Blockchain and privacy computing collaborative verification system

Abstract

The application relates to the technical field of blockchains, and discloses a blockchain and privacy calculation cooperative verification system, which comprises a data sharding module, a privacy fusion module, a cooperative verification module, a consensus correction module and a sharding graph optimization module. The data sharding module constructs a multi-level data cooperative topology through a preprocessing node and a dynamic sharding protocol; the privacy fusion module realizes privacy protection by utilizing a ciphertext analysis unit polymorphic switching link and a zero-knowledge verification unit; the cooperative verification module generates a benchmark verification strategy by a consensus modeling node multidimensional fusion model, and a state optimization unit dynamically aggregates and sorts sharding states; the consensus correction module corrects sharding states and calibrates ciphertext parameters by an isolation verification unit and a delay compensation unit; and the sharding graph optimization module constructs a stability evaluation network to monitor a synchronization rate and closed-loop adjusts a verification strategy. The system realizes dynamic sharding, privacy fusion and intelligent verification, and improves the efficiency, security and stability of blockchain data processing.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of blockchain technology, in particular to a blockchain and privacy computing collaborative verification system. BACKGROUND

[0002] In the application and development of blockchain technology, with the explosive growth of data size and the increasing complexity of cross-chain interaction scenarios, traditional blockchain systems face many challenges in data processing efficiency, privacy security, and system stability.

[0003] In terms of privacy security, although the open and transparent nature of blockchain ensures data integrity, it also poses a risk of privacy leakage. In blockchain applications involving sensitive data (such as user identity information, business secrets, etc.), traditional encryption techniques are difficult to effectively protect privacy during data processing. For example, in the process of data fusion and verification, traditional methods may need to expose sensitive data in plaintext to multiple nodes for processing, which provides the possibility of privacy leakage. In addition, with the development of privacy computing technology, how to deeply integrate it with blockchain technology to achieve "data available but invisible" is a problem that needs to be solved urgently.

[0004] In terms of system stability, traditional blockchain systems lack effective dynamic adjustment and self-repair mechanisms when facing network fluctuations, node failures, or external attacks. For example, when a certain shard node fails or is overloaded, traditional systems are difficult to quickly transfer data to other nodes, resulting in a decline in the performance of the entire system and even service interruption. At the same time, in the process of cross-chain interaction, the consensus mechanism, data format, and update frequency of different blockchains differ, which can easily cause data inconsistency problems, affecting the stability and reliability of the system.

[0005] In the prior art, although there have been some research on blockchain sharding and privacy computing, most lack a complete solution that organically combines the two and achieves collaborative verification. For example, some sharding technologies only consider the efficiency of data storage and processing, ignoring privacy protection; while some privacy computing solutions fail to fully integrate the distributed nature of blockchain, making it difficult to achieve efficient collaborative verification in practical applications. In addition, existing systems lack flexible strategy adjustment mechanisms when facing dynamic network environments and business demands, and cannot achieve dynamic optimization of data sharding status and verification strategies.

[0006] Therefore, there is an urgent need for a system that combines blockchain and privacy computing to achieve efficient data sharding, privacy protection, and collaborative verification, to address the shortcomings of traditional blockchain systems in data processing efficiency, privacy security, and system stability, and to meet the needs of increasingly complex blockchain application scenarios. SUMMARY

[0007] The present application aims to provide a blockchain and privacy computing collaborative verification system to solve the problems raised in the background.

[0008] To achieve the above-mentioned purpose, the present application provides the following technical solution: a blockchain and privacy computing collaborative verification system, the system comprising:

[0009] a data sharding module, a privacy fusion module and a collaborative verification module;

[0010] The data sharding module comprises a preprocessing node and a dynamic sharding protocol; the privacy fusion module comprises a ciphertext analysis unit and a zero-knowledge verification unit; the collaborative verification module comprises a consensus modeling node and a state optimization unit; the preprocessing node is used to collect real-time raw data hash values and cross-chain interaction parameters; according to the block height and timestamp sequence, a multi-level data collaborative topology is constructed through the dynamic sharding protocol; the dynamic sharding protocol is composed of multiple sharding nodes based on hash threshold in parallel; based on the real-time response characteristics of multiple sharding nodes, a benchmark verification strategy is generated through the consensus modeling node, and combined with the verification signal output by the zero-knowledge verification unit, the data sharding state is dynamically aggregated and sorted through the state optimization unit.

[0011] Preferably, the system further comprises a consensus correction module and a sharding graph optimization module; the consensus correction module comprises an isolated verification unit and a delay compensation unit;

[0012] According to the generated benchmark verification strategy and dynamic aggregation and sorting results, the data sharding state is corrected by the isolated verification unit in a distributed barrier, and the ciphertext parameters of the privacy fusion module are dynamically calibrated by the delay compensation unit; at the same time, the sharding synchronization rate is monitored in real time by the on-chain stability evaluation network constructed by the sharding graph optimization module, and the monitoring results are fed back to the consensus modeling node and the state optimization unit for closed-loop adjustment of the verification strategy.

[0013] Preferably, the isolated verification unit is composed of state control circuits corresponding to sharding nodes in parallel in the dynamic sharding protocol; the construction steps of the multi-level data collaborative topology comprise:

[0014] Collecting raw data hash values and cross-chain fluctuation signals and inputting them into a transfer learning model configured in each sharding node to generate data consistency compensation coefficients and store them in the corresponding sharding node;

[0015] Based on the consistency compensation coefficients stored by each level of sharding node, combined with the instantaneous change rate of the timestamp sequence, the data sharding state is distributed to multiple sharding nodes through a dynamic hash detection algorithm to form a dynamic collaborative matching topology.

[0016] Preferably, the pre-processing node is configured to build a slice node construction condition, which includes that the original data hash value is in a preset interval, the synchronization index corresponding to the slice node is greater than a threshold value, and the block height is less than an allowed deviation range.

[0017] The ciphertext analysis unit is also configured with a polymorphic switching link; the polymorphic switching link includes a steady-state slice sub-link and a verification avoidance sub-link; the steady-state slice sub-link is generated based on the matching degree of the on-chain stability signal of the privacy fusion module and the benchmark verification strategy; and the verification avoidance sub-link is constructed based on the association between the instantaneous verification risk of the data slice state and external interference events.

[0018] Preferably, the consensus modeling node is configured with a multi-dimensional fusion model; the multi-dimensional fusion model includes a hash-time mapping table corresponding to the data slice module and a verification avoidance parameter library of the privacy fusion module; and the step of generating the benchmark verification strategy includes:

[0019] According to the real-time hash data and the aggregated sorting result of the data slice module, the upper limit of the theoretical synchronization range of each slice node in the steady-state slice sub-link is calculated;

[0020] The upper limit of the theoretical synchronization range and the transient timestamp are input into the multi-dimensional fusion model to generate an on-chain verification reference graph, and the reference graph is verified for risk correction through the verification avoidance sub-link to form the benchmark verification strategy.

[0021] Preferably, the parameter updating step of the multi-dimensional fusion model includes:

[0022] When it is detected that the original data hash value is out of the preset interval or the slice synchronization rate exceeds the threshold value, a first verification avoidance instruction is triggered, that is, the alternative slice unit is started and the analysis granularity of the privacy fusion module is adjusted;

[0023] If the slice state has not recovered to the allowed interval after execution of the first verification avoidance instruction, a second verification avoidance instruction is triggered, that is, the verification avoidance sub-link is switched to through the polymorphic switching link, and the topology weight of the data slice state is redistributed based on the duration of the external interference event.

[0024] Preferably, the parameter updating step of the multi-dimensional fusion model further includes:

[0025] When the synchronization index of any slice node in the data slice module is lower than the threshold value, a third verification avoidance instruction is triggered, that is, the slice output channel of the node is blocked, and the corresponding data is transferred to other slice nodes through the isolation verification unit;

[0026] During the execution of the third verification circumvention instruction, if other shard nodes are detected to be overloaded, the fourth verification circumvention instruction is triggered, that is, the shard atlas optimization module is called to degrade the output strategy of the state optimization unit, and the synchronization requirement of the edge state is limited.

[0027] Preferably, the degradation processing logic of the shard atlas optimization module comprises:

[0028] According to the synchronization rate data output by the on-chain stability evaluation network and the off-chain ciphertext indicators, a shard state security level table is constructed.

[0029] When the fourth verification circumvention instruction is triggered, the shard path priority is dynamically degraded according to the security level table, and the analysis frequency of the privacy fusion module is synchronously adjusted to match the degraded verification strategy.

[0030] Preferably, the construction step of the shard node further comprises:

[0031] Based on the spatial distribution result of the original data hash value and the historical cross-chain fluctuation range, a hash gradient adaptability index table is generated;

[0032] According to the consistency compensation coefficient difference of each shard node in the index table, the topology weight distribution of the state regulation circuit of the plurality of shard nodes is distributed through a dynamic weight distribution algorithm, and a redundant transmission link between the shard nodes is established.

[0033] Preferably, the generation process of the on-chain verification reference atlas comprises:

[0034] After inputting the upper limit of the theoretical synchronization range and the transient timestamp into the multi-dimensional fusion model, the preset shard medium database and the external interference event layer are synchronously loaded;

[0035] Based on the encrypted peak frequency data output by the ciphertext analysis unit, the verification threshold of the data shard state in the reference atlas is corrected, and a time gradient parameter is superimposed in the verification circumvention sub-link for secondary calibration.

[0036] Compared with the prior art, the present application has the following advantages:

[0037] In terms of data processing efficiency, the data sharding module collects raw data hash values and cross-chain interaction parameters in real time through preprocessing nodes, and constructs a multi-level data collaborative topology using a dynamic sharding protocol. Based on the real-time response characteristics of multiple sharding nodes, this topology can dynamically adjust the sharding strategy according to data traffic and cross-chain interaction requirements, achieving efficient distribution and parallel processing of data. For example, by using a dynamic hash detection algorithm, data sharding states are assigned to multiple sharding nodes to form a dynamic collaborative matching topology, effectively improving the throughput and real-time performance of data processing, and reducing the delay of cross-chain data synchronization. At the same time, the consensus modeling node generates a benchmark verification strategy, and combines with the state optimization unit to dynamically aggregate and sort the data sharding states, further optimizing the data processing flow and ensuring efficient collaborative processing of data among different sharding nodes.

[0038] In terms of privacy protection, the ciphertext analysis unit of the privacy fusion module is configured with a multi-state switching link, including a steady-state sharding sub-link and a verification avoidance sub-link. The steady-state sharding sub-link is generated based on the matching degree of the on-chain stability signal of the privacy fusion module and the benchmark verification strategy, and can realize stable analysis and processing of ciphertext data under normal circumstances; the verification avoidance sub-link is constructed based on the correlation between the instantaneous verification risk of the data sharding state and external interference events, and can switch to this sub-link in time when a verification risk is detected, adjust the analysis strategy, and avoid exposure of sensitive data in plaintext, effectively protecting data privacy. In addition, the application of the zero-knowledge verification unit enables the verification process to be completed without revealing the specific content of the data, further enhancing the privacy protection capability.

[0039] In terms of system stability and reliability, the isolation verification unit of the consensus correction module is composed of state control circuits corresponding to the sharding nodes in parallel in the dynamic sharding protocol, which can perform distributed barrier correction on the data sharding state and handle node failures or abnormal situations in time. The delay compensation unit dynamically calibrates the ciphertext parameters of the privacy fusion module to ensure the accuracy and consistency of the ciphertext parameters and reduce system instability problems caused by parameter deviation. The on-chain stability evaluation network constructed by the sharding graph optimization module can monitor the sharding synchronization rate in real time and feed back the monitoring results to the consensus modeling node and the state optimization unit, realizing closed-loop adjustment of the verification strategy. The system can dynamically optimize the strategy according to real-time monitoring data to adapt to changes in network environment and business requirements. When a node is overloaded or fails, the system can trigger appropriate verification avoidance instructions, such as blocking the output channel of the failed node, transferring data to other nodes, and calling the sharding graph optimization module for degradation processing, to ensure the continuous and stable operation of the system and improve the fault tolerance and reliability of the system.

[0040] The multi-dimensional fusion model combines the hash-time mapping table of the data sharding module and the verification avoidance parameter library of the privacy fusion module, can generate an on-chain verification reference map according to real-time data, and correct the verification risk through the verification avoidance sub-chain link to form an accurate benchmark verification strategy. At the same time, the parameter updating mechanism of the model can trigger corresponding verification avoidance instructions according to different abnormal situations, dynamically adjust the sharding strategy and privacy protection parameters, so that the system can quickly adapt to changes in the hash value of the original data, abnormal sharding synchronization rate and the like, realize dynamic optimization and self-adaptive adjustment of the system strategy, and improve the running efficiency and stability of the system in complex environments. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 The working principle diagram of the blockchain and privacy computing collaborative verification system described in the application;

[0042] Figure 2 The design diagram of the data sharding module structure;

[0043] Figure 3 The design diagram of the privacy fusion module composition;

[0044] Figure 4 The working flowchart of the collaborative verification module;

[0045] Figure 5 The design diagram of the consensus correction module operation mechanism;

[0046] Figure 6 The design diagram of the multi-level data collaborative topology construction. DETAILED DESCRIPTION

[0047] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0048] Please refer to Figures 1-6 The application relates to a blockchain and privacy computing collaborative verification system, which comprises:

[0049] a data sharding module, a privacy fusion module and a collaborative verification module. The data sharding module is composed of a preprocessing node and a dynamic sharding protocol; the privacy fusion module comprises a ciphertext analysis unit and a zero-knowledge verification unit; the collaborative verification module comprises a consensus modeling node and a state optimization unit. Each module realizes the collaborative verification function of the system in the following way:

[0050] The preprocessing node collects original data hash values and cross-chain interaction parameters in real time, and constructs a multi-level data collaborative topology based on block height and timestamp sequence through a dynamic sharding protocol. The dynamic sharding protocol is composed of multiple sharding nodes based on hash threshold, and each sharding node performs data sharding tasks according to the interval division of the original data hash value. Specifically, the preprocessing node first performs hash calculation on the original data to generate a unique hash value, and collects parameters such as cross-chain transaction amount and transaction timestamp during the cross-chain interaction process. Subsequently, according to the continuity of the block height (i.e. the length of the current blockchain) and the timestamp sequence of the current blockchain, the dynamic sharding protocol is started, and the parallel sharding nodes are divided into different levels according to the hash threshold, for example, the nodes with the same first 16 bits of the hash value are divided into the same level, forming the basic framework of the multi-level data collaborative topology.

[0051] Based on the real-time response characteristics of multiple sharding nodes, the consensus modeling node generates a benchmark verification strategy. The specific process is as follows: after receiving the data sharding task, each sharding node feeds back its processing status (such as processing speed, resource occupancy, etc.) in real time, and the consensus modeling node collects these response data and analyzes the reliability and efficiency of each node through a preset algorithm (such as a neural network algorithm or a fuzzy logic algorithm) to generate a benchmark verification strategy for the current data sharding task, such as determining the verification order and verification weight of each node.

[0052] At the same time, the zero-knowledge verification unit verifies the data sharding process and outputs a verification signal. The zero-knowledge verification unit uses zero-knowledge proof technology to verify the correctness and integrity of the data sharding without revealing the specific content of the data. For example, by constructing a verification circuit, it verifies whether the data sharding conforms to the preset logical rules, such as whether the range of the hash value is correct and whether the sharded data can be completely restored, etc. The state optimization unit combines the benchmark verification strategy and the verification signal to dynamically aggregate and sort the data sharding state. The state optimization unit adjusts the data sharding state of each sharding node in real time according to the node weight and verification order set in the benchmark verification strategy and the verification result (such as pass or fail) output by the zero-knowledge verification unit. For example, nodes that pass the verification and have high processing efficiency are given higher aggregation priority, and nodes that fail the verification or have low processing efficiency are re-sharded or marked as abnormal state, ensuring that the data sharding state is always in the optimal configuration.

[0053] Embodiment 1:

[0054] The consensus correction module and the sharding graph optimization module set by the system work cooperatively, wherein the consensus correction module includes an isolated verification unit and a delay compensation unit, and the specific implementation is as follows:

[0055] After generating the benchmark verification strategy and completing the dynamic aggregation and sorting of the data shard state, the isolated verification unit starts the distributed barrier correction mechanism. The isolated verification unit is composed of state control circuits corresponding to each shard node in the dynamic sharding protocol. Each shard node's state control circuit is independently configured and has real-time monitoring and independent correction functions. When the data shard state of a shard node is abnormal (such as the original data hash value exceeding the preset interval, the synchronization index being lower than the threshold, or the block height exceeding the allowed deviation range), the state control circuit of the node performs correction through the following steps: first, isolate the current data shard task of the abnormal node and suspend its data interaction channel with other nodes to prevent the spread of abnormal state; second, start the local verification process, recalculate the original data hash value of the node, check the integrity of cross-chain interaction parameters (such as cross-chain transaction amount and timestamp matching blockchain records), and compare the consistency of block height and timestamp sequence; if a logical error is found in the data shard (such as a hash value calculation error causing incomplete shards), trigger the local data reconstruction program to generate correct shard data based on the original data hash value and cross-chain interaction parameters; if the verification result shows that the node's hardware resources are insufficient (such as CPU utilization continuously exceeding 90% causing processing delays), send a resource scheduling request to the system management module through the state control circuit to dynamically allocate additional computing resources or migrate part of the task to other idle nodes. During the entire distributed barrier correction process, the state control circuits of each shard node work in parallel without interfering with each other, ensuring the efficiency and independence of abnormal correction.

[0056] The delay compensation unit dynamically calibrates the ciphertext parameters of the privacy fusion module. When the privacy fusion module processes ciphertext data, the ciphertext parameters (such as the key version of the encryption algorithm, the ciphertext length of the data shard, and the circuit parameters of the zero-knowledge verification) may deviate due to network transmission delay, multi-node consensus delay, or cross-chain interaction delay. The delay compensation unit achieves calibration through the following mechanisms: First, a timestamp monitoring point is deployed on the ciphertext parameter transmission path to collect the transmission time from the ciphertext analysis unit of the privacy fusion module to the consensus modeling node of the collaborative verification module in real time, and calculate the difference between the actual transmission time and the preset time threshold (i.e., the delay difference). Then, according to the type of delay difference (such as network delay or computation delay), the corresponding compensation strategy is called - if it is network delay, the output buffer queue length of the ciphertext analysis unit is adjusted to increase the buffer time of the parameters at the sending end to balance network fluctuations; if it is computation delay, the verification circuit complexity of the zero-knowledge verification unit is dynamically adjusted, such as simplifying part of the non-critical verification steps to reduce the computation time. In addition, the delay compensation unit also needs to monitor the special delay in the cross-chain interaction scenario. When detecting the ciphertext parameter delay caused by cross-chain transactions, the block confirmation time of cross-chain transactions is analyzed to predict the delay trend of subsequent ciphertext parameters and adjust the analysis rhythm of the privacy fusion module in advance. For example, during the peak period of cross-chain transactions, the number of parallel processing threads of the ciphertext analysis unit is automatically reduced to avoid thread competition and delay.

[0057] The sharding atlas optimization module constructs an on-chain stability evaluation network to achieve real-time monitoring of the synchronization rate of shards through multi-dimensional data collection and analysis. The architecture of the on-chain stability evaluation network includes a data collection layer, a feature extraction layer, and a state evaluation layer. The data collection layer real-time captures the synchronization logs of each shard node, including the synchronization start time, synchronization end time, successful synchronization data volume, and synchronization failure times, etc. The feature extraction layer pre-processes the original data and calculates key feature indicators, such as synchronization frequency per unit time, synchronization success rate (successful synchronization times / total synchronization times), and average synchronization time consumption (total synchronization time consumption / synchronization times). The state evaluation layer dynamically analyzes the feature indicators based on a preset evaluation model (such as a sliding window model or an exponential smoothing model) to generate a synchronization rate stability score for each shard node. For example, the sliding window is used to calculate the synchronization frequency fluctuation amplitude in the past 10 minutes. If the fluctuation amplitude exceeds the preset threshold, it is determined that the synchronization rate of the node is unstable.

[0058] The feedback of the monitoring results and the closed-loop adjustment mechanism are as follows: the on-chain stability evaluation network transmits the synchronization rate stability score, characteristic indicators and abnormal events (such as a surge in synchronization failures) of each shard node to the consensus modeling node and the state optimization unit in real time. After receiving the feedback data, the consensus modeling node first updates the parameter library of the benchmark verification strategy. For example, if the synchronization success rate of a shard node is continuously lower than 80%, the verification weight of the node in the benchmark verification strategy is reduced, and the proportion of verification tasks undertaken by the node is reduced. If it is found that the synchronization frequency of multiple nodes presents periodic fluctuations in a certain time period (such as a significant decrease in synchronization frequency at 9 am every day), the time scheduling rules of the benchmark verification strategy are adjusted, and more tasks are automatically allocated to the nodes with stable synchronization in the time period. The state optimization unit adjusts the dynamic aggregation and sorting of data shard states in real time according to the synchronization rate data fed back: for the nodes with stable synchronization rate, the priority of the nodes in the aggregation and sorting is improved, and the data processed by the nodes is ensured to enter the subsequent verification process in priority; for the nodes with large synchronization rate fluctuations, the data shard states of the nodes are marked as “to be observed”, and the waiting time of the nodes in the aggregation queue is extended until the synchronization rate is restored to stability. In addition, the closed-loop adjustment process also involves cross-module cooperation: after the consensus modeling node adjusts the benchmark verification strategy, the strategy update signal is sent to the zero-knowledge verification unit of the privacy fusion module, and the zero-knowledge verification unit adjusts the resource allocation of the verification circuit according to the new verification weight and order, for example, more verification computing power is allocated to the data shards of the nodes with high weight, to ensure that the verification efficiency matches the strategy.

[0059] In the implementation process, the isolation verification unit, the delay compensation unit and the shard graph optimization module form a linkage mechanism. For example, when the isolation verification unit completes the correction of an abnormal node and restores the normal state of the node, a node state reset signal is sent to the on-chain stability evaluation network, the on-chain stability evaluation network re-collects the synchronization data of the node and updates the stability score of the node; if the delay compensation unit finds that the delay of the ciphertext parameter exceeds the compensable range of the system in the dynamic calibration process, the emergency monitoring process of the shard graph optimization module is triggered, and the monitoring frequency of the synchronization rate of the related node is temporarily increased, so as to timely discover potential systemic risks. Through the above-mentioned mechanisms, the whole system realizes dynamic correction of the data shard state, delay compensation of the ciphertext parameter and closed-loop optimization of the shard synchronization rate, and ensures the stability, accuracy and efficiency of the collaborative verification process of the blockchain and the privacy calculation.

[0060] Embodiment 2:

[0061] The construction process of the multi-level data collaborative topology involves multiple key steps and technical implementations, and the specific implementation is as follows:

[0062] The original data hash values and cross-chain fluctuation signals are collected. The original data hash values are generated by hashing various input data such as transaction records, asset information, and user behavior data, ensuring data uniqueness and tamper resistance. The cross-chain fluctuation signals are collected for cross-chain interaction activities, including cross-chain transaction frequency, amount fluctuation range, and transaction time interval. These parameters are collected through the deployment of monitoring probes on each blockchain node, which captures relevant cross-chain transaction data in real time and transmits it to the preprocessing node for unified processing.

[0063] The collected data is input into the transfer learning model configured on each sharding node. This transfer learning model is pre-trained based on historical data and has the ability to predict data consistency. The training process of the model uses a large amount of historical data, including original data hash values, corresponding cross-chain fluctuation signals, and final data consistency results. Through learning from these historical data, the model can identify potential correlation patterns between data hash values and cross-chain fluctuation signals, and generate data consistency compensation coefficients accordingly. For example, when the model detects that a certain hash value distribution is associated with a specific cross-chain transaction frequency pattern, it will generate the corresponding compensation coefficient based on historical experience to adjust the subsequent data sharding processing strategy.

[0064] The generated compensation coefficients are stored in the corresponding sharding nodes. Each sharding node maintains a local compensation coefficient table that is continuously updated based on real-time input data. The storage of compensation coefficients uses a distributed storage mechanism to ensure data security and accessibility. At the same time, to improve the response speed of the system, the compensation coefficient table is periodically synchronized between sharding nodes to ensure data consistency across all nodes.

[0065] Based on the consistency compensation coefficients stored in each level of sharding node, combined with the instantaneous change rate of the timestamp sequence, the data sharding state is allocated to multiple sharding nodes through a dynamic hash detection algorithm. The instantaneous change rate of the timestamp sequence reflects the real-time speed of data generation and transmission, which is an important indicator of system dynamic changes. The dynamic hash detection algorithm analyzes the change rate and consistency compensation coefficients to adjust the allocation strategy of data sharding in real time.

[0066] The specific workflow of the algorithm is as follows: First, analyze the current timestamp sequence and calculate its instantaneous change rate. The change rate calculation is based on a time window sliding mechanism, which compares the timestamp differences in different time windows to determine the speed change of data generation and transmission. Then, combined with the consistency compensation coefficients stored in each level of sharding node, evaluate the processing capacity of each sharding node for the current data sharding. Nodes with high consistency compensation coefficients are considered to have stronger data processing stability and accuracy, and therefore have higher priority in allocation.

[0067] According to the evaluation results, the dynamic hash detection algorithm assigns data shard states to multiple shard nodes, forming a dynamic collaborative matching topology. During the assignment process, the algorithm considers multiple factors, such as the load situation, processing capacity, network delay, etc. of the nodes, to ensure the balanced distribution and efficient processing of data shards. For example, nodes with lighter loads and stronger processing capabilities will be assigned more data shard tasks; while nodes with higher network delays will have fewer data shard tasks with higher real-time requirements.

[0068] The formed dynamic collaborative matching topology has adaptability and scalability. When the load of the system changes or new nodes join, the topology structure can automatically adjust to adapt to the new environment. This dynamic adjustment mechanism achieves this by periodically collecting state information (such as load situation, processing speed, network connection quality, etc.) of each shard node and recalculating the data shard distribution strategy based on this information.

[0069] During implementation, real-time monitoring and adjustment of data shard states are also involved. The system continuously monitors the data processing state of each shard node, including data reception, processing, and transmission. If a node is found to have processing delays or errors, the system will immediately take measures to adjust, such as redistributing the data shard tasks of the node, increasing the processing load of other nodes, etc.

[0070] In addition, in order to ensure the security and privacy of data, various encryption technologies are used during data sharding and transmission. The original data is encrypted before sharding to generate ciphertext data. During data transmission, secure communication protocols are used to ensure data integrity and confidentiality. At the same time, when each shard node processes data, strict privacy protection rules are followed to ensure that user data is not leaked.

[0071] The construction of the multi-level data collaborative topology also considers fault tolerance and recovery mechanisms. When a shard node fails or becomes unavailable, the system can automatically identify and transfer the data shard tasks of the node to other normal nodes. At the same time, the system records the state information of the failed node for subsequent troubleshooting and repair.

[0072] The entire construction process of the multi-level data collaborative topology is a dynamic and continuously optimized process. By continuously collecting and analyzing system running data, adjusting the parameters of the migration learning model and the strategy of the dynamic hash detection algorithm, the system can continuously improve the efficiency and accuracy of data processing, adapt to changing business requirements and network environments.

[0073] In practical applications, the construction of multi-level data collaboration topology can effectively improve the performance and reliability of the blockchain and privacy computing collaborative verification system. By reasonably allocating data sharding tasks and fully utilizing the processing capacity of each node, the system can achieve efficient data processing and verification. At the same time, the dynamic adjustment mechanism and fault recovery mechanism ensure that the system can maintain stable operation in the face of various abnormal situations, providing reliable services for users.

[0074] Embodiment 3:

[0075] The shard nodes configured by the preprocessing node need to meet strict construction conditions to ensure the rationality of data sharding and the stability of the system. Specific conditions include: the original data hash value is within the preset interval, the synchronization index corresponding to the shard node is greater than the threshold, and the block height is less than the allowed deviation range. Among them, the original data hash value is generated by the hash function H(data), and the preset interval is pre-set by the system security policy and business requirements, for example, set to [H min , H max ]. Only when H min ≤ H(data) ≤ H max , the data can enter the corresponding shard node. The synchronization index is used to measure the data synchronization ability of the shard node with other nodes, which is calculated based on the timeliness and integrity of data transmission between nodes, and the formula is:

[0076]

[0077] Where S is the synchronization index, T sync is the total time of successful synchronization of nodes, T total is the total synchronization time (including retry time), C comp is the synchronization data integrity coefficient (value range is 0-1, 1 indicates complete integrity). Only when S > S thresh (S thresh is the preset threshold), the node is allowed to participate in sharding. The allowed deviation range of block height is determined by the consensus mechanism of the blockchain network, assuming that the current main chain block height is H main , the allowed deviation is ΔH, and the block height H node of the shard node needs to satisfy H node ≤ H main + ΔH, to avoid data inconsistency caused by node lag.

[0078] The ciphertext analysis unit is configured with a polymorphic switching link including a steady-state fragmentation sub-link and a verification avoidance sub-link, both of which are generated based on different mechanisms and work cooperatively. The generation of the steady-state fragmentation sub-link relies on the matching degree of the on-chain stability signal of the privacy fusion module and the reference verification strategy. The on-chain stability signal is generated by monitoring indicators such as the transaction throughput and node online rate of the blockchain network, and the stability indicator is denoted as σ. The matching degree of the reference verification strategy is calculated by cosine similarity:

[0079]

[0080] where A is the on-chain stability indicator vector, and B is the parameter vector of the reference verification strategy. When Sim≥Sim thresh (Sim thresh is the preset matching degree threshold), the steady-state fragmentation sub-link is triggered. This link uses fixed analysis rules and verification order, such as sequentially analyzing ciphertext data according to the hash value order of the fragmentation node, and preferentially verifying the data fragments of high-stability nodes.

[0081] The construction of the verification avoidance sub-link is based on the association between the instantaneous verification risk of the data fragment state and external interference events. The instantaneous verification risk is determined by analyzing indicators such as the hash value fluctuation and ciphertext length anomaly of the data fragment, and the risk indicator is denoted as R. External interference events are generated by real-time monitoring of network attack logs, node failure reports, and other event vectors E. The system uses association rule algorithms such as the Apriori algorithm to mine the association patterns between R and E, for example, when R exceeds the threshold R thresh and there is a "network latency surge" event in the event vector, it is determined that there is a verification risk, and the verification avoidance sub-link is triggered. This link uses dynamic adjustment strategies, including:

[0082] Verification rule enhancement: increasing the number of rounds of zero-knowledge verification, such as expanding the default 1-round verification to 3 rounds, and using different verification circuit parameters in each round to improve the verification strength of the correctness of the data fragment.

[0083] Encryption level promotion: performing secondary encryption on the ciphertext data, using more complex encryption algorithms (such as switching from AES-128 to AES-256), and dynamically generating temporary keys to reduce the risk of privacy leakage caused by external attacks.

[0084] Data fragment rerouting: temporarily transferring high-risk data fragments to a dedicated verification node group for processing, which deploys independent computing resources and network channels, and is isolated from the main link to avoid interference diffusion.

[0085] In the implementation process, the preprocessing node dynamically selects the qualified shard node by monitoring the original data hash value, synchronization index and block height in real time. For example, when the original data hash value of a certain node exceeds the preset interval, the node is immediately marked as "not activated" state, and the new data shard task is rejected until the hash value returns to normal. For the node with synchronization index below the threshold, the system automatically triggers the synchronization optimization process, including reconfiguring the network connection parameters of the node, adjusting the data caching strategy, etc. If the synchronization index still does not meet the standard after optimization, the node is removed from the shard node list.

[0086] The polymorphic switching link of the ciphertext analysis unit is realized by a state machine mechanism. The system maintains a link state register to record the currently activated link type (steady state or verification avoidance) and switching conditions. When the matching degree of the steady-state shard sub-link is below the threshold, the state machine triggers the switching logic: first, pause the current ciphertext analysis task, save the unfinished analysis state to the cache; then initialize the parameters of the verification avoidance sub-link (such as encryption algorithm, verification rule); finally, import the cached analysis state into the new link and continue processing the ciphertext data. The entire switching process is completed in nanoseconds, ensuring the continuity of data processing.

[0087] In addition, the system also sets a link state rollback mechanism. When the verification avoidance sub-link processing is completed or the external interference event is removed, the system automatically detects the matching degree of the stability signal on the link and the reference verification strategy. If the matching degree rises above the threshold, the rollback logic is triggered, switching back to the steady-state shard sub-link, and the data processed during the verification avoidance sub-link is verified again to ensure data consistency.

[0088] In cross-chain interaction scenarios, the preprocessing node needs to additionally handle the impact of cross-chain fluctuations on the shard node construction conditions. For example, when it is detected that cross-chain transactions cause the block height of a certain shard node to exceed the allowed deviation range, the system immediately starts the block synchronization mechanism to quickly restore the block height of the node to the allowed range by pulling the missing block data from adjacent nodes. At the same time, the synchronization index calculation method of the node is dynamically adjusted, temporarily reducing the synchronization time weight during cross-chain transactions to avoid synchronization index misjudgment due to cross-chain delay.

[0089] The ciphertext analysis unit adopts different analysis strategies according to the type of the polymorphic switching link when processing cross-chain ciphertext data. In the steady-state shard sub-link, the cross-chain ciphertext data is processed using the standardized analysis process, i.e. extracting transaction amount, timestamp and other information according to the preset field order and comparing with the blockchain record; in the verification avoidance sub-link, a source verification step is added for cross-chain ciphertext data, which verifies whether the data comes from a trusted blockchain network by analyzing the traceability hash of cross-chain transactions, preventing fake cross-chain data from entering the verification process.

[0090] Throughout the implementation process, the pre-processing node and the ciphertext analysis unit interact with real-time data and dynamically adjust the strategy to ensure the compliance of the sharding node and the flexibility of the ciphertext analysis. The system avoids the randomness of human intervention through quantified construction conditions and switching rules, realizes the automation and intelligence of data sharding and ciphertext processing, and provides a solid underlying support for the collaborative verification of blockchain and privacy computing.

[0091] Embodiment 4:

[0092] The consensus modeling node configuration multi-dimensional fusion model integrates the hash-time mapping table of the data sharding module and the verification avoidance parameter library of the privacy fusion module to generate a benchmark verification strategy. Taking a supply chain finance scenario as an example, when the system processes a blockchain transaction containing supplier accounts receivable data, the original data is hashed to generate a unique identifier, and the hash-time mapping table records the transaction initiation time, block confirmation time, and other timestamp information corresponding to each hash value. The verification avoidance parameter library stores processing rules for different risk scenarios, such as automatically triggering verification process adjustment strategies when detecting abnormal hash value fluctuations.

[0093] The generation of the benchmark verification strategy is based on real-time hash data and aggregated sorting results of the data sharding module. Assuming that the current system receives 100,000 accounts receivable data, the pre-processing node distributes the data to 10 parallel sharding nodes through a dynamic sharding protocol, with each node processing 10,000 data. Real-time hash data shows that the hash value distribution of node A is concentrated in the interval [0x1a2b3c, 0x1a2b4c], and the hash value distribution of node B is relatively dispersed. The aggregated sorting results list nodes A to E as high priority and nodes F to J as low priority based on processing efficiency (such as data processing volume per second) and verification pass rate.

[0094] The system calculates the upper limit of the theoretical synchronization range of each sharding node in the steady-state sharding sub-link based on the above information. Taking high-priority node A as an example, it has successfully processed 5,000 data in the past 5 minutes, with an average of about 17 per second. Considering system resource fluctuations, the theoretical synchronization range upper limit is set to 20 per second. Similarly, node B has a lower processing efficiency, and the theoretical synchronization range upper limit is set to 10 per second. The theoretical synchronization range upper limit of each node forms a dynamic threshold list for subsequent strategy generation.

[0095] The upper limit of the theoretical synchronization range is input into the multi-dimensional fusion model together with the transient timestamp. Assuming that the current transient timestamp is May 22, 2025, 14:30:00, the model analyzes the hash-time mapping table and finds that the transactions corresponding to the hash values processed by node A are mostly initiated during the morning period, while the current time is in the afternoon period. It is speculated that the data traffic may decrease, so when generating the on-chain verification reference graph, the synchronization pressure expectation of node A is appropriately reduced. At the same time, the rule recorded in the verification avoidance parameter library that "network delay is higher in the afternoon period" is triggered, and the model automatically adjusts the reference graph to allocate more buffer time windows to node A and node B, etc.

[0096] The generated on-chain verification reference graph presents the synchronization range, verification priority, and risk prompts of each shard node in a visual form. For example, the graph area of node A shows that its theoretical synchronization range is 0-20 transactions per second, and the current load is 15 transactions per second, which is in the green safe interval; the theoretical synchronization range of node B is 0-10 transactions per second, and the current load is 8 transactions per second, which is close to the yellow warning interval. The reference graph also marks potential risk points, such as the hash value distribution of node C matching the historical cross-chain fluctuation data, which prompts the possibility of cross-chain interaction delay risk.

[0097] The verification risk of the reference graph is corrected by the verification avoidance sub-link. Assuming that the system detects that the hash value fluctuation of node C is associated with the "cross-chain gateway temporary failure" event in the external interference event layer, triggering the verification avoidance sub-link. This link first performs secondary verification on the data shard state of node C, adding steps such as re-generating the verification circuit to verify whether the data shard fully reflects the original accounts receivable information. At the same time, the verification priority of node C is adjusted, temporarily moving it from the high-priority queue to the low-priority queue until the cross-chain gateway failure is resolved.

[0098] In another scenario, if the system processes medical data privacy computing tasks, the original data contains the hash values of patient diagnosis records, and the hash-time mapping table records the creation time and upload time of each record. When the real-time hash data of a certain shard node shows that a large number of hash values correspond to recently created diagnosis records (transient timestamp is May 22, 2025, 16:00:00), the upper limit of the theoretical synchronization range is set to 50 transactions per second based on the historical data processing efficiency of the node (due to the high complexity of medical data encryption, the processing speed is slower). The multi-dimensional fusion model combines the "sensitive data requires enhanced verification" rule in the verification avoidance parameter library and adds a "double verification" label to the data shard of this node in the on-chain verification reference graph, requiring the zero-knowledge verification unit to perform two independent verifications on each data.

[0099] The application of the verification circumvention sub-link in this scenario is reflected as follows: when it is detected that the hash value of a certain medical record matches the preset sensitive data hash feature (such as containing a certain medical code segment), the verification circumvention process is automatically triggered, the data shard is routed to a dedicated privacy protection node for processing, and the node is deployed with an encryption algorithm and verification mechanism that meet the medical data security standards. At the same time, the processing path change of the data shard is labeled in the graph in real time to ensure that the verification process is traceable.

[0100] After generating the benchmark verification strategy, the consensus modeling node continuously monitors the execution of each shard node. For example, node D is processing e-commerce transaction data, and the theoretical synchronization range upper limit is 500 per second, but the actual load suddenly increases to 600 per second, exceeding the upper limit. The system captures this anomaly in real time through the on-chain stability evaluation network, and the multi-dimensional fusion model automatically adjusts the benchmark verification strategy, dynamically migrating part of the tasks of node D to node E with lower load. At the same time, the state of node D is updated to "overload processing" in the reference graph, and the state optimization unit is prompted to preferentially aggregate the data shards of node E.

[0101] The verification circumvention sub-link plays a role in the task migration process: to ensure the integrity of the migrated data, the sub-link starts a data verification process, compares the hash values of each data shard migrated from node D to node E, and verifies whether the data before and after migration is consistent. If it is found that the hash value of a certain data shard changes during migration (possibly due to network transmission errors), the retransmission mechanism is triggered, the data shard is reacquired from the original node, and is marked as "high risk", and the verification frequency is increased in subsequent verification.

[0102] In the cross-quarter financial data verification scenario, the system processes a large amount of historical transaction data, and the hash-time mapping table shows that the data hash values are obviously distributed by quarter. The consensus modeling node divides the processing tasks of the shard nodes according to the quarterly time interval, for example, the data of 2024Q1 is allocated to node F, and the data of 2024Q2 is allocated to node G. The theoretical synchronization range upper limit is dynamically adjusted according to the data volume of each quarter: the data volume of Q1 is small, and the upper limit of node F is set to 100 per second; the data volume of Q2 is large, and the upper limit of node G is set to 300 per second.

[0103] The multi-dimensional fusion model combines to circumvent the rule of "historical data needs to be periodically archived and verified" in the parameter library. In the on-chain verification reference map, the data shards of Q1 and Q2 are marked with "archived verification" labels, and the ciphertext analysis unit of the privacy fusion module is required to prioritize calling the archived encryption key version during analysis to ensure decryption accuracy. The verification circumvention sub-chain mainly handles the verification risks caused by data archiving and storage in this scenario. For example, when the encryption key of a certain historical data is too old to be decrypted, the key backtracking mechanism is automatically triggered to retrieve the corresponding key version from the historical blocks of the blockchain, and after decryption, the verification is completed.

[0104] Throughout the implementation process, the multi-dimensional fusion model realizes the dynamic generation and optimization of benchmark verification strategies through the combination of real-time data and historical rules. The verification circumvention sub-chain provides a flexible correction mechanism to address potential risks in different scenarios, ensuring the accuracy and security of data shard verification. The system integrates the hash-time mapping table, verification circumvention parameter library, and real-time monitoring data to form an intelligent decision-making system covering the entire data processing process, effectively improving the efficiency and reliability of blockchain and privacy computing collaborative verification.

[0105] Example 5:

[0106] The parameter update mechanism of the multi-dimensional fusion model triggers corresponding verification circumvention instructions in different abnormal scenarios. Taking a certain logistics traceability system as an example:

[0107] When the system detects that the original data hash value of a batch of goods traceability data exceeds the preset interval (e.g., the hash value prefix should be "0x4a" but appears "0x4b"), or the shard synchronization rate exceeds the threshold (e.g., a node synchronizes 500 times per second, far exceeding the average level of 200 times per second), the first verification circumvention instruction is triggered. At this time, the system automatically starts the alternative shard unit, which pre-stores several backup shard nodes, such as selecting 2 low-load backup nodes (node A and node B) from 10 main nodes to take over the data shard tasks of the abnormal node. At the same time, the analysis granularity of the privacy fusion module is adjusted from the default "block level" to "transaction level", i.e., more detailed ciphertext analysis for each transaction data, such as splitting the original aggregated ciphertext data into single transaction ciphertext, and verifying the integrity of the goods traceability information (such as logistics number, storage node, transportation timestamp, etc.) one by one.

[0108] If the first verification circumvention instruction is executed and the shard state still does not recover to the allowed interval (e.g., the hash value anomaly rate of backup node A still reaches 30%), the second verification circumvention instruction is triggered. The system switches from the stable shard sublink to the verification circumvention sublink through the polymorphic switching link, and reallocates the topology weight of the data shard state based on the duration of the external interference event (e.g., the duration of the network attack). For example, if a DDoS attack lasting 2 hours is detected, causing multiple nodes to be abnormal synchronously, the topology weight allocation tilts towards the unattacked nodes C and D, increasing their weight from the default 1.0 to 1.5, while the weight of the attacked nodes is reduced to 0.5. At the same time, the verification circumvention sublink starts the "attack mode" and adds the blockchain immutable verification step for all data shards, i.e., by comparing the data hash value with the blockchain storage hash value, to ensure that the data has not been tampered with.

[0109] When the synchronization index of any level-1 shard node in the data shard module is lower than the threshold (e.g., the synchronization index of node E decreases from 0.9 to 0.6, which is lower than the threshold 0.7), the third verification circumvention instruction is triggered. The system immediately blocks the shard output channel of node E, prohibiting it from transmitting data to the collaborative verification module, and transfers the corresponding data to other shard nodes (e.g., nodes F and G) through the isolation verification unit. During the transfer process, the isolation verification unit performs packet-by-packet verification on the data, such as comparing the original data hash value with the transferred data hash value to ensure consistency. Assuming that node E is responsible for processing warehouse data in a certain region, when it is transferred to node F (which processes data in an adjacent region), the system automatically adjusts the analysis parameters of node F to make it compatible with the data format differences of the two regions.

[0110] During the execution of the third verification circumvention instruction, if it is detected that nodes F and G are overloaded (e.g., CPU utilization exceeds 85%), the fourth verification circumvention instruction is triggered. The shard graph optimization module calls the pre-constructed shard state security level table, which classifies nodes into "high priority" (node G), "medium priority" (node F), and "low priority" (node E) based on the on-chain synchronization rate (e.g., the synchronization rate of node F decreases from the stable 100 times / sec to 50 times / sec) and off-chain ciphertext indicators (e.g., the ciphertext analysis delay increases from 20ms to 50ms). The system dynamically downgrades the shard path priority according to the security level table, prioritizing the data of high-priority node G, temporarily suspending the data of medium-priority node F, and limiting the edge state synchronization requirements of low-priority node E (e.g., only synchronizing key logistics node change data, not full-quantity transportation trajectory data). At the same time, the analysis frequency of the privacy fusion module is adjusted from real-time analysis to batch analysis, such as analyzing low-priority data once every 10 minutes, to match the downgraded verification strategy.

[0111] The construction step of the sharding node also includes generating a hash gradient adaptive index table. Taking cross-border e-commerce data as an example, the spatial distribution result of the hash value of the original data shows that the hash value of the order data in the European region is concentrated in the “0xeu” prefix interval, and the hash value of the order data in the American region is concentrated in the “0.us” prefix interval. The historical cross-chain fluctuation range shows that the interaction between the European chain and the American chain is more frequent from 9:00 to 11:00 in the morning. The system generates a hash gradient adaptive index table according to this information, divides the nodes in the “0xeu” interval into a European sharding group, and divides the nodes in the “0.us” interval into an American sharding group. The consistency compensation coefficient difference of each group is recorded in the index table (for example, the European group has a higher compensation coefficient due to frequent cross-chain interaction). Through a dynamic weight distribution algorithm, higher topological weights (for example, a weight coefficient of 1.2) are assigned to the nodes in the European sharding group, and redundant transmission links are established between the nodes in the European group and the American group, for example, by deploying a dedicated line through a cross-chain gateway, to ensure that data can be transmitted through the redundant link when the main link is congested.

[0112] The generation process of the on-chain verification reference map is further refined in combination with specific business scenarios. For example, when processing fresh food traceability data, the upper limit of the theoretical synchronization range (for example, the upper limit of the cold chain temperature data processed by node H is 50 pieces per second) and the transient timestamp (May 22, 2025, 8:00 am, peak period of fresh food transportation) are input into a multi-dimensional fusion model to synchronously load a pre-set sharding medium database (records that the storage capacity of node H is 1 TB, and the remaining space is 800 GB) and an external interference event layer (shows traffic control on a certain road on the day). Based on the encrypted peak frequency data output by the ciphertext analysis unit (for example, the encryption frequency of temperature data increases due to the increase in the number of sensors), the verification threshold of the data sharding state in the reference map is corrected, and the hash value verification threshold of the temperature data is tightened from the default ±0.5°C to ±0.2°C. In the verification avoidance sub-link, a time gradient parameter (for example, the verification time increases by 30% during the morning rush hour) is superimposed for secondary calibration to ensure that the verification accuracy of temperature data is not affected during the transportation peak period.

[0113] In another scenario, when processing financial transaction data, the system adopts a sampling verification strategy for low-priority historical transaction data when the fourth verification avoidance instruction triggers the degradation of the sharding path priority. For example, 1,000 pieces of data are randomly selected from 100,000 pieces of historical data for zero-knowledge verification, and if the verification is successful, the remaining data is considered normal by default to reduce the verification load. At the same time, the sharding map optimization module generates a degradation processing report, which records the degradation reason (for example, node overload), the processing time (for example, 4 hours), and the affected data range (for example, part of the transaction data from Q3 to Q4 of 2024) for subsequent audit tracing.

[0114] The triggering and execution of the verification circumvention instructions have strict timing logic. For example, a first verification circumvention instruction is executed in preference to a second verification circumvention instruction, and the latter is only initiated after the former is invalidated; a third and fourth verification circumvention instruction belong to the same level of response, and are dynamically selected for execution according to real-time monitoring data. The system ensures the atomicity of instruction execution through a state machine mechanism, for example, when the output channel of the blocked exception node is blocked, the operation log is recorded at the same time and the data preloading of the backup node is triggered, avoiding interruption of data processing.

[0115] Throughout the implementation process, the multi-dimensional fusion model and the sharding graph optimization module interact with real-time data to form a closed-loop feedback system. For example, the new strategy generated after parameter updating will be injected into the consensus modeling node in real time to adjust the benchmark verification strategy; the sharding state data after degradation processing will be fed back to the multi-dimensional fusion model for optimizing the historical data training set of the hash-time mapping table. This collaborative mechanism enables the system to continuously optimize the verification process in complex business scenarios, cope with dynamically changing risk challenges, while ensuring data privacy and system stability.

Claims

1. A blockchain and privacy computing collaborative verification system, characterized in that, include: Data sharding module, privacy fusion module, and collaborative verification module; The data sharding module includes a preprocessing node and a dynamic sharding protocol; the privacy fusion module includes a ciphertext parsing unit and a zero-knowledge verification unit; the collaborative verification module includes a consensus modeling node and a state optimization unit. The preprocessing node is used to collect raw data hash values ​​and cross-chain interaction parameters in real time; and to construct a multi-level data collaboration topology based on the block height and timestamp sequence through the dynamic sharding protocol. The dynamic sharding protocol consists of multiple sharding nodes connected in parallel based on hash thresholds. Based on the real-time response characteristics of multiple sharding nodes, a benchmark verification strategy is generated through a consensus modeling node, and combined with the verification signal output by the zero-knowledge verification unit, the data sharding state is dynamically aggregated and sorted through the state optimization unit.

2. The blockchain and privacy computing collaborative verification system as described in claim 1, characterized in that, The system also includes a consensus correction module and a fragmented graph optimization module; the consensus correction module includes an isolation verification unit and a latency compensation unit. Based on the generated benchmark verification strategy and dynamic aggregation sorting results, the data sharding state is corrected by the isolation verification unit through distributed barriers, and the ciphertext parameters of the privacy fusion module are dynamically calibrated by the latency compensation unit. At the same time, the sharding synchronization rate is monitored in real time by the on-chain stability evaluation network constructed by the sharding graph optimization module, and the monitoring results are fed back to the consensus modeling node and the state optimization unit to make closed-loop adjustments to the verification strategy.

3. The blockchain and privacy computing collaborative verification system as described in claim 2, characterized in that, The isolation verification unit is composed of state control circuits corresponding to the sharding nodes connected in parallel in the dynamic sharding protocol; the construction steps of the multi-level data collaborative topology include: Collect raw data hash values ​​and cross-chain fluctuation signals, input them into the transfer learning model configured in each shard node, generate data consistency compensation coefficients and store them in the corresponding shard node; Based on the consistency compensation coefficient of the storage of each level of sharding node, combined with the instantaneous change rate of the timestamp sequence, the data sharding status is allocated to multiple sharding nodes through a dynamic hash detection algorithm, forming a dynamic collaborative matching topology.

4. The blockchain and privacy computing collaborative verification system as described in claim 3, characterized in that, The conditions for constructing shard nodes configured in the preprocessing node include: the original data hash value is within a preset range, the synchronization index corresponding to the shard node is greater than a threshold, and the block height is less than the allowable deviation range. The ciphertext parsing unit is also configured with a multi-state switching link; the multi-state switching link includes a steady-state sharding sub-link and a verification avoidance sub-link; the steady-state sharding sub-link is generated based on the matching degree between the on-chain stability signal of the privacy fusion module and the benchmark verification strategy; the verification avoidance sub-link is constructed based on the correlation between the instantaneous verification risk of the data sharding state and external interference events.

5. The blockchain and privacy computing collaborative verification system as described in claim 4, characterized in that, The consensus modeling node is configured with a multi-dimensional fusion model; the multi-dimensional fusion model includes a hash-time mapping table corresponding to the data sharding module and a verification avoidance parameter library for the privacy fusion module; The steps for generating the benchmark verification strategy include: Based on the real-time hash data and aggregation sorting results of the data sharding module, calculate the upper limit of the theoretical synchronization range of each sharding node in the steady-state sharding sub-link; By inputting the theoretical synchronization range upper limit and transient timestamp into a multi-dimensional fusion model, an on-chain verification reference graph is generated. The verification risk of the reference graph is then corrected by verifying the avoidance sub-link, thus forming a benchmark verification strategy.

6. The blockchain and privacy computing collaborative verification system as described in claim 5, characterized in that, The parameter update steps of the multi-dimensional fusion model include: When the original data hash value is detected to exceed the preset range or the fragment synchronization rate exceeds the threshold, the first verification circumvention instruction is triggered, that is, the alternative fragment unit is started and the parsing granularity of the privacy fusion module is adjusted. If the fragmentation state still fails to recover to the allowed range after the execution of the first verification avoidance instruction, the second verification avoidance instruction is triggered, that is, the verification avoidance sub-link is switched through the polymorphic switching link, and the topology weight of the data fragmentation state is reallocated based on the duration of the external interference event.

7. The blockchain and privacy computing collaborative verification system as described in claim 6, characterized in that, The parameter update steps of the multi-dimensional fusion model also include: When the synchronization index of any first-level sharding node in the data sharding module is lower than the threshold, the third verification avoidance instruction is triggered, that is, the sharding output channel of the node is blocked, and the corresponding data is transferred to other sharding nodes through the isolation verification unit. If other shard nodes are detected to be overloaded during the execution of the third verification avoidance instruction, the fourth verification avoidance instruction is triggered, which calls the shard graph optimization module to downgrade the output strategy of the state optimization unit and restrict the synchronization requirements of edge states.

8. The blockchain and privacy computing collaborative verification system as described in claim 7, characterized in that, The degradation processing logic of the segmented map optimization module includes: Based on the synchronization rate data output by the on-chain stability assessment network and the off-chain ciphertext indicators, a sharding state security level table is constructed. When the fourth verification circumvention instruction is triggered, the priority of the fragment path is dynamically downgraded according to the security level table, and the parsing frequency of the privacy fusion module is adjusted simultaneously to match the downgraded verification strategy.

9. The blockchain and privacy computing collaborative verification system as described in claim 3, characterized in that, The steps for constructing the sharded nodes also include: Based on the spatial distribution of the original data hash values ​​and the historical cross-chain fluctuation range, a hash gradient adaptive index table is generated. Based on the differences in the consistency compensation coefficients of each shard node in the index table, a dynamic weight allocation algorithm is used to allocate topological weights to the state control circuits of multiple shard nodes and establish redundant transmission links between shard nodes.

10. The blockchain and privacy computing collaborative verification system as described in claim 5, characterized in that, The generation process of the on-chain verification reference map includes: After inputting the theoretical synchronization range upper limit and transient timestamp into the multi-dimensional fusion model, the preset segmented media database and external interference event layer are loaded synchronously. Based on the encrypted peak frequency data output by the ciphertext parsing unit, the verification threshold of the data fragmentation state in the reference spectrum is corrected, and the time gradient parameter is superimposed in the verification avoidance sub-link for secondary calibration.

Images