A data lifecycle management method, system, electronic device and storage medium

Abstract

The application discloses a data life cycle management method and system, electronic equipment and storage medium, which are used for improving the reliability of data. The method comprises the following steps: collecting target data through an Internet of Things device, and generating a digital signature bound to the identity of the Internet of Things device; at an edge computing node, verifying and reaching a consensus on the target data and the digital signature based on a dynamic trust evaluation mechanism; recording the data digest of the target data verified by the consensus, an off-chain storage address and verification information used for proving the validity of the consensus into a blockchain; performing compliance verification on the state or operation related to the target data through a smart contract on the blockchain; based on the data entity, the Internet of Things device and the related operation event recorded in the blockchain, a multi-modal heterogeneous graph used for representing the association relationship of the whole life cycle of the target data is constructed; and in response to a query or access request for the target data, a response is provided after taking dynamic privacy protection measures.

Description

Technical Field

[0001] This application relates to the field of blockchain technology, and in particular to a data lifecycle management method, system, electronic device, and storage medium. Background Technology

[0002] With the widespread application of IoT devices in industrial control, smart manufacturing, smart cities, and logistics monitoring, massive amounts of data are continuously collected, transmitted, and processed. This data includes sensor-collected information, equipment operating status, and various operational events, providing crucial information for business analysis, system optimization, and decision-making. Simultaneously, the increasing complexity and distribution of data management place higher demands on data security, reliability, and efficiency.

[0003] However, existing technologies have significant shortcomings in data lifecycle management, namely, they struggle to provide reliable security and trustworthiness guarantees for distributed data during transmission and storage. Data may be tampered with, lost, or illegally accessed during the collection, transmission, or storage stages, resulting in a failure to effectively guarantee data integrity and authenticity, thus affecting data reliability. Summary of the Invention

[0004] This application provides a data lifecycle management method, system, electronic device, and storage medium that can improve data reliability.

[0005] The first aspect of this application provides a data lifecycle management method, including:

[0006] Collect target data through IoT devices and generate a digital signature that is bound to the identity of the IoT devices;

[0007] At the edge computing node, the target data and its digital signature are verified and consensus is reached based on a dynamic trust evaluation mechanism;

[0008] The data digest of the target data that has been verified through consensus, the off-chain storage address, and the verification information used to prove the validity of this consensus are all recorded in the blockchain;

[0009] Compliance verification is performed on the status or operations related to the target data through smart contracts on the blockchain.

[0010] Based on the data entities recorded in the blockchain, the IoT devices, and related operation events, a multimodal heterogeneous graph is constructed to characterize the association relationships throughout the entire lifecycle of the target data.

[0011] In response to a query or access request for the target data, a response is provided after dynamic privacy protection measures are taken.

[0012] Optionally, at the edge computing node, verifying the target data and its digital signature based on a dynamic trust evaluation mechanism and reaching a consensus includes:

[0013] Periodically collect behavioral data and contextual information of the edge computing nodes participating in the verification, and calculate the dynamic trust score for each node;

[0014] Based on the dynamic trust score, a group of nodes with a trust score higher than a preset threshold are selected from all nodes to form a fast consensus cluster for this round of consensus.

[0015] The fast consensus cluster verifies and aggregates the target data and its digital signature.

[0016] Once a predetermined number of nodes in the fast consensus cluster have signed their confirmations, consensus is considered to have been reached.

[0017] Optionally, the behavioral data includes the quality of uploaded data, the validity of digital signatures, the stability of network connections, and the anomaly rate in historical behavior; the context information includes the reputation of the organization to which the node belongs; and the calculation of the dynamic trust score for each node includes:

[0018] The evaluation results are obtained by assessing the data quality uploaded by each node, the effectiveness of digital signatures, the stability of network connections, the anomaly rate in historical behavior, and the credibility of the organization to which the node belongs.

[0019] The evaluation results are combined according to preset weights to obtain a dynamic trust score for each node.

[0020] Optionally, the step of constructing a multimodal heterogeneous graph to characterize the association relationships throughout the entire lifecycle of the target data, based on the data entities recorded in the blockchain, the IoT devices, and related operational events, includes:

[0021] The data entities recorded in the blockchain, the IoT devices, and related operation events are mapped to different types of graph nodes;

[0022] Generate connecting edges based on the dependencies, temporal or spatial relationships between the graph nodes;

[0023] A multimodal heterogeneous graph is constructed based on the graph nodes and the connecting edges.

[0024] Optionally, after constructing a multimodal heterogeneous graph based on the graph nodes and the connecting edges, the method further includes:

[0025] The graph nodes are initialized with features to form corresponding feature representations;

[0026] Information is propagated on the multimodal heterogeneous graph through a graph neural network, aggregating information between the graph nodes and their neighboring nodes, and using an attention mechanism to calculate the importance weights of the relationships between the graph nodes in order to update the feature representation;

[0027] When an abnormal node is detected in the graph, based on the updated feature representation and importance weight, a reverse path tracing is performed from the abnormal node along the connecting edge to determine the root cause node and the abnormal propagation path.

[0028] Optionally, providing a response after taking dynamic privacy protection measures includes:

[0029] Identify the current lifecycle stage and access context of the target data, and dynamically select and execute corresponding data encryption or desensitization operations based on the identification results.

[0030] Optionally, generating a digital signature bound to the identity of the IoT device includes:

[0031] In the secure isolation environment of the IoT device, the private key of the IoT device is used to perform calculations on a data structure containing the device identifier, timestamp, and target data to generate the digital signature.

[0032] A second aspect of this application provides a data lifecycle management system, comprising:

[0033] The data acquisition unit is used to collect target data through IoT devices and generate a digital signature that is bound to the identity of the IoT devices.

[0034] The consensus unit is used at the edge computing node to verify the target data and its digital signature and reach a consensus based on a dynamic trust evaluation mechanism.

[0035] The recording unit is used to record the data digest of the target data that has been verified by consensus, the off-chain storage address, and the verification information used to prove the validity of this consensus into the blockchain.

[0036] The verification unit is used to perform compliance verification on the state or operation related to the target data through smart contracts on the blockchain.

[0037] The construction unit is used to construct a multimodal heterogeneous graph representing the association relationships throughout the entire lifecycle of the target data, based on the data entities recorded in the blockchain, the IoT devices, and related operation events.

[0038] The response unit is used to provide a response after taking dynamic privacy protection measures in response to a query or access request for the target data.

[0039] Optionally, the consensus unit includes:

[0040] The computing module is used to periodically collect behavioral data and context information of the edge computing nodes participating in the verification, and to calculate the dynamic trust score of each node.

[0041] The election module is used to elect a group of nodes whose trust scores are higher than a preset threshold from all nodes based on the dynamic trust score, so as to form a fast consensus cluster for this round of consensus.

[0042] The verification module is used to verify and aggregate the target data and its digital signature by the fast consensus cluster;

[0043] The confirmation module is used to determine that consensus has been reached after obtaining signature confirmations from a preset number of nodes in the fast consensus cluster.

[0044] Optionally, the behavioral data includes the quality of uploaded data, the validity of digital signatures, the stability of network connections, and the anomaly rate in historical behavior. The calculation module is specifically used for:

[0045] The evaluation results are obtained by assessing the data quality uploaded by each node, the effectiveness of digital signatures, the stability of network connections, the anomaly rate in historical behavior, and the credibility of the organization to which the node belongs.

[0046] The evaluation results are combined according to preset weights to obtain a dynamic trust score for each node.

[0047] Optionally, the building unit includes:

[0048] The mapping module is used to map the data entities recorded in the blockchain, the IoT devices, and related operation events into different types of graph nodes;

[0049] The generation module is used to generate connecting edges based on the dependencies, temporal or spatial relationships between the graph nodes;

[0050] A construction module is used to construct a multimodal heterogeneous graph based on the graph nodes and the connecting edges.

[0051] Optionally, the building unit further includes:

[0052] An initialization module is used to initialize the features of the graph nodes to form corresponding feature representations;

[0053] The update module is used to propagate information on the multimodal heterogeneous graph through a graph neural network, aggregate information of the graph node and its neighboring nodes, and use an attention mechanism to calculate the importance weights of the relationships between the graph nodes in order to update the feature representation.

[0054] The tracking module is used to, when an abnormal node is detected in the graph, perform reverse path tracing from the abnormal node along the connecting edge based on the updated feature representation and the importance weight, so as to determine the root cause node and the abnormal propagation path.

[0055] Optionally, the response unit is specifically used for:

[0056] Identify the current lifecycle stage and access context of the target data, and dynamically select and execute corresponding data encryption or desensitization operations based on the identification results.

[0057] Optionally, the acquisition unit is specifically used for:

[0058] In the secure isolation environment of the IoT device, the private key of the IoT device is used to perform calculations on a data structure containing the device identifier, timestamp, and target data to generate the digital signature.

[0059] A third aspect of this application provides an electronic device, comprising:

[0060] Processor, memory, input / output units, and bus;

[0061] The processor is connected to the memory, the input / output unit, and the bus;

[0062] The memory stores a program, and the processor calls the program to execute the method of the first aspect and any possible implementation of the first aspect.

[0063] The fourth aspect of this application provides a computer-readable storage medium storing a program that, when executed on a computer, causes the computer to perform the methods of the first aspect and any possible implementation thereof.

[0064] As can be seen from the above technical solutions, this application has the following advantages:

[0065] In this embodiment, digital signatures are bound to IoT devices during the data collection phase to ensure consistency between data and device identity. At edge nodes, a dynamic trust assessment mechanism is used to complete highly reliable consensus verification. Then, data digests, off-chain storage addresses, and verification information are recorded on the blockchain to achieve lightweight storage and traceability. At the same time, smart contracts are used to perform compliance verification on data-related operations to ensure legitimate use. Furthermore, a multimodal heterogeneous graph is constructed based on data entities, devices, and events in the blockchain to achieve full lifecycle association modeling. Finally, dynamic privacy protection measures are adopted in the data access phase to reduce the risk of sensitive information leakage. Thus, a unified guarantee of authenticity, compliance, traceability, and privacy security is achieved throughout the entire data lifecycle, which is conducive to improving data reliability. Attached Figure Description

[0066] Figure 1 This is a flowchart illustrating one embodiment of the data lifecycle management method in this application;

[0067] Figure 2 This is a flowchart illustrating a sub-implementation of the data lifecycle management method in this application;

[0068] Figure 3 This is a flowchart illustrating another sub-implementation of the data lifecycle management method in this application;

[0069] Figure 4 This is a schematic diagram of the structure of one embodiment of the data lifecycle management system in this application;

[0070] Figure 5 This is a schematic diagram of the structure of one embodiment of the electronic device in this application. Detailed Implementation

[0071] This application provides a data lifecycle management method, system, electronic device, and storage medium to improve data reliability.

[0072] The embodiments of this application will now be described with reference to the accompanying drawings.

[0073] Please see Figure 1 , Figure 1 An embodiment of the data lifecycle management method provided in this application includes:

[0074] S101. Collect target data through IoT devices and generate a digital signature bound to the identity of the IoT devices;

[0075] In the target data generation stage, IoT devices collect target data related to the business scenario in real time based on their configured sensing modules (such as sensors, cameras, or data acquisition terminals). When the system receives the raw data, it automatically generates a digital signature bound to the IoT device's identity. This signature uses asymmetric encryption, employing the device's private key to encrypt the collected data and generate a unique identifier. This not only ensures a one-to-one correspondence between the target data and the device's identity but also prevents data from being tampered with or forged during transmission.

[0076] S102. At the edge computing node, the target data and its digital signature are verified and consensus is reached based on the dynamic trust assessment mechanism.

[0077] After data collection is completed and a digital signature is attached, the target data is transmitted to the edge computing node. At this point, the system initiates a dynamic trust assessment mechanism to verify the target data and its digital signature. This mechanism comprehensively considers multiple dimensions, such as interaction records between nodes, historical verification results, data quality performance, and dynamic adjustments to reputation values, to calculate the real-time trust level of the node. Based on these trust levels, multiple nodes reach a consensus judgment through voting or weighted consensus, ensuring that the data indeed originates from a trusted device and has not been tampered with. Unlike traditional static assessments, the dynamic mechanism can update node trust levels according to real-time environmental changes, improving the flexibility and accuracy of data verification.

[0078] S103. Record the data digest of the target data that has been verified by consensus, the off-chain storage address, and the verification information used to prove the validity of this consensus into the blockchain.

[0079] After consensus verification is completed, the system generates a data digest of the target data. This digest, typically obtained through a hash algorithm, uniquely represents the data content. Simultaneously, the system obtains the address of the target data in an off-chain storage system (such as distributed file storage) and packages the validity proof information generated during the verification process, along with the data digest and storage address, into the blockchain. The data digest ensures data integrity, the storage address alleviates the storage pressure caused by large-scale data uploads, and the verification information ensures the traceability of the consensus process. These three elements, collectively written into the blockchain, achieve a "lightweight on-chain, robust storage" model.

[0080] S104. Compliance verification of the status or operation related to the target data is performed through smart contracts on the blockchain.

[0081] As records on the blockchain are continuously updated, smart contracts are automatically triggered to perform compliance checks on the state or operations related to the target data. These checks include verifying the legitimacy of the visitor's identity, whether data access complies with established permission rules, and whether processing operations meet industry standards or regulatory requirements. Smart contracts automatically execute these checks based on preset rules, requiring no manual intervention, thus achieving transparency and automation in the data usage process. Any requests or operations that do not meet the conditions are rejected or recorded on the blockchain, ensuring that the target data complies with legal and security requirements at every stage of its lifecycle.

[0082] S105. Based on the data entities, IoT devices and related operation events recorded in the blockchain, construct a multimodal heterogeneous graph to represent the association relationships throughout the entire lifecycle of the target data;

[0083] After compliance is verified, the system further integrates the data entities recorded in the blockchain, IoT devices, and related operational events to construct a multimodal heterogeneous graph. This graph includes not only the information modality of the target data itself, but also the identity modality of IoT devices, the reputation modality generated by consensus verification, the storage modality of blockchain records, and the operational modality of compliance verification. Through the construction of this graph structure, the entire process of data acquisition, verification, storage, retrieval, and regulation can be intuitively represented, achieving unified modeling of cross-modal information. The multimodal heterogeneous graph provides structured support for subsequent data analysis, traceability, and decision-making, making the target data lifecycle management more transparent and interpretable.

[0084] S106. In response to a query or access request for target data, provide a response after taking dynamic privacy protection measures.

[0085] When an external user or application initiates a query or access request for target data, the system first activates a dynamic privacy protection mechanism. These measures may include anonymizing sensitive data fields, introducing noise into statistical results using differential privacy technology, or allocating access scope based on user roles and permission levels. Simultaneously, the system will also perform risk analysis on requests based on access history; if abnormal requests are detected, additional verification or access denial will be initiated. Only after privacy protection is completed will the system return a response to the legitimate requester, thereby meeting business needs while minimizing the risk of privacy breaches.

[0086] In this embodiment, digital signatures are bound to IoT devices during the data collection phase to ensure consistency between data and device identity. At edge nodes, a dynamic trust assessment mechanism is used to complete highly reliable consensus verification. Then, data digests, off-chain storage addresses, and verification information are recorded on the blockchain to achieve lightweight storage and traceability. At the same time, smart contracts are used to perform compliance verification on data-related operations to ensure legal use. Furthermore, a multimodal heterogeneous graph is constructed based on data entities, devices, and events in the blockchain to achieve full lifecycle association modeling. Finally, dynamic privacy protection measures are adopted in the data access phase to reduce the risk of sensitive information leakage, thereby achieving unified protection of authenticity, compliance, traceability, and privacy security throughout the entire data lifecycle.

[0087] Furthermore, step S101 may specifically include:

[0088] In a secure, isolated environment for IoT devices, the private key of the IoT device is used to perform calculations on a data structure containing device identifiers, timestamps, and target data to generate a digital signature.

[0089] In a secure, isolated environment for IoT devices, the system invokes the device's internal private key to perform calculations on a data structure containing a device identifier, a timestamp, and target data to generate a digital signature. The device identifier uniquely identifies the IoT device, ensuring clear traceability of the data's origin. The timestamp records the exact moment the data was generated, preventing data corruption and ensuring order confirmation. The target data is the original content collected by the IoT device, serving as the core input to the digital signature. By performing this calculation in a secure, isolated environment, the private key is protected from external theft, ensuring the generated digital signature possesses authenticity, integrity, and non-repudiation. This signature is then bound to the target data, providing a trusted foundation for subsequent verification.

[0090] By generating a digital signature that is bound to the device identity and includes a timestamp during the data collection process, the target data can have verifiable source credibility and time integrity, thereby providing security for subsequent data verification and traceability.

[0091] Furthermore, step S106 may specifically include:

[0092] Identify the current lifecycle stage and access context of the target data, and dynamically select and execute the corresponding data encryption or desensitization operations based on the identification results.

[0093] Upon receiving a query or access request for target data, the system first identifies the current lifecycle stage of the target data, such as the collection, storage, usage, or archiving stage. It also considers access context information, including the requester's identity, access purpose, access permission level, and historical access behavior. Based on these identifications, the system dynamically selects the most appropriate privacy protection strategy and performs corresponding data encryption or anonymization operations. For example, if the data is in the storage stage and the requester has low privileges, the system may partially anonymize sensitive fields; if the data is in the usage stage and involves cross-subject transmission, end-to-end encryption may be implemented to ensure the data is not intercepted during transmission. This dynamic privacy protection approach allows for flexible matching of security requirements to data access in different scenarios.

[0094] By identifying the lifecycle stage and access context of target data and dynamically selecting privacy protection strategies accordingly, the system can effectively reduce the risk of sensitive information leakage while ensuring access efficiency, thereby achieving secure and controllable data access.

[0095] Please see Figure 2 In some embodiments of this application, step S102 in the above embodiments verifies the target data and its digital signature and reaches a consensus at the edge computing node based on a dynamic trust evaluation mechanism, and may include the following steps:

[0096] S201. Periodically collect behavioral data and context information of the edge computing nodes participating in the verification, and calculate the dynamic trust score of each node.

[0097] At edge computing nodes, the system periodically collects behavioral data and contextual information from participating nodes. Behavioral data includes each node's participation frequency in historical consensus, the accuracy of signature verification results, and the number of abnormal exits or delayed responses, reflecting the node's stability and reliability. Contextual information includes the node's real-time operating status, CPU and memory usage, network latency and bandwidth, as well as the node's geographical or network location, used to assess the node's availability in the current environment. The system then performs comprehensive calculations on this data to generate a dynamic trust score for each node, which reflects the node's trustworthiness and execution capabilities in real time.

[0098] Specifically, behavioral data includes the quality of uploaded data, the validity of digital signatures, the stability of network connections, and the anomaly rate in historical behavior; contextual information includes the reputation of the organization to which the organization belongs; step S201 may include:

[0099] S2011. Evaluate the data quality uploaded by each node, the effectiveness of digital signatures, the stability of network connections, the anomaly rate in historical behavior, and the credibility of the organization to which the node belongs, and obtain the evaluation results for each item.

[0100] At the edge computing nodes, the system evaluates the behavioral data and contextual information of each participating verification node. Behavioral data includes the quality of data uploaded by the node, the validity rate of digital signatures, the stability of network connections, and the anomaly rate in historical behavior. Contextual information includes the reputation score of the organization to which the node belongs. The system independently evaluates each indicator through a contextual trust modeling module, obtaining evaluation results to comprehensively reflect the overall trustworthiness of the node. Data quality score, signature validity rate, network stability, historical anomaly rate, and organizational reputation score are core indicators that the system must use. Specific sub-items within each indicator, such as sensor data integrity rate, false alarm rate, network latency, disconnection rate, and number of abnormal uploads, are only illustrative examples used to illustrate the scoring source and method, and are not limiting requirements.

[0101] S2012. The evaluation results are combined and calculated according to the preset weights to obtain the dynamic trust score of each node.

[0102] The system comprehensively calculates the evaluation results according to preset weights to obtain a dynamic trust score for each node. The weights can be adjusted based on the business scenario to highlight the importance of key indicators or reduce the impact of secondary indicators. Through the dynamic weight scheduling module, the dynamic scores of nodes provide a basis for subsequent consensus node selection. Simultaneously, combined with the punishment and recovery mechanism module, nodes with frequent anomalies or low data quality are downgraded or isolated to ensure that the scoring results accurately reflect the trustworthiness of the nodes.

[0103] By refining and evaluating five core indicators of node behavior data and context information and calculating dynamic trust scores according to weights, the system can accurately grasp the real-time trust status of each node, providing a reliable basis for selecting high-trust nodes to participate in verification and consensus, while reducing the risk of low-trust nodes interfering with consensus, thereby improving the security and reliability of target data and digital signature verification process.

[0104] S202. Based on the dynamic trust score, select a group of nodes whose trust score is higher than the preset threshold from all nodes to form a fast consensus cluster for this round of consensus.

[0105] Based on the dynamic trust score of each node, the system selects nodes with scores higher than a preset threshold from all participating nodes to form a fast consensus cluster. Nodes in the fast consensus cluster are prioritized for verification of the target data and its digital signature in this round, ensuring that the consensus process is both efficient and reliable. Through this selection mechanism, the system can concentrate verification tasks on nodes with good recent performance and high stability, reducing the impact of potentially low-trust or unstable nodes on the consensus result.

[0106] S203, The target data and its digital signature are verified and aggregated by the fast consensus cluster;

[0107] Within a fast consensus cluster, each node independently verifies the target data and its digital signature. Each node verifies the validity of the signature, the completeness of the data, and confirms that the data was indeed generated by the corresponding IoT device. Subsequently, the nodes in the cluster aggregate the verification results to form a unified judgment. The aggregation method can employ majority voting or weighted averaging to ensure that the verification result is not affected by the anomalies or erroneous judgments of a single node, thereby improving the accuracy and reliability of the verification.

[0108] S204. Once the signatures of a preset number of nodes in the fast consensus cluster are confirmed, consensus is determined to be reached.

[0109] Once a preset number of nodes in the fast consensus cluster have completed signature confirmation, the system considers consensus achieved. This signature confirmation process is equivalent to multiple nodes endorsing the verification result; the consensus is only confirmed as valid when the required number of nodes is met, preventing a few nodes from tampering with or mishandling the data and influencing the final judgment. The system then outputs the consensus result, providing a reliable basis for subsequently writing the data digest, storage address, and verification information into the blockchain.

[0110] In this embodiment, by periodically calculating the dynamic trust score of nodes, selecting highly trustworthy nodes to form a fast consensus cluster, and performing multi-node verification and aggregation within the cluster, the consensus is confirmed by a preset number of signatures. This ensures high reliability and security for the target data and its digital signature verification process, while significantly improving consensus efficiency and reducing the risk of single-node failure or malicious node interference.

[0111] Please see Figure 3 In some embodiments of this application, step S105 in the above embodiments, based on data entities, IoT devices, and related operational events recorded in the blockchain, constructs a multimodal heterogeneous graph to characterize the association relationships throughout the entire lifecycle of the target data, and may include the following steps:

[0112] S301. Map the data entities, IoT devices, and related operation events recorded in the blockchain to different types of graph nodes;

[0113] The system first traverses all data entities in the blockchain, mapping each data record to a data entity node and attaching attribute information such as data type, generation time, source device identifier, and off-chain storage address. Then, it maps IoT devices that collect, transmit, or process data to device nodes, recording device type, unique identifier, operating status, and geographical location information. Finally, it maps data-related operation events to event nodes, including data upload, modification, access, and verification operations, recording the operation time, operator identity, and operation result. The system distinguishes different information sources and functional roles through node types, providing a basic set of nodes for constructing a multimodal heterogeneous graph.

[0114] S302. Generate connecting edges based on the dependencies, temporal or spatial relationships between graph nodes;

[0115] The system analyzes potential logical dependencies between nodes. For example, data upload operations depend on the completion of data acquisition, and data processing operations depend on the results of preceding operations. It also analyzes the temporal sequence of operation events to generate temporal edges, reflecting the chronological order of events. Spatially, the system identifies the physical or logical location relationships between nodes, such as adjacent devices or data entities in the same area, thus generating spatial edges. Each connection edge contains attribute information, such as edge type, dependency strength, time interval, or spatial distance, and is labeled with its directionality to accurately reflect the relationships between nodes in subsequent feature propagation and anomaly analysis.

[0116] S303. Construct a multimodal heterogeneous graph based on the graph nodes and connecting edges.

[0117] The system integrates all types of nodes and edges into a unified graph structure, enabling the graph to simultaneously contain data entity nodes, device nodes, and event nodes, as well as edges of various relationship types. This multimodal heterogeneous graph can reflect multi-dimensional relationships between data, devices, and operations within a single graph. For example, the relationship between device nodes and data nodes represents the data generation source relationship, while the relationship between event nodes and data nodes represents the operation execution relationship. The edge types and weights between nodes reflect the importance and strength of different relationships. This graph structure provides complete lifecycle association information for subsequent feature propagation and anomaly tracking.

[0118] S304. Initialize the features of the graph nodes to form corresponding feature representations;

[0119] The system generates an initial feature vector for each node, containing the node's own attribute information, historical state records, operation history, and contextual information. For example, for data entity nodes, this includes a content summary, creation time, and off-chain storage address; for device nodes, it includes the running status, location, and acquisition capabilities; and for event nodes, it includes the operation type, timestamp, and execution result. Feature initialization ensures that each node has a computable representation before graph neural network processing, enabling it to retain its own features and effectively integrate with neighboring nodes during information propagation.

[0120] S305. Information propagation is performed on multimodal heterogeneous graphs through graph neural networks, aggregating information between graph nodes and their neighboring nodes, and using attention mechanisms to calculate the importance weights of relationships between graph nodes in order to update feature representations.

[0121] By employing graph neural networks for information propagation on multimodal heterogeneous graphs, the system first collects feature representations of all neighboring nodes for each node, distinguishing neighbors according to edge type, such as dependency edges, temporal edges, or spatial edges. During information propagation, the system applies different aggregation functions to each edge type, summarizing neighboring node features separately by edge type. Then, an attention mechanism is used to calculate the importance weight of each edge to the current node. The attention mechanism considers the feature similarity of neighboring nodes, historical interaction frequency, and edge attributes, such as dependency strength or temporal order, thereby dynamically adjusting the contribution value of neighboring nodes in feature updates. The system updates node features through multi-level iterative updates, ensuring that the node's feature representation not only retains its own attributes but also incorporates contextual information, including the state, operation history, and key indicator changes of its neighboring nodes, reflecting the node's comprehensive state and potential impact throughout its entire lifecycle. Finally, after the node feature updates are completed, each node possesses a comprehensive representation of itself and its surrounding related nodes, providing an accurate data foundation for anomaly detection and relationship analysis.

[0122] S306. When an abnormal node is detected in the graph, based on the updated feature representation and importance weight, a reverse path tracing is performed from the abnormal node along the connecting edges to determine the root cause node and the abnormal propagation path.

[0123] When an abnormal node is detected in the graph, the system traces a reverse path from the abnormal node along the connecting edges based on the updated feature representation and the importance weights of neighboring nodes. During the tracing process, the system first identifies key neighboring nodes directly associated with the abnormal node and evaluates the similarity of their features to the abnormal node's features and the importance of the edge weights to determine the possible source of the anomaly. Then, the system traces back layer by layer along the edges to higher-level or dependent nodes, analyzing the propagation path of the anomaly among different node types, including data entity nodes, device nodes, and event nodes. The system records the node features, edge types, and importance weights on each path, and combines this with historical operation records and contextual information to assess the possible causes and scope of the anomaly propagation. Ultimately, the system can accurately identify the root cause node and the anomaly propagation path, providing precise evidence for locating the source of the problem, analyzing the cause of the anomaly, and formulating intervention and correction strategies, ensuring more timely and reliable anomaly handling in the entire lifecycle of data management.

[0124] In this embodiment, by mapping data entities, IoT devices, and operational events recorded in the blockchain to multimodal heterogeneous graph nodes, connecting edges with dependencies, temporal sequences, and spatial relationships between nodes are constructed. By aggregating information through graph neural networks and calculating node relationship weights through attention mechanisms, the system can comprehensively characterize the correlation relationships throughout the entire lifecycle of target data. At the same time, it can accurately track root cause nodes and anomaly propagation paths when anomalies occur, thereby significantly improving the efficiency, accuracy, and reliability of data management, analysis, and anomaly handling.

[0125] Please see Figure 4 , Figure 4 An embodiment of the data lifecycle management system provided in this application includes:

[0126] The acquisition unit 401 is used to acquire target data through IoT devices and generate a digital signature bound to the identity of the IoT devices.

[0127] Consensus unit 402 is used to verify the target data and its digital signature and reach a consensus at the edge computing node based on a dynamic trust evaluation mechanism;

[0128] Recording unit 403 is used to record the data digest of the target data that has been verified by consensus, the off-chain storage address, and the verification information used to prove the validity of this consensus into the blockchain.

[0129] Verification unit 404 is used to perform compliance verification on the state or operation related to the target data through smart contracts on the blockchain.

[0130] Building unit 405 is used to construct a multimodal heterogeneous graph representing the association relationships throughout the entire lifecycle of target data, based on data entities, IoT devices and related operation events recorded in the blockchain.

[0131] Response unit 406 is used to provide a response after taking dynamic privacy protection measures in response to a query or access request for target data.

[0132] Optionally, consensus unit 402 includes:

[0133] The computing module is used to periodically collect behavioral data and context information of the edge computing nodes participating in the verification, and to calculate the dynamic trust score of each node.

[0134] The election module is used to elect a group of nodes whose trust scores are higher than a preset threshold from all nodes based on dynamic trust scores, so as to form a fast consensus cluster for this round of consensus.

[0135] The verification module is used to verify and aggregate the target data and its digital signature by the fast consensus cluster;

[0136] The confirmation module is used to confirm that consensus has been reached after obtaining signature confirmations from a preset number of nodes in the fast consensus cluster.

[0137] Optionally, behavioral data includes uploaded data quality, digital signature validity, network connection stability, and anomaly rate in historical behavior. The calculation module is specifically used for:

[0138] The evaluation results are obtained by assessing the data quality uploaded by each node, the effectiveness of digital signatures, the stability of network connections, the anomaly rate in historical behavior, and the credibility of the organization to which the node belongs.

[0139] The evaluation results are combined according to preset weights to obtain a dynamic trust score for each node.

[0140] Optionally, building unit 405 includes:

[0141] The mapping module is used to map data entities, IoT devices, and related operational events recorded in the blockchain into different types of graph nodes;

[0142] The generation module is used to generate connecting edges based on the dependencies, temporal or spatial relationships between graph nodes.

[0143] The building module is used to construct multimodal heterogeneous graphs based on graph nodes and connecting edges.

[0144] Optionally, building unit 405 also includes:

[0145] The initialization module is used to initialize the features of graph nodes to form corresponding feature representations;

[0146] The update module is used to propagate information on multimodal heterogeneous graphs through graph neural networks, aggregate information of graph nodes and their neighboring nodes, and use the attention mechanism to calculate the importance weights of the relationships between graph nodes in order to update the feature representation;

[0147] The tracking module is used to trace the reverse path from the abnormal node along the connecting edges when an abnormal node is detected in the graph, based on the updated feature representation and importance weight, in order to determine the root cause node and the abnormal propagation path.

[0148] Optionally, response unit 406 is specifically used for:

[0149] Identify the current lifecycle stage and access context of the target data, and dynamically select and execute the corresponding data encryption or desensitization operations based on the identification results.

[0150] Optionally, the acquisition unit 401 is specifically used for:

[0151] In a secure, isolated environment for IoT devices, the private key of the IoT device is used to perform calculations on a data structure containing device identifiers, timestamps, and target data to generate a digital signature.

[0152] In this implementation, the functions of each unit and module are as described above. Figures 1 to 3 The steps in the illustrated embodiments are the same and will not be repeated here.

[0153] Please see Figure 5 , Figure 5 One embodiment of the electronic device provided in this application includes:

[0154] Processor 501, memory 502, input / output unit 503, and bus 504;

[0155] The processor 501 is connected to the memory 502, the input / output unit 503, and the bus 504;

[0156] The memory 502 stores a program, which the processor 501 calls to execute. Figures 1 to 3 The steps in the illustrated embodiment.

[0157] In this embodiment, the function of processor 501 is the same as described above. Figures 1 to 3 The steps in the illustrated embodiments are the same and will not be repeated here.

[0158] This application also provides a computer-readable storage medium on which a program is stored. When the program is executed on a computer, it causes the computer to perform the aforementioned actions. Figures 1 to 3 The method in any possible implementation.

[0159] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0160] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0161] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0162] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0163] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A data lifecycle management method, characterized in that, include: Collect target data through IoT devices and generate a digital signature that is bound to the identity of the IoT devices; At the edge computing node, the target data and its digital signature are verified and consensus is reached based on a dynamic trust evaluation mechanism; The data digest of the target data that has been verified through consensus, the off-chain storage address, and the verification information used to prove the validity of this consensus are all recorded in the blockchain; Compliance verification is performed on the status or operations related to the target data through smart contracts on the blockchain. Based on the data entities recorded in the blockchain, the IoT devices, and related operation events, a multimodal heterogeneous graph is constructed to characterize the association relationships throughout the entire lifecycle of the target data. In response to a query or access request for the target data, a response is provided after dynamic privacy protection measures are taken; The process of verifying and reaching consensus on the target data and its digital signature based on a dynamic trust evaluation mechanism at the edge computing node includes: Periodically collect behavioral data and contextual information of the edge computing nodes participating in the verification, and calculate the dynamic trust score for each node; Based on the dynamic trust score, a group of nodes with a trust score higher than a preset threshold are selected from all nodes to form a fast consensus cluster for this round of consensus. The fast consensus cluster verifies and aggregates the target data and its digital signature. Once a predetermined number of nodes in the fast consensus cluster have signed their confirmations, consensus is considered to have been reached. The construction of a multimodal heterogeneous graph, based on data entities recorded in the blockchain, IoT devices, and related operational events, to characterize the relationships throughout the entire lifecycle of the target data includes: The data entities recorded in the blockchain, the IoT devices, and related operation events are mapped to different types of graph nodes; Generate connecting edges based on the dependencies, temporal or spatial relationships between the graph nodes; A multimodal heterogeneous graph is constructed based on the graph nodes and the connecting edges; After constructing a multimodal heterogeneous graph based on the graph nodes and the connecting edges, the method further includes: The graph nodes are initialized with features to form corresponding feature representations; Information is propagated on the multimodal heterogeneous graph through a graph neural network, aggregating information between the graph nodes and their neighboring nodes, and using an attention mechanism to calculate the importance weights of the relationships between the graph nodes in order to update the feature representation; When an abnormal node is detected in the graph, based on the updated feature representation and importance weight, a reverse path tracing is performed from the abnormal node along the connecting edge to determine the root cause node and the abnormal propagation path.

2. The method according to claim 1, characterized in that, The behavioral data includes uploaded data quality, digital signature validity, network connection stability, and anomaly rate in historical behavior. The context information includes the reputation of the organization to which the node belongs. The calculation of the dynamic trust score for each node includes: The evaluation results are obtained by assessing the data quality uploaded by each node, the effectiveness of digital signatures, the stability of network connections, the anomaly rate in historical behavior, and the credibility of the organization to which the node belongs. The evaluation results are combined according to preset weights to obtain a dynamic trust score for each node.

3. The method according to claim 1 or 2, characterized in that, The response provided after taking dynamic privacy protection measures includes: Identify the current lifecycle stage and access context of the target data, and dynamically select and execute corresponding data encryption or desensitization operations based on the identification results.

4. The method according to claim 1 or 2, characterized in that, The generation of a digital signature bound to the identity of the IoT device includes: In the secure isolation environment of the IoT device, the private key of the IoT device is used to perform calculations on a data structure containing the device identifier, timestamp, and target data to generate the digital signature.

5. A data lifecycle management system, characterized in that, For implementing the method of any one of claims 1 to 4, comprising: The data acquisition unit is used to collect target data through IoT devices and generate a digital signature that is bound to the identity of the IoT devices. The consensus unit is used at the edge computing node to verify the target data and its digital signature and reach a consensus based on a dynamic trust evaluation mechanism. The recording unit is used to record the data digest of the target data that has been verified by consensus, the off-chain storage address, and the verification information used to prove the validity of this consensus into the blockchain. The verification unit is used to perform compliance verification on the state or operation related to the target data through smart contracts on the blockchain. The construction unit is used to construct a multimodal heterogeneous graph representing the association relationships throughout the entire lifecycle of the target data, based on the data entities recorded in the blockchain, the IoT devices, and related operation events. The response unit is used to provide a response after taking dynamic privacy protection measures in response to a query or access request for the target data.

6. An electronic device, characterized in that, include: Processor, memory, input / output units, and bus; The processor is connected to the memory, the input / output unit, and the bus; The memory stores a program, and the processor calls the program to execute the method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains a program that, when executed on a computer, causes the computer to perform the method as described in any one of claims 1 to 4.

Images