Power Internet of Things traffic filtering method and device, storage medium and electronic equipment
By receiving data packets in the power Internet of Things (IoT) system and judging the list of secure devices and risky devices, and combining IP address verification and intrusion behavior verification, the risky device list is updated in real time using a neural network model. This solves the security risk of unrestricted forwarding of data packets in the power IoT system and improves the system's security and operation and maintenance efficiency.
Patent Information
- Application Number
- CN202411384740.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-30
- Publication Date
- 2024-11-05
AI Technical Summary
The unrestricted forwarding of data packets in existing power Internet of Things (IoT) systems poses security risks. The use of external data gateways and the deployment of additional access control software result in increased maintenance workload, high computing power costs, and poor compatibility, making it difficult to meet the dynamic defense requirements of new power systems.
By receiving data packets from IoT devices, it determines whether the device belongs to the list of secure or risky devices, and performs corresponding IP address verification and intrusion behavior verification. The list of risky devices is updated in real time using a neural network model, thereby enabling the filtering and interception of data packets.
It improves the security of data packet forwarding in IoT systems, solves the security risks of unrestricted forwarding, reduces the number of devices and maintenance pressure, enhances the elastic expansion capability of boundaries, and adapts to the dynamic defense requirements of new power systems.
Smart Images

Figure CN118900207A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of traffic filtering, and more specifically, to a method, device, storage medium and electronic device for traffic filtering in an electric power Internet of Things. Background Art
[0002] New power system businesses such as distributed new energy and load aggregation involve data interaction between multiple entities inside and outside the enterprise. Power system businesses need to operate between multiple entities. When operating power system businesses, the demand for cross-domain circulation of different types of data has increased dramatically. The data of power system businesses presents the characteristics of wide, multiple and frequent data interaction. Therefore, it is urgent to improve the security protection level of cross-domain data circulation under the new power system. For data security protection, relevant technologies focus on the security protection of big data and data assets, and lack security protection mechanisms for dynamic data in circulation. The security protection of data circulation is mainly achieved through external data gateways and deployment of additional access control software.
[0003] However, data security protection methods such as external data gateways and deployment of additional access control software have many problems, such as the addition of new equipment leading to more operation and maintenance work, high computing power overhead and poor compatibility. In the related technology, the access security protection of the power grid on the edge side often adopts security isolation devices, 1-star / 2-star / 3-star gateways and other measures. Such devices belong to external or reinforced modes, which not only increase the number of network devices and increase the operation and maintenance pressure, but also lack the ability to elastically expand the boundaries and it is difficult to provide dynamic defense capabilities. At the same time, this method does not have the function of data security protection. In addition, the edge IoT agents deployed on a large scale on the distribution side have insufficient computing power and do not have their own security protection capabilities. The external or reinforced security protection in the related technology is difficult to meet the growing development needs of new power systems.
[0004] With regard to the problem that the unlimited forwarding of data packets in related technologies poses a security risk, no effective solution has been proposed so far. Summary of the invention
[0005] The main purpose of this application is to provide a method, device, storage medium and electronic device for filtering traffic in an electric power Internet of Things to solve the problem of security risks in the unlimited forwarding of data packets in related technologies.
[0006] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a method for filtering traffic of an electric power Internet of Things is provided. The method comprises: receiving a data packet sent by a target Internet of Things device, judging whether the target Internet of Things device belongs to the list of safe devices based on the information of the data packet; if the target Internet of Things device belongs to the list of safe devices, performing an IP address verification operation on the target Internet of Things device, and forwarding the data packet to a cache database if the IP address verification operation of the target Internet of Things device passes the verification; if the target Internet of Things device does not belong to the list of safe devices, judging whether the target Internet of Things device belongs to the list of risk devices based on the information of the data packet; if the target Internet of Things device does not belong to the list of risk devices, performing an intrusion behavior verification operation and an IP address verification operation on the target Internet of Things device, and forwarding the data packet to a cache database if both the intrusion behavior verification operation and the IP address verification operation of the target Internet of Things device pass the verification; if the target Internet of Things device belongs to the list of risk devices, intercepting the data packet, or, if the target Internet of Things device does not belong to the list of risk devices, if the intrusion behavior verification operation or the IP address verification operation of the target Internet of Things device fails to pass the verification, intercepting the data packet.
[0007] Optionally, the list of risky devices is obtained in the following manner: collecting behavioral data of each IoT device at preset intervals, wherein the behavioral data includes log data and traffic data; normalizing the log data and traffic data to obtain behavioral feature data of the IoT device; inputting the behavioral feature data into a target neural network model to obtain a trust assessment value of the IoT device, wherein the target neural network model is trained by multiple groups of training samples, and each group of training samples includes historical behavioral feature data and historical trust assessment values; determining whether the trust assessment value is less than or equal to a trust assessment value threshold; and adding the IoT device to the list of risky devices if the trust assessment value is less than or equal to the trust assessment value threshold.
[0008] Optionally, before collecting the behavioral data of each IoT device at every preset period, the method also includes: obtaining all data of the IoT device, and extracting basic behavioral characteristics, temporal behavioral characteristics and spatial behavioral characteristics of the IoT device from all the data; wherein the basic behavioral characteristics include at least one of the following: data size, data repetition rate, data consistency, data reliability and response time; the temporal behavioral characteristics include at least one of the following: the duration of the risky behavior and the usage time of the IoT device; the spatial behavioral characteristics include at least one of the following: the location, category and identity of the IoT device; and determining the basic behavioral characteristics, temporal behavioral characteristics and spatial behavioral characteristics as behavioral data.
[0009] Optionally, normalizing the log data and traffic data to obtain behavioral characteristic data of the Internet of Things device includes: structuring the log data to obtain structured log data, parsing the structured log data to obtain an event sequence, wherein the event sequence contains multiple element values, the element values are non-negative integers, and each element value is used to represent an event corresponding to a log in the log data; mapping the element values in the event sequence to a data domain within a first interval range to obtain mapped log data; mapping the traffic data to a data domain within a second interval range to obtain mapped traffic data, wherein the first interval range does not intersect with the second interval range; and filling the mapped log data and the mapped traffic data respectively to obtain behavioral characteristic data of a target length.
[0010] Optionally, the target neural network model is trained in the following manner: for each region, sample data of all IoT devices in the region are collected, and multiple groups of training samples for the region are constructed based on the sample data; a preset neural network model is trained based on the multiple groups of training samples for the region to obtain a sub-neural network model for the region, and model parameters of the sub-neural network model are determined; model parameters of the sub-neural network models of all regions are obtained to obtain a set of model parameters, and the variance of a set of model parameters is calculated to obtain target model parameters; the model parameters of the preset neural network model are adjusted to the target model parameters to obtain the target neural network model.
[0011] Optionally, the target neural network model includes: a mapping module and an analysis module; wherein the mapping module is used to input the behavior feature data into a first preset number of dilated convolutional layers, extract the relationship information in the behavior feature data through the first preset number of dilated convolutional layers, generate a relationship feature tensor based on the relationship information, and process the relationship feature tensor through a preset activation function to obtain a behavior vector matrix, wherein the relationship information is used to characterize the relationship between different types of data in the behavior feature data, and the behavior vector matrix contains multiple behavior vectors, each behavior vector is used to characterize an operation behavior of the Internet of Things device; the analysis module is used to input the behavior vector matrix into a second preset number of target convolutional layers, extract the behavior feature information in the behavior vector matrix through the second preset number of target convolutional layers, generate a behavior feature vector based on the behavior feature information, and input the behavior feature vector into the fully connected layer and the target neural network layer to obtain a trust evaluation value, wherein the size of the perception field of each target convolutional layer is different.
[0012] Optionally, performing an IP address verification operation on a target IoT device includes: sending a response data packet to the target IoT device, and detecting whether a verification data packet of the target IoT device is received, wherein the verification data packet includes signature information of the target IoT device; when a verification data packet of the target IoT device is detected, determining that the IP address verification operation of the target IoT device has been verified; when a verification data packet of the target IoT device is not detected, determining that the IP address verification operation of the target IoT device has not been verified.
[0013] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a power Internet of Things traffic filtering device is provided. The device includes: a receiving unit, which is used to receive a data packet sent by a target Internet of Things device, and judge whether the target Internet of Things device belongs to the list of safe devices based on the information of the data packet; a first execution unit, which is used to perform an IP address verification operation on the target Internet of Things device when the target Internet of Things device belongs to the list of safe devices, and forward the data packet to a cache database when the IP address verification operation of the target Internet of Things device passes the verification; a judgment unit, which is used to judge whether the target Internet of Things device belongs to the list of risk devices based on the information of the data packet when the target Internet of Things device does not belong to the list of safe devices; a second execution unit, which is used to perform an intrusion behavior verification operation and an IP address verification operation on the target Internet of Things device when the target Internet of Things device does not belong to the list of risk devices, and forward the data packet to the cache database when the intrusion behavior verification operation and the IP address verification operation of the target Internet of Things device pass the verification; an interception unit, which is used to intercept the data packet when the target Internet of Things device belongs to the list of risk devices, or, when the target Internet of Things device does not belong to the list of risk devices, if the intrusion behavior verification operation or the IP address verification operation of the target Internet of Things device fails to pass the verification, intercept the data packet.
[0014] Through the present application, the following steps are adopted: receiving a data packet sent by a target IoT device, and judging whether the target IoT device belongs to a list of safe devices based on the information in the data packet; if the target IoT device belongs to the list of safe devices, performing an IP address verification operation on the target IoT device, and if the IP address verification operation of the target IoT device passes, forwarding the data packet to a cache database; if the target IoT device does not belong to the list of safe devices, judging whether the target IoT device belongs to a list of risky devices based on the information in the data packet; if the target IoT device does not belong to the list of risky devices, performing an intrusion behavior verification operation and an IP address verification operation on the target IoT device, and if both the intrusion behavior verification operation and the IP address verification operation of the target IoT device pass, forwarding the data packet to a cache database; if the target IoT device belongs to the list of risky devices, intercepting the data packet, or, if the target IoT device does not belong to the list of risky devices, if the intrusion behavior verification operation or the IP address verification operation of the target IoT device fails to pass, intercepting the data packet, thereby solving the problem of security risks in the unlimited forwarding of data packets in the related art. By setting up a safe device list and a risky device list to filter data packets, only IP address verification is performed on data packets that belong to the safe device list, intrusion behavior verification and IP address verification are performed on data packets that do not belong to the safe device list, and data packets that belong to the risky device list and data packets that fail the intrusion behavior verification operation or IP address verification operation are intercepted, thereby achieving the effect of improving the security of data packet forwarding in the Internet of Things. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0016] Figure 1 It is a flow chart of the power Internet of Things traffic filtering method provided according to an embodiment of the present application;
[0017] Figure 2 It is a schematic diagram of filtering network traffic in an edge IoT proxy according to an embodiment of the present application;
[0018] Figure 3 It is a structural diagram of a terminal dynamic continuous security assessment engine provided according to an embodiment of the present application;
[0019] Figure 4 is a schematic diagram of the structure of a target neural network model provided according to an embodiment of the present application;
[0020] Figure 5is a schematic diagram of a flow filtering device for an electric power Internet of Things provided according to an embodiment of the present application;
[0021] Figure 6 is a schematic diagram of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION
[0022] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0023] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.
[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0025] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for display, data for analysis, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties.
[0026] It should be noted that the information collected is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data comply with the relevant laws, regulations and standards of the relevant regions, take necessary confidentiality measures, do not violate public order and good customs, and provide corresponding operation entrances for users to choose to authorize or refuse.
[0027] In order to solve the problem of security risks in the unlimited forwarding of data packets in related technologies, the following methods can be used:
[0028] According to an embodiment of the present application, a protocol parsing method for traffic filtering based on a data plane development tool is provided, the method comprising: resource initialization, including a timer library, memory resources and a network interface; reading a configuration filter file, building and updating a filter model; collecting and reading data packets from a network interface, cyclically parsing each data packet, matching it with a filter model, filtering data packets that meet the filter conditions, and exporting the retained data packets to a database for storage. The filter model is matched before the corresponding service, and the filter rules are updated in real time, which can simultaneously meet the filter rules of multiple services.
[0029] In order to overcome the problem that fixed filtering rules cannot meet multiple business needs at the same time and the application scenarios are limited, the embodiments of the present application configure the filtering file, build and update the filtering model, collect and read data packets from the network interface, parse each data packet in a loop, match it with the filtering model, filter the data packets that meet the filtering conditions, export the retained data packets to the database for storage, match the filtering model before the corresponding business, update the filtering rules in real time, and can meet the filtering rules of multiple businesses at the same time.
[0030] However, the filtering model in the method provided in the embodiment of the present application needs to be updated manually, which makes it difficult to adapt to the unknown security attack protection requirements caused by the massive data interaction of the new power system. At the same time, this method requires technical support from data plane development tools and has the defect of complex deployment.
[0031] According to another embodiment of the present application, a traffic filtering method, device and storage medium are also provided, including: obtaining a data packet sent by an access terminal to an external service terminal for requesting a connection; judging whether the access terminal belongs to a list of secure devices based on the data packet; if the access terminal belongs to the list of secure devices, granting the access terminal connection rights to the external service terminal; if the access terminal does not belong to the list of secure devices, obtaining an authentication information sequence contained in the data packet; judging whether the authentication information sequence is correct; if the authentication information sequence is correct, adding the access terminal to the list of secure devices. Since an illegal access terminal without an authentication information sequence will have its data packet sent to an external service terminal for requesting a connection intercepted and rejected by the firewall each time, the malicious traffic of the illegal access terminal can be effectively intercepted.
[0032] However, the method provided in the embodiment of the present application only adopts the method of traffic filtering based on the safety device list, and the new power system involves many subjects, and it is difficult to ensure that all subjects are on the safety device list, which can easily cause business stagnation and affect the stable operation of the power grid.
[0033] According to another embodiment of the present application, a method and device for filtering abnormal traffic of an authentication system is also provided, including: in response to receiving an access request to the authentication system, obtaining the message type and target user identifier carried by the access request; in response to the message type matching the preset message type, based on the target binary tree that records the user access information, performing abnormal traffic judgment on the access request to obtain a judgment result; when the judgment result indicates that the access request is an abnormal access request, filtering the access request as abnormal traffic. This method realizes recording the user access information of the user to the authentication system based on the binary tree, and judging the current access request to identify abnormal traffic for filtering based on the user access information, the receiving time of the access request, and the pre-configured filtering judgment threshold for the user access information, thereby effectively reducing the consumption of the authentication system performance by frequent malicious access to the authentication system, and ensuring the safe and stable operation of the authentication system.
[0034] However, the filtering conditions of the method provided in the embodiment of the present application take into account too few factors, and the relevant thresholds rely on manual configuration, and lack the integration of cross-domain information, resulting in defects in the accuracy of malicious traffic filtering, which makes it difficult to adapt to the security protection requirements of cross-domain data interaction in new power systems.
[0035] Based on this, the present application hopes to provide a solution that can solve the above-mentioned technical problems, the details of which will be described in the subsequent embodiments.
[0036] The present invention is described below in conjunction with preferred implementation steps. Figure 1 is a flow chart of a method for filtering electric power Internet of Things traffic according to an embodiment of the present application, such as Figure 1 As shown, the method comprises the following steps:
[0037] Step S101, receiving a data packet sent by a target IoT device, and determining whether the target IoT device belongs to a list of safe devices based on information in the data packet.
[0038] Specifically, the target IoT device can be a device connected to the power IoT system, for example, a distributed new energy vehicle charging pile. The safety device list is a list of devices set by the network administrator based on expert experience. The execution subject of the embodiment of the present application can be an edge IoT agent in the power IoT system. The edge IoT agent receives data packets sent by each network node in the IoT (that is, the target IoT device), passes the data packets to the traffic filtering module based on the safety device list, and determines whether the target IoT device meets the requirements of the safety device list based on the information in the data packet.
[0039] The traffic filtering module based on the list of security devices includes the IP list of the list of security devices and a lookup table for matching, wherein the lookup table records all key information that needs to be compared with the data packet, such as: data packet flag, data packet service type, data packet source IP, data packet destination IP, etc. It is determined whether the information of the data packet matches all the conditions in the lookup table, that is, it is determined whether there is information in the matching table that is the same as the information of the data packet. If so, the data packet can be directly transmitted to the IP verification module; otherwise, the data packet must be delivered to the data packet filtering module based on the risk device list for inspection.
[0040] For example, Figure 2 is a schematic diagram of filtering network traffic in an edge IoT proxy according to an embodiment of the present application, such as Figure 2 As shown, the network traffic forwarded by the target IoT device in the electric power logistics network system is obtained by the edge IoT agent, which has a built-in data processing unit. The traffic filtering module based on the security device list in the data processing unit first filters the data packet.
[0041] Step S102, when the target IoT device belongs to the list of secure devices, an IP address verification operation is performed on the target IoT device. When the IP address verification operation of the target IoT device passes, the data packet is forwarded to the cache database.
[0042] Specifically, Figure 2 As shown in the figure, if the target IoT device belongs to the list of safe devices, the data packet is transferred to the IP verification module, and the IP verification module performs the IP address verification operation. If the IP address verification operation of the target IoT device is verified, the data packet is forwarded to the protected data node in the power IoT system, that is, the cache database. The IP address verification operation is to determine whether the data packet is sent by a real node. If the verification is passed, the data packet is forwarded to the cache database. If the verification is not passed, the data packet is blocked and an alarm is sent to the IDS (Intrusion Detection System).
[0043] Step S103: When the target IoT device does not belong to the list of safe devices, it is determined whether the target IoT device belongs to the list of risky devices based on the information of the data packet.
[0044] Specifically, Figure 2 As shown in the figure, if the target IoT device does not belong to the safe device list, the data packet is passed to the traffic filtering module based on the risk device list, which determines whether the target IoT device belongs to the risk device list. The risk device list is updated in real time by the dynamic continuous security assessment engine.
[0045] For example, when a data packet reaches the traffic filtering module based on the risk device list, it is determined whether the IP and source MAC of the data packet are on the risk device list. If so, it is determined that the target IoT device belongs to the risk device list. If the IP and source MAC of the data packet are not on the risk device list, it is determined that the target IoT device does not belong to the risk device list.
[0046] Step S104, when the target IoT device does not belong to the risk device list, perform intrusion behavior verification operation and IP address verification operation on the target IoT device. When both the intrusion behavior verification operation and the IP address verification operation of the target IoT device pass, forward the data packet to the cache database.
[0047] Specifically, the intrusion behavior verification operation refers to detecting whether there is intrusion behavior on the target IoT device through IDS, such as Figure 2 As shown, if the target IoT device does not belong to the list of risky devices, the data packet will be handed over to the IDS module and the IP verification module for intrusion behavior verification and IP address verification. If both verifications pass, the data packet will be forwarded to the protected data node, that is, the cache database.
[0048] Step S105, if the target IoT device belongs to the risk device list, intercept the data packet, or, if the target IoT device does not belong to the risk device list, if the intrusion behavior verification operation or IP address verification operation of the target IoT device fails to pass the verification, intercept the data packet.
[0049] Specifically, Figure 2 As shown, if the target IoT device belongs to the risk device list, the data packet is directly intercepted. If the target IoT device does not belong to the risk device list, but the intrusion behavior verification operation or IP address verification operation of the target IoT device fails to pass the verification, the data packet is also intercepted.
[0050] It should be noted that the embodiments of the present application focus on the idea of lightweight intrinsic security, utilize the high-speed traffic processing capability of the data processing unit embedded in the edge IoT agent, reuse the existing computing resources such as the edge IoT agent (converged terminal), security gateway, network equipment, etc., to achieve embedded traffic filtering, and solve the problems of a large number of security devices, high cost, difficult operation and maintenance, and increased exposure caused by methods such as hanging data security gateways. At the same time, it breaks through the problem of limited computing power of the edge IoT agent, and achieves accurate filtering of malicious traffic without affecting the efficient and stable operation of the business.
[0051] The electric power Internet of Things traffic filtering method provided in the embodiment of the present application receives a data packet sent by a target Internet of Things device, and determines whether the target Internet of Things device belongs to the list of safe devices based on the information in the data packet; if the target Internet of Things device belongs to the list of safe devices, performs an IP address verification operation on the target Internet of Things device, and if the IP address verification operation of the target Internet of Things device passes, forwards the data packet to a cache database; if the target Internet of Things device does not belong to the list of safe devices, determines whether the target Internet of Things device belongs to the list of risk devices based on the information in the data packet; if the target Internet of Things device does not belong to the list of risk devices, performs an intrusion behavior verification operation and an IP address verification operation on the target Internet of Things device, and if both the intrusion behavior verification operation and the IP address verification operation of the target Internet of Things device pass, forwards the data packet to the cache database; if the target Internet of Things device belongs to the list of risk devices, intercepts the data packet, or, if the target Internet of Things device does not belong to the list of risk devices, if the intrusion behavior verification operation or the IP address verification operation of the target Internet of Things device fails to pass, intercepts the data packet, thereby solving the problem of security risks in the unlimited forwarding of data packets in the related art. By setting up a safe device list and a risky device list to filter data packets, only IP address verification is performed on data packets that belong to the safe device list, intrusion behavior verification and IP address verification are performed on data packets that do not belong to the safe device list, and data packets that belong to the risky device list and data packets that fail the intrusion behavior verification operation or IP address verification operation are intercepted, thereby achieving the effect of improving the security of data packet forwarding in the Internet of Things.
[0052] By constructing a list of risky devices to filter risky data packets, optionally, in the electric power Internet of Things traffic filtering method provided in the embodiment of the present application, the list of risky devices is obtained in the following manner: collecting behavior data of each Internet of Things device at preset intervals, wherein the behavior data includes log data and traffic data; normalizing the log data and traffic data to obtain behavior feature data of the Internet of Things device; inputting the behavior feature data into a target neural network model to obtain a trust evaluation value of the Internet of Things device, wherein the target neural network model is trained by multiple groups of training samples, and each group of training samples includes historical behavior feature data and historical trust evaluation values; judging whether the trust evaluation value is less than or equal to a trust evaluation value threshold; and adding the Internet of Things device to the list of risky devices when the trust evaluation value is less than or equal to the trust evaluation value threshold.
[0053] Specifically, the list of risky devices can be updated in real time by a terminal dynamic continuous security assessment engine based on a neural network. By regularly collecting the behavioral data of each IoT device in the power IoT system, the trust assessment value of the IoT device is determined based on the behavioral data based on the target neural network model, and the risky IoT devices are determined based on the relationship between the trust assessment value and the trust assessment value threshold.
[0054] For example, Figure 3 is a schematic diagram of the structure of a terminal dynamic continuous security assessment engine provided according to an embodiment of the present application, such as Figure 3 As shown in the figure, the terminal dynamic continuous security assessment engine includes a data acquisition module, a data processing module and a credibility assessment module. The data acquisition module integrates an edge IoT agent, which is a device with communication capabilities and can use the local communication interface of the device to extract, aggregate and store the business data of various IoT devices. Therefore, the data acquisition module uses the edge IoT agent to collect the behavior data generated by each IoT device connected to it, and the behavior data includes log data and traffic data.
[0055] The data processing module normalizes the acquired behavior data. Due to the differences in the types of behavior data and source devices, there is no unified format between the log data and traffic data it contains, so format normalization is required to be used for local training of the target neural network model and evaluation of the trust assessment value.
[0056] The credibility assessment module uses the behavioral feature data obtained by the data processing module to train and update the local model parameters, and sends the updated model parameters to the aggregation module. At the same time, it receives the global model parameters from the aggregation module to train the target neural network model. Then, it uses the trained target neural network model to perform credibility scoring on the behavioral data of each IoT device, and adds IoT devices with trust assessment values lower than the trust assessment value threshold to the risk device list for traffic filtering.
[0057] The embodiment of the present application constructs a list of risky devices and updates the list of risky devices in real time through a target neural network model, thereby achieving rapid and accurate screening and blocking of data packets, and ensuring the security and reliability of data interaction in the power Internet of Things system.
[0058] The behavioral data used to evaluate the trust evaluation value needs to be screened out from the massive data of the Internet of Things devices. Optionally, in the power Internet of Things traffic filtering method provided in the embodiment of the present application, before collecting the behavioral data of each Internet of Things device at every preset period, the method also includes: obtaining all data of the Internet of Things devices, and extracting basic behavioral characteristics, temporal behavioral characteristics and spatial behavioral characteristics of the Internet of Things devices from all the data; wherein the basic behavioral characteristics include at least one of the following: data size, data repetition rate, data consistency, data reliability and response time; the temporal behavioral characteristics include at least one of the following: the duration of the risky behavior and the usage time of the Internet of Things device; the spatial behavioral characteristics include at least one of the following: the location, category and identity of the Internet of Things device; the basic behavioral characteristics, temporal behavioral characteristics and spatial behavioral characteristics are determined as behavioral data.
[0059] Specifically, the embodiment of the present application uses the edge IoT proxy to periodically send requests to the IoT device connected to the edge IoT proxy to obtain the behavior data generated by the IoT device in daily activities within a period of time as evaluation data reflecting the behavior and device status of the terminal device during this period. The behavior of IoT devices is diverse and complex, but the data characteristics of abnormal IoT devices must be different from those of normal IoT devices. For example, when a zombie IoT device launches an attack, its main goal is to destroy network devices and tamper with data. Therefore, from network QoS characteristics (Quality of Service, which refers to some key characteristics used to measure and monitor service quality in a network or communication system) such as throughput, availability, success rate, response time, etc., select and refine risk feature factors from network QoS characteristics as basic behavior features to adapt to the risk assessment of internal attack behaviors of malicious IoT devices; in addition, the time-dependent and space-dependent characteristics of the behavior of IoT devices need to be considered to increase the effectiveness of risk assessment and the generalization ability of the assessment model. The time behavior characteristics can include the duration of the risk behavior and the service life of the IoT device, and the spatial behavior characteristics can include the location, category, and identity of the IoT device. Finally, the behavior data used to assess the risk of IoT devices is composed of three parts: basic behavior characteristics, time behavior characteristics, and spatial behavior characteristics.
[0060] Among them, the basic behavior characteristics select factors such as data size, data repetition rate, data consistency, data reliability and response time to measure the basic behavior risk. The malicious attacks on the terminal mainly include theft, tampering with information perception, and injecting error messages. Therefore, data size is one of the important indicators for directly detecting abnormal behavior of IoT devices. The data packet size can be used to determine whether the data content has been tampered with or injected with error messages by malicious terminals. The data repetition rate can effectively determine whether the IoT device exhibits abnormal behavior. When the data repetition rate is low, as the packet repetition rate increases, the direct trust value of the data packet sent from the IoT device will gradually decrease. Due to the poor quality of the data transmission link, the behavior of repeatedly transmitting data packets may be a replay attack or data packet retransmission. Therefore, whether the behavior of the IoT device is normal depends on the growth rate of the repeatedly transmitted data packets. The covariance and correlation coefficient between repeated data packets are calculated to measure the degree of correlation between the data content between the data packets.
[0061] In order to prevent malicious IoT devices from forging data packets, it is necessary to analyze the consistency of data from neighboring IoT devices. In some local areas of the network, the data between neighboring IoT devices is usually highly correlated. Data consistency is the similarity and correlation between a single data and all monitoring data. The data difference corresponds to the difference between the actual received data and the theoretical data. In order to prevent data loss or anomalies caused by attacks or failures of IoT devices, it is necessary to analyze the reliability of data from IoT devices. In the power IoT scenario, the data sensed and collected by IoT devices usually have a certain confidence interval. For example, the confidence interval of the power data collected by distributed photovoltaic power monitoring equipment in a certain area is [548W-552W]. The data reliability is determined by the frequency of the data collected by the terminal appearing in the confidence interval.
[0062] Channel and socket packet sizes are correlated with other variables such as timestamp, network latency, bandwidth, etc. For example, a positive correlation indicates the relationship between the socket packet size and network latency. The amplitude and fluctuation range of the packet indicate that the socket packet size has a large volatility and poor load stability. Based on the time correlation of the malicious behavior of IoT devices, the mean and variance of the number, size, and time interval of packet arrival in the traffic data characteristics of IoT devices are related to the frequency and duration of use of IoT devices. When IoT devices send a large number of packets in a short period of time, it means that the frequency of use of IoT devices is high; when the variance of the packet size sent by IoT devices is small and the number of packets is relatively large, it means that the continuous use time of IoT devices is long. Based on the spatial correlation of IoT device behavior, features such as IP address and MAC address can reflect the location and category of the terminal, such as mobile phones, computers, routers, etc.
[0063] In addition, packet-related features such as sockets, MAC, and IP can reflect the types and quantities of applications and services running on IoT devices, further revealing the identity and category of IoT devices. Therefore, the traffic features of the above IoT devices can be used to calculate the trust evaluation value. By collecting the original network traffic of the target IoT device, considering the network activity level of the IoT device, from the statistical information of the network data traffic, the statistics of the data traffic of the IoT device include packet size, mean, variance, correlation, amplitude, radius, and covariance. Define a time window to describe the time range for capturing network flow statistics, capture data packets from different data layers, obtain the statistical features of the data in the corresponding time window, and form a data set. Based on the collected behavioral data, the terminal dynamic continuous security assessment engine can effectively evaluate the credibility of IoT devices and determine whether the IoT device that generates the data is still trustworthy.
[0064] The embodiment of the present application collects data from IoT devices and screens out behavioral data that can be used to calculate trust evaluation values from massive data, thereby ensuring the accuracy of trust evaluation values and improving the accuracy of screening risky device lists.
[0065] In order to facilitate the training of the target neural network model, it is necessary to normalize the log data and the flow data. Optionally, in the electric power Internet of Things flow filtering method provided in the embodiment of the present application, the log data and the flow data are normalized to obtain the behavioral characteristic data of the Internet of Things device, including: structuring the log data to obtain structured log data, parsing the structured log data to obtain an event sequence, wherein the event sequence contains multiple element values, the element values are non-negative integers, and each element value is used to characterize an event corresponding to a log in the log data; mapping the element values in the event sequence to a data domain within a first interval range to obtain mapped log data; mapping the flow data to a data domain within a second interval range to obtain mapped flow data, wherein the first interval range does not intersect with the second interval range; and filling the mapped log data and the mapped flow data respectively to obtain behavioral characteristic data of a target length.
[0066] Specifically, the behavioral data obtained by the edge IoT agent from the IoT device is heterogeneous data consisting of traffic data and log data. For the target neural network model, the heterogeneous data needs to be normalized. First, the traffic data and log data need to be consistent in form. At the same time, the target neural network model can distinguish whether the data is of traffic type or log type when the data form is consistent. The embodiment of the present application adopts the idea of frequency division multiplexing, maps the traffic data and log data to different data domains respectively, and processes the traffic data and log data accordingly.
[0067] For log data, the log data is first structured, and then the structured log data is parsed. Through parsing, the event sequence of the IoT device can be obtained from the log data. The event sequence is represented by a feature vector whose element value is an integer, and its vector element represents the identifier corresponding to the event, which is a non-negative integer. For traffic data, after obtaining the key statistical data of network traffic, each feature dimension of the traffic data is normalized, and the values of each feature in the feature vector are scaled to the closed interval of [0, 1]. After the above processing, the traffic data is mapped to a data domain composed of decimals, 0 and 1 within the second interval range, such as [0, 1], and the log data is mapped to a non-negative integer data domain within the first interval range, such as (1, +∞). In this way, when the data is input into the target neural network model, the target neural network model can effectively distinguish whether the received data is traffic data or log data through the value range of the feature value. At the same time, since the length of the event sequence is not fixed, and the target neural network model accepts a fixed-length data format, it is necessary to fill the processed traffic data and log data. The data filling can be completed by the following formula:
[0068] (1)
[0069] In the formula, s represents the original data (log data or traffic data), Represents the input vector accepted by the target neural network model, Indicates the maximum length of the input raw data. express Length, Indicates the length Zero-filled fields.
[0070] The embodiment of the present application facilitates training of a target neural network model based on the behavior data by normalizing the behavior data.
[0071] In order to improve the accuracy of the target neural network model in identifying IoT devices with risks, the parameters of the target neural network model are obtained by aggregating the parameters of the sub-neural network models of each region. Optionally, in the electric power IoT traffic filtering method provided in the embodiment of the present application, the target neural network model is trained in the following manner: for each region, sample data of all IoT devices in the region are collected, and multiple groups of training samples for the region are constructed based on the sample data; a preset neural network model is trained based on the multiple groups of training samples for the region to obtain a sub-neural network model for the region, and the model parameters of the sub-neural network model are determined; the model parameters of the sub-neural network models of all regions are obtained to obtain a set of model parameters, and the variance of a set of model parameters is calculated to obtain the target model parameters; the model parameters of the preset neural network model are adjusted to the target model parameters to obtain the target neural network model.
[0072] Specifically, the target neural network model in the embodiment of the present application is applied to the power Internet of Things system. The Internet of Things devices connected to the power Internet of Things system span multiple regions. When training the target neural network model, each region first trains the local sub-neural network model, and finally aggregates the model parameters of the local sub-neural network models of multiple edge Internet of Things agents from various regions through the aggregation module in the terminal dynamic continuous security assessment engine. For example, the variance of the model parameters of each sub-neural network model is calculated, and the variance is used as the global model parameter, that is, the target model parameter. The target model parameter is sent again to the data processing unit of multiple edge Internet of Things agents to improve the accuracy of the target neural network model in evaluating Internet of Things devices.
[0073] The embodiment of the present application obtains the target model parameters by aggregating the model parameters of the local sub-neural network models in each region, and finally sends the target model parameters to the sub-neural network models in each region, thereby improving the accuracy of the trust evaluation value of the Internet of Things device calculated by the target neural network model.
[0074] Optionally, in the electric power Internet of Things traffic filtering method provided in the embodiment of the present application, the target neural network model includes: a mapping module and an analysis module; wherein the mapping module is used to input the behavior feature data into a first preset number of dilated convolutional layers, extract the relationship information in the behavior feature data through the first preset number of dilated convolutional layers, generate a relationship feature tensor based on the relationship information, and process the relationship feature tensor through a preset activation function to obtain a behavior vector matrix, wherein the relationship information is used to characterize the relationship between different types of data in the behavior feature data, and the behavior vector matrix contains multiple behavior vectors, each behavior vector is used to characterize an operation behavior of the Internet of Things device; the analysis module is used to input the behavior vector matrix into a second preset number of target convolutional layers, extract the behavior feature information in the behavior vector matrix through the second preset number of target convolutional layers, generate a behavior feature vector based on the behavior feature information, and input the behavior feature vector into the fully connected layer and the target neural network layer to obtain a trust evaluation value, wherein the size of the perception field of view of each target convolutional layer is different.
[0075] Specifically, the relationship information refers to the similarity of timestamps between log data, the relationship between log categories, the similarity of key information, the similarity of data packet headers and data packet content in traffic data, and other information. The relationship information is used to identify the specific behavior of IoT devices, such as replay attack behavior, small-scale multiple-drag attacks, etc. The target neural network model used in the embodiment of the present application may include a mapping module and an analysis module. The main function of the mapping module is to reconstruct the original behavior feature data and map it to the corresponding information matrix according to the relationship information provided by the data. The data processing process can be divided into extraction and reconstruction: extraction refers to feature extraction of behavior feature data, filtering out redundant and interfering data, and extracting the relationship information between statistical features contained in the data. This part of information is for the detection of subsequent tasks. Reconstruction refers to selecting corresponding behavior vectors to form a behavior sequence based on the extracted relationship information. Through reconstruction, the model can select a given number of behavior vectors from the information matrix to form a behavior sequence and construct the behavior vector matrix required by the subsequent module. The analysis module extracts abnormal information in the behavior of IoT devices from the behavior matrix provided by the mapping module, and performs abnormal detection tasks based on this.
[0076] For example, Figure 4 is a schematic diagram of the structure of the target neural network model provided according to the embodiment of the present application, such as Figure 4As shown, the mapping module extracts the relationship information between different dimensions in the input vector through three (first preset number) dilated convolution layers with different perception field sizes for the behavioral feature data of the input model. Dilated convolution is also called hole convolution. Unlike the standard convolution kernel, dilated convolution adds some holes in the convolution kernel, which can give the target neural network model a larger perception field. The convolution kernels of these three dilated convolution layers are of different sizes, and the number of convolution kernels increases successively, and the degree of dilation also increases successively. After the behavioral feature data passes through the convolution layer, the model can obtain a shape of The relational feature tensor of , where n represents the number of behavior vectors in the subsequent behavior sequence (behavior sequence length), Represents the number of convolution kernels in the previous dilated convolution layer, and also corresponds to the number of behavior vectors in the relationship information matrix.
[0077] After obtaining the relationship feature tensor, The corresponding dimensions (a normalization function) operation, using the position of the maximum value to represent the dimension for unique hot encoding, resulting in a shape of The mapping tensor is used to obtain the corresponding behavior vector from the relationship information matrix. At this point, the target neural network model obtains a shape of The dense matrix (also known as the behavior vector matrix) contains the behavior feature information contained in the behavior feature data, which can effectively reflect the behavior pattern of the IoT device for the target neural network model. The behavior pattern refers to the performance pattern of the operation behavior of the IoT device. For example, several log data represent the operation behavior of an IoT device, and a log data plus a flow data represent the operation behavior of another IoT device.
[0078] The analysis module uses three (the second preset number) target convolution layers with different perception field sizes to convolve the dense matrix obtained in the mapping module to extract the behavioral feature information contained in the dense matrix. The three convolution layers have convolution kernels of different sizes, which enables the model to sense and extract dimensional information of different fields of view and obtain more comprehensive and detailed behavioral features contained in the data, which is conducive to improving the detection effect of the model. Finally, the model combines the feature information obtained from the three different convolution layers to form a feature vector, and then inputs the feature vector into the fully connected layer. After being processed by the fully connected layer and the sigmoid layer (a commonly used neural network layer for binary classification), the trust evaluation value is finally obtained, and the trust evaluation value is used to decide whether to add the IoT device to the risk device list.
[0079] It should be noted that in order to enable the target neural network model to learn more different behavior patterns of IoT devices, the embodiment of the present application analyzes and reconstructs the behavior feature data, converts the data information into a behavior sequence, and uses the newly constructed behavior sequence to represent the various behavior patterns in the data. For example, by maintaining an information matrix , used in the training and learning phase To record the relationship information between behavioral feature data, in the application analysis phase, based on the relationship information extracted from the original data ,from Select n behavior vectors 1 from the list to form a behavior sequence of length 1 [ l 1 , ⋯ , ln ] T , so that we can finally get a shape that can show the data behavior pattern The dense matrix ,in is the dimension of the behavior vector. The specific expression formula for the dilated convolution layer that inputs the behavior feature data is as follows:
[0080] (2)
[0081] in, represents the number of behavior vectors in the information matrix, Represents the input data The relationship information obtained from Represents a neural network consisting of three dilated convolutional layers. The expression formula of operation processing is as follows: (3)
[0082] in, represent function, It is relationship information conduct The intermediate result after function processing. The index vector in the information matrix is extracted by the following formula:
[0083] (4)
[0084] in, represent function, yes go through The index vector of the behavior vector in the information matrix obtained after the function processing. The dense matrix is constructed by the following formula: (5)
[0085] in, Represents the information matrix maintained by the model, according to The index and information matrix in You can get the behavior vector corresponding to the index to form a behavior matrix .
[0086] The embodiment of the present application constructs a neural network model to evaluate the IoT devices, determine the trust evaluation value of the IoT devices, and then screen the risky devices based on the trust evaluation value and the trust evaluation value threshold to add them to the risky device list.
[0087] Optionally, in the electric power Internet of Things traffic filtering method provided in the embodiment of the present application, performing an IP address verification operation on the target Internet of Things device includes: sending a response data packet to the target Internet of Things device, and detecting whether a verification data packet of the target Internet of Things device is received, wherein the verification data packet includes signature information of the target Internet of Things device; when the verification data packet of the target Internet of Things device is detected, determining that the IP address verification operation of the target Internet of Things device has been verified; when the verification data packet of the target Internet of Things device is not detected, determining that the IP address verification operation of the target Internet of Things device has not been verified.
[0088] Specifically, since the use of list technology is extremely susceptible to IP spoofing attacks, network attackers can use forged IP sources to create data packets. In order to solve this problem, the embodiment of the present application constructs an IP verification module to verify whether the access requesting node (that is, the target IoT device) is a real node. For example, when the edge proxy receives the access request, its embedded IP verification module sends a response data packet to the target IoT device. If a valid response from the target IoT device is received, that is, a verification data packet containing signature information, it is considered that the IP address verification operation of the target IoT device has been verified. Otherwise, the relevant data packet is intercepted, and the MAC and IP of the data packet are recorded in the IDS intelligence.
[0089] The embodiment of the present application verifies the IP address of the data packet to ensure that the target IoT device is a real node and avoids IP spoofing attacks.
[0090] The power Internet of Things traffic filtering method provided in the embodiment of the present application filters malicious traffic in the new power communication multi-domain collaborative computing scenario through a lightweight embedded mode, thereby ensuring the data security and network security of the new power system. Taking the demand-side response of distributed new energy and the distributed new energy vehicle charging piles in the vehicle-grid interaction as examples, the demonstration application is summarized as follows: 1) Power demand response. Power demand response is an important supporting means to improve the flexibility of power grid control, promote the consumption of renewable energy and the coordinated optimization of power supply and demand. In this business scenario, frequent interactions of power sales data, energy storage data, and demand response data, as well as multiple communications and negotiations with aggregators are required. At the same time, access to multiple devices is required, and there are risks of data leakage and malicious node threats. The power Internet of Things traffic filtering method provided in the embodiment of the present application is deployed on the edge Internet of Things proxy device to ensure data exchange security and network environment stability, realize the endogenous protection enhancement of new power cross-domain circulation data, and ensure the security and stability of the business.
[0091] 2) Distributed new energy vehicle charging piles. Distributed new energy vehicle charging pile applications are threatened by network attacks. Attackers steal sensitive user information by means of packet capture, eavesdropping, etc., and illegally tamper with the content, which can easily cause changes in billing data, charging card stored value data, and current data, which can easily cause economic losses and safety accidents. The power Internet of Things traffic filtering method provided in the embodiment of the present application realizes network traffic filtering, and the security capabilities of the new power system are built into the border gateway device, preventing malicious users from accessing and stealing data, and improving the security protection capabilities of the new power system communication infrastructure.
[0092] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0093] The embodiment of the present application also provides a power Internet of Things flow filtering device. It should be noted that the power Internet of Things flow filtering device of the embodiment of the present application can be used to execute the power Internet of Things flow filtering method provided in the embodiment of the present application. The power Internet of Things flow filtering device provided in the embodiment of the present application is introduced below.
[0094] Figure 5 Schematic diagram of a flow filtering device for the power Internet of Things provided according to an embodiment of the present application. Figure 5 As shown, the device comprises:
[0095] The receiving unit 501 is used to receive a data packet sent by a target IoT device, and determine whether the target IoT device belongs to a list of safe devices based on information in the data packet;
[0096] The first execution unit 502 is used to perform an IP address verification operation on the target IoT device if the target IoT device belongs to the list of secure devices, and forward the data packet to the cache database if the IP address verification operation of the target IoT device passes;
[0097] A determination unit 503 is used to determine whether the target IoT device belongs to the risk device list based on the information of the data packet when the target IoT device does not belong to the safety device list;
[0098] The second execution unit 504 is used to perform an intrusion behavior verification operation and an IP address verification operation on the target IoT device when the target IoT device does not belong to the risk device list, and forward the data packet to the cache database when both the intrusion behavior verification operation and the IP address verification operation of the target IoT device are verified.
[0099] The interception unit 505 is used to intercept data packets when the target IoT device belongs to the risk device list, or, when the target IoT device does not belong to the risk device list, if the intrusion behavior verification operation or the IP address verification operation of the target IoT device fails to pass the verification, intercept data packets.
[0100] The electric power Internet of Things traffic filtering device provided in the embodiment of the present application receives a data packet sent by a target Internet of Things device through a receiving unit 501, and determines whether the target Internet of Things device belongs to the list of safe devices based on the information in the data packet; the first execution unit 502 performs an IP address verification operation on the target Internet of Things device when the target Internet of Things device belongs to the list of safe devices, and forwards the data packet to a cache database when the IP address verification operation of the target Internet of Things device passes; the judgment unit 503 determines whether the target Internet of Things device belongs to the list of risky devices based on the information in the data packet when the target Internet of Things device does not belong to the list of safe devices; the second execution unit 504 performs an intrusion behavior verification operation and an IP address verification operation on the target Internet of Things device when the target Internet of Things device does not belong to the list of risky devices, and performs an intrusion behavior verification operation and an IP address verification operation on the target Internet of Things device. When the address verification operation is verified successfully, the data packet is forwarded to the cache database; the interception unit 505 intercepts the data packet when the target IoT device belongs to the risk device list, or, when the target IoT device does not belong to the risk device list, if the intrusion behavior verification operation or the IP address verification operation of the target IoT device fails to pass the verification, the data packet is intercepted, thereby solving the security risk problem of unlimited forwarding of data packets in the related technology, filtering the data packets by setting the security device list and the risk device list, performing only IP address verification on the data packets belonging to the security device list, performing intrusion behavior verification and IP address verification on the data packets that do not belong to the security device list, intercepting the data packets belonging to the risk device list and the data packets that have failed the intrusion behavior verification operation or the IP address verification operation, thereby achieving the effect of improving the security of data packet forwarding in the Internet of Things.
[0101] Optionally, in the electric power Internet of Things flow filtering device provided in the embodiment of the present application, the device also includes: a collection unit, used to collect behavior data of each Internet of Things device at preset intervals, wherein the behavior data includes log data and flow data; a processing unit, used to normalize the log data and flow data to obtain behavior feature data of the Internet of Things device; an input unit, used to input the behavior feature data into a target neural network model to obtain a trust evaluation value of the Internet of Things device, wherein the target neural network model is trained by multiple groups of training samples, and each group of training samples includes historical behavior feature data and historical trust evaluation values; a judgment unit, used to judge whether the trust evaluation value is less than or equal to a trust evaluation value threshold; an adding unit, used to add the Internet of Things device to the risk device list when the trust evaluation value is less than or equal to the trust evaluation value threshold.
[0102] Optionally, in the electric power Internet of Things flow filtering device provided in the embodiment of the present application, the device also includes: an acquisition unit, used to acquire all data of the Internet of Things device, and extract basic behavior characteristics, temporal behavior characteristics and spatial behavior characteristics of the Internet of Things device from all the data; wherein the basic behavior characteristics include at least one of the following: data size, data repetition rate, data consistency, data reliability and response time; the temporal behavior characteristics include at least one of the following: the duration of the risky behavior and the usage time of the Internet of Things device; the spatial behavior characteristics include at least one of the following: the location, category and identity of the Internet of Things device; a determination unit, used to determine the basic behavior characteristics, temporal behavior characteristics and spatial behavior characteristics as behavior data.
[0103] Optionally, in the electric power Internet of Things flow filtering device provided in the embodiment of the present application, the processing unit includes: a processing module, which is used to perform structured processing on log data to obtain structured log data, parse the structured log data, and obtain an event sequence, wherein the event sequence contains multiple element values, the element values are non-negative integers, and each element value is used to represent an event corresponding to a log in the log data; a first data mapping module, which is used to map the element values in the event sequence to a data domain within a first interval range to obtain mapped log data; a second data mapping module, which is used to map the flow data to a data domain within a second interval range to obtain mapped flow data, wherein the first interval range does not intersect with the second interval range; a filling module, which is used to fill the mapped log data and the mapped flow data respectively to obtain behavioral characteristic data of a target length.
[0104] Optionally, in the electric power Internet of Things flow filtering device provided in the embodiment of the present application, the target neural network model is trained in the following manner: for each region, sample data of all Internet of Things devices in the region are collected, and multiple groups of training samples for the region are constructed based on the sample data; a preset neural network model is trained based on the multiple groups of training samples for the region to obtain a sub-neural network model for the region, and model parameters of the sub-neural network model are determined; model parameters of the sub-neural network models of all regions are obtained to obtain a set of model parameters, and the variance of a set of model parameters is calculated to obtain target model parameters; the model parameters of the preset neural network model are adjusted to the target model parameters to obtain the target neural network model.
[0105] Optionally, in the electric power Internet of Things flow filtering device provided in the embodiment of the present application, the target neural network model includes: a mapping module and an analysis module; wherein the mapping module is used to input the behavior feature data into a first preset number of dilated convolutional layers, extract the relationship information in the behavior feature data through the first preset number of dilated convolutional layers, generate a relationship feature tensor based on the relationship information, and process the relationship feature tensor through a preset activation function to obtain a behavior vector matrix, wherein the relationship information is used to characterize the relationship between different types of data in the behavior feature data, and the behavior vector matrix contains multiple behavior vectors, each behavior vector is used to characterize an operation behavior of the Internet of Things device; the analysis module is used to input the behavior vector matrix into a second preset number of target convolutional layers, extract the behavior feature information in the behavior vector matrix through the second preset number of target convolutional layers, generate a behavior feature vector based on the behavior feature information, and input the behavior feature vector into the fully connected layer and the target neural network layer to obtain a trust evaluation value, wherein the size of the perception field of each target convolutional layer is different.
[0106] Optionally, in the electric power Internet of Things traffic filtering device provided in the embodiment of the present application, the first execution unit 502 includes: a sending module, used to send a response data packet to the target Internet of Things device, and detect whether a verification data packet of the target Internet of Things device is received, wherein the verification data packet includes the signature information of the target Internet of Things device; a first determination module, used to determine that the IP address verification operation of the target Internet of Things device has been verified if the verification data packet of the target Internet of Things device is detected; and a second determination module, used to determine that the IP address verification operation of the target Internet of Things device has not been verified if the verification data packet of the target Internet of Things device is not detected.
[0107] The electric power Internet of Things flow filtering device includes a processor and a memory. The above-mentioned receiving unit 501, the first execution unit 502, the judgment unit 503, the second execution unit 504 and the interception unit 505 are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.
[0108] The processor contains a kernel, which calls the corresponding program unit from the memory. One or more kernels can be set, and the security of data packet forwarding in the Internet of Things can be improved by adjusting kernel parameters.
[0109] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.
[0110] An embodiment of the present invention provides a computer-readable storage medium having a program stored thereon, which implements a method for filtering flow in an electric power Internet of Things when the program is executed by a processor.
[0111] An embodiment of the present invention provides a processor, which is used to run a program, wherein the power Internet of Things traffic filtering method is executed when the program is running.
[0112] Figure 6 is a schematic diagram of an electronic device provided according to an embodiment of the present application. Figure 6 As shown, the electronic device 601 includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: receiving a data packet sent by a target IoT device, judging whether the target IoT device belongs to the list of safe devices based on the information in the data packet; if the target IoT device belongs to the list of safe devices, performing an IP address verification operation on the target IoT device, and forwarding the data packet to a cache database if the IP address verification operation of the target IoT device passes the verification; if the target IoT device does not belong to the list of safe devices, judging whether the target IoT device belongs to the list of risk devices based on the information in the data packet; if the target IoT device does not belong to the list of risk devices, performing an intrusion behavior verification operation and an IP address verification operation on the target IoT device, and forwarding the data packet to a cache database if both the intrusion behavior verification operation and the IP address verification operation of the target IoT device pass the verification; if the target IoT device belongs to the list of risk devices, intercepting the data packet, or, if the target IoT device does not belong to the list of risk devices, if the intrusion behavior verification operation or the IP address verification operation of the target IoT device fails to pass the verification, intercepting the data packet. The device in this article can be a server, PC, PAD, mobile phone, etc.
[0113] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that initializes the following method steps: receiving a data packet sent by a target IoT device, and judging whether the target IoT device belongs to a list of safe devices based on information in the data packet; if the target IoT device belongs to the list of safe devices, performing an IP address verification operation on the target IoT device, and if the IP address verification operation of the target IoT device passes, forwarding the data packet to a cache database; if the target IoT device does not belong to the list of safe devices, judging whether the target IoT device belongs to a list of risky devices based on information in the data packet; if the target IoT device does not belong to the list of risky devices, performing an intrusion behavior verification operation and an IP address verification operation on the target IoT device, and if both the intrusion behavior verification operation and the IP address verification operation of the target IoT device pass, forwarding the data packet to a cache database; if the target IoT device belongs to the list of risky devices, intercepting the data packet, or, if the target IoT device does not belong to the list of risky devices, intercepting the data packet if the intrusion behavior verification operation or the IP address verification operation of the target IoT device fails to pass.
[0114] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.
[0115] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0116] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0117] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0118] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0119] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.
[0120] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0121] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0122] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.
[0123] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.
Claims
1. A method for filtering traffic in an electric power Internet of Things, characterized in that: include: Receive a data packet sent by a target IoT device, and determine whether the target IoT device belongs to a list of safe devices based on information in the data packet; In the case where the target IoT device belongs to the list of security devices, performing an IP address verification operation on the target IoT device, and in the case where the IP address verification operation of the target IoT device passes, forwarding the data packet to a cache database; In the case where the target IoT device does not belong to the list of safe devices, determining whether the target IoT device belongs to the list of risky devices based on information of the data packet; If the target IoT device does not belong to the risk device list, perform an intrusion behavior verification operation and an IP address verification operation on the target IoT device, and if both the intrusion behavior verification operation and the IP address verification operation of the target IoT device pass the verification, forward the data packet to the cache database; In the case where the target IoT device belongs to the risk device list, the data packet is intercepted; or, in the case where the target IoT device does not belong to the risk device list, if the intrusion behavior verification operation or the IP address verification operation of the target IoT device fails to pass the verification, the data packet is intercepted.
2. The method according to claim 1, characterized in that The risk equipment list is obtained by: Collecting behavior data of each IoT device at a preset period, wherein the behavior data includes log data and traffic data; Normalizing the log data and the traffic data to obtain behavior feature data of the IoT device; Inputting the behavior feature data into a target neural network model to obtain a trust evaluation value of the IoT device, wherein the target neural network model is trained by multiple sets of training samples, each set of training samples including historical behavior feature data and historical trust evaluation values; Determining whether the trust evaluation value is less than or equal to a trust evaluation value threshold; When the trust evaluation value is less than or equal to the trust evaluation value threshold, the Internet of Things device is added to the risk device list.
3. The method according to claim 2, characterized in that Before collecting the behavior data of each IoT device at every preset period, the method further includes: Acquire all data of the IoT device, and extract basic behavior characteristics, temporal behavior characteristics, and spatial behavior characteristics of the IoT device from all the data; The basic behavior characteristics include at least one of the following: data size, data repetition rate, data consistency, data reliability and response time; The time behavior feature includes at least one of the following: the duration of the risk behavior and the usage time of the IoT device; The spatial behavior feature includes at least one of the following: the location, category and identity of the IoT device; The basic behavior feature, the temporal behavior feature and the spatial behavior feature are determined as the behavior data.
4. The method according to claim 2, characterized in that: The log data and the flow data are normalized to obtain the behavior characteristic data of the IoT device, including: Structuring the log data to obtain structured log data, parsing the structured log data to obtain an event sequence, wherein the event sequence includes a plurality of element values, the element values are non-negative integers, and each element value is used to represent an event corresponding to a log in the log data; Mapping the element values in the event sequence to a data domain within a first interval to obtain mapped log data; Mapping the flow data to a data domain within a second interval range to obtain mapped flow data, wherein the first interval range does not intersect with the second interval range; The mapped log data and the mapped traffic data are filled respectively to obtain behavior feature data of a target length.
5. The method according to claim 2, characterized in that: The target neural network model is trained in the following way: For each region, sample data of all IoT devices in the region are collected, and multiple sets of training samples for the region are constructed based on the sample data; Training a preset neural network model based on multiple groups of training samples in the region to obtain a sub-neural network model of the region, and determining model parameters of the sub-neural network model; Obtaining model parameters of the sub-neural network models of all regions to obtain a set of model parameters, calculating the variance of the set of model parameters to obtain target model parameters; The model parameters of the preset neural network model are adjusted to the target model parameters to obtain the target neural network model.
6. The method according to claim 2, characterized in that The target neural network model includes: a mapping module and an analysis module; The mapping module is used to input the behavior feature data into a first preset number of dilated convolutional layers, extract the relationship information in the behavior feature data through the first preset number of dilated convolutional layers, generate a relationship feature tensor based on the relationship information, and process the relationship feature tensor through a preset activation function to obtain a behavior vector matrix, wherein the relationship information is used to characterize the relationship between different types of data in the behavior feature data, and the behavior vector matrix includes multiple behavior vectors, each behavior vector is used to characterize an operation behavior of the Internet of Things device; The analysis module is used to input the behavior vector matrix into a second preset number of target convolution layers, extract the behavior feature information in the behavior vector matrix through the second preset number of target convolution layers, generate a behavior feature vector based on the behavior feature information, input the behavior feature vector into a fully connected layer and a target neural network layer, and obtain the trust evaluation value, wherein the size of the perception field of each target convolution layer is different.
7. The method according to claim 1, characterized in that Performing IP address verification operations on the target IoT device includes: Sending a response data packet to the target IoT device, and detecting whether a verification data packet of the target IoT device is received, wherein the verification data packet includes signature information of the target IoT device; In the case where a verification data packet of the target IoT device is detected, determining that an IP address verification operation of the target IoT device has been verified; In the case where no verification data packet of the target IoT device is detected, it is determined that the IP address verification operation of the target IoT device has not been verified.
8. A flow filtering device for electric power Internet of Things, characterized in that: include: A receiving unit, configured to receive a data packet sent by a target IoT device, and determine whether the target IoT device belongs to a list of safe devices based on information in the data packet; A first execution unit is configured to perform an IP address verification operation on the target IoT device if the target IoT device belongs to the security device list, and forward the data packet to a cache database if the IP address verification operation of the target IoT device passes; A judging unit, configured to judge whether the target IoT device belongs to a risky device list based on information of the data packet when the target IoT device does not belong to the safe device list; A second execution unit is configured to, if the target IoT device does not belong to the risk device list, perform an intrusion behavior verification operation and an IP address verification operation on the target IoT device, and if both the intrusion behavior verification operation and the IP address verification operation of the target IoT device are verified to pass, forward the data packet to the cache database; The interception unit is used to intercept the data packet when the target IoT device belongs to the risk device list, or, when the target IoT device does not belong to the risk device list and the intrusion behavior verification operation or IP address verification operation of the target IoT device fails to pass the verification, intercept the data packet.
9. A non-volatile storage medium, characterized in that: The non-volatile storage medium includes a stored program, wherein when the program is running, the device where the non-volatile storage medium is located is controlled to execute the power Internet of Things traffic filtering method described in any one of claims 1 to 7.
10. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the electric power Internet of Things traffic filtering method described in any one of claims 1 to 7.
Citation Information
Patent Citations
Internet of things DDoS attack defense method, device and system and storage medium
CN109905361A
Security protection method and device of Internet of Things equipment, equipment and storage medium
CN112769847A
Electric power Internet of Things security defense terminal and control system thereof
CN116938507A
Network intrusion attack identification system and method based on Internet of Things
CN118337540A
Network attack interception method, computer device and computer readable storage medium
CN118353722A
Cited By
Access control method and device of power system, equipment, storage medium and program product
CN121000420A
