Heterogeneous dynamic data right confirmation and infringement arbitration method based on semantic hash tree

Abstract

The application discloses a kind of heterogeneous dynamic data based on semantic hash tree and infringe arbitration method of right, comprising: according to the classification rule based on business semantic label, the data set to be right containing several data items is divided into multiple data subsets, and each data item is converted into metadata array, the metadata array at least includes the unique identifier and content digest of the data item;Respectively for each data subset, its corresponding metadata array is stored in a hash table, and a semantic hash tree is constructed based on the hash table, to obtain the subtree authentication value corresponding to each data subset, and each subtree authentication value is aggregated to generate the global authentication value representing the current state of the data set to be right;The global authentication value is written into blockchain for right storage evidence.In the application, only one global authentication value and necessary metadata need to be stored on the chain, and when the data is dynamically updated, only the global authentication value needs to be updated, so that any size data set can be efficiently righted.

Description

Technical Field

[0001] This invention relates to the fields of blockchain and data security technology, and in particular to a method for confirming ownership and arbitrating infringement of heterogeneous dynamic data based on semantic hash trees. Background Technology

[0002] With the rapid development of the data element market, the confirmation of data asset ownership has become a key issue restricting data circulation and trading. In existing technologies, blockchain-based data ownership confirmation schemes typically write data fingerprints, metadata, or registration records into the blockchain one by one, leveraging the blockchain's immutability and traceability to provide evidentiary support for data ownership confirmation. However, when the dataset to be confirmed contains a large number of data items, registering each data item separately on the blockchain leads to a linear increase in on-chain storage overhead with the number of data items, making storage and transaction costs unsustainable. Furthermore, real-world data element assets are typically in a state of continuous dynamic change, with frequent additions, modifications, and deletions of data items. Most existing solutions adopt the method of re-registering new versions of data as entirely new data, causing the on-chain records to continuously expand with each update. Summary of the Invention

[0003] This invention provides a heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees, in order to solve the technical problems of high storage cost and insufficient support for dynamic high-frequency updates in existing data ownership confirmation methods.

[0004] To address the aforementioned technical problems, this invention provides a method for heterogeneous dynamic data ownership confirmation and infringement arbitration based on semantic hash trees, comprising: According to the classification rules based on business semantic tags, the dataset to be confirmed is divided into multiple data subsets containing several data items, and each data item is converted into a metadata array, wherein the metadata array contains at least the unique identifier and content summary of the data item. For each data subset, its corresponding metadata array is stored in a hash table, and a semantic hash tree is constructed based on the hash table to obtain the subtree authentication value corresponding to each data subset. The subtree authentication values ​​are then aggregated to generate a global authentication value representing the current state of the dataset to be confirmed. The global authentication value is written into the blockchain for confirmation and evidence storage.

[0005] The aforementioned heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees first divides the dataset to be confirmed, containing several data items, into multiple data subsets according to classification rules based on business semantic tags. Each data item is then converted into a metadata array in a unified format, freeing the data ownership confirmation process from dependence on physical hierarchy or random arrangement. This achieves universal adaptation to cross-platform and multi-form data assets. Furthermore, for each data subset, its corresponding metadata array is stored in a hash table. A semantic hash tree is then constructed on top of the hash table and aggregated into a global authentication value. This compresses datasets of arbitrary size into a constant-length cryptographic fingerprint, achieving deep integration of data business logic and cryptographic authentication. Finally, only this global authentication value is used... By writing data to the blockchain to complete the confirmation and notarization, regardless of the number of data items in the dataset, only a global authentication value and necessary metadata need to be stored on the chain. The on-chain storage overhead is completely decoupled from the data size, achieving a constant-size on-chain storage. At the same time, since only one updatable global authentication value is stored on the chain, only the global authentication value needs to be updated when the data is dynamically updated, without the need to add new on-chain records for each update. This avoids the problem of on-chain storage expanding with the number of updates. It can be seen that the heterogeneous dynamic data confirmation and infringement arbitration method based on semantic hash trees in this invention constructs an efficient authentication data structure that deeply integrates business logic and cryptography. It can efficiently confirm the rights of datasets of any size, support dynamic data updates, and the on-chain cost does not increase with the data size or the number of updates. Attached Figure Description

[0006] Figure 1 This is a flowchart illustrating the heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees provided in the first embodiment of the present invention.

[0007] Figure 2 yes Figure 1 The diagram shows the specific process flow of S120 in the heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees.

[0008] Figure 3 yes Figure 2 The diagram shows the specific process flow of S121 in the heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash tree.

[0009] Figure 4 yes Figure 1 The diagram shows the semantic hash tree constructed in step S120 of the heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees.

[0010] Figure 5 yes Figure 1 The diagram shows the specific process flow of S150 in the heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees.

[0011] Figure 6 This is a flowchart illustrating the heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees provided in the second embodiment of the present invention.

[0012] Figure 7 yes Figure 6 The diagram shows a specific process flow diagram of S160 in the heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash tree.

[0013] Figure 8 yes Figure 6 The diagram shows another specific process flow diagram of S160 in the heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash tree. Detailed Implementation

[0014] To better understand the technical content of this invention, the technical solution of this invention will be further introduced and explained below with reference to the schematic diagram, but it is not limited thereto.

[0015] Reference Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees according to the present invention. In the embodiment shown in the figure, the heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees includes the following steps S110-S150: S110. According to the classification rules based on business semantic tags, the dataset to be determined is divided into multiple data subsets containing several data items, and each data item is converted into a metadata array, wherein the metadata array contains at least the unique identifier and content summary of the data item.

[0016] In this invention, the business semantic tag refers to a classification identifier that reflects the business logic attributes of data assets, including but not limited to the data source channel, business line, physical format type, time version range, or usage classification (such as training set, validation set, test set).

[0017] This step involves data abstraction and logical partitioning, where the dataset uploaded by the data owner to be confirmed is abstracted and categorized. Specifically, to facilitate flexible data management and access, the complete dataset to be confirmed can be divided into m subsets according to classification rules based on business semantic tags. , ... Furthermore, it can abstract each independent data item (such as an image, an audio clip, or a log record) in the dataset to be confirmed, and convert it into a metadata array in a unified format. That is, it can unify heterogeneous data of varying sizes and complex forms into a lightweight metadata array. This metadata array contains at least a unique identifier (URL) for locating the data item and a content digest for anchoring the integrity of the data content.

[0018] In this implementation, to support dynamic management throughout the entire lifecycle, the metadata array is preferably designed as follows:<URL,State,Digest> The triple, or metadata array, consists of a unique identifier field, a state field, and a content digest field. The unique identifier (URL) is used to uniquely identify and locate the data item. It can use a hierarchical naming format like "protocol: / / organization name / resource category / resource name," corresponding to both local file system paths and object-based cloud storage services. The state field (State) is a discrete value indicating whether the data item is currently valid or deleted; for example, State=0 indicates valid, and State=1 indicates deleted. The content digest field (Digest) is a digest of the data item's content, calculated using a cryptographic hash function, i.e., Digest=Hash(d). The Digest ensures the integrity of the data content; any modification to the data content will change the Digest. The Hash function can be a cryptographic hash function such as Keccak-256.

[0019] S120. For each data subset, store its corresponding metadata array into a hash table, and construct a semantic hash tree based on the hash table to obtain the subtree authentication value corresponding to each data subset. Then, aggregate the subtree authentication values ​​to generate a global authentication value representing the current state of the dataset to be confirmed.

[0020] This step involves constructing an off-chain hash table and a semantic hash tree. For each data subset divided in step S110, an independent hash table is constructed, and all metadata arrays within the corresponding data subset are mapped as entries into the slots of the hash table. Then, a semantic hash tree is constructed on each hash table and aggregated into a global root hash to achieve compact authentication for datasets of any size.

[0021] Therefore, in this invention, the semantic hash tree is a cryptographic authentication data structure constructed hierarchically based on business semantic tags. Its construction logic is as follows: a heterogeneous massive dataset is divided into multiple logically isolated semantic subsets (i.e., data subsets) according to the business semantic tags; then, an independent hash table and semantic hash tree are constructed for each semantic subset. The semantic hash tree constructed in this invention is freed from the constraints of physical layers, achieving a deep integration of data business logic and cryptographic authentication. While ensuring cryptographic tamper-proof characteristics, it can also achieve efficient local dynamic updates and ownership confirmation of heterogeneous massive data based on semantic isolation.

[0022] Specifically, such as Figure 2 As shown, step S120 includes the following steps S121-S124: S121. For each subset of data, construct a hash table based on the open addressing strategy, and map the metadata array to the corresponding slot in the hash table using the unique identifier in the metadata array as the index key.

[0023] In this step, a hash table is constructed for each subset of data. (i=1,2,3...), create a corresponding hash table Hash table capacity Based on the size of the data subset | | and the preset load factor α are determined, that is = This ensures the hash table's capacity is a power of 2, facilitating the construction of a perfect binary semantic hash tree; the load factor α is used to balance space utilization and query efficiency, and can be set to, for example, α = 0.75. Furthermore, the hash table... Each slot is initialized as an empty slot, indicating a state where no data items are stored. Subsequently, a subset of data is stored using the URL of the data item as the index key. The metadata array corresponding to each data item is inserted into the hash table one by one.

[0024] In this embodiment, in order to achieve constant-level fast lookup, the hash table preferably adopts an open addressing strategy, that is, the hash table is an off-chain hash table based on the open addressing strategy. During the data insertion process, when a hash collision occurs, a dual hashing method including a main hash function and a step hash function is used to calculate the alternative slot on the collision path, so as to realize the fast location and lookup of the metadata array in the corresponding data subset, thereby ensuring retrieval efficiency.

[0025] Understandably, since the hash tables corresponding to each subset of data are independent of each other, multiple hash tables can be built in parallel to improve efficiency. Furthermore, semantic hash trees can be built on top of each hash table in parallel.

[0026] Specifically, such as Figure 3 As shown, step S121, which maps the metadata array to the corresponding slots in the hash table using the unique identifier in the metadata array as the index key, specifically includes the following steps S1211-S1213: S1211. Determine the initial slot in the hash table for the unique identifier of the metadata array according to the master hash function.

[0027] Specifically, the main hash function The initial slot can be calculated using a polynomial rolling hash function, which can be achieved using the formula... Calculate the initial slots, where k is the input string with length n, and M is the total capacity of the hash table.

[0028] S1212. If the initial slot is an available slot, store the metadata array into the initial slot.

[0029] In this embodiment, available slots include slots in the hash table where the lookup status field indicates a deleted status and empty slots that do not store any data items.

[0030] S1213. If the initial slot is occupied, calculate the probe step size according to the step size hash function, and use the probe formula to probe the alternative slots sequentially based on the probe step size until an available slot is found, and store the metadata array into the available slot.

[0031] In this step, if the initial slot is already occupied, the hash function based on the step size is used. Calculate the detection step size and, combined with the slots where collisions occurred in the previous step, calculate the detection position for the next step. Then, for the i-th (i=1,2,3....) detection, the slot index is calculated as follows: Probe sequentially until an available slot is found in the hash table and the data is inserted.

[0032] Preferably, the step-size hash function can be the FNV-1a hash function or a variant of the FNV-1a hash function. Furthermore, to ensure that the probe sequence covers all slots in the hash table, the step size needs to be ensured. The hash table capacity is coprime to the hash table capacity. Since the hash table capacity in this embodiment is a power of 2, it is only necessary to ensure... It can be an odd number.

[0033] Furthermore, in some embodiments, empty slots in the hash table can be filled with preset dummy values, allowing the empty slots to participate as leaf nodes in the hash concatenation calculation of the semantic hash tree. For example, empty slots in the hash table can also be filled with "0".

[0034] For steps S1211-S1213 above, the double hash open addressing strategy can effectively reduce clustering and ensure the space utilization and query efficiency of the hash table.

[0035] S122. Construct corresponding semantic hash trees on each of the hash tables, and calculate the hash value of the metadata array in each slot of each hash table as the leaf node hash value.

[0036] In this embodiment, a semantic hash tree is constructed on each hash table to obtain the subtree authentication value corresponding to each data subset.

[0037] Specifically, for hash tables Each slot in the hash table is treated as a leaf node. For a slot containing a stored metadata array, the hash value of that metadata array is calculated using the formula Hash(URL). State Digest), as the hash value of the leaf node, where This indicates a data concatenation operation; for empty slots, a preset dummy value is used to fill the slot and a hash value is calculated as the value of the leaf node.

[0038] S123. Calculate the hash value of the parent node layer by layer from bottom to top until the hash value of the root node corresponding to the hash table is obtained, and use the hash value of the root node as the subtree authentication value of the data subset.

[0039] In this step, for each hash table, using the leaf node hash value as the bottom layer, adjacent leaf node hash values ​​are concatenated pairwise for hash calculation from bottom to top until the top layer corresponding to the hash table is obtained. The hash value of the node, wherein the hash value of the top-level node is the same as the hash value of the root node.

[0040] Specifically, the hash values ​​of each pair of adjacent leaf nodes are concatenated and hashed again to obtain the hash value of the node at the next higher level, i.e., the hash value of the parent node; the above process is repeated until the root node is calculated, and the value of the root node is the subtree authentication value of the data subset.

[0041] S124. The authentication values ​​of each subtree are concatenated in a preset order, and the concatenation result is hashed to obtain the global authentication value.

[0042] In this step, the authentication values ​​of each subtree are... , , ..., Aggregate into a global authentication value R, i.e., R = Hash( … ).

[0043] Based on the above steps, a hash table-based cryptographic authentication data structure is constructed to realize the authentication of massive heterogeneous data elements. Specifically, instead of using the original data as the input for authentication, a hash table storing the metadata array corresponding to all original data items is used as the data to be authenticated. A semantic hash tree is constructed on top of the hash table corresponding to each data subset, and each leaf node of the semantic hash tree corresponds sequentially to each entry in the hash table, with the leaf node hash value being the hash value of the hash table entry. The hash value of the parent node is recursively calculated from bottom to top using the same cryptographic hash function Hash(·) until the hash value of the root node of the hash tree is calculated. For m hash tables, m hash trees are constructed, resulting in m root values. Each root value is considered as the subtree authentication value of a data subset of a certain category. Finally, a top-level root node connects these m tree roots, as shown below. Figure 4 As shown, this is equivalent to merging m subtrees into a hash tree, and finally calculating a root node hash value R that can authenticate the entire large dataset.

[0044] In this invention, the global authentication value R is the unique cryptographic commitment to the current state of the entire dataset to be authenticated. If the content or state of any data item changes, or if any subset of data is altered, the recalculated R will also change accordingly.

[0045] S130. Write the global authentication value into the blockchain for confirmation and evidence storage.

[0046] This step is an on-chain registration and rights confirmation step, which can be written by calling the smart contract deployed on the blockchain.

[0047] Preferably, in some embodiments, to prevent malicious entities from preemptively registering using copies of others' data, a compliance review can be conducted before the on-chain registration and rights confirmation step. This can be done by a trusted institution conducting a compliance review of the data source of the dataset to be registered. After the compliance review is passed, a globally unique dataset identifier assigned by the trusted institution is obtained. The global authentication value, along with the dataset identifier, data owner identity information, and timestamp, is then written into the blockchain for rights confirmation and evidence storage. This pre-review mechanism cuts off the path for malicious entities to preemptively register using copies of others' data at the source.

[0048] Specifically, data source compliance materials are submitted to a trusted institution (such as the Data Element Market Infrastructure (DEMI)). These materials can include relevant supporting documentation for the data's production, processing, or collection, proof of business system operation, technical reports on data anonymization, privacy agreements, and third-party compliance audit reports. The trusted institution reviews the submitted materials for compliance. Upon approval, a globally unique dataset ID is assigned to the dataset, along with metadata including the dataset ID, dataset description, timestamp, and institution signature. This metadata, along with the globally authenticated value, is then written to the blockchain via a smart contract for notarization and ownership verification. During the writing process, the smart contract first verifies the uniqueness of the dataset ID and the authoritative signature. After successful verification, the on-chain registration and ownership verification steps are executed. Furthermore, after the smart contract completes the notarization, an on-chain event is triggered, facilitating off-chain synchronization of the ownership status.

[0049] Based on the above steps, since only a global authentication value R and a small amount of necessary metadata need to be stored on the chain, the on-chain storage overhead is completely decoupled from the data scale, realizing constant-size on-chain storage, which is suitable for large-scale, heterogeneous and frequently updated data ownership confirmation scenarios.

[0050] S140. In response to the ownership verification instruction containing the target data item, based on the location path of the target data item in the hash table and the cascading path of the semantic hash tree, generate a membership relationship proof path containing the set of sibling nodes in the off-chain.

[0051] This step is for generating proof of data item ownership. When a data ownership dispute occurs, the data owner needs to prove that a target data item belongs to a dataset with confirmed ownership. This invention achieves ownership verification through the generation of proof of data item ownership and lightweight on-chain verification.

[0052] Specifically, when generating the data item ownership proof, the corresponding slot is located in the hash table to which the target data item belongs based on its URL, and then the leaf node position of the slot in the corresponding semantic hash tree is determined. The membership proof path includes first path information and second path information. The first path information is the set of sibling nodes on the path from the leaf node corresponding to the target data item to the root node of the subtree corresponding to the data subset in which it is located. The second path information is the subtree authentication value corresponding to each of the other data subsets in the same dataset to be confirmed, excluding the data subset in which the target data item is located.

[0053] S150. Based on the metadata array of the target data item and the membership proof path, the ownership is verified using the verification smart contract on the blockchain.

[0054] This step is a lightweight on-chain verification step, specifically, as follows: Figure 5 As shown, step S150 includes the following steps S151-S155: S151. Based on the metadata array of the target data item and the membership proof path, the authentication value is recalculated using the verification smart contract on the blockchain that stores the global authentication value of the target data item.

[0055] Specifically, in this step, a verification smart contract deployed on the blockchain can calculate the leaf node hash value based on the metadata array. Based on the leaf node hash value and the sibling node set in the first path information, a bottom-up hash calculation is performed to obtain the verification subtree authentication value. That is, using the leaf node hash value and the sibling node hash values ​​in the first path information, a bottom-up hash concatenation calculation is performed layer by layer to recursively calculate the verification subtree authentication value of the data subset containing the target data item. Then, the verification subtree authentication value is aggregated with the authentication values ​​of the remaining subtrees in the second path information to obtain the global verification authentication value. This global verification authentication value is then compared with the pre-stored global authentication value on the chain. If they match, it proves that the target data item indeed belongs to the dataset with proven ownership, and the contract automatically outputs the judgment result confirming ownership.

[0056] S152. Compare the recalculated authentication value with the global authentication value stored on the blockchain.

[0057] S153. If the comparison is consistent, it is confirmed that the target data item has been confirmed as owned, and the arbitration result confirming ownership is output.

[0058] For the verification steps consisting of S151-S153, the verification process only involves lightweight operations such as hash concatenation and comparison, avoiding complex cryptographic calculations such as zero-knowledge proofs commonly found in existing technologies, and truly realizing low-gas on-chain infringement arbitration.

[0059] S154. Record the verified status on the blockchain and trigger the event.

[0060] S155. Generate an ownership certificate based on the event. The ownership certificate shall at least include a content summary of the target data item, the global authentication value, the ownership confirmation timestamp, and the data owner's identity information, so that the data owner can submit it as arbitration evidence to judicial or regulatory agencies.

[0061] After the verification smart contract completes the verification and passes the verification, it records the verification status on the blockchain and triggers an event; the off-chain system generates an ownership certificate based on the event. The ownership certificate includes at least a content summary of the data item, the global authentication value, the ownership confirmation timestamp, and the data owner's identity information, so that the data owner can submit it to the judiciary or regulatory agency as arbitration evidence.

[0062] Based on steps S154-S155 above, ownership verification can also be linked to judicial procedures and applied to infringement arbitration scenarios. That is, after verification is successful, the verification smart contract records the successful verification status on the chain and triggers an event. After the event is captured off-chain, a legally valid ownership certificate can be generated. This certificate contains necessary metadata such as the hash value of the data item, the authentication value of the dataset, the time of ownership confirmation, and the identity information of the data owner. In subsequent judicial or arbitration proceedings, the data owner can submit the plaintext of the data item to the court or appraisal institution in a secure manner. The appraisal institution uses the same hash function to recalculate the content digest of the data item and compares it with the Digest field on the ownership certificate, thereby binding the on-chain ownership certificate with the offline plaintext evidence.

[0063] As can be seen, in the above scheme, this embodiment's heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees stores the corresponding metadata array of each partitioned data subset into a hash table, then constructs a semantic hash tree on the hash table and aggregates it into a global authentication value, thereby compressing a dataset of arbitrary size into a cryptographic fingerprint of constant length, achieving a deep integration of data business logic and cryptographic authentication; finally, only this global authentication value is written to the blockchain to complete the ownership confirmation and notarization, so that no matter how many data items the dataset contains, only one global authentication value and necessary metadata need to be stored on the chain. The overhead is completely decoupled from the data size, achieving constant-size on-chain storage. Furthermore, since only one updatable global authentication value is stored on-chain, only the global authentication value needs to be updated when the data is dynamically updated, eliminating the need to add new on-chain records for each update. This avoids the problem of on-chain storage expanding with the number of updates. Therefore, this invention, based on a semantic hash tree-based heterogeneous dynamic data ownership confirmation and infringement arbitration method, constructs a highly efficient authentication data structure that deeply integrates business logic and cryptography. It can efficiently confirm ownership of datasets of any size, support dynamic data updates, and ensure that on-chain costs do not increase with data size or the number of updates.

[0064] Reference Figure 6 , Figure 6 This is a flowchart illustrating the second embodiment of the heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees according to the present invention. Following step S130 in the above embodiment, the heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees in this embodiment may further include the following steps S160-S170: S160. In response to a dynamic change instruction for the dataset to be verified, the corresponding hash table and semantic hash tree are locally updated according to the type of the dynamic change instruction, and the updated global authentication value is obtained by re-aggregation.

[0065] In this invention, the change instruction may include at least one of the addition, modification and deletion instructions, and the dynamic change instruction may be for a data item in a data subset or for a data subset in a dataset to be determined.

[0066] When the dynamic change instruction is a change instruction for a data item, such as Figure 7 As shown, the step of locally updating the corresponding hash table and semantic hash tree, and then re-aggregating to obtain the updated global authentication value, specifically includes the following steps S161a-S163a: S161a. Update the metadata array corresponding to the data item targeted by the change instruction in the corresponding hash table.

[0067] In this step, in response to a change instruction for a data item in the dataset to be confirmed, the metadata array corresponding to that data item in the hash table is updated.

[0068] S162a. Along the path from the leaf node to the root node corresponding to the metadata array, locally update the semantic hash tree to obtain the updated subtree authentication value.

[0069] S163a. ​​The updated subtree authentication value is re-aggregated with the authentication values ​​of other subtrees in the same dataset to be verified, to obtain the updated global authentication value.

[0070] Based on the above steps S161a-S163a, the update only occurs in the local path of a single semantic hash tree, without touching the semantic hash trees corresponding to other data subsets in the same dataset to be confirmed.

[0071] Specifically, in response to a change instruction that modifies or deletes a data item in the dataset to be confirmed, the corresponding data in the slots of the metadata array is updated accordingly; or in response to a new instruction, an available slot is located in the hash table, and the metadata array corresponding to the new data item is written into the available slot to update the slot; starting from the leaf node corresponding to the updated slot, only the local paths associated with the leaf node in the semantic hash tree are hashed to obtain the updated global authentication value, and then the global authentication value is overwritten with the updated global authentication value through a smart contract on the blockchain.

[0072] For example, when a new data item needs to be added, an available slot is located in the hash table, and the metadata array corresponding to the new data item is written to that slot; when a data item needs to be modified, the slot corresponding to the data item is located in the hash table, and the content summary field of the metadata array in that slot is updated; when a data item needs to be deleted, the status field in the metadata array corresponding to the data item is updated to indicate that it has been deleted.

[0073] Based on the above design, the State field in the metadata array can support flexible and dynamic data management. While maintaining the stability of the index structure, it can meet the dynamic change requirements of data and enable logical deletion and reuse to be completed in place in the hash table slot.

[0074] When the dynamic change instruction is a change instruction for a data item, such as Figure 8 As shown, the step of locally updating the corresponding hash table and semantic hash tree, and then re-aggregating to obtain the updated global authentication value, specifically includes the following steps S161b-S162b: S161b If the dynamic change instruction is to add a new data subset, then construct a hash table of the new data subset and its corresponding semantic hash tree to obtain the new subtree authentication value, and re-aggregate the new subtree authentication value with the original other subtree authentication values ​​to obtain the updated global authentication value.

[0075] S162b. If the dynamic change instruction is to delete a subset of data, then delete the corresponding hash table and its corresponding semantic hash tree, and re-aggregate the authentication values ​​of the remaining subtrees to obtain the updated global authentication value.

[0076] Based on steps S161b-S162b, in the above-mentioned category-level dynamic changes, all other existing data subsets do not need to be recalculated, demonstrating significant flexibility and execution efficiency.

[0077] S170. Write the updated global authentication value into the blockchain to overwrite the original global authentication value.

[0078] In this step, the blockchain smart contract is invoked to replace the global authentication value stored on the chain with the updated global authentication value.

[0079] Specifically, in this embodiment, when a new data item needs to be added, the available slot is first searched in the corresponding hash table according to the URL of the data item. There are two types of available slots: one is that the slot is empty (never occupied); the other is that the State field of the slot indicates that it has been deleted and can be reused. After finding an available slot, the metadata array (URL, State=0, Digest=Hash(d)) corresponding to the data item is written to the slot. Then, the semantic hash tree is locally updated along the path from the leaf node to the root node corresponding to the slot, and the updated global authentication value R' is obtained by re-aggregation. Finally, R' is written to the blockchain.

[0080] When it is necessary to delete a data item that has been confirmed for ownership, the process first locates the corresponding slot in the hash table to which the data item belongs, using the URL as the index key. After locating the target slot, instead of directly clearing it, the State field of the metadata entry in that slot is updated to indicate a deleted state (e.g., updating State from 0 to 1). This logical deletion method maintains the continuity of the open addressing probe sequence, ensuring that other data items in the hash table can still be retrieved normally. Subsequently, along the path from the leaf node to the root node corresponding to the target slot, the hash values ​​of each node on the path are recalculated layer by layer from bottom to top. Other unaffected nodes do not need to be recalculated, thus obtaining the updated subtree authentication value of the data subset. Then, the updated subtree authentication value is re-aggregated with other unchanged subtree authentication values ​​to obtain the updated global authentication value R'. Finally, the smart contract on the blockchain is called to write R' to the blockchain to overwrite the original global authentication value R, completing the synchronous update of the ownership status on the chain.

[0081] When it is necessary to modify the content of an existing data item, first locate the slot corresponding to the data item in the hash table, then update the Digest field of the metadata entry in the slot to the hash value of the new data content; subsequently, locally update the semantic hash tree along the path from the corresponding leaf node to the root node, and re-aggregate to obtain the updated global authentication value R', and finally write R' into the blockchain.

[0082] When a new data subset needs to be added, a hash table and semantic hash tree are built only for the new data subset, and its subtree authentication value is calculated. The new subtree authentication value is then re-aggregated with the original subtree authentication values ​​to obtain the updated global authentication value.

[0083] When a subset of data needs to be deleted, the hash table and semantic hash tree corresponding to that subset are discarded, and the authentication values ​​of the remaining subtrees are re-aggregated to obtain the updated global authentication value.

[0084] Based on the above steps S160-S170, in reality, data element assets are constantly changing dynamically. When the addition, modification, and deletion of data items occur frequently, the method of this embodiment can strictly limit the computational complexity of dynamic updates to a very small local path through the local update mechanism of the State field in the metadata array and the semantic hash tree. This will not cause the reconstruction of the entire semantic hash tree, which can efficiently support dynamic data updates and adapt to the high-frequency dynamic data element market. Furthermore, the immutability of blockchain ensures that all historical state changes can still be completely traced back and audited.

[0085] As can be seen, in the above scheme, by abstracting massive data items into a unified metadata array and storing it in a hash table, and then constructing a semantic hash tree on top of the hash table and aggregating it into a global authentication value, the dataset of any size is compressed into a cryptographic fingerprint of constant length, achieving a deep integration of data business logic and cryptographic authentication. Finally, only this global authentication value is written to the blockchain to complete the confirmation and notarization, so that no matter how many data items the dataset contains, only one global authentication value and necessary metadata need to be stored on the chain. The on-chain storage overhead is completely decoupled from the data size, achieving constant-size on-chain storage. At the same time, the semantic hash tree supports local updates along the path from the leaf node to the root node. The addition, modification and deletion of data only require local recalculation of the authentication path and updating the global authentication value on the chain to complete the synchronization of the confirmation status, without generating new on-chain records, further avoiding the problem of on-chain storage expanding with the number of updates.

[0086] It should be understood that the above method embodiments are described as a series of actions for the sake of simplicity. However, those skilled in the art should know that the present invention is not limited by the described order of actions. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0087] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to depart from the present invention. The scope of the technical solutions in the embodiments should be included within the protection scope of this invention.

Claims

1. A method for determining ownership and arbitrating infringement of heterogeneous dynamic data based on semantic hash trees, characterized in that, include: According to the classification rules based on business semantic tags, the dataset to be confirmed is divided into multiple data subsets containing several data items, and each data item is converted into a metadata array, wherein the metadata array contains at least the unique identifier and content summary of the data item. For each data subset, its corresponding metadata array is stored in a hash table, and a semantic hash tree is constructed based on the hash table to obtain the subtree authentication value corresponding to each data subset. The subtree authentication values ​​are then aggregated to generate a global authentication value representing the current state of the dataset to be confirmed. The global authentication value is written into the blockchain for confirmation and evidence storage.

2. The method for heterogeneous dynamic data ownership confirmation and infringement arbitration based on semantic hash trees as described in claim 1, characterized in that, The heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees also includes: In response to a dynamic change instruction for the dataset to be verified, the corresponding hash table and semantic hash tree are locally updated according to the type of the dynamic change instruction, and then re-aggregated to obtain the updated global authentication value. The updated global authentication value is written into the blockchain to overwrite the original global authentication value.

3. The method for determining ownership and arbitrating infringement of heterogeneous dynamic data based on semantic hash trees as described in claim 2, characterized in that, When the dynamic change instruction is a change instruction for a data item, the step of partially updating the corresponding hash table and semantic hash tree, and re-aggregating to obtain the updated global authentication value, specifically includes: Update the metadata array corresponding to the data item targeted by the change instruction in the corresponding hash table; The semantic hash tree is locally updated along the path from the leaf node to the root node corresponding to the metadata array to obtain the updated subtree authentication value; The updated subtree authentication value is re-aggregated with the authentication values ​​of other subtrees in the same dataset to be verified, to obtain the updated global authentication value; When the dynamic change instruction is a change instruction at the data subset level, the step of partially updating the corresponding hash table and semantic hash tree, and re-aggregating to obtain the updated global authentication value, specifically includes: If the dynamic change instruction is to add a new data subset, then a hash table of the new data subset and its corresponding semantic hash tree are constructed to obtain the new subtree authentication value, and the new subtree authentication value is re-aggregated with the original other subtree authentication values ​​to obtain the updated global authentication value. If the dynamic change instruction is to delete a subset of data, then the corresponding hash table and its corresponding semantic hash tree are deleted, and the authentication values ​​of the remaining subtrees are re-aggregated to obtain the updated global authentication value.

4. The method for determining ownership and arbitrating infringement of heterogeneous dynamic data based on semantic hash trees as described in claim 1, characterized in that, The heterogeneous dynamic data ownership confirmation and infringement arbitration method based on semantic hash trees also includes: In response to an ownership verification instruction containing a target data item, a membership proof path containing a set of sibling nodes is generated off-chain based on the location path of the target data item in the hash table and the concatenation path of the semantic hash tree. Ownership verification is performed using the metadata array of the target data item and the membership proof path on the blockchain through a verification smart contract.

5. The method for heterogeneous dynamic data ownership confirmation and infringement arbitration based on semantic hash trees as described in claim 4, characterized in that, The step of verifying ownership using a verification smart contract on the blockchain based on the metadata array of the target data item and the membership proof path specifically includes: Based on the metadata array of the target data item and the membership proof path, the authentication value is recalculated using the verification smart contract on the blockchain that stores the global authentication value of the target data item; The recalculated authentication value is compared with the global authentication value stored on the blockchain; If the comparison is consistent, the ownership of the target data item is confirmed, and the arbitration result confirming ownership is output.

6. The method for heterogeneous dynamic data ownership confirmation and infringement arbitration based on semantic hash trees as described in claim 5, characterized in that, After confirming that the target data item has been properly assigned ownership, the process further includes: Record the verified status on the blockchain and trigger the event; A certificate of ownership is generated based on the event. The certificate of ownership includes at least a content summary of the target data item, the global authentication value, the ownership confirmation timestamp, and the data owner's identity information, so that the data owner can submit it as arbitration evidence to judicial or regulatory agencies.

7. The method for heterogeneous dynamic data ownership confirmation and infringement arbitration based on semantic hash trees as described in claim 1, characterized in that, For each data subset, the corresponding metadata array is stored in a hash table, and a semantic hash tree is constructed based on the hash table to obtain the subtree authentication value corresponding to each data subset. Specifically, this includes: For each subset of data, a hash table based on an open addressing strategy is constructed, and the metadata array is mapped to the corresponding slot in the hash table using the unique identifier in the metadata array as the index key. A corresponding semantic hash tree is constructed on each of the hash tables, and the hash value of the metadata array in each slot of each hash table is calculated as the hash value of the leaf node. The hash value of the parent node is calculated layer by layer from bottom to top until the hash value of the root node corresponding to the hash table is obtained, and the hash value of the root node is used as the subtree authentication value of the data subset.

8. The method for heterogeneous dynamic data ownership confirmation and infringement arbitration based on semantic hash trees as described in claim 7, characterized in that, The step of mapping the metadata array to the corresponding slots in the hash table using the unique identifier in the metadata array as the index key specifically includes: The initial slot of the unique identifier of the metadata array in the hash table is determined according to the main hash function; If the initial slot is an available slot, the metadata array is stored in the initial slot; If the initial slot is already occupied, the probe step size is calculated based on the step-size hash function, and alternative slots are probed sequentially using the probe formula based on the probe step size until an available slot is found. The metadata is then... The array is stored in the available slot.

9. The method for determining ownership and arbitrating infringement of heterogeneous dynamic data based on semantic hash trees as described in claim 7, characterized in that, The process of calculating the hash value of the parent node layer by layer from bottom to top until the hash value of the root node corresponding to the hash table is obtained specifically includes: For each hash table, the leaf node hash value is used as the bottom layer, and adjacent leaf node hash values ​​are concatenated and hashed from bottom to top until the top-level node hash value corresponding to the hash table is obtained. The top-level node hash value is the root node hash value.

10. The method for determining ownership and arbitrating infringement of heterogeneous dynamic data based on semantic hash trees as described in claim 1, characterized in that, Before writing the global authentication value into the blockchain for rights confirmation and evidence storage, the following steps are also included: The data sources for the data to be confirmed by trusted institutions shall be subject to compliance review. After the compliance review is passed, a globally unique dataset identifier assigned by the trusted institution is obtained. The global authentication value, the dataset identifier, the data owner's identity information, and the timestamp are then written into the blockchain for confirmation and evidence storage.

Images