Unknown asset positioning system based on situation awareness

By designing an unknown asset positioning system based on situational awareness in the power monitoring system, using real-time monitoring and ARP protocol analysis, combined with classification and neural network model, the rapid identification and positioning of unknown assets is achieved, and the problem of inefficiency of traditional systems is solved. , improves the safety and operation and maintenance efficiency of the power monitoring system.

CN119026069APending Publication Date: 2024-11-26YUNNAN POWER GRID CO LTD PUER POWER SUPPLY BUREAU

Patent Information

Application Number
CN202410909352.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-08
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

Traditional situational awareness systems are inefficient when discovering unknown assets, have a high false alarm rate, and require a large amount of manpower and material resources to investigate, which threatens the safety and stability of the power monitoring system.

Method used

Design an unknown asset positioning system based on situational awareness, including real-time monitoring module, ARP protocol analysis module, situational awareness module, unknown asset positioning module and alarm and response module. By capturing network data packets in real time, analyzing ARP packets, using classification Algorithms, association analysis and neural network models identify and locate unknown assets, and generate network topology maps to visualize the placement.

Benefits of technology

It improves the positioning efficiency and accuracy of unknown assets, reduces the time and energy of operation and maintenance personnel to investigate, enhances the safety and protection capabilities of the power monitoring system, and promotes the construction of the safety management system of the power supervision system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119026069A_ABST
    Figure CN119026069A_ABST
Patent Text Reader

Abstract

The invention provides an unknown asset positioning system based on situation awareness. The system comprises a real-time monitoring module, an ARP protocol analysis module, a situation awareness module, an unknown asset positioning module and an alarm and response module. Data in a power grid system and historical data in an integrated system are monitored in real time, transmission data of an ARP protocol in a network are deeply analyzed, related information of unknown assets and a service system to which the unknown assets belong are analyzed and predicted through a situation awareness module, unknown network equipment is dynamically monitored and found in real time, and a network topological graph is established. And rapid positioning of assets is realized. And a corresponding alarm and response module is arranged, so that the security and the stability of the network are improved. According to the system, efficient, accurate and rapid identification and management of network assets can be realized, the safety protection capability of a situation awareness system is further exerted, the safety of a service system is guaranteed, and stable operation of a power monitoring system is guaranteed.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of power monitoring systems and network security technology, and specifically to an unknown asset positioning system based on situational awareness, which aims to identify unknown assets in a power monitoring system and quickly locate the type and business of unknown assets. Background Art

[0002] As power monitoring systems become increasingly complex and intelligent, the number of devices in the network continues to increase, including known and unknown assets. These unknown assets may include unregistered devices, illegally connected devices, etc., which pose a threat to the security and stability of the power monitoring system. At present, traditional situational awareness and collection devices have a long cycle for discovering unknown assets. They rely on grassroots operation and maintenance personnel to repeatedly enter the station for inspection and invest a lot of manpower and material resources to verify unknown assets, but this method has problems such as low efficiency and high false alarm rate. Therefore, this application proposes an unknown asset positioning system based on situational awareness, which aims to improve the positioning efficiency and accuracy of unknown assets and enhance the security of the power monitoring system. Summary of the invention

[0003] To solve the above problems, the present application provides an unknown asset positioning system based on situational awareness, which can realize real-time monitoring of data packet information in the network environment, perform situational awareness by analyzing the content of ARP messages and other relevant information in the data packets, and quickly locate and identify unknown assets.

[0004] To achieve the above purpose, the technical solution adopted in this application is:

[0005] A situation awareness-based unknown asset positioning system comprises a real-time monitoring module, an ARP protocol analysis module, a situation awareness module, an unknown asset positioning module, and an alarm and response module, wherein the real-time monitoring module is respectively connected to the ARP protocol analysis module and the situation awareness module, the unknown asset positioning module is respectively connected to the ARP protocol module, the situation awareness module, and the alarm and response module, and the situation awareness module is respectively connected to the ARP protocol module and the alarm response module;

[0006] The real-time monitoring module is used to capture ARP messages and related communication data in the network environment in real time and collect and integrate system-related historical data;

[0007] The ARP protocol analysis module is used to parse the captured ARP packets, extract key information, find abnormal connections through node analysis based on ARP protocol data, and output ARP protocol analysis results for subsequent processing and verification;

[0008] The situation awareness module is used to classify data through a classification algorithm and output classification results; identify network traffic patterns and user behavior patterns related to unknown assets based on an association analysis algorithm, and output identification results; predict the business and type of unknown assets by inputting the collected data into a preset asset prediction model, and output unknown asset prediction results; conduct a comprehensive analysis based on the captured data, classification results, identification results, and prediction results, and output a situation awareness assessment report;

[0009] The unknown asset positioning module is used to build a network topology map in combination with the ARP protocol analysis results and the situation awareness assessment report, analyze and locate unknown assets in the network topology map, output the positioning information of the corresponding unknown assets and visualize the network topology map;

[0010] The alarm and response module is used to generate alarm response suggestions by combining the situation awareness assessment report information, the unknown asset location information and the relevant system historical data;

[0011] The unknown assets include: unregistered devices, external mobile devices, network nodes of unknown origin, temporarily built systems or devices, forgotten old devices and illegally accessed devices.

[0012] Furthermore, the real-time monitoring module includes a data acquisition unit and a data protection unit;

[0013] The data acquisition unit is used to receive data packets in real time through a network interface and collect and integrate system-related historical data, and extract key indicators or attributes of unknown asset behaviors and characteristics from the data packets;

[0014] The data protection unit is used to protect the transmission of data packets through an encryption algorithm.

[0015] Furthermore, the ARP protocol analysis module is used to verify the captured ARP message content, determine abnormal connections between nodes based on node analysis, and check the source of data.

[0016] Furthermore, the situation awareness module includes a data classification unit, a correlation analysis unit, an asset prediction model and a situation assessment unit; the data classification unit, the correlation analysis unit and the asset prediction model are connected in sequence, and the situation assessment unit is respectively connected to the data classification unit, the correlation analysis unit and the asset prediction model.

[0017] Furthermore, the data classification unit is used to classify and label the features of the data of the unknown asset and output the classification results; the data classification unit adopts a decision tree classification algorithm, and the specific operation process includes the following steps:

[0018] Receive data from the real-time monitoring module and the ARP protocol analysis module, and format and standardize the data based on the received data;

[0019] Select key features as decision nodes through the received known asset data, and gradually split the data into different branches according to the relevant feature values ​​to form a tree structure;

[0020] The collected data of unknown assets is input into the trained decision tree model, and the classification judgment of the unknown assets is realized through the output leaf node position.

[0021] Furthermore, the association analysis unit is used to perform association analysis in combination with known asset information to find the association relationship of the unknown assets related thereto, and output the corresponding identification result. The operation process of the association analysis unit includes the following steps:

[0022] Using known asset information as prior knowledge, generate association rules between known assets;

[0023] Receive unknown asset data from the real-time monitoring module and convert the data into a transaction data set, setting a reasonable minimum support threshold and minimum confidence level based on the size and relevant characteristics of the data;

[0024] Finding frequent item sets through the minimum support threshold and generating association rules according to the found frequent item sets;

[0025] Set anomaly thresholds to determine abnormal patterns of association rules.

[0026] Furthermore, the asset prediction model is used to analyze the data packets and log data collected by the real-time monitoring module and extract feature information to predict the type of unknown assets and their business, and output the unknown asset prediction results; the asset prediction model adopts a neural network model.

[0027] Furthermore, the situation assessment unit is used to integrate the classification results, identification results and prediction results output by the data classification unit, the association analysis unit and the asset prediction model to form a comprehensive data set, and analyze the classification, type, association relationship with known assets and corresponding business of unknown assets based on the comprehensive perception and understanding of the situation by the comprehensive data set.

[0028] Furthermore, the unknown asset positioning module is used to locate unknown assets by building an updated network topology map, combining the equipment layout on the network topology map with the location information of known assets, and analyzing the area where the location assets are located.

[0029] Furthermore, the alarm and response module is used to visualize the relevant information of unknown assets based on the constructed network topology diagram, set the relevant alarm level according to the relevant historical data of the system, and track and evaluate the effect of alarm processing to adjust and optimize the strategy.

[0030] The beneficial effects of this application are:

[0031] This application can capture the data packet information in the network environment of the power grid system in real time and process the data packet through the real-time monitoring module. This real-time monitoring method and data processing can ensure the real-time cognition of the power monitoring system, provide high-quality data support for subsequent ARP protocol analysis and situational awareness modules, and can adapt to the ever-changing network environment and asset conditions, and maintain effective monitoring of newly emerging unknown assets. The ARP protocol analysis module can parse the ARP protocol field in the real-time monitored data packet, extract the corresponding key data to provide network connection information for the construction of the network topology map. The situational awareness module learns a classification model for the training data of known categories through a classification algorithm, and identifies the unknown asset category according to the classification model. The network traffic pattern and user behavior pattern related to the unknown asset are identified through the association analysis algorithm. The type and business of the unknown asset are predicted through the neural network model. Combined with the relevant output, the overall perception of the relevant information of the unknown asset is carried out, which greatly improves the speed and accuracy of the perception of the unknown asset information.

[0032] The unknown asset positioning module of the present application generates a network topology map of the power monitoring system based on the analysis results of the situation awareness module and the parsing of the ARP protocol, and determines the specific location of the unknown asset in the power monitoring system network in combination with the network topology map. This rapid positioning method improves the efficiency of locating unknown assets, making it possible to quickly locate the specific location and business of unknown assets, facilitating on-site inspection by operation and maintenance personnel, reducing the time and effort of manual inspection and processing, and greatly improving the work efficiency and processing speed of operation and maintenance personnel. The alarm and response module provides relevant alarms and tracks and evaluates the response and disposal effects of relevant staff. By tracking and evaluating the response and disposal adaptive optimization module of relevant personnel, the accuracy of the generation of recommendations for early warning information can be improved, which is conducive to promoting the safety supervision of the power supervision system, promoting the establishment of a comprehensive and efficient power monitoring system safety management system, and improving the overall safety management level and protection capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic diagram of an unknown asset positioning system based on situation awareness of the present application.

[0034] Figure 2 It is a structural diagram of the situation awareness module of this application.

[0035] Figure 3 This is a flowchart of the steps for building a network topology diagram for the unknown asset positioning module of this application. DETAILED DESCRIPTION

[0036] See also Figure 1 As shown in FIG. 3 , the present application provides an unknown asset positioning system based on situational awareness, which can realize real-time monitoring of data packet information in a network environment, perform situational awareness by analyzing the content of ARP (Address Resolution Protocol) messages and other relevant information in data packets, and quickly locate and identify unknown assets.

[0037] The unknown asset positioning system based on situational awareness of the present application comprises a real-time monitoring module, an ARP protocol analysis module, a situational awareness module, an unknown asset positioning module, and an alarm and response module; the real-time monitoring module is respectively connected to the ARP protocol analysis module and the situational awareness module, the unknown asset positioning module is respectively connected to the ARP protocol module, the situational awareness module, and the alarm and response module, and the situational awareness module is respectively connected to the ARP protocol module and the alarm response module;

[0038] The real-time monitoring module is used to capture ARP messages and related communication data in the network environment in real time and collect and integrate system-related historical data;

[0039] The unknown assets include: unregistered devices, external mobile devices, network nodes of unknown origin, temporarily built systems or equipment, forgotten old equipment and illegally accessed equipment.

[0040] In one embodiment, the real-time monitoring module includes a data acquisition unit and a data protection unit;

[0041] The data acquisition unit is used to receive data packets in real time through a network interface and collect and integrate system-related historical data, and extract key indicators or attributes of unknown asset behaviors and characteristics from the data packets;

[0042] Specifically, the data collection workflow includes data capture, data preprocessing, and error handling and logging;

[0043] The specific workflow of data capture includes the following steps:

[0044] Monitoring settings: Network interfaces that use one or more physical network interface cards (NICs) or virtual network interfaces to capture data packets; these interfaces can be accessed through programming interfaces, such as socket programming or Winsock. Socket programming is also known as socket programming. Winsock is a specification for network programming under Windows. This specification is a widely used, open, and multi-protocol network programming interface under Windows.

[0045] Set the capture mode: In order to capture all ARP packets on the network, you need to set the network interface to promiscuous mode. In promiscuous mode, the network interface will receive all packets passing through it.

[0046] Create a socket: Create a raw socket (Raw Socket) to receive and process raw network data packets. Use the socket(AF_PACKET, SOCK_RAW, htons(ETH_P_ALL)) function. The ETH_P_ALL parameter indicates that the socket should receive data packets of all Ethernet protocol types including the ARP protocol. The AF_PACKET parameter indicates the socket type, and the SOCK_RAW parameter indicates the address family. The socket(AF_PACKET, SOCK_RAW, htons(ETH_P_ALL)) function is a raw socket programming function.

[0047] Bind the socket to the interface: Use the bind() function to bind the socket to the previously selected network interface; the bind() function is used in JavaScript to create a new function that sets its this value when called and prefixes the given parameter list to the parameter sequence of the original function when calling the new function. JavaScript ("JS" for short) is a lightweight, interpreted or just-in-time compiled programming language with function-first execution.

[0048] Receive data packets: Use the recvfrom() function to receive data packets from the socket. The received data will contain a complete Ethernet frame header and data part. The recvfrom() function is a system call used to receive data from the socket.

[0049] When capturing a large number of data packets in the power grid system, asynchronous processing is used to separate data packet capture and parsing to reduce the impact on network performance.

[0050] The specific workflow of data preprocessing is as follows: clean the data after data packet parsing, including the number of data packets, data packet size, time-related characteristics, data packet information flags, access frequency, network traffic data, log data, and system monitoring data, remove outliers, missing values, and duplicate values ​​in the data to ensure the accuracy and reliability of the data, and extract key indicators or attributes that describe the behavior and characteristics of unknown assets from the original data after data cleaning, including: network traffic patterns, access frequency, system resource usage, etc.

[0051] In the specific real-time monitoring process, you may encounter errors such as network failures and interface disconnections. Use an exception capture mechanism to record detailed information about the exception, including error codes, error descriptions, etc., and record logs for subsequent troubleshooting; regularly record data collection log information, including collection time, data volume, abnormal conditions, etc., to facilitate analysis and optimization of real-time monitoring performance.

[0052] The data protection unit is used to protect the transmission of data packets through an encryption algorithm to ensure the security of data transmission.

[0053] Specifically, encryption is performed between the sending end and the receiving end so that data security is protected during real-time monitoring. The specific workflow is as follows: an encryption algorithm is selected, and the ACS key is encrypted by RSA (asymmetric encryption algorithm) and transmitted to the receiving end, and the receiving end obtains the AES (Advanced Encryption Standard) key after decryption; before the sending end sends the data packet, the data packet is encrypted using the AES encryption algorithm and key, and the encrypted data is decrypted using the corresponding key at the receiving end to obtain the original data packet, thereby ensuring data security and privacy.

[0054] The ARP protocol analysis module is used to parse the captured ARP message, extract key information, and find abnormal connections through node analysis based on the ARP protocol data for subsequent processing and verification;

[0055] In one embodiment, the ARP protocol analysis module is used to verify the captured ARP message content, determine abnormal connections between nodes based on node analysis, and check the source of data.

[0056] Specifically, the ARP protocol analysis module is used to check and verify the ARP message in the power grid system captured by the real-time monitoring module. The ARP message contains a checksum field to verify the integrity of the message; when parsing the ARP message, the checksum of the message needs to be calculated and compared with the checksum field in the message. If the two are inconsistent, it means that an error occurred during the transmission of the message and the message should be discarded; for each field in the ARP message, verify whether it complies with the provisions of the ARP protocol, including: the value of the hardware type field should be 1, the value of the protocol type field should be 0x0800, and the value of the operation code field should be 1 or 2, indicating and reply messages respectively. If the value of the field does not comply with the regulations, it means that the message is invalid and the message should be discarded; after extracting the IP address and MAC address in the ARP message, further verification is performed, including: verifying whether the IP address belongs to the valid IPv4 address range, whether the MAC address conforms to the format of the Ethernet MAC address, etc. If the verification fails, it means that the IP address or MAC address in the message is invalid and the message should be discarded. In a network environment, there may be duplicate ARP messages. By comparing the key fields of the message such as the source MAC address, source IP address, destination MAC address and destination IP address, it is determined whether it is a duplicate message. If it is a duplicate message, it is directly discarded; otherwise, it is passed to the subsequent situation awareness module for processing. MAC address, also known as Media Access Control Address, is also called media access control address, also known as local area network address (LAN Address).

[0057] Collect and verify ARP protocol data, and analyze the ARP protocol data of different nodes to discover abnormal connections and activities. The specific steps for node analysis are: through the collected node-related information in the network, including IP address, MAC address, node type, etc., analyze ARP request and response packets, determine the relationship between each node, and clarify the requester and responder; observe the behavior pattern of the node in the network, including the frequency of the node sending and receiving ARP data packets, the size of the data packet, etc., to understand the activity of the node; by comparing the normal behavior pattern of the node with the current behavior pattern, detect possible abnormal nodes, such as nodes that frequently send ARP request or response packets; for the detected abnormal nodes, log the relevant data information and time node information to check whether the data comes from unknown assets.

[0058] Specifically, the checksum calculation of the IPv4 (Internet Protocol version 4) datagram is mainly performed on the IPv4 header, and the specific calculation steps are: setting the checksum field of the IPv4 datagram header to 0, so as to reduce the influence of these two bytes in the calculation process; dividing the rest of the IPv4 header except the checksum field into multiple fragments of 16 bits each; adding each 16-bit fragment in binary form, and adding the carry bit to the low bit; if the result of adding two 16-bit numbers exceeds the maximum value that can be represented by 16 bits, it is necessary to add the carry bit to the low bit of the next 16-bit fragment; bitwise inverting the obtained sum (i.e., performing a logical NOT operation on each bit), and the result obtained is the final checksum value.

[0059] The situation awareness module is used to classify data through a classification algorithm and output classification results; identify network traffic patterns and user behavior patterns related to unknown assets based on an association analysis algorithm, and output identification results; predict the business and type of unknown assets through a preset asset prediction model input by the collected data, and output unknown asset prediction results; conduct a comprehensive analysis based on the captured data, classification results, identification results and prediction results, and output a situation awareness assessment report.

[0060] In one embodiment, the situation awareness module includes a data classification unit, a correlation analysis unit, an asset prediction model and a situation assessment unit; the data classification unit, the correlation analysis unit and the asset prediction model are connected in sequence, and the situation assessment unit is respectively connected to the data classification unit, the correlation analysis unit and the asset prediction model.

[0061] In one embodiment, the data classification unit is used to classify and label the features of the data of the unknown asset and output the classification results; the data classification unit adopts a decision tree classification algorithm, and the specific operation process includes the following steps:

[0062] Receive data from the real-time monitoring module and the ARP protocol analysis module, and format and standardize the data based on the received data;

[0063] Select key features as decision nodes through the received known asset data, and gradually split the data into different branches according to the relevant feature values ​​to form a tree structure;

[0064] The collected data of unknown assets is input into the trained decision tree model, and the classification judgment of the unknown assets is realized through the output leaf node position.

[0065] Specifically, the data classification module uses a decision tree algorithm to classify and label the relevant data of known assets and unknown assets. The workflow for classification and labeling based on data features (such as IP address, port number, traffic pattern, etc.) is as follows:

[0066] Collect and process various network traffic, logs, device status, ARP protocol and other characteristic data through the real-time monitoring module and ARP protocol analysis module to ensure the real-time and accuracy of the data for analysis and prediction.

[0067] Put all the training data of known assets after preprocessing at the root node, select the best feature of the current known asset as the root node of the decision tree, and divide the data set into several subsets according to the feature value of the root node. Each subset corresponds to a child node. This process is recursive, that is, each child node will continue to select an optimal feature for division until the stopping condition is met.

[0068] When the data of a subset can be basically correctly classified or there are no suitable features to continue dividing, a leaf node is generated and the subset is divided into the corresponding leaf node. The leaf node represents the known asset classification labels under different switches in the final plant station.

[0069] The pre-pruning method is used to delete unnecessary nodes or subtrees and prune the generated decision tree to make it more concise and have strong generalization ability.

[0070] The constructed decision tree is evaluated using the test data set. Common evaluation indicators include accuracy, recall, F1 value, etc. The cross-validation method is used to accurately evaluate the performance of the decision tree.

[0071] The trained decision tree model is deployed in the situation awareness module, the corresponding data of the collected unknown assets is input into the decision tree model, and the classification judgment of the unknown assets is realized through the output leaf node position.

[0072] In one embodiment, the association analysis unit is used to perform association analysis in combination with known asset information to find the association relationship of the unknown assets related thereto, and output the corresponding identification result. The operation process of the association analysis unit includes the following steps:

[0073] Using known asset information as prior knowledge, generate association rules between known assets;

[0074] Receive unknown asset data from the real-time monitoring module and convert the data into a transaction data set, setting a reasonable minimum support threshold and minimum confidence level based on the size and relevant characteristics of the data;

[0075] Finding frequent item sets through the minimum support threshold and generating association rules according to the found frequent item sets;

[0076] Set anomaly thresholds to determine abnormal patterns of association rules.

[0077] In this embodiment, the association analysis unit uses the PCY algorithm to analyze the association relationship between data;

[0078] The PCY algorithm is a probability counting algorithm, which is mainly used to mine frequent item sets and association rules. The PCY algorithm was proposed to improve the Apriori algorithm. The algorithm reduces the number of times the Apriori algorithm frequently scans the database, while reducing the memory occupied by the candidate item set. The PCY algorithm uses the concept of "partial candidates" to reduce the number of candidate item sets, thereby improving efficiency. The Apriori algorithm is the first association rule mining algorithm and the most classic algorithm. It uses an iterative method of layer-by-layer search to find the relationship between item sets in the database to form rules. The process consists of connection (matrix-like operation) and pruning (removing unnecessary intermediate results).

[0079] The information data format of known assets is converted into a transaction data set. The PCY algorithm is used to analyze the association between the data, and frequent item sets and association rules are discovered. The network traffic pattern and user behavior pattern related to unknown assets can be identified, and potential anomalies can be discovered. The specific implementation process is as follows:

[0080] According to the size and related characteristics of the data set processed by the real-time monitoring module, a reasonable minimum support threshold is set, which is used to filter frequent item sets, and a minimum confidence threshold is set to evaluate the strength of association rules. According to the size and characteristics of the data set, appropriate hash function Bloom filter parameters are selected. The hash function is used to map data to the Bloom filter, and the Bloom filter is used to efficiently check whether an element exists in a set.

[0081] Specifically, the selection and setting of hash functions include: having good hashability to ensure that asset information data is evenly distributed in the hash space; since hash collisions are inevitable, a low collision rate is required, and multiple hash functions can be used; having a fast calculation speed to support real-time association relationship analysis; having determinism, that is, the same input always produces the same output;

[0082] Bloom filter parameters include: bit array size, number of hash functions, and expected false alarm rate. Among them, the bit array size directly affects the false alarm rate and space efficiency of the Bloom filter. A larger bit array can reduce the false alarm rate, but it will increase space consumption. The bit array size is estimated based on the number of known asset information data, the update rate, and the required false alarm rate. The number of hash functions and the bit array size jointly determine the performance of the Bloom filter. More hash functions can reduce the false alarm rate, but it will also increase the computational complexity and the possibility of hash conflicts. It is necessary to strike a balance between the false alarm rate and performance based on actual conditions. The expected false alarm rate is an important indicator of the Bloom filter, which indicates the probability of judging a non-existent element as existing. It is set according to the accuracy requirements of the power monitoring system for the association analysis. A lower false alarm rate requires a larger bit array and more hash functions, but it will improve the accuracy of the analysis.

[0083] Use the PCY algorithm to scan the integrated transaction data set, count the number of occurrences of each item, and build a candidate item set starting from a single item according to the support threshold. The initial candidate item set only contains a single item, and then gradually build a candidate item set containing more items.

[0084] Use a hash function to map each candidate item set to a Bloom filter, count the number of times the corresponding candidate item set appears in the transaction data set, and filter out frequent item sets that meet the conditions according to the support threshold. If the support of a candidate item set is greater than or equal to the support threshold, it is considered a frequent item set.

[0085] Analyze the relationship between the frequent itemsets found, find out the item set pairs with strong correlation by calculating the confidence, lift and other indicators between the item sets, and generate association rules through the item set pairs with strong correlation. The rules describe the association relationship between the item sets and reveal the potential connection between unknown assets and known assets.

[0086] Identify association rules related to unknown assets: Filter out the rules related to unknown assets from all generated association rules by checking whether the items of the rules contain identifiers or keywords related to unknown assets, and further analyze the filtered association rules to find patterns or behaviors related to unknown assets. These patterns may reveal how the unknown assets are used, how they are accessed, or how they are associated with other assets.

[0087] If the pattern of an association rule is significantly different from the normal pattern of a known asset, it may represent an anomaly. This may be because the behavior of the unknown asset is inconsistent with expectations, or because the way it is associated with other assets does not conform to the known pattern. Set the corresponding anomaly threshold according to actual needs. When the confidence or other measurement indicators of an association rule exceed this threshold, it is determined that the pattern of the association rule represents an anomaly.

[0088] Result reporting and subsequent processing: The detected association rules and anomalies related to unknown assets are presented in the form of reports, and the report results are output to the situation assessment unit.

[0089] The asset prediction model is used to analyze the data packets and log data collected by the real-time monitoring module and extract feature information to predict the type of unknown assets and their business, and output the unknown asset prediction results; the asset prediction model adopts a neural network model.

[0090] In this embodiment, the asset prediction model adopts a CNN-LSTM (Convolutional Long Short-Term Memory Neural Network) neural network model.

[0091] Specifically, the asset prediction model building and training process includes the following steps:

[0092] The data preprocessed by the real-time monitoring module is converted into a format suitable for processing by the neural network model. For the time-related feature data of real-time monitoring, the data needs to be reorganized to reflect the temporal sequence. The convolution layer is used to extract the spatial features of the input data. A 3x3 convolution kernel and 32 filters are used as the starting point. A pooling layer is added after the convolution layer, and the maximum pooling is used to reduce the size of the feature map and retain important features.

[0093] The feature map processed by the CNN layer (convolutional layer) is converted into a one-dimensional vector, and the Flatten layer is used to achieve the conversion. The one-dimensional feature vector is used as the input of the LSTM layer, and the corresponding number of LSTM layer units is set according to the data scale to capture the long-term dependencies in the sequence data and learn the temporal patterns in the data. The LSTM layer is a special neural network layer and is part of the Long Short-Term Memory network.

[0094] Multiple fully connected layers are added after the LSTM layer to output the model's prediction results for unknown asset types and businesses, using softmax as the activation function.

[0095] The Softmax activation function is a function used to convert the output of a neural network into a probability distribution. It is often used in the last layer of a multi-classification problem. Its main function is to map a real vector into an interval, and the sum of all elements is 1, which allows it to be interpreted as a probability distribution.

[0096] The data set consisting of the number, size, time-related features, data packet information flags, network traffic patterns, access frequency, system resource usage, and corresponding ARP protocol of known assets is divided into training set, test set, and validation set according to the ratio of 6:2:2. The model is trained using the training set, and the loss function uses the cross entropy loss function. The optimizer selects the Adam optimizer, monitors the performance on the validation set, and adjusts the hyperparameters to optimize the model. The Adam optimizer is used to adjust the parameters of the model to minimize the loss function.

[0097] In one embodiment, the situation assessment unit is used to integrate the classification results, identification results and prediction results output by the data classification unit, the association analysis unit and the asset prediction model to form a comprehensive data set, and analyze the classification, type, association relationship with known assets and corresponding business of unknown assets based on the comprehensive perception and understanding of the situation by the comprehensive data set.

[0098] Specifically, the classified data output by the data classification unit, the associated relationship data mined by the associated analysis unit, and the evaluation data given by the asset prediction model are collected and organized. For the classified data, the specific category to which the data point belongs is clarified, and the asset information is classified; the data of the associated analysis unit can reveal the relationship between different assets, such as which assets have dependencies; the evaluation data of the asset prediction model can provide prediction information about unknown assets, such as possible classifications and types. The data is integrated in a unified format and structure to form a comprehensive data set.

[0099] Based on the comprehensive data set, we can conduct a comprehensive situational awareness and understanding, and determine the classification of unknown assets by analyzing the classification data; we can use the association data to determine the association between unknown assets and known assets, and the association between known assets and unknown assets includes: collaborative work, upstream and downstream relationships, etc. We can infer the type of unknown assets and the business system they belong to by combining the output of the asset prediction model, and clarify the specific characteristics; we can infer the business corresponding to the unknown assets based on the above analysis and prediction results, including which business processes the unknown assets play a role in or which business links they are associated with.

[0100] In one embodiment, the unknown asset positioning module is used to locate unknown assets by building an updated network topology map, combining the device layout on the network topology map with the location information of known assets, and analyzing the area where the location assets are located.

[0101] Specifically, the specific working process of the unknown asset positioning module is as follows: combining the results of situational awareness, sending a specific ARP request message to the target device and receiving its response message, analyzing the response message through the ARP protocol module to obtain detailed information of the target device, querying the ARP table on the switch according to the MAC of the unknown asset, and obtaining the mapping relationship between the MAC address and the switch port through the ARP table on the switch; building a network topology map through the collected device information, listing all devices of known assets in the network, and recording the detailed information of the devices and the connection relationship between the devices. The topology shows all monitoring objects in the power monitoring system, including servers, switches, routers, vertical encryption devices, horizontal isolation devices, hardware firewall devices, etc. The detailed information of the equipment includes: equipment type, location, and related operation information. The acquired information of unknown assets is added to the network topology map through the real-time monitoring module, ARP protocol analysis module, and situational awareness module.

[0102] Specific icons are used to identify network access devices, and grayscale processing is used for network exit device icons. The network topology map enables layer-by-layer drilling of multi-layer topologies, and supports zooming in, zooming out, and panning.

[0103] By finding the connection lines and nodes related to the unknown assets on the network topology map, gradually follow the connection lines to find other devices connected to the unknown assets; combining the device layout on the network topology map with the location information of known assets, analyze the possible areas where the location assets may be located.

[0104] The analysis results are cross-validated, and the possible location of unknown assets obtained by analysis is compared with real-time monitoring data to check whether there is corresponding traffic activity or abnormality at the location; the MAC address and other information obtained by ARP protocol analysis are matched with the devices on the network topology map to verify consistency; the evaluation results output by the situation awareness module are analyzed, and the evaluation results are mutually verified with the speculation based on the network topology map to eliminate unreasonable speculation; and the corresponding unknown asset information is corrected and updated to the network topology map. The unknown asset information includes: IP address, MAC address, device type, location, etc. Based on the above, the positioning results are corrected and optimized to improve accuracy.

[0105] The alarm and response module is used to generate alarm response suggestions based on the unknown asset location information and related historical business analyzed by the module.

[0106] In one embodiment, the alarm and response module is used to visualize the relevant information of unknown assets based on the constructed network topology diagram, set the relevant alarm level according to the relevant historical data of the system, and track and evaluate the effect of alarm processing to adjust and optimize the strategy.

[0107] Specifically, the alarm and response module displays the analysis results, prediction results, response and disposal information to the user in the form of graphics, tables, etc., and displays the network topology of the power monitoring system monitored by the platform at this level in real time, as well as the corresponding equipment online rate, number of online devices, total number of devices, and corresponding alarm conditions. The alarm conditions are divided into two states: unprocessed and processed, and are counted in three alarm levels: emergency, important, and general. New alarms are prompted by color marking, sound reminders, voice broadcast of alarm content, and SMS push.

[0108] Set relevant alarm levels based on historical events, including emergency alarms, major alarms, and general alarms. Set alarm levels based on relevant historical data and the severity of events caused by unknown assets that are not handled in a timely manner, such as setting cross-region interconnection of devices as emergency alarms, setting host device risk illegal device access as major alarms, and setting abnormal user operations discovered by host devices as general alarms.

[0109] Set up the corresponding security knowledge base according to the alarm solution of the historical event, and provide the corresponding alarm solution suggestion when the alarm occurs. Track and evaluate the effect of alarm processing, eliminate the real-time alarm information after the alarm is resolved, and update the corresponding solution of the corresponding event to the security knowledge base by selecting or manually entering the solution to improve the accuracy of the generated alarm solution suggestion.

[0110] After the operation and maintenance personnel check the unknown assets entering the site based on the relevant information, they update the specific location of the unknown assets and the actual business operations performed to the network topology map, and update the specific information of the equipment to the database to improve the application index of the plant situation awareness asset registration rate. The specific information of the equipment includes: operating status information, configuration information, traffic transmission information, and corresponding business system information.

[0111] The above implementation modes are merely descriptions of the preferred implementation modes of the present application, and are not intended to limit the scope of the present application. Without departing from the design spirit of the present application, various modifications and improvements made to the technical solutions of the present application by ordinary engineering and technical personnel in the field shall fall within the protection scope determined by the claims of the present application.

Claims

1. An unknown asset positioning system based on situational awareness, characterized in that: It includes a real-time monitoring module, an ARP protocol analysis module, a situation awareness module, an unknown asset positioning module, and an alarm and response module. The real-time monitoring module is connected to the ARP protocol analysis module and the situation awareness module respectively, the unknown asset positioning module is connected to the ARP protocol module, the situation awareness module, and the alarm and response module respectively, and the situation awareness module is connected to the ARP protocol module and the alarm response module respectively; The real-time monitoring module is used to capture ARP messages and related communication data in the network environment in real time and collect and integrate system-related historical data; The ARP protocol analysis module is used to parse the captured ARP packets, extract key information, find abnormal connections through node analysis based on ARP protocol data, and output ARP protocol analysis results for subsequent processing and verification; The situation awareness module is used to classify the data through a classification algorithm and output the classification result; Identify network traffic patterns and user behavior patterns related to unknown assets based on association analysis algorithms, and output identification results; The collected data is input into the preset asset prediction model to predict the business and type of unknown assets, and the prediction results of unknown assets are output; Combine the captured data, classification results, recognition results and prediction results for comprehensive analysis and output a situation awareness assessment report; The unknown asset positioning module is used to build a network topology map in combination with the ARP protocol analysis results and the situation awareness assessment report, analyze and locate unknown assets in the network topology map, output the positioning information of the corresponding unknown assets and visualize the network topology map; The alarm and response module is used to generate alarm response suggestions by combining the situation awareness assessment report information, the unknown asset location information and the relevant historical system data; The unknown assets include: unregistered devices, external mobile devices, network nodes of unknown origin, temporarily built systems or devices, forgotten old devices and illegally accessed devices.

2. The unknown asset positioning system based on situational awareness according to claim 1, characterized in that: The real-time monitoring module includes a data acquisition unit and a data protection unit; The data acquisition unit is used to receive data packets in real time through a network interface and collect and integrate system-related historical data, and extract key indicators or attributes of unknown asset behaviors and characteristics from the data packets; The data protection unit is used to protect the transmission of data packets through an encryption algorithm.

3. The unknown asset positioning system based on situational awareness according to claim 1, characterized in that: The ARP protocol analysis module is used to verify the captured ARP message content, determine abnormal connections between nodes based on node analysis, and check the source of data.

4. The unknown asset positioning system based on situational awareness according to claim 1, characterized in that: The situation awareness module includes a data classification unit, a correlation analysis unit, an asset prediction model and a situation assessment unit; the data classification unit, the correlation analysis unit and the asset prediction model are connected in sequence, and the situation assessment unit is connected to the data classification unit, the correlation analysis unit and the asset prediction model respectively.

5. The unknown asset positioning system based on situational awareness according to claim 4, characterized in that: The data classification unit is used to classify and label the features of the data of the unknown asset and output the classification results; the data classification unit adopts a decision tree classification algorithm, and the specific operation process includes the following steps: Receive data from the real-time monitoring module and the ARP protocol analysis module, and format and standardize the data based on the received data; Select key features as decision nodes through the received known asset data, and gradually split the data into different branches according to the relevant feature values ​​to form a tree structure; The collected data of unknown assets is input into the trained decision tree model, and the classification judgment of the unknown assets is realized through the output leaf node position.

6. The unknown asset positioning system based on situational awareness according to claim 4, characterized in that: The association analysis unit is used to perform association analysis in combination with known asset information to identify the association relationship of unknown assets related thereto, and output corresponding identification results. The operation process of the association analysis unit includes the following steps: Using known asset information as prior knowledge, generate association rules between known assets; Receive unknown asset data from the real-time monitoring module and convert the data into a transaction data set, setting a reasonable minimum support threshold and minimum confidence level based on the size and relevant characteristics of the data; Finding frequent item sets by using the minimum support threshold and generating association rules according to the found frequent item sets; Set anomaly thresholds to determine abnormal patterns of association rules.

7. The unknown asset positioning system based on situational awareness according to claim 4, characterized in that: The asset prediction model is used to analyze the data packets and log data collected by the real-time monitoring module and extract feature information to predict the type of unknown assets and their business, and output the unknown asset prediction results; the asset prediction model adopts a neural network model.

8. The unknown asset positioning system based on situational awareness according to claim 4, characterized in that: The situation assessment unit is used to integrate the classification results, identification results and prediction results output by the data classification unit, the association analysis unit and the asset prediction model to form a comprehensive data set, and based on the comprehensive perception and understanding of the situation by the comprehensive data set, analyze the classification, type, association relationship with known assets and corresponding business of unknown assets.

9. The unknown asset positioning system based on situational awareness according to claim 1, characterized in that: The unknown asset positioning module is used to locate unknown assets by building an updated network topology map, combining the device layout on the network topology map with the location information of known assets, and analyzing the area where the location assets are located.

10. The unknown asset positioning system based on situational awareness according to claim 1, characterized in that: The alarm and response module is used to visualize the relevant information of unknown assets according to the constructed network topology diagram, set the relevant alarm level according to the relevant historical data of the system, and track and evaluate the effect of alarm processing to adjust and optimize the strategy.

Citation Information

Patent Citations

  • Numerical control system security situation awareness and analysis system, method, equipment and terminal

    CN115996146A

  • Topology discovery and network asset association method and device for network operation and maintenance system

    CN117579492A

  • Network asset characterization, classification, grouping and control

    US20190089741A1

Cited By

  • Production line-oriented protocol analysis method and intelligent acquisition device

    CN119520652A