IT asset intelligent supervision platform based on internet of data

Abstract

The invention relates to the technical field of IT asset intelligent supervision platforms, and discloses a data networking-based IT asset intelligent supervision platform, which comprises a data acquisition module, a data fusion module, an analysis module, a resource scheduling module and a self-healing module, the data acquisition module is used for deploying edge computing nodes on each IT asset and acquiring equipment state data, network flow data and user behavior data; and the data preprocessing module is used for carrying out preliminary cleaning and formatting on the data and generating a preprocessed data packet. According to the IT asset intelligent supervision platform based on the internet of data, edge computing nodes are deployed on each IT asset, and equipment state data, network flow data and user behavior data are acquired, so that comprehensive monitoring of the IT assets is realized, the data transmission delay is reduced through a deployment mode, the data acquisition efficiency is improved, and the user experience is improved. And meanwhile, the burden of the central server is reduced, so that the whole system is more efficient and stable.

Description

Technical Field

[0001] The present invention relates to the technical field of IT asset intelligent supervision platform, and in particular to an IT asset intelligent supervision platform based on a digital network. Background Art

[0002] The IT asset intelligent supervision platform is a system based on digital networking technology, which is designed to comprehensively monitor and manage the enterprise's IT assets. However, the existing systems often rely on simple data merging methods for data fusion and lack effective data consistency and integrity verification, resulting in low quality of fused data, affecting the accuracy of analysis results. In addition, the existing systems usually rely on traditional statistical methods for intelligent analysis and lack advanced time series prediction and classification regression capabilities, and are unable to detect and predict potential problems in a timely manner. At the same time, the existing systems usually adopt static resource allocation strategies for resource scheduling and are unable to dynamically adjust resource allocation according to system load and prediction results, resulting in low resource utilization and unstable system performance. Summary of the invention

[0003] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0004] To achieve the above object, the present invention provides the following technical solutions: An IT asset intelligent supervision platform based on a data network includes a data acquisition module, a data fusion module, an analysis module, a resource scheduling module and a self-healing module; the data acquisition module is used to deploy edge computing nodes on each IT asset to collect device status data, network traffic data and user behavior data; the data preprocessing module is used to perform preliminary cleaning and formatting of data, generate preprocessed data packets, and transmit them to a central server through a communication protocol; the data fusion module is used to verify the integrity and consistency of the data after the central server receives the data packet, and use a data fusion algorithm based on a graph model to fuse data from different data sources, and store the fused data in In the distributed database; the intelligent analysis module is used to extract historical data from the database, perform feature engineering, use deep learning models for time series prediction, use integrated learning methods for classification and regression tasks, and develop a real-time monitoring system, combined with anomaly detection algorithms, to analyze the data; the resource scheduling module is used to dynamically adjust resource allocation using reinforcement learning algorithms, and optimize resource allocation strategies based on current system load and prediction results; the self-healing module is used to regularly scan the system status and detect potential faults. Once a fault is found, the self-healing system automatically starts the repair program and tries to restore the service. If the automatic repair fails, the system will generate a work order to notify the operation and maintenance personnel to intervene manually.

[0005] As a further solution of the present invention: the edge computing nodes are deployed on each IT asset to collect device status data, network traffic data and user behavior data. The specific steps are: deploying an edge computing node on each IT asset, the edge computing node includes a data collection unit, a data preprocessing unit and a data transmission unit; collecting device status data, network traffic data and user behavior data.

[0006] As a further solution of the present invention: the data is initially cleaned and formatted to generate preprocessed data packets, which are transmitted to the central server through a secure communication protocol. The specific steps are: checking the validity of the data, removing null values, error values ​​and abnormal values; normalizing different types of data to the same scale; converting the normalized data into an agreed structured format; packaging the structured data into data packets, each data packet containing a timestamp and a set of data records; compressing the data packets, and transmitting the data packets to the central server using a secure communication protocol TLS.

[0007] As a further solution of the present invention: after the central server receives the data packet, the data integrity and consistency are verified, the data from different data sources are merged using a data fusion algorithm based on a graph model, and the merged data is stored in a distributed database. The specific steps are: the central server receives the data packet from the edge computing node through a secure communication protocol TLS; the integrity of the data packet is verified; after the verification, the data packets from different data sources in the same time period are compared, and the expression is: ; in, Score the data consistency, For the Data packets, For the Data packets, is the kth normalized data record in the ith data packet, is the kth normalized data record in the jth data packet, is the total number of data records; Use the data fusion algorithm based on the graph model to fuse data from different data sources. The expression is: ; ; in, is the fused data, For the Normalized data records of each data source, is the total number of data sources, For the The weight of the data source, For the The data consistency score between the data source and the reference data source, A data package that is a reference data source.

[0008] As a further solution of the present invention: the historical data is extracted from the database, feature engineering is performed, a deep learning model is used for time series prediction, an integrated learning method is used for classification and regression tasks, and a real-time monitoring system is developed at the same time, and the data is analyzed in combination with anomaly detection algorithms. The specific steps are: extracting historical data from a distributed database, including device status data, network traffic data, and user behavior data; performing feature extraction and transformation on the extracted historical data to generate a feature vector suitable for model input; using a long short-term memory network to perform time series prediction, using historical data to train an LSTM model to generate a prediction model; using random forests to perform classification and regression tasks, using historical data to train a random forest model to generate classification and regression models; using a monitoring system to read the latest data in the database, monitor device status, network traffic, and user behavior; and using isolation forests for anomaly detection.

[0009] As a further solution of the present invention: the long short-term memory network is used to perform time series prediction, and the LSTM model is trained using historical data to generate a prediction model, which is expressed as follows: ; in, For time The predicted value of is the activation function, is the weight matrix of the hidden state, For time The hidden state of is the input weight matrix, For time The input feature vector is is the bias term.

[0010] As a further solution of the present invention: the random forest is used for classification and regression tasks, and the random forest model is trained using historical data to generate classification and regression models, and the expression is: ; in, is the classification result, is the input feature vector, is the number of decision trees, For the The prediction category of a decision tree, is the category label, is the indicator function.

[0011] As a further solution of the present invention: the resource allocation is dynamically adjusted using a reinforcement learning algorithm, and the resource allocation strategy is optimized according to the current system load and the prediction results. The specific steps are: collecting various state data of the current system, including CPU usage, memory usage, disk I / O, and network bandwidth; representing the collected state data as a state vector; setting an initial resource allocation strategy; defining an action space to represent the resource allocation actions that can be taken; defining a reward function to evaluate the system performance after taking an action in the state; and using a reinforcement learning algorithm to update the resource allocation strategy, the expression is: ; in, is the state-action value function, is the learning rate, is the discount factor, which indicates the weight of future rewards, To take action The new state after The action in the new state.

[0012] As a further solution of the present invention: the system status is scanned regularly to detect potential faults. Once a fault is found, the self-healing system automatically starts a repair program to try to restore service. If the automatic repair fails, the system will generate a work order to notify the operation and maintenance personnel to perform manual intervention. The specific steps are: setting a scheduled task to regularly scan the system status data, including CPU usage, memory usage, disk I / O, and network bandwidth; collecting the latest status data from each IT asset and aggregating these data to a central server; using an anomaly detection algorithm to detect anomalies in the system status data; setting an anomaly threshold, if the anomaly score exceeds the threshold, it is considered that the system has a fault; locating the specific location and type of the fault based on the feature with the highest anomaly score; starting a preset self-healing program to try to automatically repair the fault; if the self-healing program fails to repair the fault, recording the fault information, including the fault time, fault type, and fault location; generating a work order, sending the fault information to the operation and maintenance personnel, and notifying them to perform manual intervention.

[0013] As a further solution of the present invention: an abnormality threshold is set. If the abnormality score exceeds the threshold, it is considered that the system has a fault. The expression is: ; in, It is a fault flag, 1 indicates a fault, 0 indicates normal. is the abnormality score, is the abnormal threshold.

[0014] Compared with the prior art, the present invention has the following beneficial effects: By deploying edge computing nodes on various IT assets and collecting device status data, network traffic data and user behavior data, comprehensive monitoring of IT assets is achieved. The deployment method reduces the delay in data transmission and improves the efficiency of data collection. It also reduces the burden on the central server, making the entire system more efficient and stable. After the central server receives the data packet, it verifies the data integrity and consistency, uses a graph-based data fusion algorithm to fuse data from different data sources, and stores the fused data in a distributed database, achieving efficient fusion and management of multi-source data. The fusion method not only improves the utilization rate of data, but also enhances the accuracy and integrity of the data, providing a solid foundation for subsequent intelligent analysis. The use of distributed databases further improves the efficiency of data storage and access, ensuring the high availability and scalability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a module connection diagram of an IT asset intelligent supervision platform based on the Internet of Things. DETAILED DESCRIPTION

[0016] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and understandable, the specific implementation modes of the present invention are described in detail below in conjunction with the accompanying drawings.

[0017] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0018] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0019] Example 1 See also Figure 1, is the first embodiment of the present invention, which provides an IT asset intelligent supervision platform based on the Internet of Things, including a data acquisition module, a data fusion module, an analysis module, a resource scheduling module and a self-healing module; the data acquisition module is used to deploy edge computing nodes on each IT asset to collect device status data, network traffic data and user behavior data; the data preprocessing module is used to perform preliminary cleaning and formatting of the data, generate preprocessed data packets, and transmit them to the central server through the communication protocol; the data fusion module is used to verify the integrity and consistency of the data after the central server receives the data packet, and use the data fusion algorithm based on the graph model to fuse the data from different data sources, and The fused data is stored in a distributed database; the intelligent analysis module is used to extract historical data from the database, perform feature engineering, use deep learning models for time series prediction, use integrated learning methods for classification and regression tasks, and develop a real-time monitoring system, combined with anomaly detection algorithms, to analyze the data; the resource scheduling module is used to dynamically adjust resource allocation using a reinforcement learning algorithm, and optimize the resource allocation strategy based on the current system load and prediction results; the self-healing module is used to regularly scan the system status and detect potential faults. Once a fault is found, the self-healing system automatically starts the repair program and tries to restore the service. If the automatic repair fails, the system will generate a work order to notify the operation and maintenance personnel to perform manual intervention.

[0020] Specifically, edge computing nodes are deployed on each IT asset to collect device status data, network traffic data and user behavior data. The specific steps are: deploy an edge computing node on each IT asset, and the edge computing node includes a data collection unit, a data preprocessing unit and a data transmission unit; collect device status data, network traffic data and user behavior data.

[0021] It should be noted that the data acquisition module can not only collect data in real time, but also dynamically adjust the data collection strategy according to the preset collection frequency and conditions to ensure the comprehensiveness and timeliness of the data. By dynamically adjusting the data collection strategy, the data acquisition module can flexibly collect data in different time periods and under different conditions to ensure the comprehensiveness and timeliness of the data. The dynamic adjustment mechanism not only improves the flexibility of data collection, but also reduces unnecessary data collection, reduces the consumption of system resources, and improves the overall efficiency of the system.

[0022] Specifically, the data is initially cleaned and formatted to generate preprocessed data packets, which are then transmitted to the central server via a secure communication protocol. The specific steps are: checking the validity of the data, removing null values, error values, and outliers; normalizing different types of data to the same scale; converting the normalized data into an agreed structured format; packaging the structured data into data packets, each of which contains a timestamp and a set of data records; compressing the data packets, and transmitting them to the central server using a secure communication protocol TLS.

[0023] It should be noted that the data preprocessing module also includes a data encryption function to ensure the security of data during transmission. The data preprocessing module encrypts the data before data transmission to ensure that the data is not stolen or tampered with during transmission. The data encryption function greatly improves the security of data transmission, protects the security of sensitive data, and enhances the security of the system.

[0024] Specifically, after the central server receives the data packet, it verifies the data integrity and consistency, uses the data fusion algorithm based on the graph model to fuse the data from different data sources, and stores the fused data in the distributed database. The specific steps are: the central server receives the data packet from the edge computing node through the secure communication protocol TLS; verifies the integrity of the data packet; after verification, compares the data packets from different data sources in the same time period, and the expression is: ; in, Score the data consistency, For the Data packets, For the Data packets, is the kth normalized data record in the ith data packet, is the kth normalized data record in the jth data packet, is the total number of data records; Use the data fusion algorithm based on the graph model to fuse data from different data sources. The expression is: ; ; in, is the fused data, For the Normalized data records of each data source, is the total number of data sources, For the The weight of the data source, For the The data consistency score between the data source and the reference data source, A data package that is a reference data source.

[0025] It should be noted that the data fusion module also includes a data quality assessment function, which regularly assesses the quality of the fused data and generates a quality report. The data fusion module ensures the accuracy and reliability of the data by regularly assessing the quality of the fused data. The quality report can help operation and maintenance personnel to timely understand the data quality status and take necessary measures to optimize it. The data quality assessment function not only improves the credibility of the data, but also provides data support for the continuous optimization of the system, ensuring the long-term stable operation of the system.

[0026] Specifically, historical data is extracted from the database, feature engineering is performed, a deep learning model is used for time series prediction, an integrated learning method is used for classification and regression tasks, and a real-time monitoring system is developed at the same time. The data is analyzed in combination with anomaly detection algorithms. The specific steps are: extract historical data from the distributed database, including device status data, network traffic data, and user behavior data; extract and transform the extracted historical data to generate feature vectors suitable for model input; use a long short-term memory network for time series prediction, use historical data to train the LSTM model, and generate a prediction model; use random forests for classification and regression tasks, use historical data to train the random forest model, and generate classification and regression models; use a monitoring system to read the latest data in the database, monitor device status, network traffic, and user behavior; and use isolation forests for anomaly detection; use a long short-term memory network for time series prediction, use historical data to train the LSTM model, and generate a prediction model. The expression is: ; in, For time The predicted value of is the activation function, is the weight matrix of the hidden state, For time The hidden state of is the input weight matrix, For time The input feature vector is is the bias term; random forest is used for classification and regression tasks, and the random forest model is trained using historical data to generate classification and regression models. The expression is: ; in, is the classification result, is the input feature vector, is the number of decision trees, For the The prediction category of a decision tree, is the category label, is the indicator function.

[0027] It should be noted that the intelligent analysis module also includes a data visualization function, which intuitively displays the analysis results in the form of charts and dashboards. The intelligent analysis module uses the data visualization function to intuitively display complex analysis results in the form of charts and dashboards, helping operation and maintenance personnel to quickly understand and make decisions. The data visualization function not only improves the comprehensibility of data analysis results, but also simplifies the workflow of operation and maintenance personnel and improves work efficiency.

[0028] Specifically, the reinforcement learning algorithm is used to dynamically adjust resource allocation, and the resource allocation strategy is optimized according to the current system load and prediction results. The specific steps are: collect various status data of the current system, including CPU usage, memory usage, disk I / O, and network bandwidth; represent the collected status data as a state vector; set an initial resource allocation strategy; define the action space to represent the resource allocation actions that can be taken; define the reward function to evaluate the system performance after taking actions in the state; use the reinforcement learning algorithm to update the resource allocation strategy, the expression is: ; in, is the state-action value function, is the learning rate, is the discount factor, which indicates the weight of future rewards, To take action The new state after The action in the new state.

[0029] It should be noted that the resource scheduling module also includes a real-time monitoring function for resource usage, which displays the resource usage of the current system in real time. The resource scheduling module uses the real-time monitoring function to display the resource usage of the current system in real time, helping operation and maintenance personnel to understand the system load status in a timely manner and dynamically adjust resources. The real-time monitoring function of resource usage not only improves the accuracy of resource scheduling, but also reduces problems caused by insufficient or over-allocation of resources, thereby improving the overall performance of the system.

[0030] Specifically, the system status is scanned regularly to detect potential faults. Once a fault is found, the self-healing system automatically starts the repair program and tries to restore the service. If the automatic repair fails, the system will generate a work order and notify the operation and maintenance personnel to intervene manually. The specific steps are: set a scheduled task to regularly scan the system status data, including CPU usage, memory usage, disk I / O, and network bandwidth; collect the latest status data from each IT asset and aggregate this data to the central server; use anomaly detection algorithms to detect anomalies in the system status data; set an anomaly threshold. If the anomaly score exceeds the threshold, the system is considered to have a fault; locate the specific location and type of the fault based on the feature with the highest anomaly score; start the preset self-healing program and try to automatically repair the fault; if the self-healing program fails to repair the fault, record the fault information, including the fault time, fault type, and fault location; generate a work order, send the fault information to the operation and maintenance personnel, and notify them to intervene manually; set an anomaly threshold. If the anomaly score exceeds the threshold, the system is considered to have a fault. The expression is: ; in, It is a fault flag, 1 indicates a fault, 0 indicates normal. is the abnormality score, is the abnormal threshold.

[0031] It should be noted that the self-healing module also includes a fault prediction function, which uses a machine learning algorithm to predict potential faults and perform preventive maintenance in advance. The self-healing module uses a machine learning algorithm to analyze historical data, predict potential faults, perform preventive maintenance in advance, and reduce the occurrence of faults. The fault prediction function not only improves the reliability of the system, but also reduces service interruptions caused by sudden faults, thereby improving the stability of the system and user experience.

[0032] In summary, by deploying edge computing nodes on various IT assets and collecting device status data, network traffic data, and user behavior data, comprehensive monitoring of IT assets is achieved. The deployment method reduces the delay in data transmission and improves the efficiency of data collection. It also reduces the burden on the central server, making the entire system more efficient and stable. After the central server receives the data packet, it verifies the data integrity and consistency, uses a graph-based data fusion algorithm to fuse data from different data sources, and stores the fused data in a distributed database, achieving efficient fusion and management of multi-source data. The fusion method not only improves the utilization rate of data, but also enhances the accuracy and integrity of the data, providing a solid foundation for subsequent intelligent analysis. The use of distributed databases further improves the efficiency of data storage and access, ensuring the high availability and scalability of the system.

[0033] Example 2 Please refer to Table 1, which shows the second embodiment of the present invention.

[0034] In order to verify the effectiveness and superiority of the IT asset intelligent supervision platform based on the Internet of Things, this embodiment was actually tested in a large enterprise data center, which has 100 servers. Edge computing nodes are deployed on each server to collect device status data, network traffic data, and user behavior data. The entire test process is divided into the following steps: S1. Data collection module deployment: Edge computing nodes are deployed on each server. The edge computing nodes include data collection units, data preprocessing units and data transmission units. The data collection unit is responsible for collecting device status data (such as CPU usage, memory usage, disk I / O, network bandwidth), network traffic data and user behavior data. The data preprocessing unit performs preliminary cleaning on the collected data, removes null values, error values ​​and outliers, and normalizes different types of data to the same scale. The data transmission unit converts the normalized data into a unified structured format and packages it into data packets. Each data packet contains a timestamp and a set of data records, and then uses the secure communication protocol TLS to transmit the data packets to the central server.

[0035] S2, data fusion module processing: The central server receives data packets from edge computing nodes through the secure communication protocol TLS, verifies the integrity of the data packets to ensure the integrity and consistency of the data, and uses a data fusion algorithm based on a graph model to fuse data from different data sources. The specific steps include comparing data packets from different data sources in the same time period, calculating data consistency scores, and then fusing data based on the scores. The fused data is stored in a distributed database for subsequent intelligent analysis.

[0036] S3. Application of intelligent analysis module: extract historical data from distributed databases, including device status data, network traffic data, and user behavior data, perform feature extraction and transformation on the extracted historical data, generate feature vectors suitable for model input, use long short-term memory networks (LSTM) for time series prediction, use historical data to train LSTM models, generate prediction models, use random forests for classification and regression tasks, use historical data to train random forest models, generate classification and regression models, develop real-time monitoring systems, read the latest data in the database, monitor device status, network traffic, and user behavior, use isolation forests for anomaly detection, and promptly discover anomalies in the system.

[0037] S4. Optimization of resource scheduling module: Collect various status data of the current system, including CPU usage, memory usage, disk I / O, and network bandwidth, represent the collected status data as a state vector, set an initial resource allocation strategy, define the action space, represent the resource allocation actions that can be taken, define the reward function to evaluate the system performance after taking actions in the state, use the reinforcement learning algorithm to update the resource allocation strategy, dynamically adjust resource allocation according to the current system load and prediction results, and optimize the resource allocation strategy.

[0038] S5. Self-healing module test: Set up scheduled tasks to regularly scan the system status data, including CPU usage, memory usage, disk I / O, and network bandwidth. Collect the latest status data from each IT asset and aggregate this data to the central server. Use anomaly detection algorithms to detect anomalies in the system status data and set an anomaly threshold. If the anomaly score exceeds the threshold, the system is considered to have a fault. Based on the features with the highest anomaly score, locate the specific location and type of the fault, start the preset self-healing program, and try to automatically repair the fault. If the self-healing program fails to repair the fault, record the fault information, including the fault time, fault type, and fault location, generate a work order, and notify the operation and maintenance personnel to perform manual intervention.

[0039] sheet: Table 1: Performance comparison of the prior art and the present invention in data centers

[0040] By comparing the data in the above table, it can be clearly seen that the present invention has significant advantages over the prior art in many aspects: The present invention effectively reduces the CPU utilization rate, reduces the system load, and improves the overall performance of the system through intelligent resource scheduling and optimization. The CPU utilization rate of the present invention is as low as 68%, while the highest in the prior art is 78%.

[0041] The present invention reduces memory usage, improves memory resource utilization, and reduces the risk of memory overflow by optimizing memory management and resource allocation. The lowest memory usage of the present invention is 73%, while the highest in the prior art is 82%.

[0042] The present invention reduces the occupation of network bandwidth, improves the efficiency of network transmission, and reduces the possibility of network congestion through data preprocessing and compression. The minimum network bandwidth of the present invention is 88 Mbps, while the maximum of the prior art is 105 Mbps.

[0043] The present invention significantly improves the accuracy of fault detection through advanced anomaly detection algorithms and self-healing modules, timely discovers and handles potential faults, and improves the reliability and stability of the system. The fault detection rate of the present invention is as high as 96%, while the highest in the prior art is only 62%.

[0044] The present invention dynamically adjusts resource allocation through a reinforcement learning algorithm, optimizes resource usage, and improves resource utilization from 70% to 87%, significantly improving the overall performance and resource utilization efficiency of the system. The resource utilization rate of the present invention is as high as 87%, while the maximum in the prior art is only 72%.

[0045] In summary, by deploying edge computing nodes on various IT assets and collecting device status data, network traffic data, and user behavior data, comprehensive monitoring of IT assets is achieved. The deployment method reduces the delay in data transmission and improves the efficiency of data collection. It also reduces the burden on the central server, making the entire system more efficient and stable. After the central server receives the data packet, it verifies the data integrity and consistency, uses a graph-based data fusion algorithm to fuse data from different data sources, and stores the fused data in a distributed database, achieving efficient fusion and management of multi-source data. The fusion method not only improves the utilization rate of data, but also enhances the accuracy and integrity of the data, providing a solid foundation for subsequent intelligent analysis. The use of distributed databases further improves the efficiency of data storage and access, ensuring the high availability and scalability of the system.

[0046] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An IT asset intelligent supervision platform based on digital network, characterized in that: include: Data acquisition module, data fusion module, analysis module, resource scheduling module and self-healing module; The data collection module is used to deploy edge computing nodes on various IT assets to collect device status data, network traffic data and user behavior data; The data preprocessing module is used to perform preliminary cleaning and formatting of the data, generate preprocessed data packets, and transmit them to the central server through the communication protocol; The data fusion module is used to verify the integrity and consistency of the data after the central server receives the data packet, fuse the data from different data sources using a data fusion algorithm based on a graph model, and store the fused data in a distributed database; The intelligent analysis module is used to extract historical data from the database, perform feature engineering, use deep learning models for time series prediction, use ensemble learning methods for classification and regression tasks, and develop a real-time monitoring system in combination with anomaly detection algorithms to analyze data; The resource scheduling module is used to dynamically adjust resource allocation using a reinforcement learning algorithm and optimize resource allocation strategies based on current system load and prediction results; The self-healing module is used to regularly scan the system status and detect potential faults. Once a fault is found, the self-healing system automatically starts the repair program and attempts to restore the service. If the automatic repair fails, the system will generate a work order to notify the operation and maintenance personnel to perform manual intervention.

2. The IT asset intelligent supervision platform based on the digital network according to claim 1 is characterized by: Deploy edge computing nodes on each IT asset to collect device status data, network traffic data, and user behavior data. The specific steps include: Deploy edge computing nodes on each IT asset. The edge computing nodes include data collection units, data preprocessing units, and data transmission units. Collect device status data, network traffic data, and user behavior data.

3. The IT asset intelligent supervision platform based on digital network according to claim 1 is characterized by: The data is initially cleaned and formatted to generate pre-processed data packets, which are then transmitted to the central server via a secure communication protocol. The specific steps include: Check the validity of data and remove null values, error values ​​and outliers; Normalize different types of data to the same scale; Convert normalized data into an agreed structured format; Pack the structured data into data packets, each of which contains a timestamp and a set of data records; The data packets are compressed and transmitted to the central server using the secure communication protocol TLS.

4. The IT asset intelligent supervision platform based on the digital network according to claim 1 is characterized by: After the central server receives the data packet, it verifies the data integrity and consistency, uses a data fusion algorithm based on a graph model to fuse data from different data sources, and stores the fused data in a distributed database. The specific steps include: The central server receives data packets from edge computing nodes through the secure communication protocol TLS; Verify the integrity of data packets; After verification, the data packets from different data sources in the same time period are compared, and the expression is: ; in, Score the data consistency, For the Data packets, For the Data packets, is the kth normalized data record in the ith data packet, is the kth normalized data record in the jth data packet, is the total number of data records; Use the data fusion algorithm based on the graph model to fuse data from different data sources. The expression is: ; ; in, is the fused data, For the Normalized data records of each data source, is the total number of data sources, For the The weight of the data source, For the The data consistency score between the data source and the reference data source, A data package that is a reference data source.

5. The IT asset intelligent supervision platform based on digital network according to claim 1 is characterized by: The steps of extracting historical data from the database, performing feature engineering, using deep learning models for time series prediction, using ensemble learning methods for classification and regression tasks, and developing a real-time monitoring system combined with anomaly detection algorithms to analyze the data include: Extract historical data from distributed databases, including device status data, network traffic data, and user behavior data; Perform feature extraction and transformation on the extracted historical data to generate feature vectors suitable for model input; Use long short-term memory networks for time series forecasting, train LSTM models using historical data, and generate forecasting models; Use random forests for classification and regression tasks, train random forest models using historical data, and generate classification and regression models; Use a monitoring system to read the latest data in the database and monitor device status, network traffic, and user behavior; And use isolation forest for anomaly detection.

6. The IT asset intelligent supervision platform based on the digital network according to claim 5 is characterized by: The long short-term memory network is used to perform time series prediction, and the LSTM model is trained using historical data to generate a prediction model, which is expressed as: ; in, For time The predicted value of is the activation function, is the weight matrix of the hidden state, For time The hidden state of is the input weight matrix, For time The input feature vector is is the bias term.

7. The IT asset intelligent supervision platform based on digital network according to claim 5 is characterized by: The random forest is used for classification and regression tasks, and the random forest model is trained using historical data to generate classification and regression models. The expression is: ; in, is the classification result, is the input feature vector, is the number of decision trees, For the The prediction category of a decision tree, is the category label, is the indicator function.

8. The IT asset intelligent supervision platform based on digital network according to claim 1 is characterized by: The method uses a reinforcement learning algorithm to dynamically adjust resource allocation and optimize resource allocation strategies according to the current system load and prediction results. The specific steps include: Collect various status data of the current system, including CPU usage, memory usage, disk I / O, and network bandwidth; Represent the collected state data as a state vector; Set an initial resource allocation strategy; Define an action space, representing the resource allocation actions that can be taken; Define a reward function to evaluate the performance of the system after taking an action in a state; Use reinforcement learning algorithm to update resource allocation strategy, the expression is: ; in, is the state-action value function, is the learning rate, is the discount factor, which indicates the weight of future rewards, To take action The new state after The action in the new state.

9. The IT asset intelligent supervision platform based on digital network according to claim 1 is characterized by: The system status is scanned regularly to detect potential faults. Once a fault is found, the self-healing system automatically starts the repair program and tries to restore the service. If the automatic repair fails, the system will generate a work order to notify the operation and maintenance personnel to intervene manually. The specific steps include: Set up scheduled tasks to regularly scan system status data, including CPU usage, memory usage, disk I / O, and network bandwidth; Collect the latest status data from each IT asset and aggregate the data to the central server; Use anomaly detection algorithms to detect anomalies in system status data; Set an anomaly threshold. If the anomaly score exceeds the threshold, the system is considered to have a fault. Locate the specific location and type of the fault based on the feature with the highest anomaly score; Start the preset self-healing program to try to automatically repair the fault; If the self-healing procedure fails to repair the fault, record the fault information, including the fault time, fault type and fault location; Generate a work order and send the fault information to the operation and maintenance personnel to notify them to perform manual intervention.

10. The IT asset intelligent supervision platform based on the digital network according to claim 9 is characterized by: An abnormality threshold is set. If the abnormality score exceeds the threshold, it is considered that the system has a fault. The expression is: ; in, It is a fault flag, 1 indicates a fault, 0 indicates normal. is the abnormality score, is the abnormal threshold.

Images