An improved verification cross-chain query method based on blockchain

Abstract

This invention discloses an improved cross-chain query verification method based on blockchain. First, the blockchain address to be queried is obtained. Then, the cross-chain dependency relationship of the address is queried in the inverted index of the dependent chains to quickly locate the relevant cross-chain transaction records. A cumulative hash verification algorithm based on RSA cryptography is used to verify the integrity of the query results by generating a global verification value and exclusion proof. Before querying the transaction chain, a bidirectional hash cross-verification algorithm is used to ensure data integrity. Intelligent sharding technology based on LSTM dynamic prediction is employed to predict transaction frequency based on the user's historical transaction sequence and dynamically allocate highly active users to the same shard to achieve load balancing. Transaction records are queried in parallel within the transaction chain shards based on the relevant address set, and integrated to generate a complete cross-chain transaction process. This invention effectively reduces the risk of single points of failure and improves the performance and reliability of cross-chain queries.

Description

Technical Field

[0001] This invention relates to the field of information security technology, and specifically to an improved cross-chain verification query method based on blockchain. Background Technology

[0002] With the rapid development of information technology, blockchain technology, with its unique advantages such as decentralization, data transparency, and immutability, has demonstrated enormous application potential in numerous fields. Taking supply chain management as an example, against the backdrop of rapid global economic development, the globalization of supply chains is becoming increasingly significant, with more and more suppliers relying on complex and vast global supply chain networks. If each supplier from different countries adopts an independent blockchain system to record and monitor key data, this fragmented data management approach will inevitably exacerbate the information silo problem within the supply chain ecosystem. Due to the lack of effective data sharing and interaction mechanisms, the transparency of the supply chain will be significantly reduced, and operational efficiency will be severely affected, posing a major challenge to the overall collaborative development of the supply chain. To achieve efficient information sharing and collaboration across multiple chains, cross-chain query technology has become a necessary solution. However, existing cross-chain query schemes suffer from low query efficiency and difficulty in verifying query integrity. For example, the Vassago scheme reduces the latency of cross-chain queries but neglects the verification of query results. Summary of the Invention

[0003] Purpose of the invention: The purpose of this invention is to provide an improved cross-chain query verification method based on blockchain to solve the problems of low query efficiency and difficulty in verifying query integrity in existing cross-chain query schemes.

[0004] Technical Solution: The present invention provides an improved cross-chain verification query method based on blockchain, comprising the following steps:

[0005] (1) Obtain the blockchain address to be queried;

[0006] (2) Query and associate the cross-chain dependency relationship of blockchain addresses in the inverted index of the dependent chain. The inverted index maps user addresses to related cross-chain transaction records to form an inverted index table from user to transaction, which can quickly and accurately locate all cross-chain transaction records of the user. If there is a cross-chain dependency, the relevant cross-chain address set and verification information are returned. The verification information includes the proof, accumulated hash value and modulus of each relevant transaction address.

[0007] (3) Use the cumulative hash verification algorithm based on the RSA cryptosystem to verify the integrity of the query results based on the verification information. The verification includes generating a global verification value and an exclusion proof, and checking the consistency between the exclusion proof and the global verification value to confirm that no data is missing.

[0008] (4) Before executing the transaction chain query, perform bidirectional hash cross-validation on the transaction chain stored data: obtain the target block from the block header on the chain. The hash value calculated from the block header information, and subsequent information to the block The recorded preceding block hash values ​​are respectively compared with the corresponding blocks stored off-chain locally. The hash values ​​calculated from the block header information are compared bidirectionally to ensure that the off-chain shard data to be queried is completely consistent with the original on-chain data, and that the data block sequence is continuous, without missing or tampered data.

[0009] (5) Based on the relevant cross-chain address set, query the corresponding transaction records in the shards of the transaction chain. The transaction chain adopts intelligent sharding technology based on LSTM dynamic prediction. By analyzing the user's historical transaction sequence, predict the future transaction frequency, and dynamically allocate or migrate the transactions of highly active users to the same shard based on the prediction results to achieve load balancing. Determine the shard number stored in each relevant address through the sharding table, query multiple shards in parallel to obtain transaction records, and integrate all the queried transaction records to generate a complete cross-chain transaction process.

[0010] Furthermore, in step (2), the inverted index is as follows: an inverted list is generated based on the user's address, the list stores information on all cross-chain transactions in which the user has participated, and fast access is achieved through hash mapping.

[0011] Furthermore, in step (2), the proof and accumulated hash value in the verification information are generated through hash operation and modulo operation.

[0012] Furthermore, in step (3), the cumulative hash verification algorithm is as follows: generate an RSA modulus based on a large prime number; select a generator; calculate a global verification value based on the user address and related cross-chain transaction address dataset; generate a proof excluding itself for each address element; and verify the integrity by checking the consistency between the proof and the global verification value during the verification phase.

[0013] Furthermore, in step (4), the bidirectional hash cross-validation specifically includes:

[0014] Forward verification: obtaining the target block on the chain. The hash value calculated from the block header information Corresponding blocks in local storage The hash value calculated from the block header information Compare and verify the data integrity;

[0015] Backward verification: obtaining subsequent blocks on the chain The parent hash stored in the block header (i.e., on-chain blocks) Forward block (hash record), and the corresponding block stored locally. The hash value calculated from the block header information A comparison is performed to verify the integrity of the data; only if the two comparisons are consistent can it be determined that the off-chain data is complete and continuous.

[0016] Furthermore, in step (5), the sharding strategy of the transaction chain is as follows: based on the intelligent sharding technology based on LSTM dynamic prediction, highly active user transactions are dynamically allocated or migrated to the same shard. Specifically, the intelligent sharding strategy based on LSTM dynamic prediction is as follows:

[0017] Data preparation and feature extraction: Collect historical transaction data of each user, construct a historical transaction quantity sequence in chronological order, and extract key features such as transaction timestamps and the number of transaction input and output addresses;

[0018] Model Training and Prediction: A one-way LSTM model is trained using historical transaction sequences; the model learns the evolution of transaction patterns over time and outputs a prediction of the number of transactions in future periods.

[0019] Sharding Decision and Migration: The system periodically runs the prediction model and lists the hot users who are about to enter a highly active state based on the prediction results. Then, based on the predicted number of transactions of users and the current load of each shard, shard allocation and user migration are performed to concentrate users who frequently transact with each other in the same shard and maintain a balanced predicted load among the shards. Finally, the record information on the shard table is updated.

[0020] Furthermore, in step (5), parallel query sharding includes: sending query requests to multiple shards simultaneously based on the user information on the sharding table, and aggregating the returned transaction records.

[0021] An electronic device according to the present invention includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the steps of any of the methods described herein.

[0022] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the methods described herein.

[0023] Beneficial Effects: Compared with existing technologies, this invention has the following significant advantages: First, this invention constructs an inverted index on the dependent chain. The inverted index maps user addresses to related cross-chain transaction records, forming an inverted index table from user to transaction. Specifically, each user address corresponds to an inverted list, which stores information about all cross-chain transactions in which the user has participated. This index structure enables the system to quickly locate a user's cross-chain transaction records without traversing the entire dependent chain, thus significantly reducing query time and resource consumption. Second, this invention innovatively designs a bidirectional hash cross-validation step in the transaction chain query process. This step establishes a bidirectional, closed-loop verification relationship between on-chain metadata and off-chain stored data, ensuring the integrity of off-chain sharded data. This solves the problem of unreliable query results caused by data tampering or damage during off-chain storage, and together with the RSA verification of the dependent chain, constitutes an end-to-end trusted cross-chain query system. Next, this invention creatively introduces a computationally efficient LSTM prediction model into the sharding management mechanism of the transaction chain. By learning the causal patterns of user transaction behavior along the time sequence, it can efficiently and quickly predict future transaction activity. The intelligent sharding based on this prediction not only proactively optimizes data distribution and reduces cross-shard transactions, but also significantly reduces the computational overhead and latency of intelligent decision-making due to the lightweight nature of the model itself, achieving an excellent balance between performance improvement and resource consumption. Finally, based on the LSTM prediction results, users who frequently transact with each other and their transactions are grouped into the same shard, maintaining a balanced prediction load across shards, and updating the record information on the sharding table. This sharding strategy not only reduces the risk of single points of failure but also optimizes the efficiency of data querying across shards, ultimately improving the performance of transaction chain queries. Attached Figure Description

[0024] Figure 1 This is a flowchart of the present invention;

[0025] Figure 2 This is a framework diagram of the present invention;

[0026] Figure 3 This is a sample diagram of the cross-chain transaction process obtained by querying according to the present invention;

[0027] Figure 4 This is an architecture diagram of the bidirectional hash cross-validation algorithm of the present invention. Detailed Implementation

[0028] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0029] like Figures 1-2 As shown, this embodiment of the invention provides a verifiable cross-chain query method supporting blockchain sharding, including the following steps:

[0030] (1) Obtain the blockchain address to be queried;

[0031] (2) The inverted index of the dependent chain queries the cross-chain dependency relationship of the blockchain address. The inverted index maps the user address to the relevant cross-chain transaction record to form an inverted index table from user to transaction, which can quickly locate the user's cross-chain transaction record. If a cross-chain dependency exists, the relevant cross-chain address set and verification information are returned. The verification information includes the proof, accumulated hash value and modulus of each relevant transaction address. The inverted index is as follows: an inverted list is generated based on the user address. The list stores the information of all cross-chain transactions in which the user has participated, and fast access is achieved through hash mapping. The proof and accumulated hash value in the verification information are generated through hash operation and modulus operation.

[0032] (3) The cumulative hash verification algorithm based on the RSA cryptosystem is used to verify the integrity of the query results based on the verification information. The verification includes generating a global verification value and an exclusion proof, and checking the consistency between the exclusion proof and the global verification value to confirm that the data is not missing. The cumulative hash verification algorithm is as follows: generate an RSA modulus based on a large prime number; select a generator; calculate the global verification value based on the user address and related cross-chain transaction address dataset; generate a proof that excludes itself for each address element; in the verification stage, the integrity is verified by checking the consistency between the proof and the global verification value.

[0033] (4) Subsequently, based on the verified cross-chain dependency information, perform bidirectional hash cross-validation on the target shard data before executing the transaction chain query: obtain the target block from the on-chain block header. The hash value calculated from the block header information, and subsequent information to the block The recorded preceding block hash values ​​are respectively compared with the corresponding blocks stored off-chain locally. The hash values ​​calculated from the block header information are compared bidirectionally to ensure that the off-chain shard data to be queried is completely consistent with the original on-chain data, and that the data block sequence is continuous, without missing or tampered data.

[0034] (5) Finally, based on the relevant cross-chain address set, the corresponding transaction records are queried in the shards of the transaction chain. The transaction chain uses an LSTM prediction strategy to predict the future transaction quantity of each user, and performs dynamic shard allocation and user migration based on the predicted transaction quantity of the user and the current load of each shard. After the migration is completed, the shard table is updated. During the query, the shard number of each relevant user address is determined through the shard table, and multiple shards are queried in parallel to obtain transaction records. All the queried transaction records are integrated to generate a complete cross-chain transaction process.

[0035] The LSTM-based prediction strategy is as follows:

[0036] Data preparation and feature extraction: Collect historical transaction data of each user, construct a historical transaction quantity sequence in chronological order, and extract key features such as transaction timestamps and the number of transaction input and output addresses;

[0037] Model Training and Prediction: A one-way LSTM model is trained using historical transaction sequences; the model learns the evolution of transaction patterns over time and outputs a prediction of the number of transactions in future periods.

[0038] LSTM Model Structure and Training: Construct an LSTM network suitable for medium-length time series prediction, with the following structure:

[0039] Input layer: Receives reshaped tensors with dimensions ( ,L,1).

[0040] LSTM layer: Contains two LSTM layers. The first layer has M hidden units (M = 128) and returns the complete sequence as input to the second layer; the second layer also has M hidden units and returns only the output of the last time step. By stacking LSTM layers, the network can learn more complex abstract features from the data.

[0041] Dropout layer: Add a Dropout layer after each LSTM layer, with a dropout probability set to 0.2 to prevent overfitting.

[0042] Fully connected output layer: Contains one neuron, uses ReLU as the activation function, and outputs the predicted number of transactions for the next time window.

[0043] Training process: Historical blockchain data is used as the training set, mean squared error is used as the loss function, and the Adam optimizer is used for iterative training until the model converges.

[0044] Predictive migration mechanism: Online prediction: For each active user address, the actual transaction count of the most recent L time windows is input into the trained LSTM model to obtain the predicted transaction count for the next period.

[0045] Sharding Decision and Migration: The system periodically runs the prediction model and lists the hot users who are about to enter a highly active state based on the prediction results. Then, based on the predicted number of transactions of users and the current load of each shard, shard allocation and user migration are performed to concentrate users who frequently transact with each other in the same shard and maintain a balanced predicted load among the shards. Finally, the record information on the shard table is updated.

[0046] Sharding table update: After the user migration is completed, the system updates the shard number corresponding to the user address in the global sharding mapping table to ensure that subsequent queries can accurately locate the address.

[0047] The specific process for shard allocation and user migration is as follows: The overall load of sharding is defined as... , representing the transaction load of shard set S within time window t, where there are n shard sets. S ={ S 1 , S 2 ,..., S n The set of transactions that need to be processed in each shard of} is L ={ L 1 , L 2 ,..., L n The average number of transactions that all shards need to process is} At time t, obtain the minimum load. The formula is shown below:

[0048] ;

[0049] The current shard load exceeds a preset threshold θ, triggering a migration mechanism. The system selects users to be migrated based on the target load of each shard. The selection criteria comprehensively consider the user's predicted transaction volume and data size, prioritizing the migration of "hot" users with large transaction volumes and small data volumes within the predicted time window t, in order to minimize migration overhead.

[0050] The specific process of this invention is as follows: Given a blockchain address that needs to be queried, input the address into the inverted index of the dependent chain to perform cross-chain transaction query, and verify the query results on the dependent chain. Then, based on the cross-chain transaction information returned by the dependent chain, query the relevant transactions in the shard of the corresponding chain information in the transaction chain. Finally, integrate all the queried transactions to form a complete transaction flow.

[0051] The batch processing RSA cumulative hash verification algorithm is used to verify the results obtained from queries on the dependency chain. The relevant parameters of the algorithm are as follows:

[0052] a. Large prime numbers p and q: Randomly generated large prime numbers p and q;

[0053] b. Modulus : Generating moduloes based on large prime numbers p and q ;

[0054] c. Generator : with modulus Coprime numbers;

[0055] d. Set of relevant cross-chain dependency addresses : These are all the cross-chain dependency addresses obtained from the query;

[0056] e. The calculated set of relevant cross-chain dependency addresses The corresponding hash value set :

[0057] f. Set of relevant cross-chain dependency addresses The corresponding verification information set π: (For each cross-chain dependency address obtained from the query, generate an address excluding itself.) Validation value );

[0058] g. Accumulated hash value The cumulative hash value of all related cross-chain dependent addresses of this user address on the dependent chain. ;

[0059] h. Finally, in the verification phase, for any element and its proof Through inspection Whether it holds true is used to verify the integrity of the data queried on the dependency chain.

[0060] like Figure 3 The image shows the complete process of participating in cross-chain transactions, including the following steps:

[0061] First, obtain the cross-chain related transaction set information obtained during the parallel query phase on the transaction chain;

[0062] Secondly, addresses on the same chain are integrated into the cross-chain related transaction set, that is, addresses 1, 2 and 3 on chain 1; addresses 4, 5 and 6 on chain 2; and addresses 7 and 8 on chain 3.

[0063] Subsequently, based on the verified cross-chain dependency information, the cross-chain behaviors of chain 1 and chain 2, and chain 2 and chain 3 are linked together, that is, the address 3 of chain 1 and the address 4 of chain 2, and the address 6 of chain 2 and the address 7 of chain 3 are connected.

[0064] This ultimately yields a complete cross-chain transaction path.

[0065] The transaction chain is verified for data integrity and authenticity using a two-way hash cross-validation algorithm. The relevant parameters of the algorithm are as follows:

[0066] Target block on the chain The block header hash value;

[0067] The corresponding block in local storage The hash value;

[0068] Off-chain subsequent blocks The parent hash value stored in the block header;

[0069] During the verification phase, it is necessary to verify the two hash values. and Perform forward verification and compare the two hash values. and Backward verification is performed; if the results of the two comparisons are consistent, the off-chain data can be determined to be complete and continuous.

[0070] like Figure 4 As shown, the bidirectional hash cross-validation algorithm is explained in detail below:

[0071] The mechanism consists of two phases: an on-chain phase and an off-chain phase. The on-chain phase contains raw data from multiple chains. Transaction data from different chains is stored in blocks on off-chain servers through full nodes on the chain.

[0072] Since the block header is the metadata part of each block, it contains the following six fields: version number (4 bytes), previous block hash (32 bytes), Merkle root hash (32 bytes), timestamp (4 bytes), block difficulty (4 bytes), and random number (4 bytes). These six fields are concatenated in order to form a serialized data set of 80 bytes. Then, the serialized data is hashed twice using the SHA-256 hash function to obtain the block header hash of block n. The detailed calculation scheme is shown in Algorithm 1.

[0073] Algorithm 1: Block Header Hash Generation

[0074] Input: Information recorded in the block header (version number: blockHeader.version, previous block hash: blockHeader.prevBlockHash, Merkle root: blockHeader.merkleRoot, timestamp: blockHeader.timestamp, block difficulty: blockHeader.bits, random number: blockHeader.nonce)

[0075] Output: 32-byte block hash value

[0076] Function CalculateBlockHash(blockHeader):

[0077] serializedData = blockHeader.version / / Version number

[0078] + blockHeader.prevBlockHash / / Previous block hash

[0079] + blockHeader.merkleRoot / / Merkle root

[0080] + blockHeader.timestamp / / Timestamp

[0081] + blockHeader.bits / / Block difficulty

[0082] + blockHeader.nonce / / Random number

[0083] hash1 = SHA256(serializedData)

[0084] hash2 = SHA256(hash1) / / Double SHA-256 hash

[0085] return hash2;

[0086] When storing each block, the first step is to sequentially concatenate the six fields contained in the block header of the off-chain record according to Algorithm 1, and then calculate the block n hash. Meanwhile, the six fields contained in the block header recorded on the chain are concatenated sequentially to calculate the block n hash. Subsequently, the block header of block n+1 on the chain is queried to obtain the hash value information of the previous block. Subsequently, the two hash values ​​were analyzed. and Perform forward verification and compare the two hash values. and Backward verification is performed. If the results of the two comparisons are consistent, the off-chain data can be determined to be complete and continuous. Otherwise, it means that the transaction data in the off-chain stored block is inconsistent with the on-chain data. After stopping the data download, the analysis is performed to determine which part of the hash value has a synchronization error.

Claims

1. An improved method for validating cross-chain queries based on blockchain, characterized in that, Includes the following steps: (1) Obtain the blockchain address to be queried; (2) Query the cross-chain dependency relationship of blockchain address in the inverted index of the dependent chain. The inverted index maps user address to corresponding cross-chain transaction record to form an inverted index table from user to transaction, which can quickly locate the user's cross-chain transaction record. If there is a cross-chain dependency, return the relevant cross-chain address set and verification information. The verification information includes the proof, accumulated hash value and modulus of each relevant transaction address. (3) Use the cumulative hash verification algorithm based on the RSA cryptosystem to verify the integrity of the query results based on the verification information. The verification includes generating a global verification value and an exclusion proof, and checking the consistency between the exclusion proof and the global verification value to confirm that no data is missing. (4) Before executing the transaction chain query, perform bidirectional hash cross-validation on the transaction chain stored data: obtain the target block from the block header on the chain. The hash value calculated from the block header information, and subsequent information to the block The recorded preceding block hash values ​​are respectively compared with the corresponding blocks stored off-chain locally. The hash values ​​calculated from the block header information are compared bidirectionally. (5) Based on the relevant cross-chain address set, query the corresponding transaction records in the shards of the transaction chain. The transaction chain adopts an intelligent sharding strategy based on LSTM dynamic prediction. By analyzing the user's historical transaction sequence, predict the future transaction frequency, and dynamically allocate or migrate high-activity user transactions to the same shard based on the prediction results to achieve load balancing. Determine the shard number stored in each relevant address through the sharding table, and query multiple shards in parallel to obtain transaction records. Integrate all queried transaction records to generate a complete cross-chain transaction process. The sharding strategy of the transaction chain is as follows: Based on the intelligent sharding technology based on LSTM dynamic prediction, dynamically allocate or migrate high-activity user transactions to the same shard. The intelligent sharding strategy based on LSTM dynamic prediction is as follows: Data preparation and feature extraction: Collect historical transaction data of each user, construct a historical transaction quantity sequence in chronological order, and extract key features such as transaction timestamps and the number of transaction input and output addresses; Model training and prediction: A one-way LSTM model is trained using historical transaction sequences; the model learns the evolution of transaction patterns over time and outputs a prediction of the number of transactions in future periods. Sharding Decision and Migration: The system periodically runs the prediction model and lists the hot users who are about to enter a high-activity state based on the prediction results. Then, based on the predicted number of transactions of users and the current load of each shard, it performs shard allocation and user migration, concentrating users who frequently transact with each other in the same shard and maintaining a balanced predicted load among the shards. Finally, the record information on the shard table is updated.

2. The improved verification cross-chain query method based on blockchain according to claim 1, characterized in that, In step (2), the inverted index is as follows: an inverted list is generated based on the user's address. The list stores information on all cross-chain transactions in which the user participates and enables fast access through hash mapping.

3. The improved verification cross-chain query method based on blockchain according to claim 1, characterized in that, In step (2), the proof and accumulated hash value in the verification information are generated through hash operation and modulo operation.

4. The improved verification cross-chain query method based on blockchain according to claim 1, characterized in that, In step (3), the cumulative hash verification algorithm is as follows: generate the RSA modulus based on large prime numbers; select generators; calculate the global verification value based on the user address and related cross-chain transaction address dataset; generate a proof excluding itself for each address element; During the verification phase, integrity is verified by checking the consistency between the proof and the global verification value.

5. The improved verification cross-chain query method based on blockchain according to claim 1, characterized in that, In step (4), the bidirectional hash cross-validation algorithm is as follows: for the target block on the chain... The hash value calculated from the block header information Corresponding blocks in local storage The hash value calculated from the block header information Perform forward data integrity verification; and simultaneously verify subsequent blocks on the chain. The parent hash stored in the block header Blocks on the blockchain Forward block The hash record and the corresponding block stored locally. The hash value calculated from the block header information Perform backward data integrity verification; if the forward data integrity verification and backward data integrity verification are consistent, it is determined that the off-chain data is complete and continuous.

6. The improved verification cross-chain query method based on blockchain according to claim 1, characterized in that, In step (5), parallel query sharding includes: sending query requests to multiple shards simultaneously based on the shard set, and aggregating the returned transaction records.

7. An electronic device comprising a processor and a memory, characterized in that The memory stores a computer program, and when the processor executes the computer program, it implements the steps of the method according to any one of claims 1-6.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1-6.

Images