Monitoring device for a data router and method for monitoring a data router

Description

[0001] The invention relates to a system comprising a data router and a monitoring device connected to or integrated into the data router. The invention also relates to a monitoring device for a data router and a method for monitoring a data router.

[0002] Data routers – or simply routers – are responsible for forwarding data packets between multiple data processing devices, such as computers or embedded systems, within a data network. Data routers are used, for example, to connect local data processing devices to the internet. Data routers can also be configured as cellular routers to transmit data packets over the mobile network. This is particularly useful where a wired data network is not available for connecting a data processing device.

[0003] Routers make their routing decisions based on information from network layer 3 (for the IP protocol, this is the network portion of the IP address). Many routers also translate between private and public IP addresses (Network Address Translation (NAT) or Port Address Translation (PAT)) or implement firewall functions through a set of rules.

[0004] Routers are used in a variety of locations, including transformer stations and charging stations that are regularly connected to the internet. Transformer stations, in particular, are often located far from populated areas, for example, near solar farms. Routers typically need to be restarted at longer intervals because they have, for example, "frozen." Unfavorable events can cause routers to "crash," also known as "freezing." In this case, the data connection established via the router is permanently interrupted.

[0005] In such cases, the manufacturer often recommends resetting the router in order to restore the data connection.

[0006] It is known that routers can be automatically restarted when a monitoring unit—for example, software—detects that a router has lost its internet connection. For this to work, the monitoring unit must track the data traffic passing through the router. Products already exist that can force a router reset (e.g., "resetplug.com"). These products connect to the router as a Wi-Fi client, limiting their use to Wi-Fi-enabled routers. Furthermore, such a concept represents a potential security vulnerability.

[0007] Because the latter is unsatisfactory, there is a need for alternative methods that can cause the router to restart automatically. One simple solution is to restart the router at regular, predetermined intervals – similar to a timer. However, if these intervals are too long, the router may still be unable to provide an internet connection for a significant period. Conversely, restarting the router too frequently has the disadvantage of requiring it to restart too often, leading to unnecessary data connection interruptions.

[0008] WO 2006 / 047583 A2 describes techniques for restarting computer system hardware from a remote location.

[0009] According to the invention, a monitoring device is proposed that monitors the energy or power consumption of the data router and triggers a restart of the data router depending on its energy or power consumption. The monitoring device according to the invention detects the crash or freezing of the data router and automatically initiates a reset or restart of the data router.

[0010] The solution according to the invention incorporates the insight that the power consumption of the data router depends on whether the data router is sending or receiving data packets. When data packets are actually being transmitted, the power consumption is higher than at times when no data is being transmitted. Typically, a data router transmits data packets repeatedly at regular short intervals of a few seconds or minutes to maintain or establish contact with a base station (in the case of a mobile router) or the nearest internet exchange point. This results in the power consumption of a functioning data router fluctuating over a period of a few minutes. If, however, the power consumption remains approximately constant—especially consistently low—this indicates that the data router is no longer operational and needs to be restarted.Accordingly, a monitoring device according to the invention can be designed such that it causes a restart of the data router if its power consumption remains constant, in particular consistently low, over an observation period of a few minutes and corresponds, for example, to a previously recorded average value for the power consumption of the data router in standby mode - i.e., without an active data connection.

[0011] The same applies to the energy consumption of the data router within a specific period. The energy consumed by the data router within a given period represents the average power consumption of the data router during that period.

[0012] Mobile routers typically require a minimum power level to establish a data connection to the nearest base station, which depends, among other things, on the router's distance from the base station. A data router that has no connection to the base station does not need to provide this minimum power level and therefore does not draw any power. Accordingly, the monitoring device can be configured to monitor whether a data router at a given location is drawing the minimum power level necessary to maintain a connection to the nearest base station. If the data router no longer draws this minimum power level—preferably location-specific—or a power level exceeding the minimum, but instead draws a lower power level, this indicates that the data router has no connection to the base station. This, in turn, can be a criterion that prompts the monitoring device to restart the data router.

[0013] A monitoring device according to the invention, which triggers a reset or restart of a data router depending on its power or energy consumption, has the advantage that it relies solely on the power consumption of the data router or a component or unit of the data router whose power consumption correlates with data transmission, and does not require any information about the data stream flowing through the data router. Therefore, such a monitoring device also does not require a data interface to the router. Rather, the monitoring device focuses solely on the power supply of the data router. Because no data connection to the data router is necessary, as required by other products, no potential security vulnerability can arise.

[0014] Furthermore, the invention is universally applicable due to the measurement principle, so that routers without WiFi or WLAN can also be monitored.

[0015] In a preferred embodiment, the monitoring device detects the power consumption of the data router at its power supply input, i.e., where the power required for its operation is supplied to the data router. Alternatively, the monitoring device can also be configured to detect the power consumption of a component or unit of the data router whose power consumption correlates with the regular operation of the data router. Such a unit or component is, for example, the modem of the data router, in particular a mobile cellular modem with a corresponding mobile cellular transmitter and receiver. Specifically, the current or power consumption of the processor unit of the data modem or data router, the transceiver module of the data modem, or the radio module of the data modem can be detected.

[0016] The monitoring device can be designed such that the monitoring criteria and the criteria for restarting or resetting the data router are adjustable, allowing them to be adapted to the current circumstances. Monitoring criteria and the criteria for restarting or resetting the data router can be values—especially limit values—for monitored or recorded parameters, such as the current power consumption or the energy consumed within a given period (corresponding to the average power consumption during that period).

[0017] It is particularly preferred if the monitoring device can be "trained" or "learned" to automatically learn the given conditions.

[0018] For example, the monitoring device can be designed to determine, during a learning phase, a value for the minimum amount of energy that the data router consumes within a specified time period and to derive from this value a limit value for the amount of energy below which a reset or restart of the data router is triggered.

[0019] The monitoring device can also be equipped with a neural network trained on training datasets representing the data router's power consumption over time, assuming the data router is operational and has an internet connection to the nearest node. The neural network can be configured as a binary classifier that assigns datasets representing the current power consumption of the data router either to the state class for which the neural network was trained (e.g., when an internet connection exists) or not to that state class. If a current dataset cannot be assigned to the state class representing a functioning data router, the neural network outputs a correspondingly low membership value, which can then be used as the criterion for restarting the data router.

[0020] A neural network configured as a binary classifier in a monitoring device can also be trained with training datasets representing the power consumption history of the data router when the router is not operational or has no data connection. In this case, the monitoring device is trained to trigger a reset or restart of the data router when the neural network—i.e., the binary classifier—assigns a dataset representing current power consumption to the state class for which the neural network is trained. A high membership value for the state class "Data router not OK" can therefore be used as a criterion for restarting or resetting the data router.

[0021] A monitoring device with a neural network offers the advantage that the test criterion that triggers a restart or reset of the data router is not a simple threshold or limit comparison, but rather represents a pattern in the data router's power consumption. This allows the monitoring device to be more precisely tailored to the site-specific characteristics of the data router and its power consumption.

[0022] The data sets – and thus also the training data sets – are preferably data tuples representing the power consumption of the data router over a predetermined period, preferably a few minutes. Accordingly, the monitoring device is preferably configured to record the power consumption of the data router over the relevant period and preferably to sample and digitize the power values. Power acquisition preferably takes place within a sliding time window.

[0023] The monitoring device preferably comprises a microcontroller-controlled electronic circuit as a data acquisition unit. This circuit is configured to continuously record the current consumption of any downstream data router. Preferably, the current consumption is analyzed and evaluated by an evaluation unit. The evaluation unit may include the aforementioned neural network.

[0024] Alternatively, the acquisition unit can also be configured to acquire the power consumption of a component or unit of the data router whose power consumption correlates with data transmission. Preferably, this is a data modem of the data router. In particular, the current or power consumption of the processor unit of the data modem or the data router, the transceiver module of the data modem, or the radio module of the data modem can be acquired.

[0025] The evaluation unit of the monitoring device is connected to the acquisition unit and is configured to analyze the output signal of the acquisition unit with respect to at least one test criterion and, if the test criterion is met, to trigger a reset or a restart of the data router. For this purpose, the evaluation unit can incorporate the aforementioned neural network, wherein the output signal supplied by the acquisition unit comprises data sets representing the current time course of the power consumption of the data router.

[0026] If the evaluation unit determines that the device has "crashed" or "frozen," the router's power supply is preferably briefly interrupted and then restored. This results in a forced reset of the router, or at least a part of it, such as the part responsible for an external data connection.

[0027] A method according to the invention for monitoring a data router comprises the following steps: Acquire at least one measurement correlated with the power consumption of a data router, check the measurement against a reference criterion; restart or reset the router depending on the test result.

[0028] The measured value correlated with the power consumption of a data router could, for example, be the current strength of the power consumed by the router.

[0029] Checking the measured value against a reference criterion can, for example, involve a comparison with a reference value. The reference value could be, for instance, the minimum energy or the minimum amount of current (electrical charge) that the router consumes in its operational state within a predetermined period, preferably a few minutes. Alternatively, the reference criterion could be generated by a neural network trained with appropriate training datasets, in particular a neural network in the form of a binary classifier.

[0030] The router can be restarted by briefly interrupting the power supply to the router or by sending a reset signal to the data router via a corresponding control line.

[0031] Another aspect is the use of a system consisting of a data router and a monitoring device connected to or integrated with it, and / or a monitoring device of the type described here, in a data processing unit, a monitoring unit, a renewable energy plant, an electric vehicle charging station, or similar equipment. Renewable energy plants, such as wind turbines or photovoltaic systems, typically have data processing units that, for example, monitor the feed-in of generated energy into the power grid. Renewable energy plants are often located in remote areas, so they depend on a reliable internet connection. Therefore, an automated restart of a frozen data router using the monitoring device presented here is very helpful.The same applies to charging stations for electric vehicles. In the case of charging stations, the additional issue is that a data router frozen for an extended period renders the charging station unusable for that period, as electricity consumed during this time cannot be billed. The invention will now be explained in more detail with reference to exemplary embodiments and the figures. Figure 1 shows: . Fig. 1 :a first variant of a system according to the invention comprising a data router with integrated power supply and a monitoring device in a simplified, schematic block representation; Fig. 2 :a first variant of a system according to the invention comprising a data router with external power supply and a monitoring device in a simplified, schematic block representation; Fig. 3:a second variant of a system according to the invention comprising a data router and a monitoring device in a simplified, schematic block diagram; Fig. 4 :a more detailed description of a monitoring device according to the invention with a neural network; and Fig. 5 :a more detailed representation of a monitoring device according to the invention with a parameter memory for an evaluation unit Fig. 6 :a third variant of a system according to the invention comprising a data router and a monitoring device in a simplified, schematic block representation; Fig. 7 :a fourth variant of a system according to the invention comprising a data router and a monitoring device in a simplified, schematic block diagram; and Fig. 8:an external power supply designed to continuously monitor the power output and, if the test criterion is met, to cause a brief interruption of the power output by the power supply.

[0032] A system 10 according to the invention comprises a data router 12 and a monitoring device 14. The data router 12 has a power supply line 16, via which the data router 12 – in particular its power supply unit 40 – is supplied with the electrical power necessary for the operation of the data router 12. The power supply unit 40 can be integrated into the data router 12. However, the power supply unit 40' can also be a separate power supply unit, which is connected to the power supply input of the data router 12 by means of a cable as a DC power supply line 16'; see Figure 2The data router 12 can, for example, be a mobile router with a mobile modem 42 and an antenna 18 to wirelessly receive or transmit data packets. Data processing devices, such as those intended to connect to the internet, can be connected to the data router 12, and in particular to its processor unit 46, via appropriate wireless or wired interfaces 44. An example of such a data line 20 is shown in Figure 1 and 2 The data processing devices connected to the data router 12 can also be connected wirelessly, for example via Wireless LAN (WiFi).

[0033] In normal operation, the data router 12 repeatedly establishes contact with, for example, a base station. This requires a minimum electrical power supply. In the event of a malfunction of the data router 12, it may "freeze" and then no longer be able to establish a data connection to a base station. As a result, the power consumption of the data router 12 remains consistently very low and does not exhibit the characteristic minimum power peaks that occur when a data connection between the data router 12 and a base station is active.

[0034] To detect a possible malfunction of the data router 12, the monitoring device 14 is provided. This device includes a sensing unit 22, which can be implemented, for example, as a preferably microcontroller-controlled electronic circuit and is configured to detect the current or power consumption of the data router 12 via the power supply line 16 and to provide measured values ​​representing this consumption. If the power supply for the data router 12 is a separate power supply connected to a power input of the data router 12 by means of a cable, the cable 16 typically carries the DC voltage required for the operation of the data router. In this case as well, the sensing unit 22 is configured to detect the current or power consumption of the data router 12 via the power supply line 16.

[0035] The acquisition unit 22 is connected to an evaluation unit 24, which is configured to evaluate the measured values ​​generated by the acquisition unit 22 and, in particular, to check them against a test criterion. The test criterion is selected such that it represents a characteristic power consumption of the data router 12 during its regular operation or – alternatively – a characteristic power consumption of the data router 12 when it has "crashed" or "frozen".

[0036] The evaluation unit 24 is configured to cause a reset or restart of the data router 12 if the test based on the test criterion shows that the data router 12 does not exhibit the power consumption characteristic of regular operation, but in particular a lower power consumption.

[0037] One test criterion can be the power consumption over a specified period – preferably a sliding time window. In effect, the test criterion in this case is the electrical energy consumed by the data router within the specified period.

[0038] Alternatively or additionally, the test criterion can also be the fluctuation in the power or current consumed by the data router 12. A "frozen" data router 12 is characterized by its power consumption—and thus its current consumption—remaining constant or nearly constant over a longer period and exhibiting no significant fluctuations. The "freezing" of the data router can therefore be recognized by the fact that its current consumption does not exhibit the fluctuations associated with regular contact with a base station.

[0039] Preferably, the test criterion is derived from a typical power consumption pattern. For this purpose, the evaluation unit 24 can use a neural network 32 (see Figure 4The training data set comprises a binary classifier that is trained on training data sets formed from data tuples representing the power consumption of a given data router 12 over a specified period—preferably the time course of the power consumption of the corresponding data router 12 over a specified period. The training data sets can represent the power consumption of the data router 12 for periods in which the router functions correctly or, alternatively, for periods in which the data router 12 does not function correctly. Preferably, the training data sets are determined individually for each data router 12, since the power consumption of a data router 12 can depend on the specific circumstances and also on the type of data router 12 and the data processing devices connected to it.

[0040] If the evaluation of the measured values ​​by the evaluation unit 24 shows that the data router 12 is no longer functioning properly, the evaluation unit 24 triggers a reset of the data router 12 either via a reset line 26 (see Figure 1 and 2 ) or it causes a switch 30 to open via a control line 28 in order to interrupt the power supply to the data router 12 and thus force a restart of the data router 12 (see Figure 3 ).

[0041] If the evaluation unit 24 is a neural network 32 configured as a binary classifier (see Figure 4The neural network, which has a certain characteristic, can assign current data sets generated from measured values ​​recorded by the acquisition unit 22 to, for example, the state "data router ok" or "data router not ok". As soon as the neural network has assigned one or a predetermined number of consecutive data sets generated by the acquisition unit 22 to the state "data router not ok", a signal generator 34 connected to the neural network 32 can generate a control signal and output it via a signal line 36, which causes a reset or restart of the data router 12. The signal line 36 thus corresponds, for example, to the reset line 26 from Figure 1 or 2 or the control line 28 from Figure 3 .

[0042] In this case, the acquisition unit 22 is configured to record measured values ​​and generate data sets from them that represent the current power consumption of the data router 12. These data sets are fed to the neural network 32 of the evaluation unit 14, which is configured as a binary classifier. The neural network classifies each data set by determining a membership value for it within a state class for which it has been trained using training data sets. These training data sets could, for example, represent the state "Data Route io" and thus define a state class corresponding to an operational data router. The neural network determines a membership value for each current data set within this state class and makes this membership value available at a node of the neural network's output layer.If the membership value for the state class "Data Router io" falls below a predefined minimum value - which thus defines a test criterion - the signal generator 34 downstream of the neural network 32 can generate the control signal and output it via a signal line 36, which causes a reset or restart of the data router 12.

[0043] Figure 5 Figure 14 shows a monitoring device in which the evaluation unit 24 is connected to a parameter memory 38. The parameter memory 38 can, for example, contain individual parameter values ​​for a given system 10, e.g., for the power consumption of the data router 12 in standby mode or the like, so that the evaluation unit 24 can trigger a reset or restart of the data router 12, e.g., depending on site-specific criteria for power consumption.

[0044] In alternative embodiments, a detection unit 22' detects the current or power consumption of a unit or component of the data router 12, in particular the current or power consumption of the mobile modem 42. The inventors have recognized that the power or current consumption of the mobile modem 42 also correlates with the operating state of the data router 12 and decreases accordingly when the data router 12 has "crashed" or "frozen".

[0045] Except for the design of the recording unit 22', the version described in Figure 6 is shown, approximately in the version shown in Figure 1 or 2 is shown. Accordingly, the design variant corresponds to the one shown in Figure 7 is shown, except for the design of the recording unit 22' of the version variant, which is in Figure 3 is shown.

[0046] Regarding the specific design of the evaluation unit 24, it should be noted that the configuration of the evaluation unit 24 varies depending on the specific configuration. Figure 6 and 7 The illustrated implementation variants are each adapted to the correspondingly configured acquisition unit 22' such that the test criteria applied by the evaluation unit 24 correspond to the output signals supplied by the acquisition unit 22' during operation. For example, if the evaluation unit 24 contains a neural network 32, the training data sets used to train it are generated with a corresponding acquisition unit 22'.

[0047] If the evaluation unit 24 has a parameter memory 38 (see Figure 5 ) contains, then the parameter values ​​stored in the parameter memory 38 are matched to the power or current values ​​that the acquisition unit 22' records during operation or in standby mode of the data router 12.

[0048] In the case of a data router 12' with an external power supply 40', the acquisition unit 22 can also be integrated into the external power supply. The external power supply can be, as in the exemplary embodiment, in Figure 3 The monitoring device also includes the acquisition unit 22 and the evaluation unit 24 and is designed to briefly interrupt the power supply to the data router if the data router does not consume a specified minimum amount of energy within a predetermined period or if the evaluation unit has otherwise determined that the data router is not functioning as intended. In this case, the data router itself can remain unchanged compared to conventional data routers with an external power supply.

[0049] An external (or integrated) power supply with a sensing unit designed to continuously measure the power output and provide an output signal correlated with that output, and with an evaluation unit connected to the sensing unit and designed to analyze the output signal of the sensing unit with respect to at least one test criterion and, if the test criterion is met, to briefly interrupt the power output by the power supply, thus constitutes an independently protectable concept. Such an external power supply, unlike other power supplies, is designed to interrupt the power output only briefly, e.g., for a few seconds to a few minutes, e.g., for 20 seconds up to a maximum of 120 seconds, in order to restart an electronic device connected to the power supply during operation.A corresponding 40" power supply is shown schematically in . Figure 8 shown. 10 System 12 Data Router 14 Monitoring Device 16 Power Supply Line 18 Antenna 20 Data Line 22 22' Acquisition Unit 24 Evaluation Unit 26 Reset Line 28 Control Line 30 Switch 32 Neural Network 34 Signal Generator 36 Signal Line (Control Line or Reset Line) 38 Parameter Memory 40 Power Supply 42 Mobile Modem 44 Wired Interface 46 Processor Unit

Claims

1. System (10) comprising a data router (12) and a monitoring device (14) connected to or integrated with the data router, the monitoring device (14) configured to monitor a power consumption of the data router (12) or a component or unit of the data router whose power consumption correlates with the transmission of data and to trigger a restart or reset of the data router (12) depending on its power consumption.

2. System according to claim 1, characterized in that the data router (12) comprises electronic components, a mobile radio modem (42) and a wired interface (44) for connecting further data processing devices via a data line, as well as a power supply unit (40) for supplying the electrical components of the data router with power.

3. System according to claim 1 or 2, characterized in that the monitoring device (14) is adapted to monitor whether the data router (12) or a component (42) or unit of the data router (12) whose power consumption correlates with the transmission of data consumes a predetermined minimum power within a predetermined period of time, and to trigger a restart or a reset of the data router (12) if the power consumed by the data router in the predetermined period of time is less than the predetermined minimum power, wherein the value for the predetermined minimum energy is represented by a variable stored in a parameter memory (38) of the monitoring device (14).

4. System according to at least one of claims 1 to 3, characterized in that the monitoring device (14) comprises a sensing unit (22) in the form of an electronic circuit (22) which is designed to continuously sense the power consumption of the downstream data router (12) or of a component or unit of the data router, power consumption of the downstream data router (12) or of a component or unit of the data router whose power consumption correlates with the transmission of data, and to provide an output signal correlated with the power consumption of the data router (12) or of a component or unit of the data router whose power consumption correlates with the transmission of data.

5. System according to at least claim 4, characterized in that the monitoring device (14) has an evaluation unit (24) which is connected to the sensing unit (22) and is designed to analyze the output signal of the sensing unit (22) with regard to at least one test criterion and to trigger a reset or a restart of the data router (12) if the test criterion is met.

6. System according to at least claim 5, characterized in that the evaluation unit (24) has a neural network (32) and the output signal supplied by the sensing unit (22) comprises data records which represent the respective current temporal course of the power consumption of the data router (12) or of a component or unit of the data router whose power consumption correlates with the transmission of data.

7. Monitoring device (14) for a system according to claims 1 to 6, characterized by a sensing unit (22) in the form of an electronic circuit which is designed to continuously sense the power consumption of the downstream data router (12) or of a component or unit of the data router whose power consumption correlates with the transmission of data and to provide an output signal correlated with the power consumption of the data router (12) or of the component or unit of the data router, the power consumption of which correlates with the transmission of data, the monitoring device (14) also having an evaluation unit (24) which is connected to the sensing unit (22) and is designed to analyze the output signal of the sensing unit (22) with regard to at least one test criterion and to trigger a reset or a restart of the data router (12) if the test criterion is fulfilled, wherein the evaluation unit comprises a neural network (32) configured as a binary classifier.

8. Use of a system according to claims 1 to 6 and / or a monitoring device according to claim7 in a data processing device of a monitoring device of a renewable energy generation plant or in a charging station for electric vehicle.

9. Procedure for monitoring a data router with the steps: - sensing at least one measured value correlated with the power consumption of a data router or a component or unit of the data router whose power consumption correlates with the transmission of data, - Checking the measured value against a reference criterion; and - Restart or reset the router depending on the test result.

10. Method according to claim 9, wherein the at least one measured value correlated with power consumption of a data router or the component or unit of the data router whose power consumption correlates with the transmission of data is the amperage of current drawn by the router or the component or unit of the data router.

11. Method according to claim 9 or 10, wherein checking the measured value against a reference criterion includes comparing the measured value to a reference value.

12. Method according to claim 11, wherein the reference value is the minimum energy or the minimum amount of current that the data router or the component or unit of the data router whose power consumption correlates with the transmission of data consumes in the operable state within a predetermined period of time, a few minutes.

13. Method according to claim 9, wherein the reference criterion is a membership value formed by a neural network trained with appropriate training data sets in the form of a binary classifier to a state class of the data router or the component or unit of the data router whose power consumption correlates with the transmission of data.

14. Method according to at least one of claims 8 to 13, wherein restarting the data router is effected by momentarily interrupting the power supply to the data router, or wherein restarting the data router is effected by delivering a reset signal to the data router.

15. Power supply with - a sensing unit adapted to continuously sense the power delivered and to provide an output signal correlated with the power delivered, and with - an evaluation unit which is connected to the sensing unit and is designed to analyze the output signal of the sensing unit with regard to at least one test criterion, wherein the power supply is configured to cause a momentary interruption of power output by the power supply for a period of time between 5 seconds and 120 seconds when the test criterion is met.