Controlled equipment protection method based on edge gateway

Abstract

The invention discloses a controlled device protection method based on an edge gateway, the protection method is realized based on an Internet of Things system composed of an Internet of Things platform, the edge gateway and a sensing layer device, the edge gateway receives a control instruction of the Internet of Things platform, and issues the control instruction to the sensing layer device to realize control of the device; a control instruction analyzing and processing unit is arranged in the edge gateway, on the basis, the protection method comprises the steps that the control instruction analyzing and processing unit in the edge gateway divides a control instruction into a low-frequency instruction and a high-frequency instruction according to the control frequency of the control instruction, and different coping strategies are adopted respectively; according to the invention, the device control instruction from each terminal can be analyzed, the control frequency of the controlled device is reduced, useless instructions are filtered, and an abnormal condition is judged to carry out active early warning under the condition that the normal control real-time performance of a user is not influenced.

Description

Technical Field

[0001] The present invention belongs to the technical field of Internet of Things system control methods, and relates to a controlled device protection method based on an edge gateway. Background Art

[0002] In the fields of smart buildings, smart parks, smart hospitals, etc., the IoT system is generally composed of cloud, edge, and end. The cloud generally refers to the platform, the edge refers to the edge gateway, and the end refers to the perception layer device. The edge gateway is located between the IoT platform and the perception layer device, and plays a connecting role when the platform accesses the terminal device. The edge gateway is connected to a variety of terminal devices in the south, such as curtain motors, air conditioning thermostats, lighting relays, etc.; it is connected to the IoT platform, edge interactive screens, third-party systems, etc. in the north, such as the attached Figure 1 As shown in the figure, the edge gateway receives control instructions from the north, analyzes the instructions, converts the protocols, and sends them to the corresponding controlled devices in the south, thus completing the control of the devices.

[0003] When multiple users control the same device at the same time, the platform device linkage logic is not configured properly, the edge interactive screen is abnormal, the user maliciously controls, the program triggers a bug, hackers attack, etc., a large number of device control instructions may be generated in a short period of time and sent to the controlled device through the edge gateway for execution. This may cause abnormal situations such as frequent operation of curtain motors, frequent switching of lamp relay modules, and frequent switching of air valves and water valves of air conditioners, causing the controlled modules and controlled devices to heat up and crash, affecting the service life of the equipment. More intuitively, lamps may turn on and off quickly, causing trouble to users.

[0004] Another situation is that after the edge gateway goes offline and comes back online, the accumulated northbound commands will be sent to the edge gateway in a short period of time, which will also cause the edge gateway to perform some repeated control on the controlled device, such as a light being turned on, then turned off, and then turned on again; the above situation will also occur when multiple users control the same device in a short period of time.

[0005] Therefore, when a large number of device control instructions are generated due to malicious human manipulation or program abnormalities, it is necessary to design a protection buffer method for the controlled device to analyze and filter the repeated instructions made by the edge gateway to the controlled device to reduce the situation where the controlled device is repeatedly or repeatedly controlled. Summary of the invention

[0006] In order to improve the life of the controlled equipment and reduce interference to users, the present invention proposes a command analysis and filtering method based on an edge gateway, which can analyze the device control commands from each end, reduce the number of controls on the controlled equipment without affecting the real-time control of the user, filter useless commands, and judge abnormal situations for active warnings.

[0007] The technical solution of the present invention is a controlled device protection method based on an edge gateway. The above protection method is implemented in an Internet of Things system composed of an Internet of Things platform, an edge gateway, and a perception layer device. The above edge gateway receives the control instructions of the Internet of Things platform and issues them to the above perception layer device to realize the control of the device. The key is to set a control instruction analysis and processing unit in the above edge gateway. On this basis, the above protection method is that the control instruction analysis and processing unit in the edge gateway divides the control instructions into low-frequency instructions and high-frequency instructions according to the control frequency of the control instructions, and adopts different response strategies respectively:

[0008] For the above-mentioned low-frequency instructions, a cache table file timing analysis strategy is implemented, an early warning threshold is set, and the data in the cache table is analyzed regularly. When the number of operations on the same device reaches the above-mentioned early warning threshold, an alarm is triggered;

[0009] For the above high-frequency instructions, an analysis and filtering strategy is adopted, specifically:

[0010] S1. Query the record of the controlled device ID in the cache table, and put the instructions with the same device ID and the same attributes into the buffer;

[0011] S2, delaying and / or merging the control instructions in the buffer;

[0012] S3. When the instruction execution time is up, issue the instruction.

[0013] The above-mentioned S2 step is specifically:

[0014] S2-1. Analysis of control instructions for waiting queues in the buffer:

[0015] When the waiting queue is empty, query the instruction waiting time in the cache strategy table according to the number of records queried in the cache table, save the current control record to the cache table, calculate the instruction execution time according to the instruction waiting time and the current time, and enter step S3;

[0016] When the waiting queue is not empty, the instructions with the same device ID and the same attributes are executed according to step S2-2;

[0017] S2-2. Analyze the attribute values:

[0018] If the attribute values ​​are the same, keep the control record in the cache table, discard the control instruction, do not enter step S3, and end;

[0019] If the attribute values ​​are different, the control record is kept in the cache table, and the control instruction is used to update the existing control instruction in the buffer waiting queue. After the update, the original waiting time is followed and the waiting time is waited for issuance.

[0020] Furthermore, saving this control instruction includes saving and recording the device ID, registration time, attributes, attribute values ​​and record source in the control instruction; the above query is based on three dimensions: device ID, registration time < current time-period, and attributes. Records that meet all conditions are valid records.

[0021] Furthermore, the above protection method also includes periodically refreshing the cache strategy table to retain valid records within a period.

[0022] Preferably, the value of the waiting time in the above cache strategy table is generated by Formula 1:

[0023] T=a×n 2 +b Formula 1

[0024] In Formula 1, T represents the instruction waiting time, a and b are parameters set according to the sensitivity requirements, and n is the number of valid cache entries with the same device ID and the same attributes within a cycle. The cycle can be 60 seconds, 120 seconds, 180 seconds, or 240 seconds.

[0025] More specifically, the specific process of setting the warning threshold mentioned above is: setting the analysis period of low-frequency instructions and the warning threshold of the same device ID, regularly analyzing the device control records of the most recent record analysis period, and triggering an alarm when the number of device control records exceeds the warning threshold; the above-mentioned warning threshold is 80 to 200 times; the above-mentioned regular analysis period is 60 seconds, 120 seconds, 180 seconds or 240 seconds.

[0026] More specifically, the above-mentioned record analysis period is 12 hours or 24 hours.

[0027] It should be clarified that the above-mentioned low-frequency instructions are monitored and analyzed in a cycle of 1 day or half a day. When the same attribute of the same device is controlled multiple times, the intervals between two adjacent instructions are above the segmentation threshold; the above-mentioned high-frequency instructions are instructions that are below the segmentation threshold when the same attribute of the same device is controlled multiple times. For the above-mentioned high-frequency instructions, real-time processing and filtering are performed; the above-mentioned segmentation threshold is any value between 3 seconds and 6 seconds.

[0028] Preferably, the above segmentation threshold is 4 seconds.

[0029] The beneficial effects of the present invention are:

[0030] The present invention divides the control instructions issued by the edge gateway to the controlled device into low-frequency instructions and high-frequency instructions, and adopts different response strategies respectively. For high-frequency instructions, an analysis and filtering strategy is adopted to establish a buffer mechanism, that is, to analyze the instruction issuance situation in real time, and discard or merge and delay the instructions to be issued. The present invention can filter out the records of instructions to be issued for the same device and the same attribute in a short time. When issuing an instruction to a certain device for the first time in a short time, no delay is made to the instruction, so that the real-time control of the user device can be guaranteed; when issuing an instruction to the device for the second time in a short time, a short wait will be made before issuing, that is, delayed issuance; if a third or more issuance requests are received during the waiting period, these instructions will be merged into one (if the attribute values ​​are the same, the later instructions will be discarded, because the same instruction can be issued to the controlled device to achieve the effect. If the attribute values ​​are different, only the latest instructions will be retained. Because the new instruction control action can replace the action of the old instruction, there is no need to issue the old instruction again). In a short time but outside the waiting period, the present invention will read the corresponding waiting time from the cache strategy table according to the number of issuances to wait.

[0031] A cache table file timing analysis strategy is adopted for low-frequency instructions. By regularly analyzing the instructions registered in the cache table file, if the number of device operations in 12 hours or 24 hours reaches a certain threshold, an early warning message is triggered.

[0032] The present invention reduces the number of times of controlling the controlled device as much as possible, filters useless instructions, and proactively warns of abnormal situations without affecting the real-time performance of normal user control, thereby increasing the life of the controlled device and reducing interference with the user. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a schematic diagram of the cloud-edge structure.

[0034] Figure 2 Distribute flow chart for control instructions.

[0035] Figure 3 Dispatches streams for buffer control instructions. DETAILED DESCRIPTION

[0036] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] The following is an explanation of some of the names involved in the present invention:

[0038] Device: refers to a specific controlled device, such as a lamp or an air conditioner.

[0039] Attribute: refers to a specific function of a device. For example, a light generally has on / off attributes and dimming attributes.

[0040] Attribute value: refers to an optional operation on an attribute. For example, the switch attribute of a light has two attribute values: on and off, and the brightness attribute of a light has a brightness value in the range of 0-100. For example, turning on a light means writing the switch attribute of the light device to the value on.

[0041] Cache table and cache table file: The cache table file is a file that records all the instructions that need to be sent through the edge gateway. When the program starts, part of the cache table file data is read into the memory, which is called the cache table. The cache table is used to record the instruction information recently sent to the controlled device, and expired data in the table will be cleared regularly. For an example of the cache table, see Table 1.

[0042] Cache strategy table: used to save cache strategies. You can adjust cache strategy data by setting parameters to affect the response time of the device being controlled multiple times and the level of instruction filtering. When the program starts, the cache strategy table will be read into the memory. For an example of the cache strategy table, see Table 2.

[0043] Example

[0044] In this implementation, when the same attribute of the same device is controlled multiple times in one day, the control instructions with an interval of more than 4 seconds between two adjacent instructions are regarded as low-frequency instructions, and the control instructions with an interval of less than 4 seconds between two adjacent instructions are regarded as high-frequency instructions, and 4 seconds is the segmentation threshold of the instruction control frequency. Half a day can also be selected as the analysis period, and other segmentation thresholds can be selected within the range of 3 seconds to 6 seconds according to actual conditions.

[0045] The response strategy for low-frequency instructions is to analyze the data in the cache table regularly and issue active alarms. For example, the cache table file is analyzed every 60 seconds, and the device control records of the past 12 hours are analyzed. If the control of the same device exceeds 100 times in the record, an alarm will be triggered; the query interval and the alarm threshold are configurable. You can select 120 seconds, 180 seconds or 240 seconds as the regular analysis cycle according to actual needs; you can select the record analysis cycle of analyzing 24 hours of device control records according to actual needs; you can also select a suitable alarm threshold within the range of 80 to 200 times according to actual needs.

[0046] For high-frequency instructions, the response strategy is to analyze the instruction issuance in real time, discard or merge, and delay the instructions to be issued, that is, filter the instructions. Figure 2 and attached Figure 3 As shown, Figure 2It is a control instruction distribution flow chart. The edge gateway receives the control instruction from the platform or mobile terminal or interactive screen terminal. After parsing the message and assembling the target instruction, the instruction distribution task is sent to the device management thread of the edge gateway. After the thread receives the task, it is attached. Figure 1 The entry point of the process ( Figure 1 The process determines whether the instruction is issued immediately or enters the cache table information. Figure 2 process), ready to issue control instructions.

[0047] Figure 3 It is the buffer instruction distribution process, mainly used to process Figure 1 The instructions that need to be filtered are transmitted, and the waiting time of this instruction is obtained in combination with the cache strategy table. The messages in the waiting queue are analyzed to determine whether this instruction needs to be discarded, and the execution time is calculated to be inserted or updated into the waiting queue. Once the execution time of the message in the waiting queue reaches the execution time, the command will be issued to the controlled device.

[0048] Table 1 is an example of a cache table, which exists in the memory and synchronizes the data to the cache table file. The cache table is used to record all recent pending instructions of the edge gateway. Index is its self-incrementing counter. The device ID uniquely identifies a specific device. The execution time is the time when the record is registered. The attribute is the attribute corresponding to the instruction. The attribute value is the value to be written to the attribute. The record source is a record made at another location in the process, which is used for later data analysis.

[0049] Table 2 is an example of a cache strategy table. The data in the cache strategy table is given by formula 1: T = a × n 2 +b is generated, where T represents the instruction waiting time; a and b are parameters set according to sensitivity requirements. a can adjust the control sensitivity, and b can fine-tune the control sensitivity during normal operation. The greater the sensitivity, the greater the delay in controlling the device multiple times, and the greater the protection of the device; n is the number of valid cache entries for records with the same device ID and the same attributes within a cycle, that is, the number of valid cache entries for records with the same device and the same attributes within a cycle. This cycle can be 60 seconds, 120 seconds, 180 seconds or 240 seconds, and can also be set according to actual conditions.

[0050] In this implementation, a=100, b=-100 are selected as recommended parameters, and the period n is 60 seconds.

[0051] According to Formula 1, within a period of 60 seconds, each time a command with the same device and attributes is issued, the waiting time for the first three commands is 0 milliseconds, 300 milliseconds, and 800 milliseconds respectively.

[0052] Table 1: Cache table example

[0053] Index Device ID Registration Time property Property Value Record source 1 d01 1724299201070 PowerSwitch 1 1 2 d01 1724299201270 PowerSwitch 0 2 3 d01 1724299201570 PowerSwitch 1 3 4 d02 1724299201770 Brightness 80 1

[0054] Table 2: Example of cache strategy table

[0055]

[0056] Table 2: Example of cache strategy table

[0057]

[0058] The following is an example of the process of the protection method for the high-frequency instruction part (a=100, b=-100, cycle is 60 seconds):

[0059] S1. Query the records in the waiting execution queue in the cache table and save the current control instruction in the cache table to obtain the instruction record number, and put the instructions with the same device ID and the same attributes into the buffer.

[0060] As shown in the cache table example in Table 1, the cache table is queried based on the device ID to filter out records with the same attributes within the valid time. The period n configured this time is 60 seconds. When querying, the three dimensions of device ID, registration time < current time - 60 seconds, and attributes are queried. Records that meet all the conditions are valid records.

[0061] When saving valid records, the device ID, registration time, attributes, attribute values ​​and record source of the control command need to be recorded. Among them, the device ID can uniquely identify a specific controlled device, the registration time records the current timestamp, and the attributes and attribute values ​​identify a specific action of this command on the device.

[0062] S2. Delay or / and merge the control instructions in the buffer to filter the instructions with the same device ID and the same attributes:

[0063] The instructions in the waiting queue are all waiting for the execution time to arrive, which is equivalent to a waiting area for instructions to be issued.

[0064] S2-1. Analysis of control instructions for waiting queues in the buffer:

[0065] When the waiting queue is empty, the instruction waiting time is queried in the cache strategy table according to the number of records queried in the cache table, the control record is saved to the cache table, the instruction execution time is calculated according to the instruction waiting time and the current time, and the step S3 is entered.

[0066] When the waiting queue is not empty, the instructions with the same device ID and the same attributes are executed according to step S2-2.

[0067] S2-2. Analyze the attribute values:

[0068] If the attribute values ​​are the same, the control record is kept in the cache table, the control instruction is discarded, and step S3 is not entered, and the process ends.

[0069] If the attribute values ​​are different, the control record is kept in the cache table, and the control instruction is used to update the existing control instruction in the buffer waiting queue. After the update, the original waiting time is followed and the waiting time is waited for issuance.

[0070] The cache strategy table is shown in Table 2. If the number of valid records found in the cache table last time is 3, and this instruction is the fourth instruction in the cycle, the waiting time found in combination with the parameters is 1500 milliseconds.

[0071] S3. When the instruction execution time is up, issue the instruction.

[0072] The above protection method also includes a cache table timed refresh strategy:

[0073] The cache table in the memory will be refreshed relatively quickly. In this embodiment, when the cycle is 60 seconds, the cache table in the memory will be refreshed every 5 minutes, and only valid records within the cycle will be retained to improve the query efficiency of the cache table in the memory.

[0074] For the low-frequency instructions, the control strategy using the above high-frequency instructions is not effective, so another coping strategy is used to implement the cache table file timing analysis strategy:

[0075] Newly added records in the cache table will be automatically synchronized to the cache table file.

[0076] For low-frequency control command analysis, the cache table file will be analyzed every 12 hours, and the device control records of the past 12 hours will be analyzed. If the control of the same device exceeds 100 times in the record, an early warning will be triggered; the query interval and early warning threshold are configurable.

[0077] Based on this embodiment, the following is a specific description of the switch control of a lamp at intervals:

[0078] The light is the controlled device. When the edge gateway program is started and the buffer mechanism receives the first control instruction, it is called instruction 1 (the subsequent instructions n are all instructions with the same device and attributes). The process to be followed is as follows. Figure 2 shown.

[0079] Issue command 1, attribute value is on. There is no relevant record in the query cache table, continue to register the command, save the record of this control to the cache table, issue the command immediately, and the light will be turned on immediately;

[0080] After 1 second, command 2 is issued, and the attribute value is off. Command 2 finds that there is a valid record and enters the attached Figure 3The buffer process shown. Because there is one valid record in the cache table, the waiting time when querying the cache strategy table N=2 is 300 milliseconds; then save this control record to the cache table, and add the current time plus the waiting time as the execution time to the waiting queue for execution. When the execution time is up, the command is issued and the light will be turned off.

[0081] After 1 second, command 3 is issued, and the attribute value is on. Command 3 finds that there are 2 valid records, and the queue is empty (1 second is enough for the previous command with a delay of 300 milliseconds to be issued). The waiting time for querying the cache strategy table is 800 milliseconds, so command 3 will be issued after a delay of 800 milliseconds, and the light will be turned on;

[0082] After 1 second, command 4 is issued, and the attribute value is off. The process of command 4 is the same as command 3. Command 4 will be issued after a delay of 1500 milliseconds.

[0083] After 1 second, command 5 is issued, and the attribute value is on. Command 5 finds that there are 4 valid records, and the waiting queue is not empty. After analyzing and filtering the records, it is found that command 4 is still waiting to be issued in the queue. The operation of command 4 is to turn off the light, and command 5 is to turn on the light. At this time, the attribute value of command 5 will be updated to the queue and replace the attribute value of command 4. In other words, turn off the light and turn it on; that is, update the command of command 4, the execution time remains unchanged, and command 5 is discarded; after 500 milliseconds, the light will not be turned off.

[0084] After 1 second, instruction 6 is issued, and the attribute value is off. Instruction 6 finds that there are 5 valid records, and the queue is empty. Instruction 6, which queries the cache strategy table, will be issued after a delay of 3500 milliseconds.

[0085] After 1 second, command 7 is issued, and the attribute value is off. The waiting queue is not empty, and the records are analyzed and filtered. It is found that command 6 is still waiting to be issued in the queue. The operation of command 6 is to turn off the light, and command 7 is also to turn off the light. They are duplicate commands, so command 7 is directly discarded after being registered.

[0086] In this way, the filtering effect of high-frequency instructions is shown in Table 3.

[0087] Table 3: Effective filtration verification results

[0088]

[0089] Table 3: Results of effective filtration verification

[0090]

[0091] (Note: Table 3 Bold line Some of them are instructions that are discarded by filtering)

[0092] Table 3 illustrates the filtering effect of the recommended parameters according to the cache strategy table in Table 2.

[0093] Table 3 lists the control instructions for sending different attribute values ​​of the same device and the same attribute at intervals of 0.2 seconds, 1 second, 2 seconds, and 4 seconds. The frequency interval and theoretical execution time are recorded on the time axis. The theoretical execution time is the time from the instruction Figure 1 The flowchart is input to the time point when the command can be sent to the controlled device. The bold line part is the command that is filtered out.

[0094] As shown in Table 3, within a 60-second period, for high-frequency control of 0.2 seconds, the protection method of the present invention filters at least 80% of the instructions; 1 second is 66%, 2 seconds is 50%, and 4 seconds is 33%. It can be seen that the present invention achieves filtering of high-frequency instructions, and reduces the control times of the controlled device without affecting the normal control of the user.

Claims

1. A controlled device protection method based on an edge gateway, wherein the protection method is implemented in an Internet of Things system consisting of an Internet of Things platform, an edge gateway, and a perception layer device, wherein the edge gateway receives control instructions from the Internet of Things platform and issues them to the perception layer device to control the device, and wherein: A control instruction analysis and processing unit is set in the edge gateway. On this basis, the protection method is that the control instruction analysis and processing unit in the edge gateway divides the control instructions into low-frequency instructions and high-frequency instructions according to the control frequency of the control instructions, and adopts different response strategies respectively: For the low-frequency instructions, a cache table file timing analysis strategy is implemented, an early warning threshold is set, and the data in the cache table is analyzed regularly. When the number of operations on the same device reaches the early warning threshold, an alarm is triggered; For the high-frequency instructions, an analysis and filtering strategy is adopted in real time, specifically: S1. Query the record of the controlled device ID in the cache table, and put the instructions with the same device ID and the same attributes into the buffer; S2, delaying and / or merging the control instructions in the buffer; S3. When the instruction execution time is up, issue the instruction.

2. According to claim 1, a controlled device protection method based on an edge gateway is characterized in that: The S2 step is specifically: S2-1. Analysis of control instructions for waiting queues in the buffer: When the waiting queue is empty, query the instruction waiting time in the cache strategy table according to the number of records queried in the cache table, save the current control record to the cache table, calculate the instruction execution time according to the instruction waiting time and the current time, and enter the S3 step; When the waiting queue is not empty, the instructions with the same device ID and the same attributes are executed according to step S2-2; S2-2. Analyze the attribute values: If the attribute values ​​are the same, keep the control record in the cache table, discard the control instruction, do not enter step S3, and end; If the attribute values ​​are different, the control record is kept in the cache table, and the control instruction is used to update the existing control instruction in the buffer waiting queue. After the update, the original waiting time is followed and the waiting time is waited for issuance.

3. A method for protecting a controlled device based on an edge gateway according to claim 2, characterized in that: Saving this control instruction includes saving and recording the device ID, registration time, attributes, attribute values ​​and record source in the control instruction; the query is based on three dimensions: device ID, registration time < current time-period, and attributes. Records that meet all conditions are valid records.

4. A controlled device protection method based on edge gateway according to claim 2, characterized in that: The protection method also includes regularly refreshing the cache strategy table to retain valid records within a period.

5. A controlled device protection method based on edge gateway according to claim 3 or 4, characterized in that: The value of the waiting time in the cache strategy table is generated by Formula 1: T=a×n 2 +b Formula 1 In Formula 1, T represents the instruction waiting time, a and b are parameters set according to the sensitivity requirements, and n is the number of valid cache entries with the same device ID and the same attributes within a cycle. The cycle can be 60 seconds, 120 seconds, 180 seconds, or 240 seconds.

6. The method for protecting a controlled device based on an edge gateway according to claim 1, characterized in that: The specific process of setting the warning threshold is: setting the timing analysis period of low-frequency instructions and the warning threshold of the same device ID, timing analysis of the device control records of the most recent record analysis period, and triggering an alarm when the number of device control records exceeds the warning threshold; the warning threshold is 80 to 200 times; the timing analysis period is 60 seconds, 120 seconds, 180 seconds or 240 seconds.

7. A method for protecting a controlled device based on an edge gateway according to claim 6, characterized in that: The recording and analysis cycle is 12 hours or 24 hours.

8. The method for protecting a controlled device based on an edge gateway according to claim 1, characterized in that: The low-frequency instructions are monitored and analyzed in a cycle of 1 day or half a day. When the same attribute of the same device is controlled multiple times, the intervals between two adjacent instructions are above the segmentation threshold. The high-frequency instructions are instructions when the same attribute of the same device is controlled multiple times, and the intervals between two adjacent instructions are below the segmentation threshold. Real-time processing and filtering are performed on the high-frequency instructions. The segmentation threshold is any value between 3 seconds and 6 seconds.

9. A method for protecting a controlled device based on an edge gateway according to claim 8, characterized in that: The segmentation threshold is 4 seconds.

Images