Digital asset dynamic management and cross-chain transaction method and system based on block chain and privacy calculation

Abstract

The invention relates to the technical field of digital asset management, in particular to a digital asset dynamic management and cross-chain transaction method and system based on a block chain and privacy computing. In the initial link of asset management, rapid investigation of potential risk assets is realized, subsequent processing is established on the basis of credible data, abnormal access fragments hidden in large-scale access behaviors are accurately captured in conjoint analysis of time sequence division of access logs and permission state and node reputation multi-dimensional parameters, and the risk assessment of the assets is realized. According to the method, accurate locking of high-risk objects is realized, and when cross-chain asset transfer is executed, screening is performed based on available duration and index validity of fragments in a path, so that data integrity and delivery certainty in a cross-chain transaction process are improved; and the security, circulation efficiency and data guarantee capability of the digital assets in the cross-chain management process are promoted to be comprehensively improved.

Description

Technical Field

[0001] This invention relates to the field of digital asset management technology, and in particular to a method and system for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing. Background Technology

[0002] The field of digital asset management technology mainly involves the technical system for storing, maintaining, authorizing, transferring, trading, and controlling the circulation of assets existing in electronic form, including cryptocurrencies, tokens, digital certificates, and on-chain contract rights. The field covers multiple aspects such as encrypted storage, access control, transaction execution, ledger recording, and audit verification. The applied technologies include distributed ledger systems, encryption algorithms, access control protocols, smart contract execution mechanisms, and cross-chain interaction mechanisms, aiming to achieve secure storage, efficient circulation, and tamper-proof traceability of digital assets across different systems and platforms.

[0003] A method for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing refers to a technical solution that uses blockchain technology to record the storage and transaction information of digital assets, and combines privacy computing technology to achieve dynamic authorization, verification and transfer of cross-chain digital assets while protecting the sensitive data of transaction participants. The aim is to solve the problems of insufficient storage security, low efficiency of cross-chain transactions and weak user privacy protection in the existing digital asset management process, thereby achieving secure storage of digital assets, improved cross-chain availability and enhanced privacy protection.

[0004] Existing technologies for storing and managing digital assets across chains primarily focus on recording transaction ledgers and execution results, lacking multi-layered risk assessment of assets and paths before transactions are triggered. This makes it difficult to identify potentially problematic assets and high-risk access behaviors in the early stages. In cross-chain path selection, evaluations are usually based on static configurations and simple historical statistics, failing to effectively reflect dynamic changes in node latency, lock-up duration, and processing time. Transaction paths are prone to performance fluctuations or even interruptions during actual execution. Link performance evaluation often only considers bandwidth and a single speed indicator, without comprehensively considering key parameters such as packet loss rate and processing speed. This results in an incomplete evaluation of path merits, potentially leading to the selection of links with hidden instability factors. In actual operation, this can cause negative consequences such as increased cross-chain transaction failure rates, reduced data consistency, and wasted link resources, affecting the stability and reliability of cross-platform circulation of digital assets. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method and system for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing, comprising the following steps:

[0007] Step 1: Based on the digital asset metadata records stored in the blockchain and the location information of the encrypted shards on the distributed nodes, extract the ownership identifier, transaction hash and sharding delay, calculate consistency and filter anomaly numbers to generate a list of potential abnormal assets;

[0008] Step 2: Based on the potential abnormal asset list, extract the encrypted shard access logs of the distributed nodes, and compare the access interval, permission level and node reputation using an isolated forest. Filter records that deviate from the rules and record the shard number to generate a high-risk access shard list.

[0009] Step 3: Based on the high-risk access shard list, extract the asset path information required for the target cross-chain transaction, and use the Long Short-Term Memory network to obtain the delay, lock-up and processing time. Eliminate paths with accumulated delay exceeding the limit to form the remaining path set, and generate an executable transaction path set.

[0010] Step 4: Based on the executable transaction path set and inter-chain transmission link data, extract processing rate, bandwidth utilization and packet loss rate, perform weighted scoring to replace low-scoring paths and retain high-scoring paths, and generate a preferred cross-chain path set;

[0011] Step 5: Based on the preferred cross-chain path set, sequentially access the nodes in the path to read the encrypted shard index and available time, eliminate paths with shorter execution cycles and missing indexes, perform shard decryption and reorganization and on-chain registration, and generate a cross-chain asset transfer registration form.

[0012] As a further embodiment of the present invention, the potential abnormal asset list includes an abnormal asset number, a corresponding ownership identifier, and a distributed node identifier; the high-risk access shard list includes a shard number, an associated access permission level, and an access node reputation score; the executable transaction path set includes a path number, the latency time of each node, and the path execution efficiency; the preferred cross-chain path set includes a path number, a comprehensive path score, and a path node performance label; and the cross-chain asset transfer registration form includes a transaction number, a target chain identifier, and the asset cross-chain time.

[0013] As a further aspect of the present invention, the specific steps for generating the potential abnormal asset list are as follows:

[0014] Based on the digital asset metadata records stored on the blockchain and the location information of encrypted shards on distributed nodes, the asset ownership identifier in the metadata is called and the corresponding transaction hash value is read. The physical address of the distributed node is located and the sharding delay time is extracted. The three items are sorted by asset number and written into a table to form a structured record, generating an asset information structure table.

[0015] Based on the asset information structure table, the consistency between the asset ownership identifier and the transaction hash value is compared and the matching status is marked. The sharding delay time is detected and compared with a preset threshold. The asset numbers that do not match or exceed the delay limit are stored in the abnormal data list to generate a potential abnormal asset list.

[0016] As a further aspect of the present invention, the specific steps for generating the high-risk access shard list are as follows:

[0017] Based on the list of potential abnormal assets, the corresponding encrypted shards of the assets are located and the access logs are read. Events are divided into multiple access cycles in chronological order. The number of permission changes and the magnitude of reputation fluctuations in each cycle are counted. An isolated forest is used to filter abnormal records. A mapping between shard numbers and statistical values ​​is established to generate a set of shard cycle fluctuation relationships.

[0018] Based on the fragmented periodic fluctuation relationship set, the time difference between adjacent periods is calculated and compared with the baseline length. The number of permission changes is matched with the rule limit, the reputation fluctuation amplitude is compared with the allowed range, abnormal periods are extracted and marked, and a suspected illegal access record set is generated.

[0019] Based on the suspected unauthorized access record set, the encrypted fragment numbers involved are extracted and duplicate numbers are removed. The remaining fragment numbers are then serialized and sorted from highest to lowest frequency of occurrence. The sorting results are then converted into a number list in a standard storage format to generate a high-risk access fragment list.

[0020] As a further aspect of the present invention, the isolated forest execution process specifically involves dividing the events into multiple access cycles in chronological order based on the located and read access log data, calculating the number of permission changes and the fluctuation range of node reputation in each cycle, constructing a multi-dimensional data set of cycles and corresponding parameter values, comparing the path lengths recorded in the isolation tree one by one in the set, marking cycles with path lengths significantly greater than most records as abnormal, and establishing a corresponding relationship between abnormal cycles and shard numbers and recording them in a sequence list.

[0021] As a further aspect of the present invention, the specific steps for generating the executable transaction path set are as follows:

[0022] Based on the high-risk access shard list, locate the asset path node corresponding to the shard number in the cross-chain transaction configuration and read the path order, obtain the delay, lock-up time and processing time of each confirmation node, summarize the parameters according to the path number and store them in a table to generate a path performance parameter table.

[0023] Based on the path performance parameter table, the cumulative delay of each path is calculated using a long short-term memory network and compared with the upper limit. The node delay, lock-up time and processing time are checked to see if they exceed the limits. The path numbers that exceed the limits are added to the removal set and removed to generate a table of remaining available paths.

[0024] Based on the remaining available paths table, the remaining path numbers and corresponding parameters are sorted according to latency and processing efficiency. An index list is established based on the sorting results and merged into a unified path set. The path set is then converted into structured data that can be read by cross-chain transactions, generating an executable transaction path set.

[0025] As a further aspect of the present invention, the execution process of the Long Short-Term Memory Network specifically involves, based on the path performance parameter table, combining the confirmation node delay time, asset lock-up duration, and inter-chain gateway processing time of each path into sequential input data in chronological order, performing multiple rounds of state transmission and calculating time steps, obtaining the cumulative delay prediction value of each path when it is executed in the future, comparing the prediction value with the preset upper limit one by one, and recording the number of the path exceeding the limit to the removal set for subsequent removal processing.

[0026] As a further aspect of the present invention, the specific steps for generating the preferred cross-chain path set are as follows:

[0027] Based on the executable transaction path set and inter-chain transmission link data, the nodes in the path are traversed sequentially and the node processing rate value, link bandwidth utilization value and data packet loss rate are extracted. The parameters are compared with the corresponding benchmark thresholds and status labels are generated for the nodes. The status labels of all nodes in the path are summarized to form a comprehensive performance identifier set and generate a path performance identifier set.

[0028] Based on the path performance identifier set, the comprehensive performance level is calculated and sorted according to the order of nodes in the path. Paths with performance levels below the average value are replaced with candidate paths with higher performance levels. The path numbers and corresponding node parameters of the peak performance level are retained and an optimal set is generated, thus generating an optimal cross-chain path set.

[0029] As a further aspect of the present invention, the specific steps for generating the cross-chain asset transfer registration form are as follows:

[0030] Based on the preferred cross-chain path set, the nodes in the path are accessed sequentially to read the encrypted shard index and available time. Nodes with shorter execution cycles and missing indexes are recorded and marked with associated paths. After removing the marked paths, the remaining numbers are summarized to generate a list of available paths.

[0031] Based on the available path list, node shards are called in order of path and index consistency is checked. Shards are decrypted in sequence and complete digital asset files are assembled. The files are uploaded to the target chain to generate transfer records and associate the records with the path numbers to generate a cross-chain asset transfer registration form.

[0032] A dynamic management and cross-chain transaction system for digital assets based on blockchain and privacy computing is provided. This system is used to execute the aforementioned dynamic management and cross-chain transaction method for digital assets based on blockchain and privacy computing. The system includes:

[0033] Anomaly asset identification module: Based on the digital asset metadata stored in the blockchain and the encrypted sharding location of distributed nodes, it verifies the correspondence between ownership identifiers and transaction hashes and analyzes the latency distribution characteristics to generate a list of potential anomaly assets;

[0034] Shard risk analysis module: Based on the potential abnormal asset list, the events in the access log are processed in a time sequence, the permission status and reputation value are jointly analyzed, and the shard number with abnormal status is determined by using an isolated forest to generate a list of high-risk access shards;

[0035] Cross-chain path filtering module: Based on the high-risk access shard list, extract the node operation parameters in the cross-chain transaction path one by one, and perform aggregation modeling with latency, lock-up, and processing duration. Use a long short-term memory network to filter out paths with total latency exceeding the range, and generate a set of executable transaction paths.

[0036] Preferred Path Generation Module: Based on the executable transaction path set and inter-chain transmission link data, it quantitatively calculates the processing efficiency, bandwidth utilization, and data packet loss ratio of each node and comprehensively generates a performance level, replaces low-performance paths, and generates a preferred cross-chain path set.

[0037] Asset cross-chain execution module: Based on the preferred cross-chain path set, it sequentially reads the shard data stored in the node and checks the index, removes paths that do not meet the execution cycle and lack shards, decrypts and combines shards in sequence, and generates a cross-chain asset transfer registration form.

[0038] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0039] 1. In this invention, by extracting ownership identifiers, transaction hashes, and sharding delays from the digital asset metadata and distributed node sharding location information stored in the blockchain, and performing consistency calculations and anomaly number screening, a rapid screening of potentially risky assets is achieved in the initial stage of asset management, so that subsequent processing is based on trusted data.

[0040] 2. In this invention, by combining the time sequence division of access logs with the multi-dimensional analysis of permission status and node reputation parameters, an isolated forest is introduced to identify rule deviation records, accurately capture abnormal access fragments hidden in large-scale access behavior, and achieve precise locking of high-risk objects.

[0041] 3. In this invention, by filtering based on the availability duration and index validity of shards in the path during cross-chain asset transfer, the data integrity and delivery certainty in the cross-chain transaction process are improved, thereby promoting a comprehensive improvement in the security, circulation efficiency and data protection capabilities of digital assets in the cross-chain management process. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the workflow of the present invention;

[0043] Figure 2 This is a system flowchart of the present invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0045] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0046] Example 1

[0047] Please see Figure 1 This invention provides a technical solution: a method for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing, comprising the following steps:

[0048] Step 1: Based on the digital asset metadata records stored in the blockchain and the location information of the encrypted shards on the distributed nodes, extract the ownership identifier, transaction hash and sharding delay, calculate consistency and filter anomaly numbers to generate a list of potential abnormal assets;

[0049] Step 2: Based on the list of potentially abnormal assets, extract the encrypted shard access logs of distributed nodes, and use the isolated forest to compare the access interval, permission level and node reputation, filter the records that deviate from the rules and record the shard number to generate a list of high-risk access shards;

[0050] Step 3: Based on the high-risk access shard list, extract the asset path information required for the target cross-chain transaction, and use the Long Short-Term Memory network to obtain the delay, lock-up and processing time. Eliminate paths with accumulated delay exceeding the limit to form the remaining path set, and generate an executable transaction path set.

[0051] Step 4: Based on the set of executable transaction paths and inter-chain transmission link data, extract processing rate, bandwidth utilization and packet loss rate, perform weighted scoring to replace low-scoring paths and retain high-scoring paths, and generate a preferred cross-chain path set;

[0052] Step 5: Based on the optimized cross-chain path set, sequentially access the nodes in the path to read the encrypted shard index and available time, eliminate paths with shorter execution cycles and missing indexes, perform shard decryption and reorganization and on-chain registration, and generate a cross-chain asset transfer registration form.

[0053] The potential abnormal asset list includes the abnormal asset number, corresponding ownership identifier, and distributed node identifier; the high-risk access shard list includes the shard number, associated access permission level, and access node reputation score; the executable transaction path set includes the path number, the latency time of each node, and the path execution efficiency; the preferred cross-chain path set includes the path number, the path comprehensive score, and the path node performance label; and the cross-chain asset transfer registration form includes the transaction number, the target chain identifier, and the asset cross-chain time.

[0054] The specific steps for generating a list of potentially anomalous assets are as follows:

[0055] Based on the digital asset metadata records stored on the blockchain and the location information of encrypted shards on distributed nodes, the asset ownership identifier in the metadata is called and the corresponding transaction hash value is read. The physical address of the distributed node is located and the sharding delay time is extracted. The three items are sorted by asset number and written into a table to form a structured record, generating an asset information structure table.

[0056] Based on the asset information structure table, the consistency between the asset ownership identifier and the transaction hash value is compared and the matching status is marked. The sharding delay time is detected and compared with the preset threshold. The asset numbers that do not match and exceed the delay limit are stored in the abnormal data list to generate a potential abnormal asset list.

[0057] Based on the digital asset metadata records stored in the blockchain and the location information of encrypted shards on distributed nodes, the asset ownership identifier in the metadata is called and the corresponding transaction hash value is read. The input data is grouped into 512 bits and filled to the full length of the group. The initial hash value is a standard constant. In the loop operation, each group of data is processed by preset shift and bitwise logical operation to generate an intermediate value. Then, it is superimposed with the initial hash in multiple rounds to obtain a fixed-length hash value. The latitude and longitude information of the distributed nodes is obtained, the geographical location between two points is calculated to obtain the network delay from the node to the central node, and the delay is recorded as a millisecond value. Finally, the ownership identifier, hash value and delay data are sorted by asset number from smallest to largest and organized into a two-dimensional data structure with three columns: number, hash and delay, to generate an asset information structure table.

[0058] Based on the asset information structure table, the consistency between the asset ownership identifier and the transaction hash value is compared and the matching status is marked. The Euclidean algorithm is used to calculate the difference between the two data, and the two are converted into binary vectors for bit-by-bit comparison and the overall difference distance is calculated. The bit difference threshold T1 obtained from historical statistics is set to 15 bits. The difference value is marked as a match if it does not exceed T1, and marked as a mismatch if it exceeds T1. The fragmentation latency is detected and compared with the threshold T2. T2 is obtained by averaging multiple network latency tests. The asset numbers that exceed T2 and are mismatched are added to the abnormal data list. Finally, the list is deduplicated to generate a list of potential abnormal assets.

[0059] The specific steps for generating a list of high-risk access shards are as follows:

[0060] Based on the list of potential abnormal assets, the corresponding encrypted shards of the assets are located and the access logs are read. Events are divided into multiple access cycles in chronological order. The number of permission changes and the magnitude of reputation fluctuations in each cycle are counted. An isolated forest is used to filter abnormal records. A mapping between shard numbers and statistical values ​​is established to generate a set of shard cycle fluctuation relationships.

[0061] Based on the fragmented periodic fluctuation relationship set, the time difference between adjacent periods is calculated and compared with the baseline length. The number of permission changes is matched with the rule limit, the reputation fluctuation amplitude is compared with the allowed range, abnormal periods are extracted and marked, and a suspected illegal access record set is generated.

[0062] Based on the suspected unauthorized access record set, the encrypted fragment numbers involved are extracted and duplicate numbers are removed. The remaining fragment numbers are then serialized and sorted from highest to lowest frequency of occurrence. The sorting results are then converted into a number list in a standard storage format to generate a high-risk access fragment list.

[0063] Based on the list of potentially abnormal assets, the corresponding encrypted shards of the assets are located and access logs are read. The Isolation Forest algorithm is used to detect anomalies in the access logs. Access events are sorted in chronological order and divided into access periods of fixed length, with the period length preset to 24 hours. The parameter is set by the period control field in the system configuration file. The number of permission level changes and the fluctuation range of node reputation are counted for each access period. The permission level is represented by five-level integer values, and the node reputation value is set to a range of 0 to 100 points and stored as an integer. When building the Isolation Forest, the number of trees is set to 200, and the number of subsamples is 256 of the total number of records. The anomaly score of the path length used for each record is calculated. Records with an anomaly score higher than 0.65 are marked as abnormal. Then, a one-to-many mapping relationship is established between the shard number of the abnormal record and the period statistical value and stored in a two-dimensional array to generate a shard period fluctuation relationship set.

[0064] Based on the fragmented periodic fluctuation relationship set, the time difference between adjacent periods is calculated and compared with the baseline length. The difference between the timestamps of the start and end times of two consecutive periods is calculated and the absolute value is taken. The result is compared with the preset baseline length threshold, which is set in the system configuration to the number of seconds corresponding to 24 hours. Then, the number of permission changes in each period is matched with the permission rule limit, which is preset to no more than 3 changes per day. The node reputation fluctuation amplitude is compared with the allowable range, which is set to no more than 15 minutes of change. Periodic data records that exceed the threshold in any of the detection items are added to the marking table to generate a suspected unauthorized access record set.

[0065] Based on the suspected unauthorized access record set, the relevant encrypted fragment numbers are extracted and duplicates are removed. Each fragment number is converted into a fixed-length string and stored in a hash set to ensure uniqueness. Then, the deduplicated fragment number list is sorted from highest to lowest according to the frequency of occurrence. The size is judged based on the integer value of the frequency of occurrence of the corresponding fragment number. After sorting, a number sequence is generated. The sequence is transcribed into a standard number format with a fixed length. The number is represented by 16 characters and written to a structured list format file to generate a high-risk access fragment list.

[0066] The execution process of the isolated forest is as follows: based on the access log data that has been located and read, the events are divided into multiple access periods in chronological order, the number of permission changes and the fluctuation range of node reputation in each period are calculated, a multi-dimensional data set of periods and corresponding parameter values ​​is constructed, the path length recorded in the isolation tree is compared one by one in the set, periods with path lengths significantly greater than most records are marked as abnormal, and the abnormal periods and shard numbers are established and recorded in the sequence list.

[0067] An isolated forest, according to the formula:

[0068]

[0069] in: This represents the improved outlier score. This represents the input access cycle sample record. This represents the path length weighting coefficient. Indicates sample Average path length across all isolated trees This represents the weighting coefficient for the number of permission changes. Indicates sample Normalized value of the number of permission changes during the access period This represents the weighting coefficient for the fluctuation range of node reputation. Indicates sample The standardized value of the node reputation fluctuation range during the access period. This represents the access latency weighting coefficient. Indicates sample The standardized value of the average access latency during the access period. This represents the function for calculating normalized constants. Indicates the number of samples. This represents the adjustment factor for the magnitude of delay fluctuation. Indicates sample The standardized value of latency fluctuation during the access period. Indicates sample Path length in an isolated tree;

[0070] Execution process: First, the data from blockchain transaction logs and distributed node access records are parsed to extract access cycle samples. In conjunction with privacy-preserving computational mechanisms, this ensures that input parameters do not reveal sensitive information, and generates path lengths for samples across multiple isolated trees. Calculate the average value And multiplied by the path length weighting factor Calculate the number of permission changes during the access period, and obtain the result after normalization. And multiply by the weighting factor for the number of permission changes. Extract node reputation changes from access cycle records and standardize them to obtain... And multiplied by the credit fluctuation weighting factor Then calculate the average latency of the access cycle, and obtain it after standardization. Multiply by the delay time weighting factor The numerator is the sum of the four parts mentioned above, and the denominator is based on the total number of samples. Calculated normalized constant And add the standardized value of the delayed fluctuation amplitude. Multiply by adjustment factor Divide the numerator by the denominator and substitute into the exponential function. In the middle, abnormal scores were obtained. .

[0071] The specific steps for generating an executable transaction path set are as follows:

[0072] Based on the high-risk access shard list, locate the asset path node corresponding to the shard number in the cross-chain transaction configuration and read the path order, obtain the delay, lock-up time and processing time of each confirmation node, summarize the parameters according to the path number and store them in a table to generate a path performance parameter table.

[0073] Based on the path performance parameter table, the cumulative latency of each path is calculated using a long short-term memory network and compared with the upper limit. The node latency, lock-up duration and processing time are checked to see if they exceed the limits. The path numbers that exceed the limits are added to the removal set and removed to generate a table of remaining available paths.

[0074] Based on the remaining available paths table, the remaining path numbers and corresponding parameters are sorted according to latency and processing efficiency. An index list is built based on the sorting results and merged into a unified path set. The path set is then converted into structured data that can be read by cross-chain transactions, generating an executable transaction path set.

[0075] Based on the high-risk access shard list, the algorithm locates the asset path nodes corresponding to the shard numbers in the cross-chain transaction configuration and reads the path order. It extracts the information of each confirmed node in the shortest order, records the node delay time in milliseconds, the lock-up time in seconds, and the processing time in milliseconds. The algorithm presets the traversal start point as the first node of the path and the end point as the last node. The algorithm updates the cumulative distance of the nodes in each traversal and stores it in a two-dimensional matrix. The parameters are categorized by path number index. The delay, lock-up, and processing time data are encapsulated into a row structure with three columns of attributes. The data is then sorted in ascending order by path number and inserted into a relational storage table to generate a path performance parameter table.

[0076] Based on the path performance parameter table, the cumulative latency of each path is calculated using a Long Short-Term Memory (LSTM) network and compared with the upper limit. The latency, lockout duration, and processing time are used as time series inputs. The length of the input sequence is set to a multiple of the total number of path nodes, the batch size is set to 64, and the time step contains a vector composed of three parameters. The network structure includes an input layer, two hidden memory layers, and an output layer. The number of hidden layer units is set to 128, the activation function is tanh, the weight initialization method is set to uniform distribution, the training epochs are set to 100, and the Adam optimization method with a learning rate of 0.001 is used. The cumulative latency prediction value is output, and the prediction value is compared with the preset upper limit of 3000 milliseconds for each path. The path number that exceeds the latency limit is added to the elimination list, the corresponding record is removed from storage, and a table of remaining available paths is generated.

[0077] Based on the remaining available paths table, the remaining path numbers and corresponding parameters are sorted comprehensively according to latency and processing efficiency. A weighted ratio is assigned to the latency and efficiency of each path, with a latency weight of 0.6 and a processing efficiency weight of 0.4. The score of each path is obtained by adding the two indicators according to their weights. A quicksort algorithm is used to sort the paths from highest to lowest score. The sorting results are written into an index list, and a hash index is created for each path using a unique number. The index is bound to the path number and the parameters of each node in the path. The index list is merged into a unified path set, which is then formatted using a JSON structure for direct reading by the cross-chain transaction module, generating an executable transaction path set.

[0078] The Long Short-Term Memory Network (LSTM) execution process is as follows: Based on the path performance parameter table, the confirmation node delay time, asset lock-up duration, and inter-chain gateway processing time of each path are combined into sequence input data in chronological order. Multiple rounds of state transmission are performed and time steps are calculated to obtain the cumulative delay prediction value of each path when it is executed in the future. The prediction value is compared with the preset upper limit one by one, and the number of the path that exceeds the limit is recorded in the removal set for subsequent removal processing.

[0079] Long Short-Term Memory (LSTM) networks, according to the formula:

[0080]

[0081] in: This represents the cumulative delay value of the predicted cross-chain transaction path. Represents the path sequence of the first... Time step index This represents the total number of time steps in the path. Indicates the corresponding hidden state input The weighting coefficients, Indicates time step The hidden state output value of the Long Short-Term Memory network. Represents the corresponding node network The weighting factor for round-trip delay. Indicates time step The standardized value of the round-trip time of the node network. Indicates the corresponding lock-up processing time. The weighting coefficients, Indicates time step The standardized value of the duration of on-chain lock-up processing events. Indicates the corresponding cross-chain data forwarding time. The weighting coefficients, Indicates time step The standardized value of cross-chain gateway data forwarding time. This represents the weighting coefficient indicating the impact of end-to-end latency fluctuations. This represents the average delay fluctuation across all time steps of the path.

[0082] Execution process: First, extract the paths from the path performance parameter table across all... The specific record of each time step, including the hidden state output of each time step. Round-trip delay Lock-up processing time Cross-chain data forwarding time After standardization, a unified dimension is formed. Through correlation analysis between historical cross-chain transaction records and various parameters, weight coefficients are obtained. , , , To ensure their sum equals 1, this is used to balance the contribution of each parameter to the prediction. During the calculation, the values ​​at each time step are considered. , , , Multiply by the corresponding weights respectively , , , The values ​​are summed and accumulated over time steps to obtain the basic delay prediction value, and then the average end-to-end delay fluctuation is introduced. and influence coefficient As a correction term, it is added to the accumulated result to generate the final predicted cumulative delay value for the cross-chain transaction path. .

[0083] The specific steps for generating the preferred cross-chain path set are as follows:

[0084] Based on the executable transaction path set and inter-chain transmission link data, the nodes in the path are traversed sequentially and the node processing rate value, link bandwidth utilization value and data packet loss rate are extracted. The parameters are compared with the corresponding benchmark thresholds and status labels are generated for the nodes. The status labels of all nodes in the path are summarized to form a comprehensive performance label set and generate a path performance label set.

[0085] Based on the path performance identifier set, the comprehensive performance level is calculated and sorted according to the order of nodes in the path. Paths with performance levels below the average value are replaced with candidate paths with higher performance levels. The path number and corresponding node parameters of the peak performance level are retained and an optimal set is generated, thus generating an optimal cross-chain path set.

[0086] Based on the executable transaction path set and inter-chain transmission link data, the nodes in the path are traversed sequentially, and the node processing rate value, link bandwidth utilization value, and data packet loss rate are extracted. The three types of parameters are standardized and calculated respectively. The global mean of each parameter is subtracted and divided by the standard deviation. The mean and standard deviation are calculated in memory through the traversed and extracted data. After standardization, the parameter values ​​are compared with the benchmark thresholds set in the system configuration file. The benchmark thresholds include the minimum processing rate, the minimum bandwidth utilization, and the maximum data packet loss rate. Each threshold is stored as a floating-point number in the global configuration object. A status label is generated for the node according to the comparison result. The label value uses an integer 1 to indicate that the requirements are met and 0 to indicate that the requirements are not met. The status labels of all nodes in the path are recorded in a two-dimensional array according to the path number. All label data are summarized to form a comprehensive performance identifier set, and a path performance identifier set is generated.

[0087] Based on the path performance identifier set, the overall performance level of the nodes is calculated and sorted according to their order in the path. The scores of the three indicators of the nodes are weighted according to the preset weight ratio, with the weights allocated as processing rate 0.5, bandwidth utilization 0.3, and packet loss rate 0.2. The weighted values ​​are calculated in floating-point format and retained to two decimal places. The overall performance level of the path is obtained by summing the weighted values ​​of all nodes in the path and dividing by the total number of nodes. After the calculation, a descending sorted list is built according to the overall performance level. The list is constructed using the quicksort method. The path number with a performance level lower than the average of the path set is replaced with the candidate path number that ranks higher in the sorting results through an index mapping table. The path number with the highest performance level and the original parameter set of the corresponding node are retained, and the data is formed into an optimal set to generate the optimal cross-chain path set.

[0088] The specific steps for generating a cross-chain asset transfer register are as follows:

[0089] Based on the preferred cross-chain path set, the nodes in the path are accessed sequentially to read the encrypted shard index and available time. Nodes with shorter execution cycles and missing indexes are recorded and marked with associated paths. After removing the marked paths, the remaining numbers are summarized to generate a list of available paths.

[0090] Based on the list of available paths, node shards are called in order of path and index consistency is checked. Shards are decrypted in sequence and complete digital asset files are assembled. The files are uploaded to the target chain to generate transfer records and the records are associated with path numbers to generate a cross-chain asset transfer registration form.

[0091] Based on the optimized cross-chain path set, the encrypted shard index and available duration are read sequentially from the nodes in the path. The validity of the node's available duration and shard index is determined. The node's available duration parameter is converted into a millisecond integer value and compared with the preset execution cycle parameter in the system configuration file. The cycle parameter is stored in a global constant table in milliseconds. The shard index is read as a fixed-length 16-byte binary data and decoded into a string format. The string length and character set conformity are used as the validity verification standard. If the available duration is less than the execution cycle and the shard index length does not match the preset value, the corresponding node number is recorded in the abnormal node set. The path number associated with it is marked through the path-to-node mapping table. The abnormal path number is recorded in the removal list. The removal list is deduplicated. The remaining path numbers after removal are sorted in ascending order and written to a structured array file to generate an available path list.

[0092] Based on the list of available paths, node shards are called in order of path and index consistency is checked. The shard content is then decrypted. The decryption key is read from the security key management module, and the key length is set to 256 bits. During decryption, the initialization parameters corresponding to each group of data blocks are used. The initialization parameters are 16-byte random numbers generated during encryption and stored together with the encrypted shards. During the decryption process, each data group is processed in a preset order. Extra bytes added during the encryption phase to pad the data length are removed from the end of the group. The decrypted data fragments are then concatenated in ascending order according to the shard index number to form a complete digital asset file. The cross-chain data transfer interface is called to upload the file to the target chain node. The generated transaction identifier value is associated with the corresponding path number, and the combination record of the path number, transaction identifier, and target chain identifier is written to the on-chain notarization module to generate a cross-chain asset transfer registration table.

[0093] Please see Figure 2 A dynamic management and cross-chain transaction system for digital assets based on blockchain and privacy computing, comprising:

[0094] Anomaly asset identification module: Based on the digital asset metadata stored in the blockchain and the encrypted sharding location of distributed nodes, it verifies the correspondence between ownership identifiers and transaction hashes and analyzes the latency distribution characteristics to generate a list of potential anomaly assets;

[0095] Shard Risk Analysis Module: Based on a list of potential abnormal assets, the module performs time-series processing on events in the access logs, jointly analyzes permission status and reputation value, and uses an isolated forest to determine the shard number of the abnormal status, generating a list of high-risk access shards.

[0096] Cross-chain path filtering module: Based on the high-risk access shard list, extract the node running parameters in the cross-chain transaction path one by one, and aggregate them with latency, lock-up, and processing duration. Use a long short-term memory network to filter out paths with total latency exceeding the range, and generate a set of executable transaction paths.

[0097] The preferred path generation module: Based on the set of executable transaction paths and inter-chain transmission link data, it quantitatively calculates the processing efficiency, bandwidth utilization and data packet loss ratio of each node and generates a comprehensive performance level, replaces low-performance paths, and generates a preferred cross-chain path set.

[0098] The cross-chain asset execution module reads the shard data stored in the nodes in sequence and checks the index based on the optimized cross-chain path set. Paths that do not meet the execution cycle or lack shards are removed. The shards are decrypted and combined in sequence to generate a cross-chain asset transfer registration form.

[0099] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for dynamic management of digital assets and cross-chain transaction based on blockchain and privacy calculation, characterized in that, Includes the following steps: Step 1: Based on the metadata records of digital assets stored on the blockchain and the location information of encrypted shards on distributed nodes, extract ownership identifiers, transaction hashes and sharding delays, calculate consistency and filter anomaly numbers to generate a list of potential abnormal assets; Step 2: Based on the potential abnormal asset list, extract the encrypted shard access logs of the distributed nodes, and compare the access interval, permission level and node reputation using an isolated forest. Filter records that deviate from the rules and record the shard number to generate a high-risk access shard list. Step 3: Based on the high-risk access shard list, extract the asset path information required for the target cross-chain transaction, and use the Long Short-Term Memory network to obtain the delay, lock-up and processing time. Eliminate paths with accumulated delay exceeding the limit to form the remaining path set, and generate an executable transaction path set. Step 4: Based on the executable transaction path set and inter-chain transmission link data, extract processing rate, bandwidth utilization and packet loss rate, perform weighted scoring to replace low-scoring paths and retain high-scoring paths, and generate a preferred cross-chain path set; Step 5: Based on the preferred cross-chain path set, sequentially access the nodes in the path to read the encrypted shard index and available time, eliminate paths with shorter execution cycles and missing indexes, perform shard decryption and reorganization and on-chain registration, and generate a cross-chain asset transfer registration form.

2. The method for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing according to claim 1, characterized in that, The potential abnormal asset list includes an abnormal asset number, corresponding ownership identifier, and distributed node identifier; the high-risk access shard list includes a shard number, associated access permission level, and access node reputation score; the executable transaction path set includes a path number, node latency time, and path execution efficiency; the preferred cross-chain path set includes a path number, path comprehensive score, and path node performance label; and the cross-chain asset transfer registration form includes a transaction number, target chain identifier, and asset cross-chain time.

3. The method for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing according to claim 1, characterized in that, The specific steps for generating the potential abnormal asset list are as follows: Based on the digital asset metadata records stored on the blockchain and the location information of encrypted shards on distributed nodes, the asset ownership identifier in the metadata is called and the corresponding transaction hash value is read. The physical address of the distributed node is located and the sharding delay time is extracted. The three items are sorted by asset number and written into a table to form a structured record, generating an asset information structure table. Based on the asset information structure table, the consistency between the asset ownership identifier and the transaction hash value is compared and the matching status is marked. The sharding delay time is detected and compared with a preset threshold. The asset numbers that do not match or exceed the delay limit are stored in the abnormal data list to generate a potential abnormal asset list.

4. The method for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing according to claim 1, characterized in that, The specific steps for generating the high-risk access shard list are as follows: Based on the list of potential abnormal assets, the corresponding encrypted shards of the assets are located and the access logs are read. Events are divided into multiple access cycles in chronological order. The number of permission changes and the magnitude of reputation fluctuations in each cycle are counted. An isolated forest is used to filter abnormal records. A mapping between shard numbers and statistical values ​​is established to generate a set of shard cycle fluctuation relationships. Based on the fragmented periodic fluctuation relationship set, the time difference between adjacent periods is calculated and compared with the baseline length. The number of permission changes is matched with the rule limit, the reputation fluctuation amplitude is compared with the allowed range, abnormal periods are extracted and marked, and a suspected illegal access record set is generated. Based on the suspected unauthorized access record set, the encrypted fragment numbers involved are extracted and duplicate numbers are removed. The remaining fragment numbers are then serialized and sorted from highest to lowest frequency of occurrence. The sorting results are then converted into a number list in a standard storage format to generate a high-risk access fragment list.

5. The method for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing according to claim 1, characterized in that, The isolated forest execution process specifically involves dividing events into multiple access cycles based on the located and read access log data, calculating the number of permission changes and node reputation fluctuations in each cycle, constructing a multi-dimensional data set of cycles and corresponding parameter values, comparing the path lengths recorded in the isolated tree one by one in the set, marking cycles with path lengths significantly greater than most records as abnormal, and establishing a correspondence between abnormal cycles and shard numbers to record in a sequence list.

6. The method for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing according to claim 1, characterized in that, The specific steps for generating the executable transaction path set are as follows: Based on the high-risk access shard list, locate the asset path node corresponding to the shard number in the cross-chain transaction configuration and read the path order, obtain the delay, lock-up time and processing time of each confirmation node, summarize the parameters according to the path number and store them in a table to generate a path performance parameter table. Based on the path performance parameter table, the cumulative delay of each path is calculated using a long short-term memory network and compared with the upper limit. The node delay, lock-up time and processing time are checked to see if they exceed the limits. The path numbers that exceed the limits are added to the removal set and removed to generate a table of remaining available paths. Based on the remaining available paths table, the remaining path numbers and corresponding parameters are sorted according to latency and processing efficiency. An index list is built based on the sorting results and merged into a unified path set. The path set is then converted into structured data that can be read by cross-chain transactions, generating an executable transaction path set.

7. The method for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing according to claim 1, characterized in that, The execution process of the Long Short-Term Memory Network is as follows: based on the path performance parameter table, the confirmation node delay time, asset lock-up time and inter-chain gateway processing time of each path are combined into sequence input data in chronological order. Multiple rounds of state transmission are performed and time steps are calculated to obtain the cumulative delay prediction value of each path when it is executed in the future. The prediction value is compared with the preset upper limit one by one, and the number of the path that exceeds the limit is recorded in the removal set for subsequent removal processing.

8. The method for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing according to claim 1, characterized in that, The specific steps for generating the preferred cross-chain path set are as follows: Based on the executable transaction path set and inter-chain transmission link data, the nodes in the path are traversed sequentially and the node processing rate value, link bandwidth utilization value and data packet loss rate are extracted. The parameters are compared with the corresponding benchmark thresholds and status labels are generated for the nodes. The status labels of all nodes in the path are summarized to form a comprehensive performance label set and generate a path performance label set. Based on the path performance identifier set, the comprehensive performance level is calculated and sorted according to the order of nodes in the path. Paths with performance levels below the average value are replaced with candidate paths with higher performance levels. The path numbers and corresponding node parameters of the peak performance level are retained and an optimal set is generated, thus generating an optimal cross-chain path set.

9. The method for dynamic management and cross-chain transactions of digital assets based on blockchain and privacy computing according to claim 1, characterized in that, The specific steps for generating the cross-chain asset transfer registration form are as follows: Based on the preferred cross-chain path set, the nodes in the path are accessed sequentially to read the encrypted shard index and available time. Nodes with shorter execution cycles and missing indexes are recorded and marked with associated paths. After removing the marked paths, the remaining numbers are summarized to generate a list of available paths. Based on the available path list, node shards are called in order of path and index consistency is checked. Shards are decrypted in sequence and the complete digital asset file is assembled. The file is uploaded to the target chain to generate a transfer record and the record is associated with the path number to generate a cross-chain asset transfer registration form.

10. A dynamic management and cross-chain transaction system for digital assets based on blockchain and privacy computing, characterized in that, The method for dynamic management and cross-chain transaction of digital assets based on blockchain and privacy computing according to any one of claims 1-9, wherein the system comprises: Anomaly asset identification module: Based on the digital asset metadata stored in the blockchain and the encrypted sharding location of distributed nodes, it verifies the correspondence between ownership identifiers and transaction hashes and analyzes the latency distribution characteristics to generate a list of potential anomaly assets; Shard Risk Analysis Module: Based on the potential abnormal asset list, the events in the access log are processed in a time sequence, the permission status and reputation value are jointly analyzed, and the shard number with abnormal status is determined by using an isolated forest to generate a list of high-risk access shards; Cross-chain path filtering module: Based on the high-risk access shard list, extract the node operation parameters in the cross-chain transaction path one by one, and aggregate them with latency, lock-up, and processing duration to form a model. Use a long short-term memory network to filter out paths with total latency exceeding the range, and generate a set of executable transaction paths. Preferred Path Generation Module: Based on the executable transaction path set and inter-chain transmission link data, it quantitatively calculates the processing efficiency, bandwidth utilization, and data packet loss ratio of each node and comprehensively generates a performance level, replaces low-performance paths, and generates a preferred cross-chain path set. Asset cross-chain execution module: Based on the preferred cross-chain path set, it sequentially reads the shard data stored in the node and checks the index, removes paths that do not meet the execution cycle and lack shards, decrypts and combines shards in sequence, and generates a cross-chain asset transfer registration form.

Images