METHOD FOR MONITORING A SEMI-DUPLEX ETHERNET NETWORK AND ASSOCIATED SEMI-DUPLEX ETHERNET NETWORK

Description

[0001] The present invention relates to a method for monitoring a Half Duplex Ethernet network. The present invention also relates to such a Half Duplex Ethernet network.

[0002] In particular, the Half Duplex Ethernet network according to the invention conforms to one of the following standards: 10BASE-T1S, 100BASE-T1S, 10BASE-T1L, 100BASE-T1L. These standards conform to the IEEE STD 802.3cg type standard.

[0003] Networks of this type, as is well known, allow the transmission of data from, for example, multiple sensors to a central switch. This type of network is a half-duplex network, as it allows frames to be transmitted in both directions between the nodes that make up this network on the same pair of conductors.

[0004] In addition, in a particular operating mode, this type of network implements a collision management technique optimized to use the maximum of available bandwidth for exchanges between different network nodes.

[0005] This mode of operation is known as PLCA (from English « Physical Layer Collision Avoidance ») and is based primarily on the presence of a master node on the network which periodically sends a synchronization signal called a BEACON.

[0006] This synchronization signal allows other nodes to synchronize and determine their own frame transmission opportunities in the network.

[0007] The absence of the synchronization signal on the network, due for example to a failure of the master node, forces the different network nodes to reconfigure the collision management mode by switching from PLCA operating mode to a degraded operating mode called CSMA / CD (from the English " Carrier Sense Multiple Access / Collision Detection »).

[0008] Switching to this mode results in a significant decrease in network performance. When the network is in this degraded operating mode, each node wishing to transmit a frame first checks if another node is already transmitting frames. If so, the first node will wait a random amount of time before performing this check again. An example of a network implementing PLCA operating mode is described in US patent application 20221 / 377343 A1.

[0009] In certain fields, and particularly in the avionics field, the aforementioned type of network presents an interesting alternative to the networks currently in use.

[0010] For example, in the avionics field, the use of ARINC 429 type networks is known to connect sensors to a centralized switch. However, the type of network conforming to this standard has a very low data rate, typically around 100 Kbit / s, and is only unidirectional.

[0011] On the contrary, for example, a 10BASE-T1S type network has a throughput of 10 Mbit / s and a 100BASE-T1S type network has a throughput of 100 Mbit / s.

[0012] Furthermore, a half-duplex network, which allows frames to be transmitted in both directions using the same pair of conductors, would reduce the amount of cabling required. Indeed, for existing networks, a separate pair of conductors is necessary to ensure frame transmission in each direction.

[0013] However, replacing existing networks with this type of network presents a number of difficulties.

[0014] In particular, when the network, for example of the 10BASE-T1S type, enters its degraded operating mode, the frame transmission time can no longer be limited.

[0015] This makes the network of this type non-deterministic. The use of non-deterministic networks is problematic in certain fields, particularly in avionics.

[0016] The present invention aims to solve this problem and make the Ethernet Half Duplex type network (such as 10BASE-T1S, 100BASE-T1S, 10BASE-T1L, 100BASET-1L) deterministic.

[0017] This will then allow its use in many fields, particularly in the avionics field.

[0018] To this end, the invention relates to a method for monitoring an Ethernet Half Duplex network comprising a plurality of nodes and having a PLCA operating mode and a CSMA / CD operating mode, each node being defined by an identifier.

[0019] In PLCA operating mode, a master node sends a synchronization signal to all other nodes for the transmission of their frames.

[0020] In CSMA / CD operating mode, all nodes transmit their frames at random time intervals.

[0021] The process includes the following steps implemented by each node: retrieval of an internal state of the node describing the mode of operation of the network; transmission of the internal state to all other nodes; analysis of the external states received corresponding to the internal states transmitted by other nodes; based on this analysis, reconfiguration of the node or maintenance of its current configuration.

[0022] According to other advantageous aspects of the invention, the method comprises one or more of the following features taken individually or in all technically possible combinations: Each internal state indicates the network's operating mode as seen by the corresponding node; each internal state also indicates whether the corresponding node receives or transmits a synchronization signal; the internal state transmission step includes integrating this internal state into each frame transmitted by the corresponding node; the node reconfiguration step includes disabling that node or shifting its identifier; the identifier shift includes: assigning a new identifier chosen based on the maximum number of nodes in the network when the corresponding node is the master node; reducing the identifier by "1" in all other cases; the received external state analysis step includes a validation substep comprising selecting a valid state from among all received states; the analysis step includes analyzing at least three external states from different nodes;The validation substep includes comparing the received external states and selecting the valid state corresponding to a common state in a majority of the received external states; the valid state indicates: PLCA operating mode, the current configuration of the corresponding node is maintained; CSMA / CD operating mode, the corresponding node is reconfigured by shifting its identifier; neither of these operating modes, the corresponding node is reconfigured by disabling it; the analysis step includes a substep of activating CSMA / CD operating mode when the corresponding node does not receive any frames in a first time interval; following the activation of CSMA / CD operating mode, the reconfiguration step includes shifting the identifier of the corresponding node when this node receives at least one frame in a second time interval and disabling the node otherwise.

[0023] The invention also relates to an Ethernet Half Duplex type network comprising a plurality of nodes and having a PLCA operating mode and a CSMA / CD operating mode, each node being defined by an identifier; In PLCA mode, a master node is capable of sending a synchronization signal to all other nodes for the transmission of their frames; in CSMA / CD mode, all nodes are capable of transmitting their frames at random time intervals; each node is configured to implement the process as defined previously.

[0024] The invention will become clearer upon reading the following description, given solely by way of non-limiting example and with reference to the drawings in which: [ Fig.1 ] there figure 1 is a schematic view of a Half Duplex Ethernet network according to the invention, the network comprising in particular a plurality of nodes; [ Fig.2 ] there figure 2 is a flowchart of a network monitoring process of the figure 1 ; Fig.3 ] ] Fig.4 ] ] Fig.5 ] ] Fig.6 ] ] Fig.7 ] THE figures 3 à 7 are different illustrations of the implementation of the process of the figure 2 .

[0025] There figure 1 illustrates indeed a network 10 according to the invention.

[0026] This network 10 is a Half Duplex Ethernet network that conforms to one of the following standards: 10BASE-T1S, 100BASE-T1S, 10BASE-T1L, or 100BASE-T1L. Generally, such a standard is defined in the IEEE STD 802.3cg type standard. "Type X standard" refers to the current version of the X standard, or any earlier or later version that enables the implementation of the features described in relation to that standard below.

[0027] As illustrated on the figure 1 , network 10 comprises a plurality of nodes 12-0, ... 12-N which are connected by a cable 13. Cable 13 advantageously presents a cable with a differential pair, for example a twisted pair cable.

[0028] Each of the nodes 12-0, ... 12-N is integrated into a device capable of generating digital data to send to other devices via network 10 and / or receiving digital data from these devices. The digital data is transferred via network 10 in the form of frames.

[0029] In particular, according to one embodiment of the invention, the network 10 is usable in the avionics field, for example the network 10 is carried on board an aircraft.

[0030] An aircraft is defined as any flying machine that can be piloted at least partially manually and / or at least partially automatically.

[0031] According to one embodiment, at least some of the devices integrating nodes 12-0, ... 12-N have avionics sensors 14 and a device integrating at least one of these nodes has a switch 15. Such a switch 15 is configured to transmit data generated by the sensors to a centralized computer.

[0032] According to this example, the data generated by the sensors 14 include measurement data which are then transmitted to the central computer via the network 10 and in particular the switch 15.

[0033] Thus, in the example in the main figure, node 12-0 is integrated into such a switch 15 and the other nodes 12-1 to 12-N are integrated into sensors 14.

[0034] Furthermore, node 12-0 is also capable of transmitting data to other nodes. This data includes, for example, configuration data for the corresponding sensors 14.

[0035] As is well known, each 12-0, ... 12-N node has a physical interface, often implemented by a component called a PHY. This PHY physical interface includes, among other things, an internal register storing various operating parameters of the node.

[0036] Furthermore, each node 12-0, ... 12-N is associated with a unique identifier in the network 10. For example, the identifiers of nodes 12-0, ... 12-N correspond to the number of the node assigned within the network and ranging from 0 to N.

[0037] The node identifiers can be changed, as will be described in more detail later.

[0038] Network 10, as is known in itself, has two modes of operation.

[0039] In a first mode of operation, also called PLCA mode of operation, all frame exchanges in the network 10 are controlled by a synchronization signal transmitted to all nodes by the master node.

[0040] This master node typically has the identifier "0," which distinguishes it from other nodes in the network. This master node is configured to periodically send a synchronization signal, also called a beacon. All other nodes are able to receive this synchronization signal and transmit their frames within a predetermined time slot for the corresponding node. This time slot is determined locally by the corresponding node using the synchronization signal. The master node is also configured to behave like the other nodes; that is, it is also configured to transmit frames within a specific time slot.

[0041] When the corresponding node has no frame to send in its slot, that slot remains empty and no frame is sent into the network during that slot.

[0042] Advantageously, each slot can have a variable duration. For example, each slot can correspond to a period during which the node can begin transmitting. Each slot therefore has a minimum duration corresponding to the absence of a frame, but this duration can extend to the duration of the frame to be transmitted when such a frame is present. In such a case, the next transmitting node will only begin transmitting once the first node has finished transmitting its frame; that is, the next transmitting node will only begin transmitting after a period of "silence" is detected in the network.

[0043] The second operating mode of the 10 network, also known as CSMA / CD, is triggered when no synchronization signal is received by the nodes after a cycle, that is, after all nodes have had their turn to transmit. In such a case, each node is able to send its frames at random intervals to avoid collisions. Specifically, when a node tries to send a frame while another node is sending another frame at the same time, the first node will wait a random amount of time before attempting to send its frame again.

[0044] According to the invention, each node is further configured to implement a monitoring method according to the invention, which reconfigures the network when it switches to CSMA / CD operating mode, in order to return to PLCA operating mode. This method is, for example, implemented periodically in network 10. In other words, steps 110 to 140 described below can be repeated at predetermined intervals.

[0045] This process will now be explained with reference to the figure 2 presenting an organizational chart of its stages.

[0046] Initially, it is assumed that network 10 operates according to PLCA operating mode and the set of nodes 12-0, ..., 12-N are capable of sending and receiving frames.

[0047] It is further assumed that the steps described below are carried out by each of the nodes 12-0, ..., 12-N. Thus, these steps will be described with reference to a single node (for example, node 12-1) knowing that all the other nodes are capable of implementing the same steps.

[0048] During an initial step 110, node 12-1 retrieves an internal state allowing the current operating mode of network 10 to be described from the point of view of that node.

[0049] Specifically, this internal state indicates the network's operating mode, either PLCA or CSMA / CD. This internal state also indicates whether this node is receiving or transmitting a synchronization signal.

[0050] More specifically, the internal state can indicate the PLCA operating mode and whether this node is transmitting a synchronization signal. In other words, this internal state indicates that the corresponding node is the master node in network 10.

[0051] In another example, this internal state may indicate that the network is in PLCA operating mode and the corresponding node is receiving a synchronization signal. In such a case, the corresponding node is a slave node that then receives a synchronization signal from a master node.

[0052] The internal state may also indicate that the operating mode of network 10 is CSMA / CD and that the corresponding node is not receiving any signaling signal. This case corresponds to normal operation of the node in this CSMA / CD operating mode of network 10. In yet another example, the internal state of the node may indicate that network 10 is in CSMA / CD operating mode while the corresponding node is receiving or transmitting a synchronization signal. This latter state indicates a network malfunction, which will then be addressed as explained in the following steps of the procedure.

[0053] Advantageously, the internal state of the node is kept in its internal register.

[0054] Retrieving such a state therefore involves retrieving that state from this internal register of node 12-1.

[0055] In the next step 120, node 12-1 transmits its internal state to all other nodes 12-0, 12-2 to 12-N.

[0056] To do this, node 12-1 integrates its identifier and internal state into each frame emitted by this node in network 10.

[0057] To do this, various techniques are possible.

[0058] According to a first technique illustrated on the figure 3 The 12-1 node integrates additional data (D) into a header of each frame. This header corresponds, for example, to the MAC header type. This additional data (D) can be integrated more specifically into the field indicating the frame's origin, called the MAC Source field.

[0059] According to a second technique illustrated on the figure 4 The additional data D is integrated into the payload of the corresponding frame.

[0060] This payload is known by the English abbreviation Payload.

[0061] According to yet another technique (not shown), the additional data D can be integrated into a field independent of the useful application data Payload field. This then implies a reduction in the maximum size of the Payload field.

[0062] In the following step 130, node 12-1 analyzes external states received from other nodes in network 10, or, if necessary, in cases where such external states have not been received. Specifically, external states are defined as internal states determined by other network nodes and then transmitted to the corresponding node using frames emitted by those nodes.

[0063] Analysis step 130 determines whether it is necessary to reconfigure node 12-1 or to maintain its current configuration during step 140, in order to restore the PLCA operating mode of network 10 if necessary.

[0064] Several examples of implementing steps 130 and 140 are possible. Furthermore, as will be explained later, in some cases it is necessary to implement steps 130 and 140 several times in order to achieve the PLCA operating mode of network 10.

[0065] An example of a possible implementation of these steps is illustrated on the figure 5 .

[0066] With reference to this figure 5 The analysis step 130 includes a first substep 131 during which node 12-1 waits for the reception of external states from the other nodes. When no external state, i.e., no frame, is received from the other nodes after an initial time interval, node 12-1 implements substep 132 during which it forces a switch to CSMA / CD operating mode. Then, node 12-1 again waits for the reception of external states, i.e., frames from the other nodes, during substep 133.

[0067] When, after a second time interval that is, for example, equal to the first time interval, node 12-1 still does not receive a frame from the other nodes, it is considered to be a failed node and should then be deactivated during step 140. On the figure 5 The deactivation of the node then corresponds to the action designated by reference 141. Conversely, when node 12-1 receives frames from other nodes, the node implements a validation substep 134 of the received external statuses. This validation substep 134 will be explained in more detail later.

[0068] When, at the end of step 134, the network status indicating the CSMA / CD operating mode is considered invalid, the node is again considered to be faulty and is then deactivated during step 140 by action 141.

[0069] Conversely, when the status indicating the CSMA / CD network operating mode is considered valid during step 134, it is considered that network 10, and in particular the corresponding nodes 12-0 to 12-N, must be reconfigured during step 140 to restore the PLCA operating mode.

[0070] With reference to the figure 5 This action is then noted by reference 142 and will be explained in more detail later.

[0071] When, during substep 131, node 12-1 receives frames from other nodes, i.e., external states, it implements substep 134 of validating the PLCA status of network 10.

[0072] When this status is deemed valid at the end of substep 134, it is assumed that network 10, and in particular node 12-1, is operating normally in PLCA operating mode and during step 140, the current network configuration is maintained.

[0073] With reference to the figure 5 This action of maintaining the current network configuration corresponds to reference 143.

[0074] When, on the contrary, after the first implementation of validation step 134, the state indicating the PLCA operating mode of network 10 is considered invalid, node 12-1 implements validation step 134 again, during which it checks this time the validity of the state indicating the CSMA / CD operating mode.

[0075] When, at the end of this second implementation of substep 134, the state indicating the CSMA / CD operating mode is considered valid, it is considered that network 10 and in particular node 12-1 must be reconfigured during step 140 to restore the PLCA operating mode of network 10.

[0076] To do this, node 12-1 implements action 142 during step 140, which will be explained in detail later.

[0077] When, on the other hand, at the end of the second implementation of the validation substep 134, the state indicating the CSMA / CD operating mode is considered invalid, it is considered that the node 12-1 is faulty and must be deactivated during step 140. Thus, during step 140, action 141 including the deactivation of the node is then implemented.

[0078] There figure 6 illustrates an example of the implementation of substep 134 of validation of external states received by the corresponding node 12-1.

[0079] In general, validation substep 134 includes comparing external states received from different nodes and selecting the valid state corresponding to a common state according to a majority of external states received.

[0080] Thus, according to the example of the figure 6 , substep 134 includes the verification of three conditions.

[0081] The first condition (Condition 1 on the figure 6 ) is respected when node 12-1 receives two external statuses from two different nodes.

[0082] When this is not the case, node 12-1 waits until it receives external statuses from two different nodes.

[0083] When two statuses from two different nodes are received, node 12-1 checks the second condition (Condition 2 on the figure 6 ) which consists of comparing the statuses received.

[0084] When the statuses are identical, node 12-1 concludes that this status is valid. Conversely, when the second condition is not met, that is, the statuses received from the two different nodes are not identical, node 12-1 implements the third condition (Condition 3 on the figure 6 ) during which it checks if it has received an external status from a node other than the first two nodes. When this is not the case, node 12-1 waits for such a status from a third node other than the first two nodes.

[0085] When this is the case, the 12-1 node selects the valid state corresponding to the common state based on the majority of external statuses received. In other words, in such a case, the 12-1 node decides the valid status based on the majority.

[0086] Of course, other methods of validating received states are also possible. For example, it is possible to perform a double check of states received from the same node.

[0087] There figure 7 illustrates the implementation of action 142 implemented during step 140.

[0088] In particular, this action 142 includes a shift in the node identifier in order to restore the PLCA operating mode of network 10.

[0089] This offset involves assigning a new identifier determined by the maximum number of nodes in the network when the corresponding node is the master node. In the case of other nodes, the offset involves reducing that node's identifier by one.

[0090] In its part A, the figure 7 illustrates the normal operation of network 10 according to the PLCA operating mode for nodes with identifiers from 0 to N. Among these nodes, the node with identifier 0 corresponds to the master node which is able to send a synchronization signal to the other nodes.

[0091] In its part B, the figure 7 This illustrates the case where the node with identifier 0 needs to be reconfigured and a new master node needs to be selected. For this, and as illustrated in Part C of the figure 7 The previous master node receives the identifier corresponding, for example, to the highest identifier in the network and thus becomes a slave node. In such a case, the identifiers of all other nodes are decreased by one, so that the previous node with identifier number 1 becomes the node with identifier number 0. Consequently, this node becomes the new master node.

[0092] The other nodes, that is to say the previous nodes having identifiers ranging from 2 to N, remain the slave nodes but change their identifiers which now range from 1 to N-1.

[0093] When, after a first shift of the identifiers, there is no node with the identifier 0, the shift operation is carried out again until such a node with the identifier 0 is obtained. This new shift operation is carried out during a new implementation of step 140 and action 142, which are then carried out following a new implementation of analysis step 130.

[0094] Thus, network 10 can operate again according to PLCA operating mode. In such a case, the new master node receives and transmits the corresponding synchronization signal.

[0095] It is therefore understandable that the present invention offers a number of advantages.

[0096] In particular, the invention first enables the reliable detection of a network failure. This detection is considered reliable because it relies on the redundancy of different states originating from different nodes. This makes the probability of a false detection practically negligible. Second, the invention allows the network to be reconfigured to maintain and find the PLCA mode, which is then the desired operating mode.

[0097] The invention also implements a technique for verifying transmitted external states in order to protect the integrity of the network.

[0098] During network reconfiguration, the invention makes it possible to disable a malfunctioning node to prevent any anomaly on the network caused by that node.

[0099] Finally, the invention also allows for the offsetting of node identifiers across the entire network to enable the fastest possible recovery of PLCA operating mode. This operation allows the offset to be repeated multiple times in the event of multiple failures. This multiplicity of reconfiguration possibilities makes the network more reliable in an avionics context, since the probability of a system experiencing multiple similar failures is low and decreases with the number of failures.

[0100] The network according to the invention can thus be effectively used, particularly in the avionics field, to replace existing networks, for example networks according to the ARINC 429 standard.

Claims

1. A method for monitoring a Half Duplex Ethernet network comprising a plurality of nodes (12-0, ..., 12-N) and having a PLCA operating mode and a CSMA / CD operating mode, each node (12-0, ..., 12-N) being defined by an identifier; in the PLCA operating mode, a master node (12-0) sending to all the other nodes (12-1, ..., 12-N) a synchronization signal to send the frames thereof; in the CSMA / CD operating mode, all nodes (12-0, ..., 12-N) sending the frames thereof at random time intervals; the method comprising the following steps implemented by each node (12-0, ..., 12-N): - retrieving (110) an internal state of the node (12-0, ..., 12-N) describing the network operation mode; - transmission (120) of the internal state to all the other nodes (12-0, ..., 12-N); - analysis (130) of the external states received corresponding to the internal states sent by other nodes; - depending on such analysis, reconfiguration (140) of the node (12-0, ..., 12-N) or conservation of the current configuration thereof; the method being characterized in that the step of transmission (120) of the internal state comprising the integration of the internal state into each frame sent by the corresponding node (12-0, ..., 12-N).

2. The method according to claim 1, wherein each internal state indicates the operating mode of the network seen by the corresponding node (12-0, ..., 12-N).

3. The method according to claim 2, wherein each internal state further indicates whether the corresponding node (12-0, ..., 12-N) receives or sends a synchronization signal.

4. The method according to any of the preceding claims, wherein the step of reconfiguration (140) of each node (12-0, ..., 12-N) comprises a deactivation (141) of the node or a shift (142) of the identifier thereof.

5. The method according to claim 4, wherein the shift (142) of the identifier comprises: - the allocation of a new identifier chosen according to the maximum number of nodes (12-0, ..., 12-N) in the network when the corresponding node (12-0, ..., 12-N) is the master node; - the reduction of the identifier by "1" for all other cases.

6. The method according to any of the preceding claims, wherein the step of analysis (130) of the received external states comprises a validation sub-step (134) comprising the selection of a valid state from the set of received states.

7. The method according to claim 6, wherein: - the analysis step (130) comprises an analysis of at least three external states from different nodes (12-0, ..., 12-N); - the validation sub-step (134) comprises the comparison of the external states received and the selection of the valid state corresponding to a common state of a majority of the external states received.

8. The method according to claim 6 or 7, wherein when the valid state indicates: - the PLCA operating mode, the current configuration of the (12-0,..., 12-N) corresponding node is conserved; - the CSMA / CD operating mode, the corresponding node (12-0, ..., 12-N) is reconfigured by a shift of the identifier thereof; - none of the operating modes, the corresponding node (12-0, ..., 12-N) is reconfigured by a deactivation of the node (12-0, ..., 12-N).

9. The method according to any of the preceding claims, wherein the analyzing step (130) comprises a sub-step (132) of activation of the CSMA / CD operating mode when the corresponding node (12-0, ..., 12-N) receives no frames in a first time interval.

10. The method according to claim 9, wherein following the activation of the CSMA / CD operating mode, the reconfiguration step (140) comprises a shift (142) of the identifier of the corresponding node (12-0, ..., 12-N) when the node (12-0, ..., 12-N) receives at least one frame in a second time interval and a deactivation (142) of the node (12-0, ..., 12-N) otherwise.

11. A Half Duplex Ethernet network (10) comprising a plurality of nodes (12-0, ..., 12-N) and having a PLCA operating mode and a CSMA / CD operating mode, each node (12-0, ..., 12-N) being defined by an identifier; in the PLCA operating mode, a master (12-0) node being apt to send a synchronization signal to all the other nodes (12-1, ..., 12-N) to issue the frames thereof; in the CSMA / CD operating mode, all nodes (12-0, ..., 12-N) being apt to issue the frames thereof at random time intervals; each node (12-0, ..., 12-N) being configured to implement the method according to any of the preceding claims.