Network security protection operation and maintenance management system and method
Through AI algorithms, the access traffic of power applications is time-series segmented and implicit correlation encoding. Using Hamming distance comparison, the problem of difficult to understand complex network traffic patterns in the existing technology is solved, and adaptability to dynamic network environments is achieved. and automatic abnormal traffic judgment.
Patent Information
- Application Number
- CN202411570630.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-06
- Publication Date
- 2024-12-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing power operation and maintenance network security monitoring and early warning methods are difficult to understand complex network traffic patterns, lack adaptability to dynamic and complex modes, and rely on manual setting of fixed traffic thresholds, making it difficult to cope with special needs in different situations.
By obtaining the time series of access traffic values of the power application, using AI-based data processing and analysis algorithms for timing segmentation and implicit association encoding, using the Hamming distance between the features of the first time scale and the second time scale Compared with the preset threshold, the abnormal traffic is automatically judged.
Deeply understand the inherent patterns and correlations of traffic data, reduce dependence on manually setting fixed traffic thresholds, and the system can dynamically adapt to changing network environments and continuously optimize and adjust.
Smart Images

Figure CN119094246A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of network security, and more specifically, to a network security protection operation and maintenance management system and method. Background Art
[0002] Although the existing operation and maintenance process management scheme has been fine-tuned, it has not fully reflected the operation mode and information operation and maintenance service system of the main business of the power grid. There are problems such as database customization, data interaction difficulties, lack of multi-dimensional information support for IT assets, and operation and maintenance cost collection. In response to the above technical problems, China's public patent CN112766517A proposes a monitoring and early warning method for power operation and maintenance network security, which collects operation and maintenance data by establishing a basic database, collects security information in real time, and cleans and processes these data, and ensures the network security of the power system through identity authentication, authority management, network access control, and traffic monitoring and alarm for illegal requests.
[0003] However, in the above-mentioned monitoring and early warning method for power operation and maintenance network security, the network traffic usage within a preset period is combined with the acquired network traffic usage and the pre-set traffic threshold to obtain abnormal traffic behavior. This method may be difficult to understand and consider the complex correlation patterns in the traffic, and lacks adaptability to dynamic and complex network traffic patterns or special needs in different situations.
[0004] Therefore, an optimized network security protection operation and maintenance management solution is desired. Summary of the invention
[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a network security protection operation and maintenance management system and method, which obtains the time series of access flow values of power applications within a predetermined time period, and uses AI-based data processing and analysis algorithms to perform time series segmentation and implicit association encoding on the time series of the access flow values, thereby automatically judging whether there is abnormal traffic based on the comparison between the Hamming distance between the access flow time series characteristics obtained by reasoning at the first time scale and the access flow time series characteristics at the second scale and the preset threshold. In this way, the inherent patterns and correlations of traffic data can be more deeply understood, while the dependence on manually set fixed traffic thresholds can be reduced. The system can continuously optimize and adjust according to new data to dynamically adapt to the changing network environment.
[0006] According to one aspect of the present application, a network security protection operation and maintenance management system is provided, which includes: An access flow value time series acquisition module is used to obtain the time series of access flow values of the power application program within a predetermined time period; An access traffic value time series arrangement and segmentation module, configured to arrange the time series of the access traffic values according to the time dimension to obtain an access traffic time series input vector, and then segment the access traffic time series input vector to obtain a first access traffic local time series input vector and a second access traffic local time series input vector; A first access traffic local time sequence association encoding module, configured to input the first access traffic local time sequence input vector into a sequence encoder to obtain a sequence of first access traffic local time sequence implicit association feature vectors; A first traffic local time series aggregation module, configured to input the sequence of the first access traffic local time series implicitly associated feature vectors into a node feature propagation network based on a node energy decay mechanism to obtain a first access traffic local time series feature propagation aggregation representation vector; An access traffic timing reasoning module, used for inputting the first access traffic local timing feature propagation aggregation representation vector into a decoder-based access traffic simulation reasoner to obtain a decoded access traffic timing input vector; An access traffic difference calculation module, used for calculating the Hamming distance between the decoded access traffic timing input vector and the second access traffic local timing input vector; The abnormal traffic determination module is used to determine whether abnormal traffic exists based on the comparison between the Hamming distance and a preset threshold.
[0007] In the above-mentioned network security protection operation and maintenance management system, the access traffic value timing arrangement and segmentation module includes: an access traffic value data structuring unit, used to arrange the time series of the access traffic value according to the time dimension to obtain the access traffic timing input vector; an access traffic timing input vector segmentation unit, used to segment the access traffic timing input vector to obtain the first access traffic local timing input vector and the second access traffic local timing input vector, wherein the length of the first access traffic local timing input vector is greater than twice that of the second access traffic local timing input vector.
[0008] In the above-mentioned network security protection operation and maintenance management system, the first access traffic local timing association encoding module is used to: input the first access traffic local timing input vector into a sequence encoder based on the 1D-CNN model to obtain a sequence of the first access traffic local timing implicit association feature vectors.
[0009] In the above-mentioned network security protection operation and maintenance management system, the first traffic local timing aggregation module includes: a first access traffic local timing feature energy statistical paradigm value calculation unit, which is used to calculate the energy statistical paradigm value of each first access traffic local timing implicitly associated feature vector based on the maximum value, average value and variance of each first access traffic local timing implicitly associated feature vector in the sequence of the first access traffic local timing implicitly associated feature vector to obtain a sequence of first access traffic local timing implicitly associated energy statistical paradigm values, wherein the energy statistical paradigm value corresponding to the current first access traffic local timing implicitly associated feature vector in the sequence of the first access traffic local timing implicitly associated energy statistical paradigm values is used as The current node energy statistical paradigm value and other energy statistical paradigm values are used as historical node energy statistical paradigm values to obtain a sequence of the current first access traffic local time series implicitly associated energy statistical paradigm value and the historical first access traffic local time series implicitly associated energy statistical paradigm value; a first access traffic local time series node propagation space span value calculation unit, used to count the node propagation space span values between each other first access traffic local time series implicitly associated feature vector in the sequence of the first access traffic local time series implicitly associated feature vector and the current first access traffic local time series implicitly associated feature vector to obtain a sequence of the first access traffic local time series implicitly associated propagation space span values; the first access traffic local time series energy propagation attenuation coefficient A value calculation unit is used to determine the energy propagation attenuation coefficient values of other first access traffic local time series implicitly associated feature vectors in the sequence of the first access traffic local time series implicitly associated feature vectors based on the sequence of the first access traffic local time series implicitly associated propagation space span values and the sequence of the historical first access traffic local time series implicitly associated energy statistical paradigm values to obtain a sequence of first access traffic local time series implicitly associated energy propagation attenuation coefficient values, wherein the first access traffic local time series implicitly associated energy propagation attenuation coefficient value is inversely correlated with the historical first access traffic local time series implicitly associated energy statistical paradigm value; a first access traffic local time series feature weighting unit is used to weight the first access traffic local time series implicitly associated energy propagation attenuation coefficient value based on the first access traffic local time series implicitly associated propagation space span values and the historical first access traffic local time series implicitly associated energy statistical paradigm values. A sequence of partial timing implicit correlation energy propagation attenuation coefficient values is used as a weight sequence, and the weighted sum of all other first access traffic local timing implicit correlation feature vectors in the sequence of the first access traffic local timing implicit correlation feature vector is calculated to obtain a historical first access traffic local timing implicit correlation energy attenuation timing aggregate feature vector; a first access traffic local timing feature aggregation unit is used to fuse the current first access traffic local timing implicit correlation energy statistical paradigm value to calculate the weighted sum of the historical first access traffic local timing implicit correlation energy attenuation timing aggregate feature vector and the current first access traffic local timing implicit correlation feature vector to obtain the first access traffic local timing feature propagation aggregate representation vector.
[0010] In the above-mentioned network security protection operation and maintenance management system, the first access traffic local time series feature energy statistical paradigm value calculation unit is used to: extract the maximum value of each position in the first access traffic local time series implicit correlation feature vector to obtain the maximum value of the first access traffic local time series implicit correlation feature; respectively calculate the average value and variance of the first access traffic local time series implicit correlation feature vector to obtain the average value of the first access traffic local time series implicit correlation feature and the variance of the first access traffic local time series implicit correlation feature; multiply the value obtained by adding the variance of the first access traffic local time series implicit correlation feature and the preset hyperparameter by four to obtain the first energy statistical paradigm value of the first access traffic local time series implicit correlation feature; calculate the first access traffic local time series implicit correlation feature; The square of the difference between the maximum value of the association feature and the average value of the first access traffic local timing implicit association feature is calculated to obtain the difference value of the first access traffic local timing implicit association feature; the variance of the first access traffic local timing implicit association feature obtained by multiplying the variance of the first access traffic local timing implicit association feature by two is added to the value obtained by multiplying the preset hyperparameter by two and the difference value of the first access traffic local timing implicit association feature to obtain the second energy statistical paradigm value of the first access traffic local timing implicit association feature; the first energy statistical paradigm value of the first access traffic local timing implicit association feature is divided by the second energy statistical paradigm value of the first access traffic local timing implicit association feature to obtain the first access traffic local timing implicit association energy statistical paradigm value.
[0011] In the above-mentioned network security protection operation and maintenance management system, the first access traffic local time series energy propagation attenuation coefficient value calculation unit is used to: multiply each first access traffic local time series implicitly associated propagation space span value in the sequence of the first access traffic local time series implicitly associated propagation space span value by a first weight hyperparameter to obtain a sequence of first weighted propagation space span values; use each first access traffic local time series implicitly associated propagation space span value in the sequence of the first access traffic local time series implicitly associated propagation space span value as an exponent of a natural constant to calculate an exponential function value with the natural constant as the base according to the position to obtain a sequence of first access traffic local time series implicitly associated propagation class support space span values; use the first access traffic local time series implicitly associated propagation space span value as an exponent of a natural constant Each first access traffic local timing implicitly associated propagation class supported space span value in the sequence of associated propagation class supported space span values is multiplied by the second weight hyperparameter to obtain a sequence of second weighted propagation space span values; the positional sum of the sequence of the first weighted propagation space span values and the sequence of the second weighted propagation space span values is calculated to obtain a sequence of weighted sum propagation space span values; the sequence of the historical first access traffic local timing implicitly associated energy statistical paradigm values and the sequence of the weighted sum propagation space span values are divided by each group of corresponding historical first access traffic local timing implicitly associated energy statistical paradigm values and the weighted sum propagation space span values to obtain a sequence of the first access traffic local timing implicitly associated energy propagation attenuation coefficient values.
[0012] According to another aspect of the present application, a network security protection operation and maintenance management method is provided, which includes: Obtaining a time series of access flow values of a power application within a predetermined time period; After arranging the time series of the access traffic values according to the time dimension to obtain an access traffic time series input vector, the access traffic time series input vector is segmented to obtain a first access traffic local time series input vector and a second access traffic local time series input vector; Inputting the first access traffic local time sequence input vector into a sequence encoder to obtain a sequence of first access traffic local time sequence implicitly associated feature vectors; Inputting the sequence of the first access traffic local time series implicitly associated feature vectors into a node feature propagation network based on a node energy decay mechanism to obtain a first access traffic local time series feature propagation aggregation representation vector; Input the first access traffic local timing feature propagation aggregation representation vector into a decoder-based access traffic simulation reasoner to obtain a decoded access traffic timing input vector; Calculating a Hamming distance between the decoded access traffic timing input vector and the second access traffic local timing input vector; Based on the comparison between the Hamming distance and a preset threshold, it is determined whether abnormal traffic exists.
[0013] In the above-mentioned network security protection operation and maintenance management method, after arranging the time series of the access traffic values according to the time dimension to obtain the access traffic timing input vector, the access traffic timing input vector is divided to obtain the first access traffic local timing input vector and the second access traffic local timing input vector, including: arranging the time series of the access traffic values according to the time dimension to obtain the access traffic timing input vector; dividing the access traffic timing input vector to obtain the first access traffic local timing input vector and the second access traffic local timing input vector, wherein the length of the first access traffic local timing input vector is greater than 2 times that of the second access traffic local timing input vector.
[0014] Compared with the prior art, the network security protection operation and maintenance management system and method provided by the present application obtains the time series of the access flow value of the power application within a predetermined time period, and uses an AI-based data processing and analysis algorithm to perform time series segmentation and implicit association coding on the time series of the access flow value, thereby automatically judging whether there is abnormal flow based on the comparison between the Hamming distance between the access flow time series characteristics obtained by reasoning at the first time scale and the access flow time series characteristics at the second scale and the preset threshold. In this way, the inherent pattern and correlation of the flow data can be more deeply understood, and at the same time, the dependence on manually set fixed flow thresholds can be reduced. The system can continuously optimize and adjust according to new data to dynamically adapt to the changing network environment. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] Figure 1 This is a system block diagram of a network security protection operation and maintenance management system according to an embodiment of the present application.
[0016] Figure 2 Schematic diagram of the architecture of a network security protection operation and maintenance management system according to an embodiment of the present application.
[0017] Figure 3 This is a block diagram of the first traffic local timing aggregation module in the network security protection operation and maintenance management system according to an embodiment of the present application.
[0018] Figure 4 Flow chart of the network security protection operation and maintenance management method according to an embodiment of the present application. DETAILED DESCRIPTION
[0019] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.
[0020] A Chinese patent CN112766517A proposes a new power operation and maintenance network security monitoring and early warning method, which collects operation and maintenance data by building a basic database and obtains security information in real time. Through data cleaning and processing, combined with identity authentication, authority management, network access control, and monitoring and alarming of abnormal traffic, the network security of the power system is ensured.
[0021] However, the monitoring and early warning method described in the patent has limitations in some aspects. It mainly relies on monitoring network traffic within a preset time period and comparing it with a preset traffic threshold to identify abnormal traffic behavior. This method may have difficulty in fully understanding and capturing complex correlation patterns in traffic, and may also lack the ability to adapt to dynamically changing and complex network traffic patterns, as well as respond to special needs in different situations.
[0022] Therefore, in response to the above technical problems, the technical concept of this application is to obtain the time series of the access flow value of the power application within a predetermined time period, and use AI-based data processing and analysis algorithms to perform time series segmentation and implicit association coding on the time series of the access flow value, so as to automatically determine whether there is abnormal traffic based on the comparison between the Hamming distance between the access flow time series characteristics obtained by reasoning at the first time scale and the access flow time series characteristics at the second scale and the preset threshold. In this way, the inherent pattern and correlation of the traffic data can be more deeply understood, and the dependence on manually set fixed traffic thresholds can be reduced. The system can continuously optimize and adjust according to new data to dynamically adapt to the changing network environment.
[0023] Figure 1 This is a system block diagram of a network security protection operation and maintenance management system according to an embodiment of the present application. Figure 2 FIG. 1 is a schematic diagram of the architecture of a network security protection operation and maintenance management system according to an embodiment of the present application. Figure 1 and Figure 2As shown, in the network security protection operation and maintenance management system 100, it includes: an access flow value timing collection module 110, which is used to obtain the time series of the access flow value of the power application program within a predetermined time period; an access flow value timing arrangement and segmentation module 120, which is used to arrange the time series of the access flow value according to the time dimension to obtain the access flow timing input vector, and then segment the access flow timing input vector to obtain a first access flow local timing input vector and a second access flow local timing input vector; a first access flow local timing association encoding module 130, which is used to input the first access flow local timing input vector into a sequence encoder to obtain a sequence of the first access flow local timing implicit association feature vector; the first flow local timing input vector A sequence aggregation module 140 is used to input the sequence of the first access traffic local timing implicitly associated feature vectors into a node feature propagation network based on a node energy decay mechanism to obtain a first access traffic local timing feature propagation aggregation representation vector; an access traffic timing reasoning module 150 is used to input the first access traffic local timing feature propagation aggregation representation vector into a decoder-based access traffic simulation reasoner to obtain a decoded access traffic timing input vector; an access traffic difference calculation module 160 is used to calculate the Hamming distance between the decoded access traffic timing input vector and the second access traffic local timing input vector; an abnormal traffic determination module 170 is used to determine whether there is abnormal traffic based on a comparison between the Hamming distance and a preset threshold.
[0024] In an embodiment of the present application, the access traffic value time series acquisition module 110 is used to obtain the time series of the access traffic value of the power application within a predetermined time period. It should be understood that the time series data of the access traffic value can help monitor the network activity pattern of the power application, including the traffic changes during peak and trough periods, and then by modeling and analyzing the access traffic time series data, the normal access traffic pattern can be identified, thereby helping to detect abnormal traffic that does not match it. In particular, in a specific embodiment of the present application, the access traffic value time series data of the power application within a predetermined time period can be obtained from the database of the power application.
[0025] In an embodiment of the present application, the access traffic value timing arrangement and segmentation module 120 is used to arrange the time series of the access traffic value according to the time dimension to obtain the access traffic timing input vector, and then segment the access traffic timing input vector to obtain the first access traffic local timing input vector and the second access traffic local timing input vector. Specifically, the access traffic value timing arrangement and segmentation module includes: an access traffic value data structuring unit, which is used to arrange the time series of the access traffic value according to the time dimension to obtain the access traffic timing input vector; an access traffic timing input vector segmentation unit, which is used to segment the access traffic timing input vector to obtain the first access traffic local timing input vector and the second access traffic local timing input vector, wherein the length of the first access traffic local timing input vector is greater than 2 times that of the second access traffic local timing input vector. In particular, in a specific embodiment of the present application, the length of the first access traffic local timing input vector divided by the length of the second access traffic local timing input vector is greater than or equal to 4.
[0026] It should be understood that the time series data of the access traffic value has time series characteristic information on the time scale. Therefore, in order to better capture the fluctuation and change trend of the access traffic in the time dimension, in the technical solution of the present application, the time series of the access traffic value is arranged according to the time dimension to obtain the access traffic time series input vector. Then, considering that different time scales in the access traffic time series input vector have different access traffic time series characteristic information, such as local fluctuations of traffic in the short term and change trends and periodic patterns in the long term. Therefore, in the technical solution of the present application, the access traffic time series input vector is segmented to obtain a first access traffic local time series input vector and a second access traffic local time series input vector, wherein the length of the first access traffic local time series input vector is greater than 2 times that of the second access traffic local time series input vector, which means that the first local time series input vector covers a coarser time scale (longer time span), while the second local time series input vector covers a finer time scale (shorter time span), thereby helping the system to more comprehensively understand the behavior of network traffic and more effectively identify anomalies.
[0027] In an embodiment of the present application, the first access traffic local time series association encoding module 130 is used to input the first access traffic local time series input vector into a sequence encoder to obtain a sequence of the first access traffic local time series implicit association feature vectors. Specifically, in an embodiment of the present application, the first access traffic local time series association encoding module is used to: input the first access traffic local time series input vector into a sequence encoder based on a 1D-CNN model to obtain a sequence of the first access traffic local time series implicit association feature vectors. It should be understood that the first access traffic local time series input vector contains the time series feature information of the access traffic in the local time period, and the 1D-CNN model is good at extracting features from sequence data, that is, by using the convolution layer of a one-dimensional convolution kernel, the model can capture local patterns and trends in the access traffic time series data. Based on this, in the technical solution of the present application, the first access traffic local time series input vector is input into a sequence encoder based on a 1D-CNN model to capture and mine the access traffic feature information in the local time series, respectively, so as to obtain a sequence of the first access traffic local time series implicit association feature vectors.
[0028] In an embodiment of the present application, the first traffic local timing aggregation module 140 is used to input the sequence of the first access traffic local timing implicitly associated feature vectors into a node feature propagation network based on a node energy decay mechanism to obtain a first access traffic local timing feature propagation aggregation representation vector. Accordingly, considering that the first access traffic timing implicitly associated features at different local times in the sequence of the first access traffic local timing implicitly associated feature vectors are different, therefore, in order to be able to more accurately identify and capture the timing correlation between the various first access traffic local timing implicitly associated feature vectors, thereby forming a more complete and comprehensive first access traffic local timing feature representation, in the technical solution of the present application, the sequence of the first access traffic local timing implicitly associated feature vectors is input into a node feature propagation network based on a node energy decay mechanism to obtain a first access traffic local timing feature propagation aggregation representation vector.
[0029] It is worth mentioning that the node feature propagation network based on the node energy decay mechanism adopts an innovative method to focus on analyzing the evolution and propagation of node features in time series data. By imitating the natural phenomenon of the gradual decay of node energy, the network can aggregate node features at different time points to form a comprehensive feature description of the node in the time series continuous space, thereby improving the model's ability to understand and analyze dynamic data. In detail, first, by calculating the energy statistical paradigm value of each local time series implicit association feature vector of the first access traffic, in particular, the energy statistical paradigm value is related to the maximum value, average value and variance of the local time series implicit association feature vector of the first access traffic, which is used to represent the significance or activity of the local time series implicit association feature of the first access traffic. Then, by calculating the propagation intervals between the local time series implicit association feature vectors of each first access traffic, the connection pattern and spatial association between nodes are revealed, and based on these interval values and the historical node energy statistical paradigm values, the energy propagation attenuation coefficients of the local time series implicit association features of the first access traffic are calculated. These coefficients are inversely correlated with the historical node energy, imitating the decreasing phenomenon of energy in duration and scope of action. Next, these decay coefficients are used as weighting coefficients to perform weighted accumulation on each first access traffic local time series implicit correlation feature vector, generating a decay time series aggregate feature vector that integrates historical node features. Finally, by combining the energy statistical paradigm value of the current first access traffic local time series implicit correlation feature vector with this historical aggregate vector and performing weighted summation, a first access traffic local time series feature propagation aggregate representation vector that integrates changes in the time dimension and structural information in the space dimension is obtained.
[0030] Specifically, Figure 3 FIG. 1 is a block diagram of a first flow local time series aggregation module in a network security protection operation and maintenance management system according to an embodiment of the present application. Figure 3As shown, the first traffic local timing aggregation module 140 includes: a first access traffic local timing characteristic energy statistical paradigm value calculation unit 141, which is used to calculate the energy statistical paradigm value of each first access traffic local timing implicitly associated feature vector based on the maximum value, average value and variance of each first access traffic local timing implicitly associated feature vector in the sequence of the first access traffic local timing implicitly associated feature vector to obtain a sequence of first access traffic local timing implicitly associated energy statistical paradigm values, wherein the energy statistical paradigm value corresponding to the current first access traffic local timing implicitly associated feature vector in the sequence of the first access traffic local timing implicitly associated energy statistical paradigm values is used as the energy statistical paradigm value of the current node The energy statistical paradigm value is used and other energy statistical paradigm values are used as historical node energy statistical paradigm values to obtain a sequence of the current first access traffic local time series implicitly associated energy statistical paradigm value and the historical first access traffic local time series implicitly associated energy statistical paradigm value; the first access traffic local time series node propagation space span value calculation unit 142 is used to count the node propagation space span values between each other first access traffic local time series implicitly associated feature vector in the sequence of the first access traffic local time series implicitly associated feature vector and the current first access traffic local time series implicitly associated feature vector to obtain a sequence of the first access traffic local time series implicitly associated propagation space span values; the first access traffic local time series energy propagation attenuation coefficient value calculation unit 14 3. It is used to determine the energy propagation attenuation coefficient values of other first access traffic local time series implicitly associated feature vectors in the sequence of the first access traffic local time series implicitly associated feature vectors based on the sequence of the first access traffic local time series implicitly associated propagation space span values and the sequence of the historical first access traffic local time series implicitly associated energy statistical paradigm values to obtain a sequence of first access traffic local time series implicitly associated energy propagation attenuation coefficient values, wherein the first access traffic local time series implicitly associated energy propagation attenuation coefficient value is inversely correlated with the historical first access traffic local time series implicitly associated energy statistical paradigm value; the first access traffic local time series feature weighting unit 144 is used to weight the first access traffic local time series implicitly associated energy propagation attenuation coefficient value based on the first access traffic local time series implicitly associated energy statistical paradigm value. A sequence of sequential implicit correlation energy propagation attenuation coefficient values is used as a weight sequence, and the weighted sum of all other first access traffic local timing implicit correlation feature vectors in the sequence of the first access traffic local timing implicit correlation feature vector is calculated to obtain a historical first access traffic local timing implicit correlation energy attenuation timing aggregate feature vector; a first access traffic local timing feature aggregation unit 145 is used to fuse the current first access traffic local timing implicit correlation energy statistical paradigm value to calculate the weighted sum of the historical first access traffic local timing implicit correlation energy attenuation timing aggregate feature vector and the current first access traffic local timing implicit correlation feature vector to obtain the first access traffic local timing feature propagation aggregate representation vector.
[0031] More specifically, in an embodiment of the present application, the first access traffic local timing feature energy statistical paradigm value calculation unit is used to: extract the maximum value of each position in the first access traffic local timing implicit correlation feature vector to obtain the maximum value of the first access traffic local timing implicit correlation feature; respectively calculate the mean value and variance of the first access traffic local timing implicit correlation feature vector to obtain the mean value of the first access traffic local timing implicit correlation feature and the variance of the first access traffic local timing implicit correlation feature; multiply the value obtained by adding the variance of the first access traffic local timing implicit correlation feature and the preset hyperparameter by four to obtain the first energy statistical paradigm value of the first access traffic local timing implicit correlation feature; calculate the first access traffic local timing implicit correlation feature. The square of the difference between the maximum value of the feature and the average value of the first access traffic local timing implicit correlation feature is calculated to obtain the difference value of the first access traffic local timing implicit correlation feature; the variance of the first access traffic local timing implicit correlation feature obtained by multiplying the variance of the first access traffic local timing implicit correlation feature by two is added to the value obtained by multiplying the preset hyperparameter by two and the difference value of the first access traffic local timing implicit correlation feature to obtain the second energy statistical paradigm value of the first access traffic local timing implicit correlation feature; the first energy statistical paradigm value of the first access traffic local timing implicit correlation feature is divided by the second energy statistical paradigm value of the first access traffic local timing implicit correlation feature to obtain the first access traffic local timing implicit correlation energy statistical paradigm value.
[0032] More specifically, in an embodiment of the present application, the first access traffic local timing energy propagation attenuation coefficient value calculation unit is used to: multiply each first access traffic local timing implicitly associated propagation space span value in the sequence of the first access traffic local timing implicitly associated propagation space span value by a first weight hyperparameter to obtain a sequence of first weighted propagation space span values; use each first access traffic local timing implicitly associated propagation space span value in the sequence of the first access traffic local timing implicitly associated propagation space span value as an exponent of a natural constant to calculate an exponential function value with the natural constant as the base according to the position to obtain a sequence of first access traffic local timing implicitly associated propagation class support space span values; use the first access traffic local timing implicitly associated propagation space span value Each first access traffic local timing implicitly associated propagation class support space span value in the sequence of propagation class support space span values is multiplied by the second weight hyperparameter to obtain a sequence of second weighted propagation space span values; the positional sum of the sequence of the first weighted propagation space span values and the sequence of the second weighted propagation space span values is calculated to obtain a sequence of weighted sum propagation space span values; the sequence of the historical first access traffic local timing implicitly associated energy statistical paradigm values and the sequence of the weighted sum propagation space span values are divided by each group of corresponding historical first access traffic local timing implicitly associated energy statistical paradigm values and the weighted sum propagation space span value to obtain a sequence of the first access traffic local timing implicitly associated energy propagation attenuation coefficient values.
[0033] In the embodiment of the present application, specifically, the first traffic local time series aggregation module is used to: input the sequence of the first access traffic local time series implicitly associated feature vectors into the node feature propagation network based on the node energy decay mechanism, and process it with the following aggregation formula to obtain the first access traffic local time series feature propagation aggregation representation vector; wherein the aggregation formula is: in, is a sequence of implicitly associated feature vectors of the local timing of the first access traffic, and are respectively the first access traffic local time series implicit correlation feature vector sequence and The first access traffic local time series implicit correlation feature vector, The number of feature vectors in the sequence representing the implicit correlation feature vector of the local time series of the first access traffic, and are respectively the first access traffic local time series implicit correlation feature vectors in the sequence The mean and variance of the implicit correlation feature vector of the local time series of the first access traffic, For the The maximum value of each position in the implicit correlation feature vector of the local time series of the first access traffic, is the preset hyperparameter, It is The energy statistical norm value of the implicit correlation feature vector of the local time series of the first access traffic, It is The first access traffic local time series implicit correlation feature vector The eigenvalues at the positions, is the number of eigenvalues in the implicit correlation eigenvector of each first access traffic local time series, and They are and The energy statistical norm value of the implicit correlation feature vector of the local time series of the first access traffic, express and The number of eigenvectors between , , and is the weight hyperparameter, It is the propagation aggregation representation vector of the local timing feature of the first access traffic.
[0034] In an embodiment of the present application, the access traffic timing reasoning module 150 is used to input the first access traffic local timing feature propagation aggregation representation vector into a decoder-based access traffic simulation reasoner to obtain a decoded access traffic timing input vector. Accordingly, it is considered that the first access traffic local timing feature propagation aggregation representation vector aggregates the traffic-related feature information of different local timings in the first scale. Therefore, in order to more reasonably obtain the timing pattern of the access traffic, in the technical solution of the present application, the first access traffic local timing feature propagation aggregation representation vector is input into a decoder-based access traffic simulation reasoner to obtain a decoded access traffic timing input vector. That is, the decoder can be used to reconstruct the timing pattern of the access traffic by decoding the aggregate representation of the local timing traffic features to generate a decoded access traffic timing input vector that is close to the original traffic pattern.
[0035] In an embodiment of the present application, the access traffic difference calculation module 160 is used to calculate the Hamming distance between the decoded access traffic timing input vector and the second access traffic local timing input vector. Accordingly, in order to more accurately evaluate the difference between the actual second access traffic local timing input vector and the inferred decoded access traffic timing input vector, so as to more accurately determine whether there is abnormal traffic, in the technical solution of the present application, the Hamming distance between the decoded access traffic timing input vector and the second access traffic local timing input vector is calculated. The Hamming distance is a metric used to measure the difference between two equal-length input data. In this way, by calculating the Hamming distance between the two, the differences between the actual traffic and the simulated traffic can be identified, and these differences may indicate traffic anomalies.
[0036] Here, considering that the decoded access traffic timing input vector represents the decoded inference timing distribution of the one-dimensional local timing correlation characteristics of the access traffic in the first local time domain based on the timing node feature propagation aggregation, and the second access traffic local timing input vector represents the source timing distribution of the access traffic in the second local time domain, therefore, when calculating the Hamming distance between the decoded access traffic timing input vector and the second access traffic local timing input vector to map the decoded access traffic timing input vector and the second access traffic local timing input vector to the common Hamming distance space, it is expected to further improve the overall expression logical dependency based on the distribution of their respective feature sets, thereby improving the accuracy of the calculated Hamming distance.
[0037] Based on this, the present application jointly optimizes the decoded access traffic timing input vector and the second access traffic local timing input vector. That is, in a preferred example, the Hamming distance between the decoded access traffic timing input vector and the second access traffic local timing input vector is calculated, including the following steps: Cascading the decoded access traffic timing input vector and the second access traffic local timing input vector to obtain a decoded access traffic-second access traffic cascade timing vector; Determine a first weight matrix and a second weight matrix, wherein the first weight matrix and the second weight matrix are The characteristic values are respectively the first and second The eigenvalue and The square root of the correlation value and the L2 distance value of the eigenvalue: in, and respectively represent the first and second access traffic concatenation timing vectors of the decoded access traffic and the second access traffic The eigenvalue and The square root of the correlation value and the L2 distance value of the eigenvalue, represents the first weight matrix Eigenvalues, represents the second weight matrix Eigenvalue.
[0038] Respectively matrix-multiplying the decoded access traffic-second access traffic cascade timing vector with the first weight matrix to obtain a first intermediate vector, matrix-multiplying the decoded access traffic-second access traffic cascade timing vector with the second weight matrix to obtain a second intermediate vector, and dot-multiplying the first weight matrix with the second weight matrix and then matrix-multiplying the decoded access traffic-second access traffic cascade timing vector to obtain a third intermediate vector; and The first intermediate vector, the second intermediate vector and the third intermediate vector are added together to obtain an optimized decoding access traffic-second access traffic cascade timing vector.
[0039] For example, the decoded access traffic-second access traffic cascade timing vector is a row vector, then the optimized decoding access traffic-second access traffic cascade timing vector It is expressed as: in, represents the decoding access traffic-second access traffic cascade timing vector, represents the first weight matrix, represents the second weight matrix, Indicates point addition, represents matrix multiplication, , and represent the first intermediate vector, the second intermediate vector and the third intermediate vector respectively, The optimized decoding access traffic-second access traffic cascade timing vector is represented.
[0040] Then, the optimized decoding access traffic-second access traffic cascade timing vector is split into an optimized decoding access traffic timing input vector and an optimized second access traffic local timing input vector according to the cascade of the decoding access traffic timing input vector and the second access traffic local timing input vector, and the Hamming distance between the optimized decoding access traffic timing input vector and the optimized second access traffic local timing input vector is calculated.
[0041] That is, by using the decoded access traffic-second access traffic cascade timing vector as its statistical reference-free distribution response framework based on the self-correlation matrix and the self-distance matrix, the simple repetition of feature distribution is avoided based on the retrieval-enhanced reverse response construction for the decoded access traffic-second access traffic cascade timing vector, and the surface combination mapping is avoided by ensuring the intrinsic mapping logic of the decoded access traffic-second access traffic cascade timing vector based on the retrieval-response context, so as to achieve the semantic logic dependency mapping of the decoded access traffic-second access traffic cascade timing vector while maintaining the intuitive response correlation, improve the overall expression logic dependency of the optimized decoded access traffic timing input vector and the optimized second access traffic local timing input vector, and improve the calculation accuracy of the Hamming distance between them. In this way, the intrinsic pattern and correlation of traffic data can be understood more deeply, and the dependence on manually set fixed traffic thresholds can be reduced. The system can continuously optimize and adjust according to new data to adapt to the changing network environment in a dynamic manner.
[0042] In the embodiment of the present application, the abnormal traffic determination module 170 is used to determine whether there is abnormal traffic based on the comparison between the Hamming distance and the preset threshold. That is, the Hamming distance obtained by calculating the difference between the decoded access traffic timing input vector and the second access traffic local timing input vector and the preset threshold are compared to automatically determine whether there is abnormal traffic. In this way, the inherent pattern and correlation of traffic data can be more deeply understood, and the dependence on manually set fixed traffic thresholds can be reduced. The system can continuously optimize and adjust according to new data to dynamically adapt to the changing network environment.
[0043] In summary, the network security protection operation and maintenance management system 100 based on the embodiment of the present application is explained, which obtains the time series of the access flow value of the power application within a predetermined time period, and adopts an AI-based data processing and analysis algorithm to perform time series segmentation and implicit association coding on the time series of the access flow value, so as to automatically determine whether there is abnormal traffic based on the comparison between the Hamming distance between the access flow time series characteristics obtained by reasoning at the first time scale and the access flow time series characteristics at the second scale and the preset threshold. In this way, the inherent pattern and correlation of the traffic data can be more deeply understood, and at the same time, the dependence on manually set fixed traffic thresholds can be reduced. The system can continuously optimize and adjust according to new data to dynamically adapt to the changing network environment.
[0044] As described above, the network security protection operation and maintenance management system 100 according to the embodiment of the present application can be implemented in various terminal devices, such as a server for network security protection operation and maintenance. In one example, the network security protection operation and maintenance management system 100 according to the embodiment of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the network security protection operation and maintenance management system 100 can be a software module in the operating system of the terminal device, or it can be an application developed for the terminal device; of course, the network security protection operation and maintenance management system 100 can also be one of the many hardware modules of the terminal device.
[0045] Alternatively, in another example, the network security protection operation and maintenance management system 100 and the terminal device may also be separate devices, and the network security protection operation and maintenance management system 100 may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.
[0046] Figure 4 Flowchart of the network security protection operation and maintenance management method according to the embodiment of the present application. Figure 4 As shown, in the network security protection operation and maintenance management method, it includes: S110, obtaining the time series of access flow values of the power application program within a predetermined time period; S120, arranging the time series of the access flow values according to the time dimension to obtain the access flow timing input vector, and then dividing the access flow timing input vector to obtain a first access flow local timing input vector and a second access flow local timing input vector; S130, inputting the first access flow local timing input vector into a sequence encoder to obtain a sequence of implicitly associated feature vectors of the first access flow local timing; S140, The sequence of implicitly associated feature vectors of the local timing of the first access traffic is input into a node feature propagation network based on a node energy decay mechanism to obtain a first access traffic local timing feature propagation aggregation representation vector; S150, the first access traffic local timing feature propagation aggregation representation vector is input into an access traffic simulation reasoner based on a decoder to obtain a decoded access traffic timing input vector; S160, the Hamming distance between the decoded access traffic timing input vector and the second access traffic local timing input vector is calculated; S170, based on the comparison between the Hamming distance and a preset threshold, it is determined whether there is abnormal traffic.
[0047] Here, those skilled in the art can understand that the specific operations of each step in the above network security protection operation and maintenance management method have been referred to above. Figures 1 to 3 The network security protection operation and maintenance management system has been introduced in detail, and therefore, its repeated description will be omitted.
[0048] To summarize, a network security protection operation and maintenance management method based on an embodiment of the present application is explained, which obtains the time series of access traffic values of power applications within a predetermined time period, and uses AI-based data processing and analysis algorithms to perform time series segmentation and implicit association encoding on the time series of the access traffic values. In this way, it is automatically determined whether there is abnormal traffic based on the comparison between the Hamming distance between the access traffic time series characteristics obtained by reasoning at the first time scale and the access traffic time series characteristics at the second scale and a preset threshold.
[0049] The above is only a preferred embodiment of the present application and does not constitute any form of limitation to the present application. Although the present application has been disclosed as a preferred embodiment as above, it is not intended to limit the present application. Any technician familiar with this profession can make slight changes or modifications to the technical contents disclosed above into equivalent embodiments with equivalent changes without departing from the scope of the technical solution of the present application.
Claims
1. A network security protection operation and maintenance management system, characterized in that: include: An access flow value time series acquisition module is used to obtain the time series of access flow values of the power application program within a predetermined time period; An access traffic value time series arrangement and segmentation module, configured to arrange the time series of the access traffic values according to the time dimension to obtain an access traffic time series input vector, and then segment the access traffic time series input vector to obtain a first access traffic local time series input vector and a second access traffic local time series input vector; A first access traffic local time sequence association encoding module, configured to input the first access traffic local time sequence input vector into a sequence encoder to obtain a sequence of first access traffic local time sequence implicit association feature vectors; A first traffic local time series aggregation module, configured to input the sequence of the first access traffic local time series implicitly associated feature vectors into a node feature propagation network based on a node energy decay mechanism to obtain a first access traffic local time series feature propagation aggregation representation vector; An access traffic timing reasoning module, used for inputting the first access traffic local timing feature propagation aggregation representation vector into a decoder-based access traffic simulation reasoner to obtain a decoded access traffic timing input vector; An access traffic difference calculation module, used for calculating the Hamming distance between the decoded access traffic timing input vector and the second access traffic local timing input vector; The abnormal traffic determination module is used to determine whether abnormal traffic exists based on the comparison between the Hamming distance and a preset threshold.
2. The network security protection operation and maintenance management system according to claim 1, characterized in that: The access traffic value time sequence arrangement and segmentation module includes: An access flow value data structuring unit, used for arranging the time series of the access flow values according to the time dimension to obtain the access flow time series input vector; An access traffic timing input vector splitting unit is used to split the access traffic timing input vector to obtain the first access traffic local timing input vector and the second access traffic local timing input vector, wherein the length of the first access traffic local timing input vector is greater than twice that of the second access traffic local timing input vector.
3. The network security protection operation and maintenance management system according to claim 2 is characterized in that: The first access traffic local timing association encoding module is used to: input the first access traffic local timing input vector into a sequence encoder based on a 1D-CNN model to obtain a sequence of the first access traffic local timing implicit association feature vectors.
4. The network security protection operation and maintenance management system according to claim 3 is characterized in that: The first traffic local time series aggregation module includes: A first access traffic local timing characteristic energy statistical paradigm value calculation unit is used to calculate the energy statistical paradigm value of each first access traffic local timing implicitly associated feature vector in the sequence of the first access traffic local timing implicitly associated feature vector based on the maximum value, average value and variance of each first access traffic local timing implicitly associated feature vector to obtain a sequence of first access traffic local timing implicitly associated energy statistical paradigm values, wherein the energy statistical paradigm value corresponding to the current first access traffic local timing implicitly associated feature vector in the sequence of the first access traffic local timing implicitly associated energy statistical paradigm values is used as the current node energy statistical paradigm value and other energy statistical paradigm values are used as historical node energy statistical paradigm values to obtain a sequence of the current first access traffic local timing implicitly associated energy statistical paradigm value and the historical first access traffic local timing implicitly associated energy statistical paradigm value; A first access traffic local time series node propagation space span value calculation unit, used for counting the node propagation space span values between each other first access traffic local time series implicitly associated feature vector in the sequence of the first access traffic local time series implicitly associated feature vector and the current first access traffic local time series implicitly associated feature vector to obtain a sequence of first access traffic local time series implicitly associated propagation space span values; The first access traffic local time series energy propagation attenuation coefficient value calculation unit is used to determine the energy propagation attenuation coefficient values of other first access traffic local time series implicitly associated feature vectors in the sequence of the first access traffic local time series implicitly associated feature vectors based on the sequence of the first access traffic local time series implicitly associated propagation space span values and the sequence of the historical first access traffic local time series implicitly associated energy statistical paradigm values to obtain the sequence of the first access traffic local time series implicitly associated energy propagation attenuation coefficient values, wherein the first access traffic local time series implicitly associated energy propagation attenuation coefficient value is inversely correlated with the historical first access traffic local time series implicitly associated energy statistical paradigm value; A first access traffic local time series feature weighting unit is configured to calculate the weighted sum of all other first access traffic local time series implicitly associated feature vectors in the sequence of the first access traffic local time series implicitly associated feature vectors by using the sequence of the first access traffic local time series implicitly associated energy propagation attenuation coefficient values as a weight sequence to obtain a historical first access traffic local time series implicitly associated energy attenuation time series aggregated feature vector; The first access traffic local timing feature aggregation unit is used to fuse the current first access traffic local timing implicit correlation energy statistical paradigm value to calculate the weighted sum of the historical first access traffic local timing implicit correlation energy attenuation timing aggregation feature vector and the current first access traffic local timing implicit correlation feature vector to obtain the first access traffic local timing feature propagation aggregation representation vector.
5. The network security protection operation and maintenance management system according to claim 4 is characterized in that: The first access traffic local time series feature energy statistical paradigm value calculation unit is used to: Extracting the maximum value of each position in the local time series implicit correlation feature vector of the first access traffic to obtain the maximum value of the local time series implicit correlation feature of the first access traffic; Respectively calculating the average value and variance of the first access traffic local time series implicit correlation feature vector to obtain the first access traffic local time series implicit correlation feature average value and the first access traffic local time series implicit correlation feature variance; The value obtained by adding the variance of the first access traffic local time series implicit correlation feature and the preset hyperparameter is multiplied by four to obtain the first energy statistical paradigm value of the first access traffic local time series implicit correlation feature; Calculate the square of the difference between the maximum value of the first access traffic local time series implicit correlation feature and the average value of the first access traffic local time series implicit correlation feature to obtain the first access traffic local time series implicit correlation feature difference value; The variance of the modulated first access traffic local time series implicit correlation feature obtained by multiplying the variance of the first access traffic local time series implicit correlation feature by two is added to the value obtained by multiplying the preset hyperparameter by two and the difference value of the first access traffic local time series implicit correlation feature to obtain the second energy statistical paradigm value of the first access traffic local time series implicit correlation feature; The first energy statistical normal value of the local timing implicit correlation feature of the first access traffic is divided by the second energy statistical normal value of the local timing implicit correlation feature of the first access traffic to obtain the energy statistical normal value of the local timing implicit correlation feature of the first access traffic.
6. The network security protection operation and maintenance management system according to claim 5, characterized in that: The first access traffic local time series energy propagation attenuation coefficient value calculation unit is used to: Multiplying each first access traffic local timing implicitly associated propagation space span value in the sequence of the first access traffic local timing implicitly associated propagation space span values by a first weight hyperparameter to obtain a sequence of first weighted propagation space span values; Using each first access traffic local timing implicitly associated propagation spatial span value in the sequence of the first access traffic local timing implicitly associated propagation spatial span values as an exponent of a natural constant to calculate an exponential function value based on the natural constant according to the position to obtain a sequence of first access traffic local timing implicitly associated propagation class support spatial span values; Multiplying each first access traffic local time series implicit association propagation class support space span value in the sequence of the first access traffic local time series implicit association propagation class support space span value by a second weight hyperparameter to obtain a sequence of second weighted propagation space span values; Calculating a position-wise sum of the sequence of the first weighted propagation spatial span values and the sequence of the second weighted propagation spatial span values to obtain a sequence of weighted sum propagation spatial span values; Divide each group of corresponding historical first access traffic local timing implicit correlation energy statistical paradigm values and weighted sum propagation space span values in the sequence of the historical first access traffic local timing implicit correlation energy statistical paradigm values and the sequence of the weighted sum propagation space span values to obtain a sequence of the first access traffic local timing implicit correlation energy propagation attenuation coefficient values.
7. A network security protection operation and maintenance management method, characterized in that: include: Obtaining a time series of access flow values of a power application within a predetermined time period; After arranging the time series of the access traffic values according to the time dimension to obtain an access traffic time series input vector, the access traffic time series input vector is segmented to obtain a first access traffic local time series input vector and a second access traffic local time series input vector; Inputting the first access traffic local time sequence input vector into a sequence encoder to obtain a sequence of first access traffic local time sequence implicitly associated feature vectors; Inputting the sequence of the first access traffic local time series implicitly associated feature vectors into a node feature propagation network based on a node energy decay mechanism to obtain a first access traffic local time series feature propagation aggregation representation vector; Input the first access traffic local timing feature propagation aggregation representation vector into a decoder-based access traffic simulation reasoner to obtain a decoded access traffic timing input vector; Calculating a Hamming distance between the decoded access traffic timing input vector and the second access traffic local timing input vector; Based on the comparison between the Hamming distance and a preset threshold, it is determined whether abnormal traffic exists.
8. The network security protection operation and maintenance management method according to claim 7 is characterized in that: After arranging the time series of the access traffic values according to the time dimension to obtain an access traffic time series input vector, segmenting the access traffic time series input vector to obtain a first access traffic local time series input vector and a second access traffic local time series input vector, including: Arranging the time series of the access traffic values according to the time dimension to obtain the access traffic time series input vector; The access traffic timing input vector is split to obtain the first access traffic local timing input vector and the second access traffic local timing input vector, wherein the length of the first access traffic local timing input vector is greater than twice that of the second access traffic local timing input vector.
Citation Information
Patent Citations
Monitoring and early warning method for power operation and maintenance network security
CN112766517A
Pregnancy risk early warning management system
CN118553424A
Intelligent monitoring system and method for early exercise rehabilitation nursing of stroke patient
CN118576166A
Intelligent management system and method for power marketing archives
CN118656535A
Intelligent management and control system and method for maintenance operation of natural gas purification device
CN118672193A
Cited By
Multi-node heterogeneous server networking data secure intercommunication method and system
CN122069112A
Methods and systems for secure data interoperability in multi-node heterogeneous server networking
CN122069112B
