A method and system for communication security protection in the power Internet of Things
By working together with edge security proxy devices and cloud security analysis platforms, the problem of cross-layer collaborative attacks being difficult to detect in a timely manner in power Internet of Things (IoT) communication security protection has been solved. This has enabled cross-layer collaborative communication security protection, reduced terminal authentication latency, and met the latency requirements of real-time power control services.
Patent Information
- Application Number
- CN202611581788.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-13
- Publication Date
- 2026-11-17
AI Technical Summary
Existing power Internet of Things (IoT) communication security protection methods implement threat detection independently at each level, making it difficult to detect cross-layer collaborative attacks in a timely manner and failing to form a cross-layer collaborative response capability.
The edge security agent receives access requests from terminal devices, determines the security level based on the device fingerprint, selects a matching authentication protocol for authentication, and pre-distributes a credential pool to the terminal device after successful authentication. It also stores the chain tail anchor, performs local threat detection and data packet processing, generates a unified format threat intelligence object, and uploads it to the cloud security analysis platform for cross-layer attack path analysis and coordinated response.
It achieves cross-layer collaborative communication security protection, reduces terminal authentication latency, adapts to the latency requirements of real-time power control services, and can promptly identify cross-layer collaborative attacks, forming a cross-layer collaborative handling capability.
Smart Images

Figure CN123123016A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power system communication security technology, and in particular to a method and system for protecting communication security in the Internet of Things (IoT) for power systems. Background Technology
[0002] With the advancement of smart grid and new power system construction, the power Internet of Things has been widely applied to equipment access and business communication in all aspects of power generation, transmission, transformation, distribution and consumption, becoming an important technical means to ensure the safe and stable operation of the power system.
[0003] Existing power IoT communication security protection methods typically deploy security devices at each level to implement threat detection. However, the security devices at each level operate independently, and the detection information is scattered and in different formats. Alarms generated by the same attack at different levels cannot be correlated and identified, making it difficult to detect cross-layer collaborative attacks in a timely manner. This makes it difficult for power IoT communication security protection to form a cross-layer collaborative response capability. Summary of the Invention
[0004] This invention provides a method and system for protecting communication security in the power Internet of Things (IoT), which can solve the problem that existing power IoT communication security protection methods typically implement threat detection independently at each level, making it difficult to detect cross-layer collaborative attacks in a timely manner.
[0005] An embodiment of the present invention provides a power Internet of Things (IoT) communication security protection method, executed by an edge security agent device, comprising: Receive access requests from terminal devices seeking to connect to the power Internet of Things, and determine the security level of the terminal devices based on their device fingerprints; According to the security level, a matching authentication method is selected from the preset authentication protocol family to complete the authentication. After the authentication is successful, a credential pool is pre-distributed to the terminal device and the tail anchor of the credential pool is stored. Extract temporary credentials and service type identifiers from the service data packets of the terminal device, and perform verification on the temporary credentials according to the chain tail anchor point; After the verification is passed, the real-time level of the service data packet is determined according to the service type identifier, and the service data packet is processed according to the real-time level. Local threat detection is performed on the processed business data packets to obtain threat information. The threat information is then converted into a threat intelligence object in a preset unified format and uploaded to the cloud security analysis platform. Receive the linkage response command generated by the cloud security analysis platform based on the threat intelligence object, and execute the corresponding security protection operation according to the linkage response command.
[0006] Furthermore, according to the security level, a matching authentication method is selected from a preset family of authentication protocols to complete the authentication, including: When the security level is a preset first security level, a first two-way authentication method based on public key cryptography is selected from a preset authentication protocol family. The edge security proxy device generates a random number as a first challenge value and sends the first challenge value to the terminal device. The first authentication response returned by the terminal device based on the first challenge value is received, and the first authentication response is verified according to the first two-way authentication method. When the first authentication response passes the verification, the authentication of the terminal device is completed. When the security level is the preset second security level, a second two-way authentication method based on a pre-shared key is selected from a preset authentication protocol family. The edge security proxy device generates a random number as a second challenge value and sends the second challenge value to the terminal device. The second authentication response returned by the terminal device based on the second challenge value is received, and the second authentication response is verified according to the second two-way authentication method. When the second authentication response passes the verification, the authentication of the terminal device is completed. When the security level is the preset third security level, an implicit authentication method based on a physical unclonable function is selected from the preset authentication protocol family. The edge security proxy device sends the pre-stored physical unclonable function challenge value to the terminal device. The physical unclonable function response value returned by the terminal device is received and compared. When the physical unclonable function response values match, the authentication of the terminal device is completed. The first security level has a higher priority than the second security level, and the second security level has a higher priority than the third security level.
[0007] Furthermore, after successful authentication, the credential pool is pre-distributed to the terminal device and the tail anchor of the credential pool is stored, including: After authentication is successful, a set of random numbers is generated as the tail anchor point of the credential pool; Starting from the tail anchor point, perform hash operations sequentially backward to obtain a certificate chain including multiple sets of first temporary certificates; wherein, each first temporary certificate in the multiple sets of first temporary certificates has a certificate serial number, and the head of the certificate chain is the first temporary certificate with the smallest certificate serial number. The credential chain is distributed to the terminal device as the credential pool, and the tail anchor point of the credential pool is stored at the same time.
[0008] Furthermore, the temporary credential is validated based on the chain tail anchor point, including: The first temporary credential retrieved by the terminal device from the credential pool is carried in the business data packet as the second temporary credential. Extract the second temporary credential and its serial number from the business data packet; The second temporary voucher is subject to time window verification and a one-time consumption verification of the voucher serial number. Perform chain hash verification on the second temporary credential to obtain the theoretical value of the chain tail anchor point, and compare the theoretical value of the chain tail anchor point with the chain tail anchor point; If both the time window verification and the one-time consumption verification pass, and the theoretical value of the chain tail anchor point is consistent with the chain tail anchor point, then the second temporary voucher verification is deemed to have passed. If the time window verification fails, the one-time consumption verification fails, or the theoretical value of the chain tail anchor point is inconsistent with the chain tail anchor point, the service data packet is rejected and the abnormal event is recorded.
[0009] Furthermore, the real-time level of the service data packet is determined based on the service type identifier, and the service data packet is processed according to the real-time level, including: The real-time level of the service data packet is obtained by querying a preset service security policy table using the service type identifier. When the real-time level is Level 1, after performing an integrity check based on pre-distributed credentials on the service data packet, the service data packet is forwarded, and the hash digest of the service data packet is uploaded to the blockchain for evidence storage. When the real-time level is Level 2, the service data packet is forwarded after performing integrity verification and symmetric encryption based on pre-distributed credentials. When the real-time level is Level 3, the business data packet is subjected to integrity verification based on pre-distributed credentials, symmetric encryption and digital signature. After obtaining the digital signature value, the business data packet is forwarded and the digital signature value and the hash digest of the business data packet are uploaded to the blockchain for evidence storage. When the real-time level is Level 4, the business data packet is forwarded after performing complete security processing including integrity verification based on pre-distributed credentials, symmetric encryption, digital signature and blockchain notarization.
[0010] Furthermore, local threat detection is performed on the processed business data packets to obtain threat information, which is then converted into a threat intelligence object in a preset unified format, including: Perform local threat detection on the processed business data packets to obtain threat events; The time of occurrence of the threat event is used as the timestamp of the threat occurrence, and the threat type and threat confidence level of the threat event are determined. The threat type is matched with a preset threat type and attack chain stage mapping rule base; When the threat type matches the mapping rule base, the attack chain stage in the mapping rule base corresponding to the threat type is determined as the attack chain stage corresponding to the threat event; When the threat type does not match the mapping rule base, the abnormal behavior sequence of the terminal device is collected, and the abnormal behavior sequence is matched with the preset attack chain stage behavior pattern base to obtain the attack chain stage corresponding to the threat event. Based on the timestamp of the threat occurrence and the device identifier of the terminal device, perform correlation analysis on multiple threat events belonging to the terminal device within a preset association time window to obtain a list of associated threat identifiers; The original alarm information of the threat event is digested to obtain the original alarm summary; The threat occurrence timestamp, the power IoT layer identifier to which the edge security agent device belongs, the device identifier of the terminal device, the threat type, the attack chain stage, the threat confidence level, the list of associated threat identifiers, and the original alarm summary are encapsulated into the threat intelligence object in a preset unified format.
[0011] Furthermore, the cloud-based security analysis platform generates a coordinated response command based on the threat intelligence object in the following manner: The cloud-based security analysis platform constructs a global threat correlation graph based on the threat intelligence objects, and analyzes the global threat correlation graph to obtain attack paths and comprehensive threat confidence. Based on the attack path and the overall threat confidence level, a linkage response command is generated and sent to the edge security agent device.
[0012] Furthermore, the cloud-based security analysis platform constructs a global threat correlation graph based on the threat intelligence objects, and analyzes the global threat correlation graph to obtain attack paths and comprehensive threat confidence, including: The cloud-based security analysis platform constructs the global threat association graph by using the devices and communication links involved in the threat intelligence objects as nodes and the threat relationships represented by the threat intelligence objects as edges. Perform temporal correlation analysis on the threat intelligence objects in the global threat correlation graph, and determine multiple threat intelligence objects whose threat occurrence timestamps are within the same preset time window as having temporal correlation; Perform causal association analysis on the threat intelligence objects in the global threat association graph, and identify multiple threat intelligence objects that satisfy the preset pre-order and post-order transition relationship in the attack chain stage as having causal correlation. Perform homology association analysis on the threat intelligence objects in the global threat association graph to identify multiple threat intelligence objects with the same or similar attack characteristics as having homology correlation; Multiple threat intelligence objects that simultaneously satisfy the temporal correlation, the causal correlation, and the common origin correlation are associated and bound through the associated threat identifier list to obtain a cross-layer attack chain, and the cross-layer attack chain is determined as the attack path.
[0013] Furthermore, the cloud-based security analysis platform generates a coordinated response command based on the attack path and the overall threat confidence level, including: The cloud-based security analysis platform compares the comprehensive threat confidence level with a preset first threshold and a second threshold. When the overall threat confidence level is greater than or equal to the first threshold, a global linkage response instruction is generated. The global linkage response instruction is used to instruct the edge security agent device to block communication, update the blacklist, and adjust the authentication policy of the affected device. When the overall threat confidence level is greater than or equal to the second threshold and less than the first threshold, a local early warning instruction is generated. The local early warning instruction is used to instruct the edge security agent device in the affected area to perform enhanced monitoring. When the overall threat confidence level is less than the second threshold, the linkage response command is not generated.
[0014] Another embodiment of the present invention provides a power Internet of Things communication security protection system, including an edge security proxy device and a cloud security analysis platform, wherein the edge security proxy device and the cloud security analysis platform are communicatively connected; The edge security proxy device is used to receive access requests from terminal devices seeking to access the power Internet of Things, determine the security level of the terminal device based on the device fingerprint, select a matching authentication method from a preset authentication protocol family according to the security level to complete authentication, pre-distribute a credential pool to the terminal device after successful authentication and store the tail anchor of the credential pool, extract temporary credentials and service type identifiers from the service data packets sent by the terminal device, perform verification on the temporary credentials according to the tail anchor, determine the real-time level of the service data packets according to the service type identifier after successful verification, process the service data packets according to the real-time level, perform local threat detection on the processed service data packets to obtain threat information, convert the threat information into a threat intelligence object in a preset unified format, and upload the threat intelligence object to the cloud security analysis platform. The cloud-based security analysis platform is used to generate a linkage response command based on the threat intelligence object, and to send the linkage response command to the edge security agent device; The edge security agent device is also used to receive linkage response instructions issued by the cloud security analysis platform and to perform corresponding security protection operations according to the linkage response instructions.
[0015] The embodiments of the present invention have the following beneficial effects: This invention provides a method and system for secure communication protection in the power Internet of Things (IoT). An edge security proxy device determines the security level of the device to be connected based on its device fingerprint. Authentication is then performed by selecting a matching authentication method from a preset authentication protocol family according to the security level. Upon successful authentication, a credential pool is pre-distributed to the device, and the tail anchor of the credential pool is stored. This allows the terminal to submit only a temporary credential corresponding to the current time window for each subsequent communication. The edge security proxy device completes the verification through a single hash calculation and comparison, eliminating the need for online key negotiation. This overcomes the problems of inconsistent terminal authentication mechanisms, high authentication overhead, and low efficiency, and reduces online authentication latency in terminal communication. Furthermore, by determining the real-time level of the service data packet based on the service type identifier and processing the service data packet according to the real-time level, the security protection mechanism adapts to real-time power control. The deterministic delay requirement of the control business avoids the problem of uncontrollable delays in critical business caused by the unified security processing process. Furthermore, the edge security agent device performs local threat detection on the processed business data packets, converts the obtained threat information into a pre-set unified format threat intelligence object, and uploads it to the cloud security analysis platform. The cloud security analysis platform analyzes the threat intelligence object to obtain the attack path and comprehensive threat confidence and issues a linkage response command. The edge security agent device executes the corresponding security protection operation, so that the threat information that is scattered and has different formats at the perception layer, network layer, platform layer, and application layer can be unified and correlated to restore the same attack path. This overcomes the problem that alarms generated by the same attack at different levels cannot be correlated and identified, and cross-layer collaborative attacks are difficult to detect in a timely manner. This enables the power Internet of Things communication security protection to form a cross-layer collaborative handling capability. Attached Figure Description
[0016] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0017] Figure 1 This is a flowchart illustrating a power Internet of Things (IoT) communication security protection method according to an embodiment of the present invention.
[0018] Figure 2This is a schematic diagram of the structure of a power Internet of Things communication security protection system provided in an embodiment of the present invention. Detailed Implementation
[0019] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.
[0021] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.
[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0023] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.
[0024] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).
[0025] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.
[0026] Specifically, in the step descriptions of the various embodiments below, the execution entities involved in this method are the edge security agent device and the cloud security analysis platform, respectively. The two are communicatively connected and cooperate with each other to jointly constitute a power Internet of Things communication security protection system. It should be understood that the edge security agent device does not specifically refer to a single physical computing device, but rather to the logical functional entity that implements the edge-side processing flow in the solution of this application. In actual physical deployment, the edge security agent device can be deployed in edge computing nodes or communication gateways such as station-side edge computing nodes or distribution terminal units, covering functional units such as device fingerprint recognition and classification, authentication method selection and execution, credential generation and management, business security processing, and threat detection and intelligence reporting. It can also run in software form on the computing platform of the aforementioned edge computing nodes or communication gateways. The cloud-based security analysis platform is deployed in a power dispatch center or a network security monitoring center. It includes functional units such as threat intelligence object reception and parsing, global threat association graph construction and maintenance, cross-layer attack chain identification, comprehensive threat confidence calculation, linkage response strategy distribution, and blockchain evidence storage and auditing. It is also connected to the edge security agent device through the power dispatch data network.
[0027] Those skilled in the art should understand that the following steps, including device fingerprint extraction and comparison, security level determination, authentication challenge value generation and verification, session master key derivation, hash chain credential pool generation and maintenance, temporary credential time window verification and one-time consumption verification and chain hash verification, secure processing of business data packets (integrity verification, symmetric encryption, digital signature and blockchain evidence storage), local threat detection and attack chain stage determination, threat intelligence object encapsulation, threat correlation analysis, comprehensive threat confidence calculation and linkage response command generation, can be flexibly configured to be executed in the processor and security encryption module of the edge computing node where the edge security agent device is located, or configured to be executed in the server processor, server cryptographic machine and cloud computing platform cluster where the cloud security analysis platform is located, without departing from the core logic of this application, based on the hardware conditions such as the processing power of the edge computing node, the bandwidth and latency constraints of the communication network, the real-time requirements of the business, and the security level requirements of the power equipment in the actual deployment environment.
[0028] Therefore, the phrases "the edge security agent device performs a certain operation" and "the cloud security analysis platform performs a certain operation" used below for the sake of brevity include situations where the edge security agent device or the cloud security analysis platform independently completes its respective side processing flow, situations where the edge side processing flow is completed collaboratively by multiple functional modules within the edge computing node, and situations where the cloud side processing flow is completed collaboratively by multiple devices and multiple levels of distributed systems such as threat intelligence servers, correlation analysis servers, policy distribution servers, and blockchain evidence storage nodes. These should not be narrowly interpreted as strict limitations on the specific physical execution location or specific hardware entity of the processing action.
[0029] To address the problem that existing power IoT communication security protection methods typically implement threat detection independently at each level, making it difficult to detect cross-layer collaborative attacks in a timely manner, an embodiment of the present invention provides a power IoT communication security protection method executed by an edge security proxy device, comprising: Step S1: Receive the access request from the terminal device to be connected to the power Internet of Things, and determine the security level of the terminal device based on the device fingerprint of the terminal device; In a preferred embodiment, receiving the access request from a terminal device seeking to access the power Internet of Things, and determining the security level of the terminal device based on its device fingerprint, specifically includes: Receive an access request sent by a terminal device that wants to access the power Internet of Things, wherein the access request carries the device type identifier and resource capability description of the terminal device; A preset probe data packet is sent to the terminal device and a response data packet is received from the terminal device. The protocol stack implementation characteristics of the terminal device are obtained based on the response data packet, and the response delay distribution characteristics of the terminal device are obtained based on the round-trip delay between the probe data packet and the response data packet. Multidimensional signal fusion is performed on the protocol stack implementation characteristics and the response delay distribution characteristics to obtain the device fingerprint vector of the terminal device; The device fingerprint vector is compared with each fingerprint template in the locally stored lightweight fingerprint database to obtain the device type of the terminal device. The security level of the terminal device is determined based on the device type.
[0030] Preferably, the access request is received in the following manner: the access request is a message sent by the terminal device to the edge security agent device when the terminal device first accesses the power IoT or when the communication connection with the edge security agent device is re-established, for example, when the terminal device re-accesses after maintenance, replacement, or power outage and restart; the terminal device is equipped with a terminal security module, which, when the terminal device first accesses, cooperates with the detection request of the edge security agent device to return the protocol stack characteristics and response latency data corresponding to the terminal device; the device type identifier carried in the access request is used to indicate the device category to which the terminal device belongs, and the resource capability description is used to indicate one or more of the terminal device's computing power, storage resources, communication capabilities, and power supply method. For example, the terminal device can be a relay protection device deployed in a substation, a smart meter deployed on a distribution line, or a passive temperature sensor deployed at a transmission line monitoring point; wherein, the device type identifier and resource capability description carried in the access request of the relay protection device correspond to a preset first security level, the device type identifier and resource capability description carried in the access request of the smart meter correspond to a preset second security level, and the device type identifier and resource capability description carried in the access request of the passive temperature sensor correspond to a preset third security level; the relay protection device has relatively abundant computing and communication capabilities, the smart meter has conventional computing and communication capabilities, and the passive temperature sensor is powered by an energy harvesting device, and its computing, storage resources, and communication capabilities are limited.
[0031] Preferably, the preset probe data packet is sent in the following manner: the preset probe data packet refers to a probe message pre-configured in the edge security agent device to trigger the terminal device to return the response data packet, and its message type, number of transmissions, and transmission interval are all preset; for example, the probe data packet may include one or more of Transmission Control Protocol synchronization messages, User Datagram Protocol messages, and Internet Control Message Protocol echo request messages; the number of transmissions may be set to three to five times, and the transmission interval may be set to fifty milliseconds to two hundred milliseconds.
[0032] Preferably, the protocol stack implementation features are obtained by parsing the response data packet and extracting the feature information presented by the terminal device at the protocol stack implementation level. The feature information includes one or more of the following: window size, content and order of option fields, maximum segment length, and time-to-live (TTL) value in the Transmission Control Protocol (TCP) message. The window size refers to the value of the field in the message header used to announce the receive window size; the content and order of option fields refer to the control information carried by the optional fields in the message header and their arrangement; the TTL value refers to the maximum number of hops the message is allowed to traverse in the network. The default values and processing methods of the above fields vary depending on the processor, operating system, protocol stack library, and firmware implementation used by the device. For example, relay protection devices deployed in substations and smart meters deployed on distribution lines typically use different protocol stacks, and their value combinations in terms of window size and option field order differ, which can help distinguish between the two types of devices.
[0033] Preferably, the response latency distribution characteristics are obtained by: continuously sending multiple probe data packets to the terminal device, and measuring the round-trip latency between each probe data packet and the corresponding response data packet; the round-trip latency refers to the time elapsed from the moment the edge security proxy device sends the probe data packet to the moment it receives the corresponding response data packet returned by the terminal device; statistical analysis is performed on the measured round-trip latencies, including calculating one or more of the mean, variance, and values at preset quantiles; the preset quantile refers to a pre-set quantile used for statistically analyzing the latency distribution, for example, it can be taken as the 95th percentile. Due to differences in processing capabilities, protocol stack scheduling methods, and current load conditions of different terminal devices, their response latency distribution characteristics vary; for example, when the processing capability of the terminal device is relatively abundant, its mean round-trip latency is small and fluctuates less; when the processing capability of the terminal device is limited or in a sleep / wake-up state, its mean round-trip latency is large and fluctuates more significantly. In a preferred embodiment, the edge security proxy device completes the extraction of the device fingerprint within a preset time threshold after receiving the access request. The preset time threshold can be, for example, 100 milliseconds, to avoid affecting the access efficiency of the terminal device.
[0034] Preferably, the device fingerprint vector is obtained by: numericalizing and normalizing each dimension of the protocol stack implementation features and the response delay distribution features, respectively, to obtain normalized feature values corresponding to each dimension; numericalization refers to converting non-numerical features into numerical representations, for example, the order of the option fields can be encoded as numerical values according to their order; normalization refers to mapping the values of each dimension of features to a unified value range to eliminate differences in units and value ranges between different features; for example, for the first... 3D features can be normalized using the following formula: in, For the first The actual values of the dimensional features, and The first one is respectively the second one. The lower and upper bounds of the values of the dimensional features are statistically obtained during the construction of the lightweight fingerprint database. For the first The normalized feature values corresponding to each dimension feature are obtained; the normalized feature values corresponding to each dimension feature are combined according to a preset feature order to obtain the device fingerprint vector of the terminal device, which can be represented as: in, The total number of feature dimensions included in the device fingerprint vector; preferably, the total number of feature dimensions can be preset according to the accuracy requirements of device type identification and the computing resource constraints of the edge security agent device; for example, when it is necessary to distinguish more types of terminal devices, more features can be included in the device fingerprint vector, and the total number of feature dimensions can be increased accordingly; the preset feature order refers to the arrangement order of each feature dimension in the device fingerprint vector, for example, it can be arranged in the order of protocol stack implementation features first and response delay distribution features last.
[0035] Preferably, the lightweight fingerprint database is constructed in the following manner: During the deployment and operation and maintenance phase of the power Internet of Things, for various legitimate terminal devices, their protocol stack implementation characteristics and response latency distribution characteristics are collected in the same manner as the above steps to obtain fingerprint samples of each legitimate terminal device, and each fingerprint sample is labeled with the corresponding device type to form a fingerprint template; each fingerprint template is stored in the local storage space of the edge security agent device to form the lightweight fingerprint database. The lightweight fingerprint database is a pre-built local reference database for device type identification; the term "lightweight" means that the lightweight fingerprint database organizes each fingerprint template at the device type granularity, and its storage capacity and retrieval overhead are lower than the full fingerprint database at the device individual granularity, so as to adapt to the storage and computing resource constraints of the edge security agent device; for example, the fingerprint templates stored in the lightweight fingerprint database include typical fingerprint templates corresponding to device types such as relay protection devices, smart meters, and passive temperature sensors.
[0036] Preferably, the device type is obtained by: calculating the similarity between the device fingerprint vector and each fingerprint template in the lightweight fingerprint database; the similarity can be measured using cosine similarity or Euclidean distance, etc., and this embodiment does not limit this; the fingerprint template with the highest similarity and exceeding a preset similarity threshold is determined as the target template matching the terminal device, and the device type labeled on the target template is determined as the device type of the terminal device; the preset similarity threshold refers to a pre-set threshold used to determine whether the device fingerprint vector matches the fingerprint template, for example, it can be set to 0.9.
[0037] In a preferred embodiment, when the similarity between the device fingerprint vector and all the fingerprint templates does not exceed the preset similarity threshold, the edge security proxy device marks the terminal device as an unknown device and performs subsequent processing on the terminal device according to the preset third security level, or rejects the access request of the terminal device and records the abnormal event.
[0038] Preferably, the security level is determined in the following manner: the security level includes a preset first security level, a preset second security level, and a preset third security level, wherein the first security level has a higher priority than the second security level, and the second security level has a higher priority than the third security level; the edge security agent device is pre-configured with a correspondence between the device type and the security level, and the correspondence is a pre-set security level classification relationship based on the importance of various devices in the power Internet of Things; for example, when the device type is a relay protection device, the security level of the terminal device is determined to be the first security level; when the device type is a smart meter, the security level of the terminal device is determined to be the second security level; when the device type is a passive temperature sensor, the security level of the terminal device is determined to be the third security level. In a preferred embodiment, the determination of the security level is further combined with the resource capability description carried in the access request; for example, when the security level corresponding to the device type is the first security level, and the resource capability description indicates that the computing power of the terminal device is lower than a preset capability threshold, the security level of the terminal device is adjusted to the second security level to avoid the overhead of subsequent authentication and security processing processes exceeding the resource capacity of the terminal device; the capability threshold refers to a pre-set capability threshold used to measure whether the resource capabilities of the terminal device meet the requirements of the corresponding security level.
[0039] It is worth noting that the extraction, comparison, and security level determination of the aforementioned device fingerprints are all completed locally on the edge security agent device, without requiring the terminal device to actively report its model or identity information. The edge security agent device can be deployed on edge computing nodes or communication gateways, such as station-side edge computing nodes or power distribution terminal units, and communicates with the terminal device via wired or wireless means. Therefore, even if an illegal device impersonates the device type identifier of a legitimate device to initiate an access request, the edge security agent device can still identify the true device type of the device initiating the access request based on the comparison results of the device fingerprints, thereby reducing the risk of illegal devices accessing the power Internet of Things with forged device type identifiers. At the same time, the storage and computational overhead of the above identification and classification process is low, which is beneficial for adapting to scenarios with a large number of terminal devices, diverse device types, and frequent access requests in the power Internet of Things.
[0040] In this embodiment, by receiving the access request from a terminal device seeking to access the power Internet of Things, the device type identifier and resource capability description carried in the request are parsed, and the device fingerprint vector obtained by multi-dimensional signal fusion based on protocol stack implementation features and response delay distribution features is compared with a local lightweight fingerprint database to complete device type identification, thereby determining the security level of the terminal device. In the above process, the extraction of device fingerprints is based on the inherent communication behavior characteristics of the terminal device itself, without the need for additional hardware modifications to the terminal device. The identification and classification process is completed locally on the edge side, which helps to reduce the latency of terminal access processing and provides a basis for dynamically selecting the matching authentication method from the preset authentication protocol family according to the security level. This helps to ensure access security while taking into account the resource capabilities and access efficiency of different types of terminal devices.
[0041] Step S2: Select a matching authentication method from the preset authentication protocol family according to the security level to complete the authentication. After successful authentication, pre-distribute the credential pool to the terminal device and store the tail anchor of the credential pool. In a preferred embodiment, authentication is performed by selecting a matching authentication method from a preset family of authentication protocols according to the security level, including: When the security level is a preset first security level, a first two-way authentication method based on public key cryptography is selected from a preset authentication protocol family. The edge security proxy device generates a random number as a first challenge value and sends the first challenge value to the terminal device. The first authentication response returned by the terminal device based on the first challenge value is received, and the first authentication response is verified according to the first two-way authentication method. When the first authentication response passes the verification, the authentication of the terminal device is completed. When the security level is the preset second security level, a second two-way authentication method based on a pre-shared key is selected from a preset authentication protocol family. The edge security proxy device generates a random number as a second challenge value and sends the second challenge value to the terminal device. The second authentication response returned by the terminal device based on the second challenge value is received, and the second authentication response is verified according to the second two-way authentication method. When the second authentication response passes the verification, the authentication of the terminal device is completed. When the security level is the preset third security level, an implicit authentication method based on a physical unclonable function is selected from the preset authentication protocol family. The edge security proxy device sends the pre-stored physical unclonable function challenge value to the terminal device. The physical unclonable function response value returned by the terminal device is received and compared. When the physical unclonable function response values match, the authentication of the terminal device is completed. The first security level has a higher priority than the second security level, and the second security level has a higher priority than the third security level.
[0042] Preferably, after successful authentication, the credential pool is pre-distributed to the terminal device and the tail anchor of the credential pool is stored, including: After authentication is successful, a set of random numbers is generated as the tail anchor point of the credential pool; Starting from the tail anchor point, perform hash operations sequentially backward to obtain a certificate chain including multiple sets of first temporary certificates; wherein, each first temporary certificate in the multiple sets of first temporary certificates has a certificate serial number, and the head of the certificate chain is the first temporary certificate with the smallest certificate serial number. The credential chain is distributed to the terminal device as the credential pool, and the tail anchor point of the credential pool is stored at the same time.
[0043] Preferably, the authentication method matching the security level is selected from the preset authentication protocol family in the following way: the preset authentication protocol family refers to a set of multiple authentication methods pre-configured in the edge security proxy device, each of which corresponds to different values of the security level; after determining the security level of the terminal device, the edge security proxy device determines and executes the authentication method matching the security level according to the correspondence between the security level and the authentication method; thus, terminal devices with different security levels adopt their corresponding authentication methods, without having to perform a unified authentication process.
[0044] Preferably, the authentication process under the first two-way authentication method is performed in the following manner: the edge security proxy device generates a 256-bit random number as the first challenge value (denoted as...). ) and the random number of the edge security agent device itself (denoted as ) ), and the first challenge value is combined with the random number The challenge is sent to the terminal device; the terminal device uses its own private key to match the first challenge value with the random number. Perform a signature operation to obtain a first signature value, and simultaneously generate a random number on the terminal device side (denoted as ). ), and the first signature value is combined with the random number The first signature value is returned to the edge security proxy device as the first authentication response. The edge security proxy device uses the pre-stored public key of the terminal device to verify the first signature value. If the verification is successful, the authentication of the terminal device is completed. Since the first challenge value and the random number generated by both parties are generated separately and jointly participate in the authentication process, the above process constitutes two-way authentication. After successful authentication, the edge security proxy device and the terminal device derive the session master key based on the random number exchanged between them. in, This is a key derivation function constructed based on the SM3 hash function. For session counter, and The random numbers generated by the terminal device and the edge security proxy device are respectively. and The symbols are the identifiers of the terminal device and the edge security proxy device, respectively. Represents the bitwise XOR operation, sign This represents a string concatenation operation; the session master key is used to protect the subsequent pre-distributed credential pool and the transmission of business data.
[0045] Preferably, the authentication process under the second two-way authentication method is performed in the following manner: the edge security proxy device and the terminal device are respectively pre-configured with the same key material, which serves as the pre-shared key used in the second two-way authentication method (denoted as...). The edge security proxy device generates a random number as the second challenge value, encrypts the second challenge value using the pre-shared key, and sends it to the terminal device. The terminal device decrypts the second challenge value using the pre-shared key and generates a response value according to a preset response rule, which is then returned to the edge security proxy device as the second authentication response. The edge security proxy device verifies the response value contained in the second authentication response according to the preset response rule. If the verification is successful, the authentication of the terminal device is completed. The preset response rule refers to the calculation rule agreed upon in advance by both parties for generating the response value based on the challenge value, such as performing a keyed hash operation on the second challenge value based on the pre-shared key. After successful authentication, the edge security proxy device and the terminal device derive a session master key based on the pre-shared key and the random number exchanged during this authentication process. in, The pre-shared key is defined as such, and the meanings of the other symbols are the same as described above; the session master key is used to protect the subsequent pre-distributed credential pool and the transmission of business data.
[0046] Preferably, the authentication process under the implicit authentication method is performed in the following manner: the physically unclonable function refers to a function that utilizes the unique physical characteristics formed by uncontrollable process deviations generated during the manufacturing process of the terminal device chip, so that the same challenge value input into the physically unclonable function will yield a unique and unclonable response value; the edge security proxy device pre-stores the correspondence between the challenge value and response value of the physically unclonable function corresponding to the terminal device; the edge security proxy device uses the pre-stored physically unclonable function challenge value (denoted as...) The challenge value is sent to the terminal device; the terminal device, through its physical unclonable function hardware module, generates a physical unclonable function response value (denoted as ) in real time based on the physical unclonable function challenge value. The data is then returned to the edge security proxy device. The above generation process does not require the terminal device to perform cryptographic operations. The edge security proxy device compares the received Physically Unclonable Function (PUC) response value with the pre-stored correspondence. If the comparison matches, the authentication of the terminal device is completed. Since the above authentication process does not require the terminal device to store keys or perform cryptographic operations, this authentication method is called implicit authentication. After successful authentication, the edge security proxy device and the terminal device derive a session master key based on the PUC response value. in, The physical non-cloning function response value. This is the physical non-cloning function challenge value used in this authentication; the meanings of the other symbols are the same as described above. In particular, the physical non-cloning function response value is generated in real time by the physical non-cloning function hardware module of the terminal device during the authentication process, without needing to be stored in the memory of the terminal device, which helps to reduce the risk of key materials being extracted in software.
[0047] Preferably, the first security level, the second security level, and the third security level are arranged in descending order of priority, and the edge security proxy device selects the corresponding authentication method according to the security level of the terminal device; for example, when the terminal device is a relay protection device and its security level is the first security level, the first two-way authentication method is selected; when the terminal device is a smart meter and its security level is the second security level, the second two-way authentication method is selected; when the terminal device is a passive temperature sensor and its security level is the third security level, the implicit authentication method is selected.
[0048] Preferably, the chain tail anchor is generated in the following manner: after the terminal device is authenticated, the edge security proxy device generates a random number (e.g., a 256-bit random number) as the chain tail anchor; the chain tail anchor is located at the end of the hash chain structure corresponding to the credential pool, serving as the starting point for the generation of the hash chain and the benchmark value for subsequent verification.
[0049] Preferably, the certificate chain is obtained by performing hash operations sequentially backward from the tail anchor point, such that each first temporary certificate in the certificate chain, except for the tail anchor point, is obtained by hashing its adjacent next value; for example, the tail anchor point is denoted as... Each of the first temporary vouchers arranged from the tail to the head in the voucher chain is sequentially denoted as follows: , ... (That is, the chain of credentials follows the direction from the head to the tail anchor point) If the hash chain is arranged in a certain way, then the generation process of the hash chain can be represented as follows: in, The preset hash function can be used, such as the SM3 hash function. The preset hash function refers to a one-way hash function that is agreed upon in advance by both parties to the authentication. Since the hash operation is one-way, it is impossible to derive the next adjacent value from any of the first temporary credentials in the credential chain. Storing the chain tail anchor point locally on the edge security agent device can meet the needs of subsequent verification.
[0050] Preferably, the multiple sets of first temporary vouchers each correspond to their respective valid time windows, and each valid time window progresses sequentially in the direction of increasing voucher serial number; the distribution time of the voucher pool is recorded as... The progressive interval between two adjacent effective time windows is denoted as Then the first The effective time window corresponding to each first temporary certificate can be represented as: in, For the first The effective time window corresponding to each first temporary certificate; for example, the number of groups of the multiple groups of first temporary certificates. The number of groups can be set from 10 to 50; in a specific example, the number is 20. The progressive interval... For example, the value can be 1 minute; in this case, the effective time window corresponding to the first temporary voucher is... The effective time window corresponding to the second first temporary certificate is And so on, the first The effective time window corresponding to each first temporary certificate is .
[0051] Preferably, the edge security proxy device can also perform rolling updates on the credential pool to maintain a continuous supply of available credentials during the use of the credential pool: when a preset trigger condition is met, the edge security proxy device actively initiates a rolling update of the credential pool; the trigger condition includes: the number of first temporary credentials used by the terminal device reaches a preset proportion of the number of the multiple sets of first temporary credentials (e.g., 80%). ,in The maximum used credential number. (This refers to the number of sets of the multiple sets of first temporary credentials); the remaining amount flag of the credential pool carried by the terminal device in the header of the service data packet is set; and a preset proportion (e.g., 80%, i.e., satisfying the requirement) has been exceeded for a period of time from the time the credential pool was distributed. ,in (Current time). The rolling update process is as follows: The edge security proxy device and the terminal device derive a new generation master key based on the currently valid session master key: in, This is the new generation master key. For predefined update salt values, To update the round counter, the edge security proxy device generates a new batch using the new generation master key as a seed. The first temporary voucher of the group, its voucher serial number starts from... to Continuous numbering, valid time window continues from the end time of the original voucher pool The process begins with the edge security proxy device using old credentials within the current valid time window as the encryption key to encrypt the entire new credential pool (e.g., using the SM4 symmetric encryption algorithm) and then sending it to the terminal device via the downlink channel. The terminal device decrypts the new credential pool using the corresponding locally stored credentials and stores it, carrying an update confirmation flag in the next uplink data. Upon receiving the update confirmation flag, the edge security proxy device releases the remaining credentials in the original credential pool, updates the session master key to the new generation master key, and resets the maximum used credential sequence number. If the edge security proxy device does not receive the update confirmation flag within a preset time period, it retransmits the new credential pool. If the retransmission fails after a preset number of times (e.g., 3 times), or if the terminal device's credential pool is completely exhausted (i.e., ...), the process continues. When the terminal device clears the locally stored credential pool and initiates re-access authentication, the edge security proxy device redistributes a brand new credential pool for it.
[0052] Preferably, the credential chain is distributed as the credential pool and the tail anchor is stored in the following manner: the edge security proxy device distributes the complete credential chain to the terminal device through a pre-established encrypted channel, and the terminal device receives and stores the complete credential chain; the edge security proxy device stores only the tail anchor locally, and does not store each of the first temporary credentials in the credential chain; since the tail anchor is the verification benchmark value of the hash chain, storing only the tail anchor is sufficient to meet the needs of subsequent chain verification of credentials submitted by the terminal device, thereby reducing the storage overhead of the edge security proxy device.
[0053] It should be noted that the above process is a pre-distribution process of the credential pool; the pre-distribution refers to the process of pre-issuing the credential pool to the terminal device and storing it locally on the terminal device before the terminal device actually sends business data; compared with the method of performing online key negotiation or online credential issuance during each business communication, the above pre-distribution method is beneficial to reducing the processing latency during the business communication stage; and the above authentication process and the credential pool pre-distribution process are connected to each other, so that the terminal device has the credential materials required for business communication when the access is completed.
[0054] In this embodiment, authentication is completed by selecting a matching authentication method from a preset authentication protocol family according to the security level. This allows terminal devices of different types and resource capabilities to access the network using an authentication method that is compatible with their security level. For terminal devices with high security levels, such as relay protection devices, a two-way authentication method based on public key cryptography is used, which helps to ensure the security strength of the authentication process. For terminal devices with medium security levels, such as smart meters, a two-way authentication method based on pre-shared keys is used, which helps to reduce the computational overhead of the authentication process. For terminal devices with limited resources, such as passive temperature sensors, an implicit authentication method based on physically non-cloning functions is used to avoid the occupation of terminal device resources by cryptographic operations. After successful authentication, a credential pool is generated using a hash chain structure and pre-distributed to the terminal devices. At the same time, the tail anchor point of the chain is stored locally, so that the credential verification in the subsequent business communication stage can be completed based on a single locally stored benchmark value. This provides a basis for security verification in the communication stage and takes into account both the security of access authentication and the resource adaptability of terminal devices.
[0055] Step S3: Extract temporary credentials and service type identifiers from the service data packets of the terminal device, and perform verification on the temporary credentials according to the chain tail anchor point; In a preferred embodiment, performing verification on the temporary credential based on the chain tail anchor includes: The first temporary credential retrieved by the terminal device from the credential pool is carried in the business data packet as the second temporary credential. Extract the second temporary credential and its serial number from the business data packet; The second temporary voucher is subject to time window verification and a one-time consumption verification of the voucher serial number. Perform chain hash verification on the second temporary credential to obtain the theoretical value of the chain tail anchor point, and compare the theoretical value of the chain tail anchor point with the chain tail anchor point; If both the time window verification and the one-time consumption verification pass, and the theoretical value of the chain tail anchor point is consistent with the chain tail anchor point, then the second temporary voucher verification is deemed to have passed. If the time window verification fails, the one-time consumption verification fails, or the theoretical value of the chain tail anchor point is inconsistent with the chain tail anchor point, the service data packet is rejected and the abnormal event is recorded.
[0056] Preferably, the second temporary credential carried in the service data packet is obtained in the following manner: the terminal device uses the first temporary credential in the credential pool window by window in ascending order of credential sequence number, uses one first temporary credential corresponding to each valid time window within each valid time window, and carries the used first temporary credential as the second temporary credential in the service data packet; for example, the second temporary credential, the credential sequence number of the second temporary credential, and the service type identifier can be carried in the extended field of the header of the service data packet; after receiving the service data packet, the edge security proxy device parses the header of the service data packet to obtain the second temporary credential, the credential sequence number of the second temporary credential, and the service type identifier.
[0057] Preferably, the second temporary credential is subjected to time window verification in the following manner: the time window verification is used to verify whether the second temporary credential is within the valid time window corresponding to its credential sequence number; the edge security agent device obtains the current time and calculates the window sequence number of the current time window based on the current time and the distribution time of the credential pool: in, This refers to the window number of the current time window. This indicates the floor function. For the current time, The distribution time of the credential pool. The progressive interval between two adjacent valid time windows; the voucher serial number of the second temporary voucher. The window number of the current time window Consistency (i.e.) When the time window verification passes, the time window validation is successful. When the second temporary credential has expired, the service data packet is rejected and an abnormal event is recorded; If the second temporary credential is used prematurely, the service data packet is rejected and the abnormal event is recorded.
[0058] In a preferred embodiment, to smoothly switch time windows and prevent replays at window boundaries, the terminal device is allowed to use the first temporary credential corresponding to the next time window in advance when the current time window is less than a preset lead time; for example, when the following conditions are met... When (where 0.2 is the proportional coefficient of the preset advance amount relative to the progressive interval; when the progressive interval When the value is 1 minute, the advance amount is 12 seconds, allowing the use of the first... A first temporary credential; in this case, when the second temporary credential obtained based on the first temporary credential used in advance passes the verification, the edge security agent device immediately updates the maximum used credential sequence number it maintains to that credential sequence number, so that the credential corresponding to the original time window will be rejected even if it is replayed at the time of window switching.
[0059] Preferably, a one-time consumption verification is performed on the credential serial number in the following manner: the one-time consumption verification is used to verify that the second temporary credential is used only once within its valid time window; the edge security proxy device maintains the maximum used credential serial number for each of the terminal devices; when the credential serial number of the second temporary credential is greater than the maximum used credential serial number (i.e., ... When the one-time consumption verification passes, the voucher sequence number of the second temporary voucher is less than or equal to the maximum used voucher sequence number (i.e., When the service deduplication cache is accessed, the service deduplication cache is further queried. The service deduplication cache is used to store the hash digest of the service payload of recently processed service data packets. Its capacity can be dynamically allocated according to the number of terminal devices managed by the edge security agent device, and its validity period can be preset. When the hash digest of the service payload of the current service data packet exists in the service deduplication cache, the service data packet is determined to be a normal retransmission, and the corresponding service response in the service deduplication cache is directly forwarded. When the hash digest of the service payload of the current service data packet does not exist in the service deduplication cache, it is determined that a replay attack has occurred, the service data packet is rejected, the abnormal event is recorded, and an alarm is reported to the cloud security analysis platform.
[0060] Preferably, chained hash verification is performed on the second temporary certificate in the following manner: starting from the second temporary certificate, hash operations are continuously performed along the certificate chain from the head to the tail anchor point, and the result of the operation is used as the theoretical value of the tail anchor point; the number of consecutive hash operations is equal to the number of the multiple sets of first temporary certificates. With the certificate number The difference (i.e.) (times); the verification process of the chain hash verification can be represented as: in, This is the theoretical value of the chain tail anchor point. This is the second temporary certificate. The number of the multiple sets of first temporary vouchers. The document number is the serial number of the document. The preset hash function is used; the edge security proxy device compares the theoretical value of the chain tail anchor with the chain tail anchor stored locally. If they match, the chain hash verification is determined to be successful, the second temporary certificate is the real certificate in the certificate chain, and the used maximum certificate sequence number is updated to the certificate sequence number of the second temporary certificate; if they do not match, the chain hash verification is determined to be unsuccessful.
[0061] Preferably, the second temporary credential verification is determined as follows: if the time window verification, the one-time consumption verification, and the chain hash verification all pass, the second temporary credential verification is determined to be successful, and subsequent processing is performed on the business data packet; if the time window verification fails, the one-time consumption verification fails, or the chain hash verification fails, the business data packet is rejected, and an abnormal event is recorded; the abnormal event is an event record that includes at least the device identifier of the terminal device, the credential serial number of the second temporary credential, the reason for the verification failure, and the time of occurrence.
[0062] Preferably, credential revocation is performed on abnormal terminal devices in the following manner: when the edge security agent device detects abnormal events of the same terminal device continuously for a preset period of time (e.g., 5 minutes) and the number of such events reaches a preset number (e.g., 3 times) or more, a credential revocation process is triggered; in the credential revocation process, the edge security agent device marks the terminal device as undergoing credential revocation, rejects the terminal device's subsequent business requests, and reports the revocation event to the cloud security analysis platform so that the cloud security analysis platform can record the revocation event on the blockchain for evidence storage; after receiving the rejection response, the terminal device clears the locally stored credential pool and initiates re-access authentication to re-acquire the credential pool.
[0063] It is worth noting that, among the three verifications mentioned above, the time window verification and the one-time consumption verification are both completed based on the status information locally maintained by the edge security agent device, and the chain hash verification is completed based on the locally stored chain tail anchor point. The verification process does not require online key negotiation. Therefore, the processing latency of the verification process can be controlled within, for example, 5 milliseconds, which is beneficial to meeting the processing latency requirements of real-time services in the power Internet of Things. Furthermore, the one-time consumption verification, combined with the service deduplication caching, distinguishes between normal retransmissions and replay attacks, which helps to avoid misjudging normal retransmissions as attacks and to intercept replay attacks.
[0064] In this embodiment, a second temporary credential and credential number are extracted from the business data packet, and time window verification, one-time consumption verification, and chain hash verification are performed sequentially. The time window verification constrains the validity period of the credential, the one-time consumption verification constrains the number of times the credential can be used, and the chain hash verification verifies the authenticity of the credential based on the one-way nature of the hash chain. All three verifications can be completed locally on the edge security agent device without online key negotiation, enabling credential verification during the business communication phase to have both low latency and high reliability. As a result, forged credentials and replayed data packets can be identified and rejected, providing support for the secure transmission of business data packets in the power Internet of Things.
[0065] Step S4: After the verification is passed, the real-time level of the service data packet is determined according to the service type identifier, and the service data packet is processed according to the real-time level; In a preferred embodiment, determining the real-time level of the service data packet based on the service type identifier, and processing the service data packet according to the real-time level, includes: The real-time level of the service data packet is obtained by querying a preset service security policy table using the service type identifier. When the real-time level is Level 1, after performing an integrity check based on pre-distributed credentials on the service data packet, the service data packet is forwarded, and the hash digest of the service data packet is uploaded to the blockchain for evidence storage. When the real-time level is Level 2, the service data packet is forwarded after performing integrity verification and symmetric encryption based on pre-distributed credentials. When the real-time level is Level 3, the business data packet is subjected to integrity verification based on pre-distributed credentials, symmetric encryption and digital signature. After obtaining the digital signature value, the business data packet is forwarded and the digital signature value and the hash digest of the business data packet are uploaded to the blockchain for evidence storage. When the real-time level is Level 4, the business data packet is forwarded after performing complete security processing including integrity verification based on pre-distributed credentials, symmetric encryption, digital signature and blockchain notarization.
[0066] Preferably, the preset business security policy table is queried using the business type identifier in the following manner: The preset business security policy table refers to the business hierarchical processing policy table pre-configured in the edge security agent device, which uses the business type identifier as an index to record the real-time level, data integrity requirements, confidentiality requirements, and other information corresponding to each business type; After the edge security agent device parses the header of the business data packet to obtain the business type identifier, it queries the business security policy table using the business type identifier as an index to obtain the real-time level matching the business type of the business data packet; For example, the real-time level may include level 1, level 2, level 3, and level 4, wherein the processing latency requirement corresponding to level 1 is no more than 10 milliseconds, the processing latency requirement corresponding to level 2 is no more than 100 milliseconds, the processing latency requirement corresponding to level 3 is no more than 1 second, and level 4 corresponds to a second-level processing latency. For example, the business security policy table can record: the real-time level of automatic generation control commands (business type identifier 0x01) is Level 1, the integrity requirement is high, and the confidentiality requirement is medium; the real-time level of primary frequency regulation response (business type identifier 0x02) is Level 1; the real-time level of telemetry (business type identifier 0x05) is Level 2, the integrity requirement is high, and the confidentiality requirement is high; the real-time level of protection action signals (business type identifier 0x06) is Level 2; the real-time level of power quality monitoring (business type identifier 0x0A) is Level 3, the integrity requirement is medium, and the confidentiality requirement is medium; and the real-time level of equipment status inspection (business type identifier 0x0F) is Level 4, the integrity requirement is medium, and the confidentiality requirement is low.
[0067] Preferably, the integrity verification based on pre-distributed credentials refers to the process of calculating a verification value on the payload of the service data packet using an integrity verification key associated with the pre-distributed credential pool and comparing it with the corresponding verification value carried in the service data packet; for example, the integrity verification can use a keyed hash algorithm based on the SM3 hash function, and the digest length of its verification value can be 256 bits; the integrity verification key can be pre-derived and cached by the edge security agent device based on the pre-distribution process of the credential pool, so that it can be directly read and used when the service data packet arrives.
[0068] Preferably, the business data packets with a real-time level of 1 are processed in the following manner: integrity verification based on pre-distributed credentials is performed on the business data packets; after successful verification, the business data packets are immediately forwarded; and the hash digest of the business data packets is uploaded to the blockchain for evidence storage asynchronously. The asynchronous method means that the operation of uploading the blockchain for evidence storage is executed in parallel with the main processing flow, without blocking the forwarding of the business data packets. Since the 1st-level business only performs lightweight integrity verification and does not perform symmetric encryption or digital signature operations, its computational load is low. For example, its computational load can be approximately one-twentieth of the computational load of a complete digital signature, and the total security processing latency can be controlled within 5 milliseconds, which is beneficial for meeting the processing latency requirements of 1st-level businesses such as automatic power generation control commands.
[0069] For example, taking the issuance of automatic power generation control commands as an example: the dispatch center issues automatic power generation control command data packets to the protection device. The protection device marks the service type identifier 0x01 in the header of the data packet and sends it after attaching a temporary credential. After receiving the data packet, the edge security agent device performs verification on the temporary credential, and the verification time is, for example, no more than 2 milliseconds. It parses the service type identifier and queries the service security policy table to determine that the service data packet is a first-level service. It reads the integrity verification key corresponding to the terminal device in the current time window, calculates the verification value of the payload of the service data packet and compares it with the verification value carried in the service data packet. This process takes, for example, no more than 2.5 milliseconds. After the comparison is successful, the service data packet is immediately forwarded to the execution mechanism. The forwarding time is, for example, no more than 0.5 milliseconds, and the hash digest of the service data packet is asynchronously uploaded to the blockchain for evidence storage. The above-mentioned end-to-end security processing delay is, for example, 3 to 5 milliseconds, which meets the latency requirement of no more than 10 milliseconds for automatic power generation control commands.
[0070] Preferably, the service data packets with a real-time level of 2 are processed in the following manner: integrity verification based on pre-distributed credentials is performed on the service data packets; after successful verification, symmetric encryption is performed on the service data packets, and the encrypted service data packets are forwarded. For example, the symmetric encryption can adopt the stream cipher mode of the SM4 block cipher, with a key length of 128 bits, to maintain high encryption and decryption speed while ensuring data confidentiality. The total security processing delay of the 2nd level service can be controlled within 20 milliseconds, which is beneficial to meeting the processing delay requirements of services such as telemetry and protection action signals.
[0071] For example, taking the reporting scenario of telemetry data as an example: the smart meter reports telemetry data and marks the service type identifier 0x05 in the header of the data packet; after receiving it, the edge security agent device queries the service security policy table to determine that it is a second-level service, performs an integrity check based on pre-distributed credentials, and after the check passes, uses the symmetric encryption algorithm to encrypt the data payload and forwards the encrypted data to the scheduling center; the security processing delay of the above process is, for example, 15 to 20 milliseconds, which meets the latency requirement of no more than 100 milliseconds.
[0072] Preferably, the business data packets with a real-time level of 3 are processed in the following manner: integrity verification based on pre-distributed credentials, symmetric encryption, and digital signature are sequentially performed on the business data packets; the processed business data packets are forwarded; and the digital signature value and the hash digest of the business data packets are asynchronously uploaded to the blockchain for evidence storage. The integrity verification, symmetric encryption, and digital signature are performed synchronously and sequentially, and the digital signature provides non-repudiation for the business data. For example, the symmetric encryption can use the block chaining mode of the SM4 block cipher, with a key length of 128 bits. The digital signature can use a signature method based on the SM2 elliptic curve public-key cryptography algorithm, whereby the edge security agent device uses its own private key to perform a signature operation on the hash value of the encrypted payload. The total security processing latency of the 3rd level business can be controlled within 500 milliseconds, meeting the latency requirement of no more than 1 second.
[0073] For example, taking the reporting scenario of power quality monitoring data as an example: the power distribution terminal reports power quality monitoring data according to a preset cycle (e.g., once every 5 minutes) and marks the service type identifier 0x0A in the header of the data packet; after receiving it, the edge security agent device queries the service security policy table to determine that it is a third-level service, and sequentially performs integrity verification, symmetric encryption and digital signature based on pre-distributed credentials, and forwards the processed data to the target platform; in the above process, the time taken for integrity verification is, for example, about 3 milliseconds, the time taken for symmetric encryption is, for example, about 8 milliseconds, the time taken for digital signature is, for example, about 35 milliseconds, and the total security processing delay is, for example, about 46 milliseconds, which meets the delay requirement of no more than 1 second.
[0074] Preferably, the service data packets with a real-time level of four are processed in the following manner: The service data packets undergo complete security processing, which includes integrity verification based on pre-distributed credentials, symmetric encryption, digital signature, and blockchain notarization. These security operations are executed synchronously and sequentially, and the service data packets are forwarded after processing. The blockchain notarization is used to upload the hash digest of the service data packet and the digital signature value to the blockchain for notarization. Level four services allow for processing delays in the order of seconds. For example, in the scenario of reporting a device status inspection report: the terminal device reports inspection data and marks the service type identifier 0x0F in the header of the data packet. The edge security agent device queries the service security policy table to determine that it is a level four service, sequentially performs integrity verification, symmetric encryption, and digital signature, and uploads the hash digest of the service data packet and the digital signature value to the blockchain for notarization before forwarding the service data packet. The security processing delay of the above process is, for example, 500 milliseconds to 1000 milliseconds, which is within its allowed processing delay range.
[0075] Preferably, the edge security proxy device can also maintain a pre-calculation buffer to further reduce online security processing latency: the edge security proxy device forms a pre-calculation window based on the current time and extends forward and backward by a preset time period. Within the pre-calculation window, the security parameters required for subsequent communication with each terminal device are pre-calculated and cached, and the pre-calculation window is updated in a rolling manner according to a preset step size; for example, the preset time period can be 30 seconds before and 30 seconds after the current time, and the preset step size can be 10 seconds; the security parameters include at least a session key fragment and an integrity verification key fragment; through the pre-calculation buffer, when a service data packet arrives, the edge security proxy device directly reads the required security parameters without online real-time calculation, thereby further reducing online security processing latency.
[0076] It should be noted that the process of processing the service data packets according to the real-time level described above can also be understood as dynamically reducing the security processing pipeline (i.e., a set of security operations executed sequentially on the service data packets) according to the real-time level: the higher the real-time level, the fewer types of security operations the edge security agent device performs on the service data packets and the shorter the processing path; the lower the real-time level, the more complete the types of security operations performed; for example, the first-level service only performs integrity verification, the second-level service adds symmetric encryption on this basis, the third-level service further adds digital signature, and the fourth-level service performs complete security processing; thus, service data packets with different real-time requirements can achieve a balance between security and processing latency that is adapted to their business characteristics.
[0077] In this embodiment, the real-time level of a business data packet is determined by querying a preset business security policy table based on the business type identifier. Different combinations of security operations are then performed on the business data packet according to the real-time level: for high real-time services, only lightweight integrity verification is performed and blockchain notarization is completed asynchronously, which helps to ensure the processing latency of hard real-time services; for medium real-time services, security operations such as symmetric encryption and digital signatures are added step by step, which can gradually enhance data confidentiality and non-repudiation within the allowable latency range; in conjunction with the pre-calculation buffer for pre-calculation and caching of security parameters, the online processing latency is further compressed; thus, the secure processing of business data packets can adapt to the real-time requirements of different services while ensuring the security of power Internet of Things services.
[0078] Step S5: Perform local threat detection on the processed business data packets to obtain threat information, then convert the threat information into a threat intelligence object in a preset unified format, and upload the threat intelligence object to the cloud security analysis platform; In a preferred embodiment, local threat detection is performed on the processed business data packets to obtain threat information, and then the threat information is converted into a threat intelligence object in a preset unified format, including: Perform local threat detection on the processed business data packets to obtain threat events; The time of occurrence of the threat event is used as the timestamp of the threat occurrence, and the threat type and threat confidence level of the threat event are determined. The threat type is matched with a preset threat type and attack chain stage mapping rule base; When the threat type matches the mapping rule base, the attack chain stage in the mapping rule base corresponding to the threat type is determined as the attack chain stage corresponding to the threat event; When the threat type does not match the mapping rule base, the abnormal behavior sequence of the terminal device is collected, and the abnormal behavior sequence is matched with the preset attack chain stage behavior pattern base to obtain the attack chain stage corresponding to the threat event. Based on the timestamp of the threat occurrence and the device identifier of the terminal device, perform correlation analysis on multiple threat events belonging to the terminal device within a preset association time window to obtain a list of associated threat identifiers; The original alarm information of the threat event is digested to obtain the original alarm summary; The threat occurrence timestamp, the power IoT layer identifier to which the edge security agent device belongs, the device identifier of the terminal device, the threat type, the attack chain stage, the threat confidence level, the list of associated threat identifiers, and the original alarm summary are encapsulated into the threat intelligence object in a preset unified format.
[0079] Preferably, local threat detection is performed on the processed service data packets in the following manner: the edge security agent device continuously performs local threat detection during operation, and the detection methods include, but are not limited to, one or more of the following: monitoring of abnormal events such as credential verification failure, detection of abnormal service data packet formats, detection of abnormal communication frequencies, and detection of device behavior deviating from the baseline; when any of the above detections detects an anomaly, a corresponding threat event is generated; the threat event refers to an event record that describes the result of a threat detection, such as including the detection time, the data packets or behaviors involved, and an anomaly description.
[0080] Preferably, the threat occurrence timestamp, threat type, and threat confidence level of the threat event are determined in the following manner: the time of occurrence of the threat event is used as the threat occurrence timestamp, and the threat occurrence timestamp is unified to Coordinated Universal Time (UTC) to facilitate subsequent cross-level and cross-device threat correlation analysis; the threat type is used to indicate the category to which the threat behavior belongs, such as including one or more of tampering, eavesdropping, denial of service, forgery, replay, and unauthorized access; the threat confidence level is used to characterize the degree of certainty about the threat event, and is a floating-point number between 0 and 1, calculated by the edge security agent device according to a preset confidence level calculation rule, for example, it can be calculated according to the following formula: in, The threat confidence level of the threat event (also known as the initial confidence level, which is used by the cloud security analysis platform for subsequent calculation of the comprehensive threat confidence level). To test credibility, For feature significance, For historical credibility, For resource criticality, to These are the weights of the four parameters mentioned above; for example: Based on the reliability of the detection methods, the weights are determined as follows: 0.95 for signature-based detection, 0.70 for anomaly-based detection, and 0.55 for heuristic-based detection. The weight is determined based on the fact that the lower the false alarm rate of the detection method, the higher the weight of this parameter in terms of confidence. This reflects the degree of matching between threat characteristics and a known attack signature database. A value of 1.0 indicates a perfect match, 0.5 indicates a partial match, and 0.2 indicates a weak match. The weights are as follows: ; This represents the alarm accuracy rate of the terminal device over a set period of time. The initial value can be 0.5, and it is dynamically adjusted based on alarm confirmation status. Its weight... ; Based on the importance of the affected equipment, the values are determined as follows: relay protection devices are assigned a value of 1.0, smart meters are assigned a value of 0.7, and sensors are assigned a value of 0.4. The aforementioned weights can be determined using the analytic hierarchy process combined with expert experience and passed a consistency test.
[0081] Preferably, the threat type is matched with the preset threat type and attack chain stage mapping rule base in the following way: the preset threat type and attack chain stage mapping rule base refers to the rule base pre-built in the edge security agent device that records the correspondence between threat types and attack chain stages; the attack chain stages include seven stages: reconnaissance and tracking, weaponization, payload delivery, vulnerability exploitation, installation and implantation, command and control, and target achievement; for example, the mapping rule base records the following correspondences: port scanning and abnormal detection requests correspond to the reconnaissance and tracking stage; abnormal file uploads and firmware update requests correspond to the weaponization stage; abnormal substation event messages and sampled value message injection and forged messages correspond to the payload delivery stage; buffer overflows and abnormal instruction sequences correspond to the vulnerability exploitation stage; abnormal process creation and configuration tampering correspond to the installation and implantation stage; abnormal external connections, covert channel communication, and periodic heartbeats correspond to the command and control stage; abnormal control commands, malfunction signals, and data leakage correspond to the target achievement stage.
[0082] In a preferred embodiment, when the threat type does not match the mapping rule base, the attack chain stage corresponding to the threat event is obtained in the following way: Collecting abnormal behavior sequences of the terminal device; the abnormal behavior sequence refers to a sequence of one or more abnormal behaviors occurring on the terminal device within a preset time window, arranged in order of occurrence, such as 60 seconds; matching the abnormal behavior sequence with a preset attack chain stage behavior pattern library using edit distance; the preset attack chain stage behavior pattern library refers to a pre-constructed set of typical abnormal behavior patterns corresponding to each attack chain stage; the edit distance refers to the minimum number of edit operations required to convert one sequence into another, and a smaller edit distance indicates a greater similarity between the two sequences; when the similarity between the abnormal behavior sequence and a certain behavior pattern in the behavior pattern library is greater than or equal to a preset similarity threshold (e.g., 0.7), the attack chain stage corresponding to that behavior pattern is determined as the attack chain stage corresponding to the threat event; for example, if the abnormal behavior sequence composed of abnormal external connections and periodic heartbeats has a similarity of 0.85 with the behavior pattern of the command and control stage, then the attack chain stage corresponding to the threat event is determined to be the command and control stage.
[0083] Preferably, the associated threat identifier list is obtained by: performing association analysis on multiple threat events belonging to the terminal device within a preset association time window according to the threat occurrence timestamp and the device identifier of the terminal device; the association analysis refers to establishing a relationship between multiple threat events that are temporally related to the same terminal device. For example, when the time difference between two threat events is less than the preset association time window (e.g., 60 seconds) and both belong to the same device identifier, the identifiers of the two events are written into their respective associated threat identifier lists; the associated threat identifier list is used to record the identifiers of other threat events associated with the current threat event, for the cloud security analysis platform to use directly in subsequent association analysis.
[0084] Preferably, the original alarm summary is obtained by: performing a summary calculation on the original alarm information of the threat event; the original alarm summary refers to a fixed-length summary value obtained by calculating the original alarm information of the threat event using a summary algorithm, and the original alarm information includes, for example, the data packet fragment that triggered the detection and a description of abnormal behavior; the summary algorithm can be, for example, the SM3 hash algorithm; using the summary value instead of the original alarm information to write to the threat intelligence object can reduce the data volume of the threat intelligence object while retaining the ability to identify the original alarm information.
[0085] Preferably, the fields are encapsulated into the threat intelligence object in the following manner: The threat occurrence timestamp, the power IoT layer identifier of the edge security agent device, the device identifier of the terminal device, the threat type, the attack chain stage, the threat confidence level, the list of associated threat identifiers, and the original alarm summary are serialized and encapsulated according to a preset unified format to obtain the threat intelligence object. The preset unified format refers to a pre-agreed unified organization rule for the field composition, field order, and encoding method of the threat intelligence object, enabling the threat intelligence objects uploaded by different edge security agent devices to be parsed and summarized by the cloud security analysis platform in a consistent manner. The power IoT layer identifier of the edge security agent device is used to indicate whether the scope monitored by the edge security agent device belongs to the perception layer, network layer, platform layer, or application layer. For example, the layer identifier of the edge security agent device deployed in the substation is the application layer. The threat intelligence object can be serialized into a preset structured text format before uploading.
[0086] It is worth noting that the aforementioned local threat detection, attack chain stage determination, and threat intelligence object encapsulation are all completed locally on the edge security agent device. The edge security agent device only uploads the encapsulated threat intelligence object to the cloud security analysis platform. As a result, the threat information uploaded by each edge security agent device is presented in a unified format, providing a standardized and structured data foundation for the cloud security analysis platform to perform cross-level threat correlation analysis. At the same time, the attack chain stage field gives single-point threat information attack chain context semantics, which helps improve the accuracy of subsequent correlation analysis.
[0087] In this embodiment, threat events are obtained by performing local threat detection on the processed business data packets, and their threat occurrence timestamps, threat types, and threat confidence levels are determined. Then, the attack chain stage corresponding to the threat event is determined by matching the mapping rule base or abnormal behavior sequence. The above information, along with the associated threat identifier list and the original alarm summary, is packaged into a threat intelligence object in a preset unified format and uploaded to the cloud security analysis platform. In the above process, threat detection and attack chain stage determination are completed locally on the edge side, which helps to reduce the amount of data uploaded and the processing burden on the cloud. The unified format of the threat intelligence object and the attack chain stage field enable threat information scattered across different levels and devices to be aggregated and correlated in the cloud, providing a data foundation for the identification and handling of cross-layer collaborative attacks.
[0088] Step S6: Receive the linkage response instruction generated by the cloud security analysis platform based on the threat intelligence object, and execute the corresponding security protection operation according to the linkage response instruction.
[0089] In a preferred embodiment, the cloud-based security analysis platform generates a coordinated response command based on the threat intelligence object in the following manner: The cloud-based security analysis platform constructs a global threat correlation graph based on the threat intelligence objects, and analyzes the global threat correlation graph to obtain attack paths and comprehensive threat confidence. Based on the attack path and the overall threat confidence level, a linkage response command is generated and sent to the edge security agent device.
[0090] Preferably, the cloud-based security analysis platform constructs a global threat correlation graph based on the threat intelligence objects, and analyzes the global threat correlation graph to obtain attack paths and comprehensive threat confidence, including: The cloud-based security analysis platform constructs the global threat association graph by using the devices and communication links involved in the threat intelligence objects as nodes and the threat relationships represented by the threat intelligence objects as edges. Perform temporal correlation analysis on the threat intelligence objects in the global threat correlation graph, and determine multiple threat intelligence objects whose threat occurrence timestamps are within the same preset time window as having temporal correlation; Perform causal association analysis on the threat intelligence objects in the global threat association graph, and identify multiple threat intelligence objects that satisfy the preset pre-order and post-order transition relationship in the attack chain stage as having causal correlation. Perform homology association analysis on the threat intelligence objects in the global threat association graph to identify multiple threat intelligence objects with the same or similar attack characteristics as having homology correlation; Multiple threat intelligence objects that simultaneously satisfy the temporal correlation, the causal correlation, and the common origin correlation are associated and bound through the associated threat identifier list to obtain a cross-layer attack chain, and the cross-layer attack chain is determined as the attack path.
[0091] Preferably, the cloud-based security analysis platform generates a coordinated response command based on the attack path and the overall threat confidence level, including: The cloud-based security analysis platform compares the comprehensive threat confidence level with a preset first threshold and a second threshold. When the overall threat confidence level is greater than or equal to the first threshold, a global linkage response instruction is generated. The global linkage response instruction is used to instruct the edge security agent device to block communication, update the blacklist, and adjust the authentication policy of the affected device. When the overall threat confidence level is greater than or equal to the second threshold and less than the first threshold, a local early warning instruction is generated. The local early warning instruction is used to instruct the edge security agent device in the affected area to perform enhanced monitoring. When the overall threat confidence level is less than the second threshold, the linkage response command is not generated.
[0092] Preferably, the global threat association graph is constructed in the following manner: the cloud security analysis platform constructs the global threat association graph using the devices and communication links involved in the threat intelligence object as nodes and the threat relationships represented by the threat intelligence object as edges; the global threat association graph is a graph structure used to describe the association relationships between each device, each communication link and the threat events occurring on it, and is dynamically updated with the received threat intelligence objects; for example, when a new threat intelligence object is received, the cloud security analysis platform includes the devices and communication links involved in it as nodes in the global threat association graph, associates the threat intelligence object as an edge with the corresponding node, and establishes a new association relationship when the threat intelligence object meets the association conditions with existing nodes or edges.
[0093] Preferably, the temporal correlation analysis is performed in the following manner: the temporal correlation analysis refers to the temporal correlation analysis of multiple threat events within the same time window; the cloud security analysis platform adopts a temporal sliding window mechanism, and the window size and sliding step of the temporal sliding window are preset, for example, the window size can be set to 60 seconds and the sliding step can be set to 10 seconds; multiple threat intelligence objects occurring within the same sliding window are considered to have temporal correlation.
[0094] Preferably, the causal correlation analysis is performed in the following manner: the causal correlation analysis refers to the causal relationship analysis between preceding and subsequent threat events; the cloud security analysis platform determines the causal logic between threat events based on the attack chain stage field in each of the threat intelligence objects; the preset preceding and following transition relationship refers to the order of occurrence between each stage of the attack chain, for example, events in the reconnaissance and tracking stage usually occur before events in the payload delivery stage, events in the payload delivery stage usually occur before events in the vulnerability exploitation stage, and so on; multiple threat intelligence objects that satisfy the above preceding and following transition relationship are considered to have causal correlation.
[0095] Preferably, the homology correlation analysis is performed in the following manner: the homology correlation analysis refers to the consistency analysis of the same attack features appearing at different levels; the cloud security analysis platform compares the attack features in different threat intelligence objects, the attack features including, for example, one or more of source address ranges, packet feature fingerprints and malicious payload features; when the above attack features are the same or highly similar, the corresponding multiple threat intelligence objects are considered to have homology correlation.
[0096] Preferably, the cross-layer attack chain is obtained in the following way: the cloud security analysis platform associates and binds multiple threat intelligence objects that simultaneously satisfy the temporal correlation, the causal correlation, and the homogeneity correlation through the associated threat identifier list, that is, the identifiers of mutually related threat intelligence objects are written into each other's associated threat identifier lists to form a cross-layer attack chain; the cross-layer attack chain reflects the complete process of an attack behavior at multiple power Internet of Things layers, and the cross-layer attack chain is determined as the attack path.
[0097] Preferably, the comprehensive threat confidence score can be obtained using a hierarchical calculation model that combines weighted fusion of single threat confidence scores with enhanced correlation analysis. The hierarchical calculation model refers to a two-layer calculation structure that first performs weighted fusion of the threat confidence scores of individual threat intelligence objects, and then enhances the results by combining them with correlation analysis. The enhanced correlation analysis refers to applying an enhancement to the weighted fusion result when multiple threat intelligence objects are determined to belong to the same attack. For example, the correlation enhancement coefficient can be calculated using the following formula: in, The correlation enhancement coefficient is... , and These are the weighting coefficients for the strength of each association (for example, they can all be set to 0.15, i.e.) ), For temporal correlation strength, For the strength of causal relationship, For the homology association strength, the values of the above three association strengths are all in the range of 0 to 1; when at least two association analyses hit simultaneously, the association enhancement coefficient is greater than 1, otherwise the association enhancement coefficient is 1.
[0098] Furthermore, the comprehensive threat confidence level can be calculated according to the following formula: in, The overall threat confidence level is... The correlation enhancement coefficient is... The number of threat intelligence targets belonging to the same attack. Let i be the threat confidence level of the i-th threat intelligence target. The time decay weight (which can be calculated according to...) Calculation, where Let be the time difference between the i-th threat intelligence target and the latest threat intelligence target. This is a preset attenuation coefficient, for example, it can be set to 0.01, i.e. The contribution of earlier threat events to the current threat situation decays exponentially over time; min indicates the smaller of 1 and the enhanced weighted average, meaning the overall threat confidence does not exceed 1.
[0099] Taking the three threat intelligence objects in this embodiment as examples: the threat confidence levels of the three threat intelligence objects are 0.75, 0.82, and 0.70, respectively; the timestamp intervals of their threat occurrences are 0 seconds, 3 seconds, and 5 seconds, respectively; and the weighted average after time decay weighted fusion is approximately 0.756; the temporal correlation strength is 0.95, the causal correlation strength is 1.0, and the homology correlation strength is 0.90. The correlation enhancement coefficient is calculated according to the aforementioned correlation enhancement coefficient formula. The value is 1.4275; the final calculated comprehensive threat confidence level is... When the value is 1.0 (i.e., the upper limit of the value is reached after association enhancement), the first threshold is reached, triggering a global linkage response.
[0100] In a specific example, consider a cross-layer attack scenario: The attacker first compromises an environmental monitoring sensor within the station using phishing techniques. This sensor is located at the perception layer. Using this sensor as a springboard, the attacker sends forged substation event messages targeting general objects to the internal network. This attack occurs at the network layer. The attacker then attempts to mislead the protection device into malfunctioning. This attack occurs at the application layer. After detecting an abnormal external connection from a sensor, the perception layer edge security agent generates the first threat intelligence object. Its threat source level is the perception layer, the threat type is abnormal external connection, the attack chain stage is the command and control stage, and the threat confidence level is 0.75. After detecting an abnormal packet, the network layer edge security agent generates the second threat intelligence object. Its threat source level is the network layer, the threat type is forged packet, the attack chain stage is the payload delivery stage, and the threat confidence level is 0.82. Its threat occurrence timestamp is 3 seconds after the threat occurrence timestamp of the first threat intelligence object. After detecting that the protection device receives an abnormal command, the application layer edge security agent generates the third threat intelligence object. Its threat source level is the application layer, the threat type is abnormal command, the attack chain stage is the target achievement stage, and the threat confidence level is 0.70. Its threat occurrence timestamp is 5 seconds after the threat occurrence timestamp of the first threat intelligence object. After receiving the three threat intelligence objects, the cloud security analysis platform performs correlation analysis within the preset time window: In terms of temporal correlation, the three threat events occurred within 5 seconds and are highly correlated in time; in terms of causal correlation, the command and control phase, payload delivery phase, and target achievement phase constitute a complete attack chain transfer relationship; in terms of homogeneity correlation, the three threat events have the same source address range and matching message feature fingerprints; based on the above three types of correlation analysis results, the cloud security analysis platform determines that the three threat intelligence objects belong to the same cross-layer attack, and binds the three threat intelligence objects into a cross-layer attack chain through the associated threat identifier list.
[0101] Preferably, the linkage response command is generated and issued in the following manner: The cloud security analysis platform compares the comprehensive threat confidence with the preset first threshold and second threshold, wherein the first threshold is greater than the second threshold; for example, the value range of the first threshold can be 0.7 to 0.9 (0.8 in a specific example), and the value range of the second threshold can be 0.3 to 0.6 (0.5 in a specific example); when the comprehensive threat confidence is greater than or equal to the first threshold, a global linkage response command is generated; for example, for the aforementioned cross-layer attack scenario, the cloud security analysis platform issues a blocking command to all edge security agent devices, adds the sensor device identifier used as an attack springboard to the permanent blacklist, and temporarily upgrades the authentication policy of the affected devices (for example, temporarily upgrading the authentication policy of terminal devices with standard security level across the entire site to high security level authentication for a preset duration, such as 30 minutes), and stores the complete attack chain information in the blockchain to generate an immutable audit record.
[0102] When the overall threat confidence level is greater than or equal to the second threshold and less than the first threshold, the cloud security analysis platform generates a local early warning instruction. The local early warning instruction is used to instruct the edge security agent device in the affected area to perform enhanced monitoring. For example, the enhanced monitoring may include increasing the detection frequency of terminal device behavior in the affected area and expanding the detection range, without triggering a global blockade, so as to avoid affecting the overall operation of the power Internet of Things. When the overall threat confidence level is less than the second threshold, only logs are recorded for audit analysis, and the linkage response instruction is not generated.
[0103] In a preferred embodiment, the first threshold and the second threshold are determined as follows: based on historical alarm data and false alarm rate statistics of power system safe operation, with a false alarm rate of less than 5% and a false alarm rate of less than 3% as constraints, a cutoff point is selected through subject operating characteristic curve analysis, and finally determined in combination with expert experience.
[0104] Preferably, the security protection operation corresponding to the linkage response instruction is executed in the following manner: After receiving the linkage response instruction issued by the cloud security analysis platform, the edge security agent device executes the corresponding security protection operation according to the linkage response instruction; for example, when receiving the global linkage response instruction, the edge security agent device performs operations such as blocking communication, updating the local blacklist, and re-performing authentication checks on the affected devices according to the adjusted authentication policy; when receiving the local early warning instruction, the edge security agent device performs enhanced monitoring, such as increasing the frequency of behavior detection of relevant terminal devices and strengthening the recording and reporting of abnormal behaviors.
[0105] It should be noted that the cloud-based security analysis platform and each of the edge security agent devices can be connected via a power dispatch data network; the cloud-based security analysis platform can also record key security events on the blockchain for evidence storage, including authentication events, threat events, linkage response events, and cross-layer attack chain information, to generate tamper-proof audit records for post-event auditing and traceability; the entire process from the generation of the first threat intelligence object to the completion of the linkage response can be controlled within, for example, within 20 seconds.
[0106] In this embodiment, the cloud-based security analysis platform constructs a global threat correlation graph based on threat intelligence objects in a unified format uploaded by various edge security agent devices. It then identifies cross-layer attack chains and obtains attack paths through three types of analysis: temporal correlation, causal correlation, and homology correlation. Finally, it calculates a comprehensive threat confidence level by combining single threat confidence level weighted fusion with correlation analysis enhancement. Based on this, it generates and sends coordinated response instructions to the edge side in a hierarchical manner. In the above process, threat information scattered at different levels is aggregated and correlated in the cloud, enabling cross-layer collaborative attacks to be identified and triggering hierarchical responses adapted to their threat levels, thus providing support for cross-layer collaborative handling of the power Internet of Things.
[0107] Another embodiment of the present invention provides a power Internet of Things (IoT) communication security protection system, applied to the power IoT communication security protection method described in the above embodiments. The system includes an edge security proxy device and a cloud security analysis platform, wherein the edge security proxy device is communicatively connected to the cloud security analysis platform. The edge security proxy device is used to execute the processing flow executed by the edge security proxy device as described in steps S1 to S6 of the above embodiments, and the cloud security analysis platform is used to execute the processing flow executed by the cloud security analysis platform as described in steps S5 to S6 of the above embodiments, so as to achieve cross-layer collaborative security protection for power IoT communication.
[0108] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.
Claims
1. A method for protecting communication security in the power Internet of Things, characterized in that, Performed by the edge security agent device, including: Receive access requests from terminal devices seeking to connect to the power Internet of Things, and determine the security level of the terminal devices based on their device fingerprints; According to the security level, a matching authentication method is selected from the preset authentication protocol family to complete the authentication. After the authentication is successful, a credential pool is pre-distributed to the terminal device and the tail anchor of the credential pool is stored. Extract temporary credentials and service type identifiers from the service data packets of the terminal device, and perform verification on the temporary credentials according to the chain tail anchor point; After the verification is passed, the real-time level of the service data packet is determined according to the service type identifier, and the service data packet is processed according to the real-time level. Local threat detection is performed on the processed business data packets to obtain threat information. The threat information is then converted into a threat intelligence object in a preset unified format and uploaded to the cloud security analysis platform. Receive the linkage response command generated by the cloud security analysis platform based on the threat intelligence object, and execute the corresponding security protection operation according to the linkage response command.
2. The power Internet of Things communication security protection method as described in claim 1, characterized in that, Authentication is completed by selecting a matching authentication method from a preset family of authentication protocols according to the security level, including: When the security level is a preset first security level, a first two-way authentication method based on public key cryptography is selected from a preset authentication protocol family. The edge security proxy device generates a random number as a first challenge value and sends the first challenge value to the terminal device. The first authentication response returned by the terminal device based on the first challenge value is received, and the first authentication response is verified according to the first two-way authentication method. When the first authentication response passes the verification, the authentication of the terminal device is completed. When the security level is the preset second security level, a second two-way authentication method based on a pre-shared key is selected from a preset authentication protocol family. The edge security proxy device generates a random number as a second challenge value and sends the second challenge value to the terminal device. The second authentication response returned by the terminal device based on the second challenge value is received, and the second authentication response is verified according to the second two-way authentication method. When the second authentication response passes the verification, the authentication of the terminal device is completed. When the security level is the preset third security level, an implicit authentication method based on a physical unclonable function is selected from the preset authentication protocol family. The edge security proxy device sends the pre-stored physical unclonable function challenge value to the terminal device. The physical unclonable function response value returned by the terminal device is received and compared. When the physical unclonable function response values match, the authentication of the terminal device is completed. The first security level has a higher priority than the second security level, and the second security level has a higher priority than the third security level.
3. The power Internet of Things communication security protection method as described in claim 2, characterized in that, After successful authentication, a credential pool is pre-distributed to the terminal device, and the tail anchor of the credential pool is stored, including: After authentication is successful, a set of random numbers is generated as the tail anchor point of the credential pool; Starting from the tail anchor point, perform hash operations sequentially backward to obtain a certificate chain including multiple sets of first temporary certificates; wherein, each first temporary certificate in the multiple sets of first temporary certificates has a certificate serial number, and the head of the certificate chain is the first temporary certificate with the smallest certificate serial number. The credential chain is distributed to the terminal device as the credential pool, and the tail anchor point of the credential pool is stored at the same time.
4. The power Internet of Things communication security protection method as described in claim 3, characterized in that, The temporary credential is validated based on the chain tail anchor point, including: The first temporary credential retrieved by the terminal device from the credential pool is carried in the business data packet as the second temporary credential. Extract the second temporary credential and its serial number from the business data packet; The second temporary voucher is subject to time window verification and a one-time consumption verification of the voucher serial number. Perform chain hash verification on the second temporary credential to obtain the theoretical value of the chain tail anchor point, and compare the theoretical value of the chain tail anchor point with the chain tail anchor point; If both the time window verification and the one-time consumption verification pass, and the theoretical value of the chain tail anchor point is consistent with the chain tail anchor point, then the second temporary voucher verification is deemed to have passed. If the time window verification fails, the one-time consumption verification fails, or the theoretical value of the chain tail anchor point is inconsistent with the chain tail anchor point, the service data packet is rejected and the abnormal event is recorded.
5. The power Internet of Things communication security protection method as described in claim 4, characterized in that, The real-time level of the service data packet is determined based on the service type identifier, and the service data packet is processed according to the real-time level, including: The real-time level of the service data packet is obtained by querying a preset service security policy table using the service type identifier. When the real-time level is Level 1, after performing an integrity check based on pre-distributed credentials on the service data packet, the service data packet is forwarded, and the hash digest of the service data packet is uploaded to the blockchain for evidence storage. When the real-time level is Level 2, the service data packet is forwarded after performing integrity verification and symmetric encryption based on pre-distributed credentials. When the real-time level is Level 3, the business data packet is subjected to integrity verification based on pre-distributed credentials, symmetric encryption and digital signature. After obtaining the digital signature value, the business data packet is forwarded and the digital signature value and the hash digest of the business data packet are uploaded to the blockchain for evidence storage. When the real-time level is Level 4, the business data packet is forwarded after performing complete security processing including integrity verification based on pre-distributed credentials, symmetric encryption, digital signature and blockchain notarization.
6. The power Internet of Things communication security protection method as described in claim 5, characterized in that, The processed business data packets undergo local threat detection to obtain threat information, which is then converted into threat intelligence objects in a preset unified format, including: Perform local threat detection on the processed business data packets to obtain threat events; The time of occurrence of the threat event is used as the timestamp of the threat occurrence, and the threat type and threat confidence level of the threat event are determined. The threat type is matched with a preset threat type and attack chain stage mapping rule base; When the threat type matches the mapping rule base, the attack chain stage in the mapping rule base corresponding to the threat type is determined as the attack chain stage corresponding to the threat event; When the threat type does not match the mapping rule base, the abnormal behavior sequence of the terminal device is collected, and the abnormal behavior sequence is matched with the preset attack chain stage behavior pattern base to obtain the attack chain stage corresponding to the threat event. Based on the timestamp of the threat occurrence and the device identifier of the terminal device, perform correlation analysis on multiple threat events belonging to the terminal device within a preset association time window to obtain a list of associated threat identifiers; The original alarm information of the threat event is digested to obtain the original alarm summary; The threat occurrence timestamp, the power IoT layer identifier to which the edge security agent device belongs, the device identifier of the terminal device, the threat type, the attack chain stage, the threat confidence level, the list of associated threat identifiers, and the original alarm summary are encapsulated into the threat intelligence object in a preset unified format.
7. The power Internet of Things communication security protection method as described in claim 6, characterized in that, The cloud-based security analysis platform generates a coordinated response command based on the threat intelligence object in the following manner: The cloud-based security analysis platform constructs a global threat correlation graph based on the threat intelligence objects, and analyzes the global threat correlation graph to obtain attack paths and comprehensive threat confidence. Based on the attack path and the overall threat confidence level, a linkage response command is generated and sent to the edge security agent device.
8. The power Internet of Things communication security protection method as described in claim 7, characterized in that, The cloud-based security analysis platform constructs a global threat correlation graph based on the threat intelligence objects, and analyzes the global threat correlation graph to obtain attack paths and comprehensive threat confidence, including: The cloud-based security analysis platform constructs the global threat association graph by using the devices and communication links involved in the threat intelligence objects as nodes and the threat relationships represented by the threat intelligence objects as edges. Perform temporal correlation analysis on the threat intelligence objects in the global threat correlation graph, and determine multiple threat intelligence objects whose threat occurrence timestamps are within the same preset time window as having temporal correlation; Perform causal association analysis on the threat intelligence objects in the global threat association graph, and identify multiple threat intelligence objects that satisfy the preset pre-order and post-order transition relationship in the attack chain stage as having causal correlation. Perform homology association analysis on the threat intelligence objects in the global threat association graph to identify multiple threat intelligence objects with the same or similar attack characteristics as having homology correlation; Multiple threat intelligence objects that simultaneously satisfy the temporal correlation, the causal correlation, and the common origin correlation are associated and bound through the associated threat identifier list to obtain a cross-layer attack chain, and the cross-layer attack chain is determined as the attack path.
9. The power Internet of Things communication security protection method as described in claim 8, characterized in that, The cloud-based security analysis platform generates a coordinated response command based on the attack path and the overall threat confidence level, including: The cloud-based security analysis platform compares the comprehensive threat confidence level with a preset first threshold and a second threshold. When the overall threat confidence level is greater than or equal to the first threshold, a global linkage response instruction is generated. The global linkage response instruction is used to instruct the edge security agent device to block communication, update the blacklist, and adjust the authentication policy of the affected device. When the overall threat confidence level is greater than or equal to the second threshold and less than the first threshold, a local early warning instruction is generated. The local early warning instruction is used to instruct the edge security agent device in the affected area to perform enhanced monitoring. When the overall threat confidence level is less than the second threshold, the linkage response command is not generated.
10. A power Internet of Things (IoT) communication security protection system, characterized in that, It includes an edge security proxy device and a cloud security analysis platform, wherein the edge security proxy device and the cloud security analysis platform are communicatively connected; The edge security proxy device is used to receive access requests from terminal devices seeking to access the power Internet of Things, determine the security level of the terminal device based on the device fingerprint, select a matching authentication method from a preset authentication protocol family according to the security level to complete authentication, pre-distribute a credential pool to the terminal device after successful authentication and store the tail anchor of the credential pool, extract temporary credentials and service type identifiers from the service data packets sent by the terminal device, perform verification on the temporary credentials according to the tail anchor, determine the real-time level of the service data packets according to the service type identifier after successful verification, process the service data packets according to the real-time level, perform local threat detection on the processed service data packets to obtain threat information, convert the threat information into a threat intelligence object in a preset unified format, and upload the threat intelligence object to the cloud security analysis platform. The cloud-based security analysis platform is used to generate a linkage response command based on the threat intelligence object, and to send the linkage response command to the edge security agent device; The edge security agent device is also used to receive linkage response instructions issued by the cloud security analysis platform and to perform corresponding security protection operations according to the linkage response instructions.
