A Consistency Inspection Method and System for Edge Environments Based on a Trusted Collaboration Center
The centralized collaborative structure, which generates temporary session keys and cloud verification tags through the central node, solves the problem of multi-source data tampering in low-security environments at the edge, and achieves highly reliable and consistent inspection and anomaly location, thereby improving the reliability and efficiency of detection.
Patent Information
- Application Number
- CN202611189016.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-06
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies are insufficient to effectively resist multi-source data tampering attacks in low-security edge environments, and consistency detection relies on the insufficient trustworthiness of edge nodes, leading to false consistency issues.
A centralized collaborative structure is adopted, which generates temporary session keys and cloud verification tags through the central node to realize data synchronization and consistency inspection between edge nodes. By using random sampling and aggregated credential verification mechanism, communication overhead is reduced and detection reliability is improved.
To improve the reliability of consistency detection in low-security environments, enhance resistance to multi-source tampering attacks, balance detection efficiency and anomaly location accuracy, and reduce communication burden.
Smart Images

Figure CN123122804A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of edge computing technology, and in particular relates to an edge environment consistency inspection method and system based on a trusted collaborative center. Background Technology
[0002] As edge computing architectures are increasingly used in industrial sites, IoT access points, campus nodes, and regional collaboration scenarios, more and more data is being distributed and collaboratively processed across multiple servers or nodes at the edge layer. To ensure the consistency of data replicas on each edge node, continuous monitoring of data consistency across multiple nodes is typically required.
[0003] Most existing consistency verification technologies are applied in dual-machine backup or redundant data storage environments. A fundamental premise is that both communicating parties have a certain security commitment, meaning that the nodes participating in the consistency verification are assumed not to have been collaboratively tampered with. In such scenarios, common solutions typically use file hash value comparison to determine data consistency. Because hash value calculation is low-cost and simple to implement, this method has some practicality in traditional trusted environments.
[0004] However, in low-security edge environments, security commitments can be broken. Attackers may simultaneously compromise multiple edge nodes and perform consistent tampering on the same data across multiple nodes, causing each node to recalculate the same hash value. In this case, while consistency detection mechanisms relying solely on hash comparison may appear consistent, this consistency is actually based on synchronized tampering and constitutes false consistency. This problem makes existing solutions ineffective against multi-source, collaborative data tampering attacks. Hash algorithms themselves are also easy to generate; if an attacker can simultaneously control multiple edge nodes and regenerate matching hash values, they can bypass traditional consistency checks. Therefore, the shortcomings of existing technologies lie in the fact that consistency detection capabilities overly rely on the trustworthiness of the edge nodes themselves, and the generation mechanism of detection tags lacks security resilience support for multi-point attack scenarios in low-security environments.
[0005] Therefore, there is an urgent need for a new data consistency inspection scheme that can still reliably detect the data status between multiple edge nodes in a low-security environment at the edge by leveraging the collaboration of a trusted center and cryptographic verification mechanisms, and locate conflicting nodes and abnormal data blocks when anomalies are detected. Summary of the Invention
[0006] The purpose of this invention is to solve the problems mentioned in the background art and to propose a highly reliable and consistent inspection method and system based on centralized collaboration.
[0007] To achieve the objectives of this invention, this application provides a consistency inspection method for edge environments based on a trusted collaboration center, the method comprising:
[0008] S1. Construct a centralized collaborative structure:
[0009] A collaborative structure is constructed, which includes a central node and multiple edge nodes in the edge layer; wherein each edge node communicates directly with the central node, but the edge nodes do not directly synchronize collaborative data with each other;
[0010] S2, Task Initialization:
[0011] The central node generates a task identifier and corresponding node identifiers for each edge node for the collaborative task, and generates a temporary session key for the collaborative task; the initial collaborative data is divided into multiple data blocks, and a cloud verification tag is calculated for each data block based on the temporary session key; the collaborative task number, node identifier, initial collaborative data and corresponding cloud verification tag are sent to each edge node.
[0012] S3, Data Collaboration:
[0013] When any edge node performs a dynamic data operation, the corresponding edge node sends the data operation Act, the target data block identifier id, and the changed content m to the central node. id The central node recalculates the corresponding cloud verification tag based on the changed target data block and synchronizes the changed target data block and the recalculated cloud verification tag to other edge nodes to achieve indirect data collaboration between edge nodes via the central node.
[0014] S4. Consistency Inspection:
[0015] The central node initiates a consistency inspection and sends a sampling inspection challenge to each edge node. The sampling inspection challenge includes at least a set of indexes of data blocks to be inspected and a challenge random number corresponding to each data block to be inspected.
[0016] S5. Generate test certificate:
[0017] Each edge node responds to the sampling and detection challenge by extracting the sampled data block from its local storage and calculating a detection certificate based on the extracted sampled data block and the corresponding local storage verification label. The detection certificate includes at least aggregated data and aggregated verification label.
[0018] S6. Document Verification and Consistency Determination:
[0019] The central node receives the detection credentials returned by each edge node and verifies the detection credentials. The verification includes at least: verifying the validity of the aggregated verification tags returned by each edge node, and comparing whether the aggregated data returned by each edge node is consistent. Only when the aggregated verification tags returned by each edge node are all verified and the aggregated data returned by each edge node is consistent is it determined that the data status between each edge node is consistent; otherwise, it is determined that there is data inconsistency.
[0020] Furthermore, the cloud verification tag is generated by the central node based on the temporary session key, and each edge node does not store the temporary session key locally, so that each edge node cannot independently generate a valid cloud verification tag.
[0021] Furthermore, when the central node calculates the cloud verification tag for each data block, the cloud verification tag is at least identical to the content of the corresponding data block; and the collaborative task ID. task Node Identifier ID i Timestamp; Temporary session key session Related.
[0022] Furthermore, the edge nodes do not directly communicate with each other for data synchronization or consistency detection; all collaborative data streams and inspection control streams are forwarded or scheduled through the central node.
[0023] Furthermore, the set of data blocks to be tested specified in the sampling inspection challenge is determined by random sampling, and the challenge random number is used to drive each edge node to perform aggregation calculation on the sampled data blocks and their corresponding verification labels.
[0024] Furthermore, the detection credentials include aggregated data calculated by each edge node after extracting the sampled data block from its local storage, and aggregated verification tags calculated based on the local storage verification tags corresponding to the sampled data block.
[0025] Furthermore, the central node verifies whether the aggregated verification labels returned by each edge node match the verification results calculated based on the corresponding sampling detection challenge and cloud-side verification information, and compares whether the aggregated data returned by each edge node is consistent; the consistency test is deemed to have passed only when all aggregated verification labels have been verified and all aggregated data are completely consistent.
[0026] Furthermore, when data inconsistency is detected, the central node marks the abnormal edge nodes and triggers a secondary fine inspection based on single data block granularity to locate the tampered data block.
[0027] Furthermore, the secondary fine inspection includes: the central node sending fine inspection challenges to the abnormal edge nodes block by block, verifying the validity of the verification tag and the consistency of the data content of the corresponding individual data blocks, so as to determine the location of the abnormal data blocks.
[0028] This application also provides an edge environment consistency inspection system based on a trusted collaboration center, the system comprising:
[0029] The central node and multiple edge nodes;
[0030] The central node and each edge node form a star-shaped network collaborative structure.
[0031] The central node is used for:
[0032] For each collaborative task, a task identifier and a corresponding node identifier for each edge node are generated, and a temporary session key for the collaborative task is generated.
[0033] The initial collaborative data is divided into multiple data blocks, and a cloud verification tag is calculated for each data block based on the temporary session key.
[0034] Send the task identifier, the node identifier, the initial collaboration data, and the corresponding cloud verification tag to each edge node;
[0035] In response to data change requests sent by edge nodes, recalculate the corresponding cloud verification tag based on the changed target data block, and synchronize the changed target data block and the recalculated cloud verification tag to other edge nodes; initiate consistency inspection, send sampling inspection challenges to each edge node; receive and verify the inspection credentials returned by each edge node;
[0036] The edge nodes are used for:
[0037] Receive the task identifier, node identifier, initial collaboration data and corresponding cloud verification tag sent by the central node and store them locally;
[0038] When a dynamic data operation occurs, the data operation type, target data block identifier, change content, and timestamp are sent to the central node.
[0039] Receive the modified target data block and corresponding cloud verification tag issued by the central node and update it locally; respond to the sampling inspection challenge, generate inspection credentials based on the locally sampled data block and corresponding locally stored verification tag and return them to the central node.
[0040] Furthermore, when data inconsistency is detected, abnormal edge nodes are marked, and a secondary fine inspection based on single data block granularity is initiated.
[0041] Furthermore, each edge node does not store the temporary session key locally.
[0042] The significant advancement of this invention compared to existing technologies lies in:
[0043] 1. Improve the reliability of consistency detection in low-security environments; This invention no longer relies on security commitments between edge nodes, but introduces a central node to uniformly generate or control the generation of verification tags, thereby reducing the risk of false consistency after multiple edge nodes are compromised at the same time.
[0044] 2. Improve resistance to multi-source collaborative tampering attacks; Since the edge nodes do not store session keys locally, even if attackers control multiple edge nodes and tamper with data simultaneously, it is difficult to re-forge legitimate labels. Therefore, it can effectively make up for the defect that ordinary hash comparison is easily bypassed.
[0045] 3. Reduce inspection communication overhead; through random sampling challenge and aggregated credential return mechanism, it is not necessary to transmit all raw data for each inspection, which can reduce communication burden while ensuring the reliability of detection.
[0046] 4. Balancing consistency assessment and anomaly location: This invention first quickly assesses global consistency through sampling inspection, and then locates specific abnormal data blocks through secondary fine inspection at the single data block granularity when anomalies are detected, thereby balancing detection efficiency and location accuracy.
[0047] To more clearly illustrate the functional characteristics and structural parameters of the present invention, further explanation is provided below in conjunction with the accompanying drawings and specific embodiments. Attached Figure Description
[0048] Figure 1 This is a flowchart of the method provided in this application;
[0049] Figure 2 This application provides a diagram of an edge layer collaborative architecture based on centralized collaboration. Detailed Implementation
[0050] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0051] The method provided by this invention uses the central node as a trusted control center, and each edge node completes data synchronization and consistency inspection through the central node, thereby avoiding the security risks caused by direct collaboration between edge nodes.
[0052] In one embodiment, the method includes the following steps:
[0053] S1. Construct a centralized collaborative structure;
[0054] A collaborative structure is constructed, consisting of a central node and multiple edge nodes in the edge layer, wherein each edge node communicates directly with the central node. Preferably, the edge nodes do not directly communicate with each other for data synchronization or consistency detection; all collaborative data flows and detection control flows are forwarded, scheduled, and managed through the central node, thus forming a star network collaborative structure.
[0055] By centralizing collaborative control at the central node, the risk of edge nodes colluding to bypass detection after being compromised can be reduced.
[0056] S2, Task initialization;
[0057] The collaboration can be initiated by any edge node. The edge node registers the collaboration service with the central node and configures initial data. The central node generates a collaboration task ID corresponding to this collaboration task. task The node identifier ID of the edge node participating in this collaborative task i Initial collaborative data is then sent to the registrant and other participants.
[0058] At the same time, the central node generates a temporary session key for this task. session =( Temporary session keys are generated or protected based on a public-private key hierarchy, where sk and pk represent the private and public keys, respectively. The public-private key hierarchy is used for key generation, distribution protection, and authentication.
[0059] In one implementation, the central node divides the initial data file F into blocks {m1, m2, ..., m}. n} and calculate the corresponding cloud verification tag for each data block;
[0060] The calculation process is as follows:
[0061] S i = ;
[0062] ;
[0063] =ID i ID task TimeStamp;
[0064] Each data block's corresponding cloud verification tag is associated with at least the following information: the content of the corresponding data block; and the collaborative task ID.task Node Identifier ID i Timestamp; Temporary session key session .
[0065] The cloud-based verification tag is not a simple hash value, but a controlled verification tag calculated by the central node based on the session key and task context information. Each participating party obtains a task ID for this collaborative task after registration. task Node Identifier ID i The initial data file F and the corresponding label set S i .
[0066] Since each edge node does not store the session key n locally, even if the edge node is compromised, it is difficult for an attacker to independently forge a legitimate verification label that is consistent with the central node locally.
[0067] S3, Data Collaboration;
[0068] When any edge node performs a dynamic data operation, such as a synchronization event including revision, addition, deletion, or replacement, the edge node sends the data operation Act, the target data block identifier id, and the changed content m to the central node. id And the timestamp.
[0069] After receiving the synchronization event, the central node uses the session key. session =( Recalculate the corresponding cloud verification tag S for the changed target data block. i The central node then broadcasts the modified target data block and the newly calculated corresponding labels to other participants. Upon receiving the change event broadcast by the central node, each participant completes its local data update and local label update.
[0070] Through this mechanism, all data changes are confirmed by the central node and the corresponding labels are recalculated, thereby preventing edge nodes from privately synchronizing and tampering with the data.
[0071] S4. Consistency Inspection;
[0072] The central node performs consistency checks, sending consistency check requests to all participants during the check. To avoid the high communication burden and computational overhead caused by full data transmission, it is preferable to use random sampling to select the data blocks to be checked, rather than transmitting and comparing all data blocks in every round of checks.
[0073] In one implementation, the central node generates a challenge random number set: Challenge = {j, v} j} j∈JWhere j represents the random offset in the data block index set, v j This represents the corresponding challenge random number; the central node sends the challenge set to each participant.
[0074] S5. Generate the test certificate;
[0075] After receiving the challenge set, each participant extracts the data block to be inspected from its local machine and determines the data block based on the challenge random number v. j The corresponding data blocks and local storage tags are aggregated and processed to generate a detection certificate, which is then sent to the central node.
[0076] The aggregated data block and the aggregated data generated voucher are as follows: ;
[0077] The aggregated data block is obtained by a verifiable aggregation method of multiple sampled data blocks according to the challenge random number; the aggregated verification label is obtained by the corresponding aggregation calculation of the label of the sampled data block; the edge node generates an aggregated detection certificate that can be verified by the central node based on the sampled data block and the corresponding label according to the challenge set.
[0078] S6. Document verification and consistency determination;
[0079] The central node verifies all received verification credentials. The verification process includes the following two aspects:
[0080] S61, Label validity verification: The following methods shall be used for verification:
[0081] ,v);
[0082] The central node first verifies the validity of the aggregated verification labels returned by each edge node. If the verification passes, it means that the corresponding label matches the data block, and the label source conforms to the constraints generated by or controlled by the central node.
[0083] S62, Aggregated data consistency comparison: Based on the successful verification of label validity, the central node compares whether the aggregated data returned by each edge node is consistent; if the aggregated data returned by multiple edge nodes is consistent, it indicates that the data status on the corresponding sampled data block set is consistent.
[0084] The consistency inspection is considered successful only if all aggregate verification tags returned by all participants are verified and all aggregate data are completely consistent; otherwise, it is determined that there is data inconsistency or abnormal nodes.
[0085] The key to the above mechanism is that even if an attacker controls multiple edge nodes at the same time, it is difficult to regenerate a legitimate aggregate verification tag for the tampered data without having the session key. Therefore, compared with the simple hash comparison method, the present invention has higher adaptability to low-security environments.
[0086] S7. Anomaly location and secondary fine inspection;
[0087] When the inspection results show that there is data inconsistency, the central node first marks the abnormal edge nodes based on the verification results of the detection certificate and identifies the set of conflicting nodes.
[0088] Furthermore, to pinpoint the specific tampered data block, the central node triggers a secondary fine-grained inspection based on the granularity of a single data block. This secondary fine-grained inspection issues more granular challenges to the anomalous edge nodes, either block by block or in batches, to verify the validity of the label and the consistency of the data content for each individual data block, thereby locating the position of the anomalous data block.
[0089] Through the above-mentioned hierarchical inspection strategy, the present invention can simultaneously ensure global inspection efficiency and anomaly location accuracy while keeping communication burden under control.
[0090] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A method for edge environment consistency inspection based on a trusted coordination center, characterized in that, The method includes: S1. Construct a centralized collaborative structure: A collaborative structure is constructed, which includes a central node and multiple edge nodes in the edge layer; wherein each edge node communicates directly with the central node, but the edge nodes do not directly synchronize collaborative data with each other; S2, Task Initialization: The central node generates a task identifier and corresponding node identifiers for each edge node for the collaborative task, and generates a temporary session key for the collaborative task; the initial collaborative data is divided into multiple data blocks, and a cloud verification tag is calculated for each data block based on the temporary session key; the collaborative task number, node identifier, initial collaborative data and corresponding cloud verification tag are sent to each edge node. S3, Data Collaboration: When any edge node occurs data dynamic operation, the corresponding edge node sends data operation Act, target data block identification id, change content m and time stamp TimeStamp to the center node id The center node re-computes the corresponding cloud check label according to the changed target data block, and synchronizes the changed target data block and the re-computed cloud check label to other edge nodes, so as to realize the indirect data cooperation of each edge node via the center node. S4. Consistency Inspection: The central node initiates a consistency inspection and sends a sampling inspection challenge to each edge node. The sampling inspection challenge includes at least a set of indexes of data blocks to be inspected and a challenge random number corresponding to each data block to be inspected. S5. Generate test certificate: Each edge node responds to the sampling and detection challenge by extracting the sampled data block from its local storage and calculating a detection certificate based on the extracted sampled data block and the corresponding local storage verification label. The detection certificate includes at least aggregated data and aggregated verification label. S6. Document Verification and Consistency Determination: The central node receives the detection credentials returned by each edge node and verifies the detection credentials. The verification includes at least: verifying the validity of the aggregated verification tags returned by each edge node, and comparing whether the aggregated data returned by each edge node is consistent. Only when the aggregated verification tags returned by each edge node are all verified and the aggregated data returned by each edge node is consistent is it determined that the data status between each edge node is consistent; otherwise, it is determined that there is data inconsistency.
2. The method of claim 1, wherein, The cloud verification tag is generated by the central node based on the temporary session key. Each edge node does not store the temporary session key locally, so that each edge node cannot independently generate a valid cloud verification tag.
3. The method of claim 1, wherein, The center node calculates a cloud check tag for each data block, and the cloud check tag is associated with at least the content of the corresponding data block, a task number ID, a node ID, and a time stamp TimeStamp and a temporary session key Key. task i session 4. The method of claim 1, wherein, The edge nodes do not communicate directly with each other for data synchronization or consistency detection. All collaborative data streams and inspection control streams are forwarded or scheduled through the central node.
5. The method of claim 1, wherein, The set of data blocks to be tested specified in the sampling inspection challenge is determined by random sampling. The challenge random number is used to drive each edge node to perform aggregation calculation on the sampled data blocks and their corresponding verification labels.
6. The method of claim 1, wherein, The detection credentials include aggregated data calculated by each edge node after extracting the sampled data blocks from its local storage, and aggregated verification tags calculated based on the local storage verification tags corresponding to the sampled data blocks.
7. The method of claim 6, wherein, The central node verifies whether the aggregated verification labels returned by each edge node match the verification results calculated based on the corresponding sampling detection challenge and cloud-side verification information, and compares whether the aggregated data returned by each edge node is consistent; the consistency test is deemed to have passed only when all aggregated verification labels have passed verification and all aggregated data are completely consistent.
8. The method of claim 1, wherein, When data inconsistency is detected, the central node marks the abnormal edge nodes and triggers a secondary fine inspection based on single data block granularity to locate the tampered data block.
9. The method of claim 7, wherein, The secondary fine inspection includes: the central node sending fine inspection challenges to the abnormal edge nodes block by block, verifying the validity of the verification tag and the consistency of the data content of the corresponding individual data blocks, so as to determine the location of the abnormal data blocks.
10. An edge environment consistency inspection system based on a trusted collaboration center, the system being used to implement any one of the methods of claims 1 to 9, characterized in that the system... include: The central node and multiple edge nodes; The central node and each edge node form a star-shaped network collaborative structure. The central node is used for: For each collaborative task, a task identifier and a corresponding node identifier for each edge node are generated, and a temporary session key for the collaborative task is generated. The initial collaborative data is divided into multiple data blocks, and a cloud verification tag is calculated for each data block based on the temporary session key. Send the task identifier, the node identifier, the initial collaboration data, and the corresponding cloud verification tag to each edge node; In response to data change requests sent by edge nodes, recalculate the corresponding cloud verification tag based on the changed target data block, and synchronize the changed target data block and the recalculated cloud verification tag to other edge nodes; initiate consistency inspection, send sampling inspection challenges to each edge node; receive and verify the inspection credentials returned by each edge node; The edge nodes are used for: Receive the task identifier, node identifier, initial collaboration data and corresponding cloud verification tag sent by the central node and store them locally; When a dynamic data operation occurs, the data operation type, target data block identifier, change content, and timestamp are sent to the central node. Receive the modified target data block and corresponding cloud verification tag issued by the central node and update it locally; respond to the sampling inspection challenge, generate inspection credentials based on the locally sampled data block and corresponding locally stored verification tag and return them to the central node.
