latency. The intention is to prevent dynamic optimization of a network after, for example, a system change. If changes occur in the system, the common target latency must be recalculated and transmitted to all nodes.

[0074] After, as in Fig. 5 As shown, the second data network node 110B, acting as the master data network node 110M, receives all hop latencies d H_XZZ / d H_ZZY from the one-hop forwarding paths of all slave data network nodes 110S to be measured, and has determined a common hop latency d CH, the determined common hop latency d CH is transmitted to all slave data network nodes 110S, which then store the value d CH in a memory of the latency matching unit LMU after it has been transmitted to it by the communication control unit LCU / CONFIG.

[0075] Since the latency matching unit (LMU) has now stored all important latencies for forwarding data packets—namely, the intrinsic node latency d SRX_B, the common hop latency d CH, and the latency for each relevant one-hop forwarding path in which the second data network node 110B can participate—the LMU can now calculate a corresponding adjustment period d ARX specific to the one-hop forwarding path and transmit it to the delay buffer CDB, which then delays data packets RDP_fwd accordingly. The delay buffer CDB can be a standard FIFO (First-In-First-Out) buffer that temporarily stores data and reads and forwards it after a predetermined delay.Furthermore, the latency balancing unit LMU is in a data connection with the communication control unit LCU / CONFIG in order to inform the LCU / CONFIG of its latency balancing status LMS.

[0076] In the following, a synchronization process of the synchronized data network system 10 according to the invention will be described.

[0077] Since, after the initialization process, the data forwarding latency on arbitrary routes through the network is now an integer multiple of the defined common deterministic hop latency d CH, this situation can be exploited to implement an extremely simplified synchronization procedure of the local time of all nodes in the network to a local time of a specific master node with a master clock CLK_M. Synchronization occurs implicitly through a periodic transmission of packets from the timing master node to all other nodes, taking arbitrary routes through the network. These packets can carry a timestamp reflecting the local time at the timing master node at the start of the transmission (t tx timestamp or "time stamp").Furthermore, these data packets or synchronization data packets (SDP) can contain a unique identifier of the timing master node as the sender address (Node ID) and can also indicate how many hops (#hop) have been made on the current route through the network to the node that is currently receiving the packet.

[0078] Based on the synchronization data packets SDP sent by the timing master node, the current local time at the timing master node when such a synchronization data packet arrives at the node is equal to the timestamp t tx contained in the synchronization data packet SDP plus the number of hops h times the specified common deterministic hop latency d CH : T master = t tx + d CH * h .

[0079] This method according to the invention is an implicit synchronization method, whereby no specific protocol needs to be implemented to synchronize the time between a timing master and slave nodes in a network, since the synchronization can occur implicitly during any type of communication between the timing master and the slave node, provided this communication occurs periodically or within a specific time window. Thus, the SDP data packet can be dedicated to synchronizing the entire data network, but it is also conceivable that, at periodic intervals or within a specific time window, the synchronization data packet information, containing the timestamp, the number of hops, and the node ID of the master data network node 110M, is inserted into the header of normal data packets sent to the corresponding slave data network node 110S.The repetition duration of the periodic sending of the synchronization data packet SDP can be chosen in accordance with the accuracy requirements, the drift of the slave clocks over time, and other factors.

[0080] Dedicated synchronization data packets are therefore only necessary if regular communication between the timing master and the other nodes in the network does not already occur, or if communication within a specific time window does not already take place. Synchronization occurs implicitly through any data packet, as the necessary information for synchronization is embedded in each packet header. These fields are, of course, only relevant if the originating node is the timing master. A synchronization data packet (SDP) is then simply a normal packet that does not necessarily have to carry data.

[0081] Each node uses the currently valid local time as the TX timestamp (t tx) for all packets it generates. This timestamp remains unchanged when packets are forwarded by subsequent nodes; only the hop count (h) is incremented. Potentially, any node is capable of assuming the role of the timing master. This is configured in advance without affecting the nodes' sending behavior. The only crucial factor for receiving is that synchronization information is only extracted from packets originating from the timing master (TX ID = Master). This highly abstracts the time synchronization process, eliminating the need for specific control beyond predefining the roles and ensuring that the application layer schedules network traffic between the timing master and all other nodes at regular intervals, which then implicitly maintains time synchronization across the network.Synchronization information includes the combination of the TX timestamp (t tx), hop count (h), and the TX ID, all of which can be transported, for example, as part of the packet header (fields), as well as the common deterministic hop latency (dCH) stored locally at each node. Specifically, the synchronization control unit CONFIG instructs the second data network node 110B, when operating as a slave data network node 110S, to perform the following steps.

[0082] Thus, this node receives synchronization data packets SDP with a master clock timestamp t tx from a master data network node 110M with the master clock CLK_M. As in Fig. 5 As shown, after the data receiving unit RX receives the synchronization data packet, either a delay of the synchronization data packet SDP (path (a) in Fig. 5 ) if the one-hop data forwarding path is configured according to the first embodiment or as described in the Fig. 3A or4A shown, is defined. In the case that the one-hop forwarding path according to the second embodiment, i.e. as in the Fig. 3B and 4B As shown, and defined, no delay needs to occur through the delay buffer CDB (path (b) in Fig. 5 ), since the specific latency adjustment time d ARX has already been added by the preceding sending data network node 110 for the input transmission path x. The secondary clock CLK_S of the second data network node 110B or 110S is then synchronized by means of a control unit LCU / CONFIG. The secondary clock CLK_S of the second data network node 110B is thus synchronized with the main clock according to the number of hops h completed by the synchronization data packet SDP between the master data network node 110M and the second data network node 110B, and by means of the common deterministic hop latency d CH according to the formula t master = t tx + d CH * h.

[0083] In the event that the second data network node 110B operates as the master data network node 110M, the synchronization control unit CONFIG instructs it to perform the following steps: The synchronization control unit CONFIG ensures that at a time t tx, synchronization data packets SDP with a master clock timestamp t tx are sent to the slave data network node 110S, whereby the synchronization data packet SDP can contain, in addition to the master clock timestamp t tx, the number of hops h and an identifier ID_M of the master data network node 110M.

[0084] According to the invention, a method for initializing a synchronized data network system 10 is also provided, comprising the following steps, as shown in Fig. 7 As shown, in step S210, a communication control unit (LCU) of a slave data network node 110S sends a round-trip latency measurement command (RTMR) to the data transmitter (TX) of the slave data network node 110S. In step S220, the data transmitter (TX) of the slave data network node 110S sends a round-trip signal or ping signal via a corresponding direct data transmission link (x) of a one-hop data forwarding link (x XBB, x BBY) to a data receiver (RX) of a corresponding further slave data network node 110S.

[0085] In this process, a timer is started in step S230 of the slave data network node 110S when the loop signal is sent by a latency adjustment unit (LMU) of the slave data network node 110S. After receiving (step S240) the loop signal sent back by the corresponding other slave data network node 110S via the corresponding direct data transmission path x by the data receiving unit (RX) of the slave data network node 110S, the timer is stopped in step S250 by the latency adjustment unit (LMU) of the slave data network node 110S upon receipt of the returned loop signal. In step S260, the loop latency d RT is read from the timer of the slave data network node 110S, and in step S270, the specific hop latency d H_XBB , d H_BBY of the one-hop data forwarding path x XBB , x BBY to be measured is calculated by adding the intrinsic node latency d SRX_B to half the loop latency d RT.After repeating the steps (S280) for all one-hop data forwarding links x XBB , x BBY to be measured, the determined hop latencies d H_XBB , d H_BBY of all one-hop data forwarding links x XBB , x BBY to be measured are stored (S290).

[0086] Optionally, if during initialization the common hop latency from all determined hop latencies of all one-hop data forwarding links to be measured is determined by a master data network node 110M, the following further steps can be performed: In a step S2100, the determined hop latencies d H_XBB , d H_BBY of all one-hop data forwarding links to be measured x XBB , x BBY are sent to the master data network node 110M. After the master data network node 110M has received the respective determined hop latencies d H_XZZ , d H_ZZY of the one-hop data forwarding path x XZZ , x ZZY to be measured of all slave data network nodes 110S in a step S2110, it determines the common hop latency d CH from the transmitted hop latencies d H_XZZ , d H_ZZY in a step S2120, where the predetermined fixed common hop latency d CH is greater than all transmitted hop latencies d H_XZZ , d H_ZZY.

[0087] In Fig. 8 Furthermore, the method provided by the invention for synchronizing a synchronized data network system 10 with at least one master data network node 110M and slave data network node 110S is shown, which comprises the following steps:

[0088] In step S310, the master data network node 110M sends synchronization data packets (SDP) with a master clock timestamp t tx to the slave data network nodes 110S at time t tx. In step S320, the slave data network nodes 110S receive synchronization packets (SDP) with a master clock timestamp t tx, sent by the master data network node 110M with a master clock CLK_M. After receiving a synchronization data packet SDP, the slave clock CLK_S of the slave data network node 110S is synchronized in step S330 with the main clock time according to the number of hops h completed of the synchronization data packet SDP between master data network node 110M and slave data network node 110S, and by means of the common deterministic hop latency d CH according to the form t master = t tx + d CH * h.

[0089] According to the invention, a synchronized data network system 10, a master data network node 110M and a slave data network node 110S are also provided, which are adapted to perform the above initialization or synchronization methods according to the invention.

Claims

1. A synchronized data network system (10), comprising: - a data network (100) that connects data network nodes (110) to one another via direct bidirectional data transmission links (x) between neighboring data network nodes (110) for an exchange of data; - a node chain of the data network nodes (110) consisting of successive first (110A), second (110B) and third (110C) data network nodes, wherein the second data network node (110B) is adapted to receive data transmitted on a first direct data transmission link (xAB) between the first (110A) and the second (110B) data network nodes from the first data network node (110A) on a data forwarding link (xABBC) and to forward said data in transmit mode to the third data network node (110C) on a second direct data transmission link (xBC) between the second (110B) and the third (110C) data network nodes, characterized in that the second data network node (110B) is adapted to delay forwarding of the data on a single-hop data forwarding link (xABB, xBBC) between the first and the third data network nodes (110A, 110C) such that the hop latency (dH_ABB, dH_BBC) is equal to a hop latency (dCH), which is identical, predetermined, fixed and common for all data network nodes of the data network system that forward data between data network nodes of the data network system and which is not dependent on network-specific or node-specific latencies.

2. The synchronized data network system (10) according to claim 1, further comprising: - a master data network node (110M) with a master clock (CLK_M) that is adapted to transmit synchronization data packets (SDP) with a master clock time stamp ttx to slave data network nodes (110S) and to synchronize the slave clocks (CLK_S) of the slave data network nodes (110S) with the master clock time in accordance with the completed number of hops h of the synchronization data packet (SDP) between master data network nodes (110M) and the respective slave data network node (110S), as well as by means of the common hop latency period dCH according to the formula tmaster = ttx + dCH * h.

3. The synchronized data network system (10) according to claim 1 or 2, characterized in that the single-hop data forwarding link (xABB, xBBC) corresponds to a data forwarding link (xABB), which, starting on the transmitter side of the first data network node (110A), passes through the second data network node (110B) via the direct data transmission link (xAB) between the first (110A) and the second (110B) data network nodes and ends on the transmitter side of the second data network node (110B).

4. The synchronized data network system (10) according to claim 1 or 2, characterized in that the single-hop data forwarding link (xABB, xBBC) corresponds to a data forwarding link (xBBC), which, starting on the receiver side of the second data network node (110B), passes through the second data network node (110B) and ends, via the direct data transmission link (xBC) between the second (110B) and the third (110C) data network nodes, on the receiver side of the third data network node (110C).

5. The synchronized data network system (10) according to any one of the preceding claims, characterized in that the second data network node (110B) comprises: - a data transmitting unit (TX) which is connected to a direct data transmission link (x) in order to transmit data (ODP) to other data network nodes (110); - a data receiving unit (RX) which is connected to a direct data transmission link (x) in order to receive data (IDP) from other data network nodes (110); - a latency matching unit (LMU) which is adapted to calculate, by means of a stored specific hop latency (dH_ABB, dH_BBC) of the current single-hop data forwarding link (xABB, xBBC) and by means of the predetermined fixed common hop latency period (dCH), a latency matching period (dARX) specific to the current single-hop data forwarding link (xABB, xBBC); and - a delay buffer (CDB) that is interposed between the data receiving unit (RX) and the data transmitting unit (TX) to delay forwarding of data packets (RDP_fwd) on the current single-hop data forwarding link (xABB, xBBC) by the specific latency matching period (dARX) received from the latency matching unit (LMU).

6. The synchronized data network system (10) according to claim 5, characterized in that the latency matching unit (LMU) comprises: - a timer for measuring a roundtrip latency (dRT) on a single-hop data forwarding link (xABB, xBBC) to be measured; - a memory for storing the intrinsic node latency (dSRX_B), the roundtrip latencies (dRT) of the single-hop data forwarding links (xABB, xBBC) of the second data network node (110B) and the predetermined fixed common hop latency (dCH); and - a calculation unit which is adapted to calculate the latency matching period (dARX) using the formula dARX = dCH - dRT / 2- dSRX_B.

7. The synchronized data network system (10) according to any one of the preceding claims, characterized in that the second data network node (110B) comprises a communication control unit (LCU / CONFIG) for controlling initialization and synchronization, including: - an initialization control unit (LCU) that controls the initialization process of the data network nodes (110); and - a synchronization control unit (CONFIG) that controls synchronization of the data network nodes (110).

8. The synchronized data network system (10) according to claim 7, characterized in that the initialization control unit (LCU) prompts the second data network node (110B) as the slave data network node (110S) to perform the steps of: - transmitting a roundtrip measurement command (RTMR) from the communication control unit (LC) to the data transmitting unit (TX); - transmitting, by the data transmitting unit (TX), a roundtrip signal or ping signal to a data receiving unit (RX) of a corresponding receiving data network node (110) via a corresponding direct data transmission link (x) of a single-hop data forwarding link (xXBB, xBBY) to be measured; - starting a timer by the latency matching unit (LMU) upon transmission of the roundtrip signal; - receiving, by the data receiving unit (RX), the roundtrip signal sent back from the corresponding data network node (110) via the corresponding direct data transmission link (x); - stopping the timer by the latency matching unit (LMU) upon receipt of the roundtrip signal sent back; - reading the roundtrip latency (dRT) from the timer; - calculating the specific hop latency (dH_XBB, dH_BBY) of the single-hop data forwarding link (xXBB, xBBY) to be measured by adding the intrinsic node latency (dSRX_B) to half the roundtrip latency (dRT); - repeating the steps for all the single-hop data forwarding links (xXBB, xBBY) to be measured; and - storing the ascertained hop latencies (dH_XBB, dH_BBY) of all the single-hop data forwarding links (xXBB, xBBY) to be measured.

9. The synchronized data network system (10) according to claim 7, characterized in that the synchronization control unit (CONFIG) prompts the second data network node (110B) as the slave data network node (110S) to perform the steps of: - receiving, from a master data network node (110M) with a master clock (CLK_M), synchronization data packets (SDP) with a master clock time stamp ttx; and - synchronizing the slave clock (CLK_S) of the second data network node (110B) with the master clock time in accordance with the completed number of hops h of the synchronization data packet (DLP) between master data network nodes (110M) and the second data network node (110B), as well as by means of the common deterministic hop latency period dCH according to the formula tmaster = ttx + dCH * h.

10. The synchronized data network system (10) according to claim 7, characterized in that the synchronization control unit (CONFIG) prompts the second data network node (110B) as the master data network node (110M) to perform the steps of: - transmitting, at a time ttx, synchronization data packets (SDP) with a master clock time stamp ttx to the slave data network nodes (110S).

11. The synchronized data network system (10) according to claim 9 or 10, characterized in that the synchronization data packet (SDP), in addition to the master clock time stamp (ttx), includes the number of hops (h) and an identifier (ID_M) of the master data network node (110M).

12. The synchronized data network system (10) according to any one of the preceding claims, characterized in that it is an avionics system data network for deterministic communication between the data network nodes (110).

13. A method for initializing a synchronized data network system (10), comprising the steps of: - transmitting (S210) a roundtrip measurement command (RTMR) from a communication control unit (LC) of a slave data network node (110S) to the data transmitting unit (TX) of the slave data network node (110S); - transmitting (S220), by the data transmitting unit (TX) of the slave data network node (110S), a roundtrip signal or ping signal to a data receiving unit (RX) of a corresponding additional slave data network node (110S) via a corresponding direct data transmission link (x) of a single-hop data forwarding link (xXBB, xBBY) to be measured; - starting (S230) a timer of the slave data network node (110S) upon transmission of the roundtrip signal by a latency matching unit (LMU) of the slave data network node (110S); - receiving (S240), by the data receiving unit (RX) of the slave data network node (110S), the roundtrip signal sent back from the corresponding additional slave data network node (110S) via the corresponding direct data transmission link (x); - stopping (S250) the timer by the latency matching unit (LMU) of the slave data network node (110S) upon receipt of the roundtrip signal sent back; - reading (S260) the roundtrip latency (dRT) from the timer of the slave data network node (110S); - calculating (S270) the specific hop latency (dH_XBB, dH_BBY) of the single-hop data forwarding link (xXBB, xBBY) to be measured by adding the intrinsic node latency (dSRX_B) to half the roundtrip latency (dRT); - repeating (S280) the steps for all the single-hop data forwarding links (xXBB, xBBY) to be measured; - storing (S290) the ascertained hop latencies (dH_XBB, dH_BBY) of all the single-hop data forwarding links (xXBB, xBBY) to be measured; and - delaying forwarding of data on the single-hop data forwarding links (xXBB, xBBY) such that the hop latency (dH_XBB, dH_BBY) becomes equal to a hop latency (dCH), which is identical, predetermined, fixed and common for all data network nodes of the data network system that forward data between data network nodes of the data network system, and which is not dependent on network-specific or node-specific latencies.

14. A method for synchronizing a synchronized data network system (10) comprising at least one master data network node (110M) and slave data network nodes (110S), comprising the steps of: - transmitting (S310), through the master data network node (110M) at a time ttx, synchronization data packets (SDP) with a master clock time stamp ttx to the slave data network nodes (110S), - receiving (S320), through the slave data network nodes (110S), synchronization data packets (SDP) with a master clock time stamp ttx from the master data network node (110M) with a master clock (CLK_M), - synchronizing (S330) the slave clock (CLK_S) of the slave data network node (110S) with the master clock time in accordance with the completed number of hops h of the synchronization data packet (SDP) between master data network nodes (110M) and slave data network nodes (110B), as well as by means of a hop latency dCH, which is identical, predetermined, fixed and common for all data network nodes of the data network system that forward data between data network nodes of the data network system and which is not dependent on network-specific or node-specific latencies, according to the formula tmaster = ttx + dCH * h.

15. A synchronized data network system (10) comprising at least one master data network node (110M) and slave data network nodes (110S), which is adapted to carry out the methods according to claim 13 or 14.