A Blockchain Cross-Network Anchoring Protection Method for Archive Channels Based on a Governance Model

Abstract

This invention relates to the field of information security and trusted data storage, and discloses a cross-network anchoring protection method for archive channels based on a governance big data model. On the sending side, the business data stream is content-defined and fragmented at the cross-network gateway, generating fragment fingerprints and forming KZG vector commitments. Anchor messages are constructed and written to the consortium blockchain via HotStuff-like consensus. On the receiving side, the commitments are recalculated and their consistency with on-chain records is verified to achieve cross-regional authentication. The governance big data model reads the local immutable evidence offline, outputs risk assessment and explanatory audit reports, and triggers human-machine review according to uncertainty and sovereignty rules, achieving traceable and third-party verifiable cross-network transmission protection.

Description

Technical Field

[0001] This invention relates to the field of information security and trusted data storage, and in particular to a cross-network anchoring protection method for archive channels based on a governance model. Background Technology

[0002] Electronic archives (such as map spatial data and physical data archives based on a two-way kernel of business and archives) typically contain core entity information, circulation information, protection and restoration records, appraisal and evaluation conclusions, and multi-source business data associated with the archived object (such as cultural relics / physical objects), possessing high confidentiality and sovereignty attributes. To meet security management requirements, related business systems are often deployed on government intranets or private networks, maintaining physical isolation from the Internet or other business networks. Existing cross-network archiving and exchange generally rely on cross-network gateways, security isolation and information exchange systems, and unidirectional transmission channels to transmit data from the sending side to the receiving side, while simultaneously employing access control, log auditing, electronic signatures, and cryptographic hash verification to achieve secure control and post-transmission traceability. On the other hand, to enhance evidence preservation and non-repudiation, some scenarios have begun to introduce consortium blockchain evidence preservation, uploading key business summaries or timestamp information to the blockchain to provide consistent reconciliation evidence across institutions.

[0003] However, the aforementioned existing technologies still have shortcomings in scenarios involving cross-network archiving and transmission of documents, mainly in the following aspects: 1. Under physical isolation and unidirectional channel conditions, it is difficult for the sending and receiving sides to form an aligned and verifiable chain of transmission evidence. Existing methods based on single-point logs or single digests are difficult to support off-site authentication, and are not easy to be penetrated for verification when link interception, replacement or replay occurs.

[0004] 2. The existing archiving and auditing mechanisms are insufficient for internal personnel risk response, which may result in the physical loss or logical deletion of locally stored evidence, leading to the loss or untraceability of key evidence, making it difficult to meet the needs of cross-network supervision and liability determination.

[0005] 3. Existing single-track archiving is prone to creating blind spots in cross-network supervision, while dual-track archiving lacks a unified mechanism to balance the consistency between the physical association information of the archived object (such as cultural relics / physical objects) and electronic archives, cross-network security protection, and audit interpretability. It is difficult to simultaneously achieve highly reliable evidence storage, fine-grained tamper location, and verifiable audit.

[0006] Therefore, there is a need for a blockchain cross-network anchoring protection method for archive channels that can address the shortcomings of the existing technologies. Summary of the Invention

[0007] One objective of this invention is to propose a blockchain-based cross-network anchoring protection method for archival channels based on a governance model. This addresses the problems of existing technologies, which, under conditions of high confidentiality, strong sovereignty attributes, and physical isolation of archived objects (such as cultural relics / physical objects) and their associated information, are prone to physical or logical loss by internal personnel, transmission link interception and tampering during cross-network archiving and transmission, making off-site authentication and penetrating auditing difficult. Furthermore, this leads to blind spots in cross-network supervision for single-track archiving and difficulty in simultaneously ensuring cross-network security protection for both physical and electronic information in dual-track archiving. The proposed method involves performing content definition fragmentation and generating fragment fingerprints after the cross-network gateway obtains the business data stream of the archived object. The invention employs a vector-based approach, where a vector commitment value is generated based on KZG vector commitments, and an anchor message is constructed. This anchor message is then written into the consortium blockchain via a HotStuff-like Byzantine fault-tolerant consensus mechanism to obtain an anchor receipt. The receiving side recalculates the vector commitment and verifies its consistency with the on-chain anchor record to generate a cross-network authentication result. Simultaneously, a governance model is deployed offline on the government intranet to read locally stored evidence records in read-only mode and outputs a risk assessment and explanatory audit report based on the authentication result. Through a technical solution that triggers human-machine collaborative review using uncertainty thresholds and core sovereignty change rules, this invention achieves the technical effects of cross-network transmission with authentication capabilities, traceable and difficult-to-destroy evidence, interpretable audits, and support for third-party penetration verification.

[0008] This invention provides a method for cross-network anchoring protection of archive channels based on a governance model, comprising: S1. Obtain the archived object business data stream to be transmitted across networks at the physically isolated cross-network gateway; S2. Perform content definition fragmentation on the archived object business data stream to generate a fragmentation sequence; S3. Generate a fragmentation fingerprint set on the fragmentation sequence, construct a fingerprint vector, and generate a vector commitment value; S4. Generate an anchor message based on the vector commitment value. The anchor message includes the vector commitment value and transmission evidence metadata corresponding to the archived object business data stream. Associate the transmission evidence metadata, fragmentation fingerprint set, and vector commitment value to form a local evidence record and write it to immutable storage; S5. Broadcast the anchor message to multiple consortium blockchain nodes, perform Byzantine fault-tolerant consensus on the anchor message, write it to the blockchain ledger, generate an on-chain anchor record, and output an anchor receipt, including transaction hash and block height; S6. Transmit the archived object business data stream through the one-way cross-network channel of the cross-network gateway. The data flow to the receiving side is repeated, and steps S2 and S3 are executed to generate the receiving side vector commitment value. Based on the anchor receipt, the on-chain anchor record in the blockchain ledger is obtained, and the consistency between the receiving side vector commitment value and the vector commitment value in the on-chain anchor record is verified to generate a cross-network authentication result; S7, the governance model is deployed offline on the government intranet, the local evidence record is read in read-only mode, and the cross-network authentication result and anchor receipt are received to generate a risk assessment result and an explanatory audit report, and the uncertainty index corresponding to the risk assessment result is calculated; S8, the uncertainty index is compared with a preset threshold, and the risk assessment result is matched with the preset sovereign core change risk rule. When the uncertainty index exceeds the preset threshold or matches the sovereign core change risk rule, a human-machine collaborative review task is output; otherwise, an explanatory audit report is output.

[0009] Optionally, S1 includes: The sending side of the cross-network gateway accesses the output interface of the archive object service system to receive the archive object service data stream; The archived object business data stream is parsed into data content and metadata, the metadata including archive identifier, archived object identifier, data category, archive mode, generation time, sending system identifier and receiving system identifier; The archiving mode is used to indicate whether the business data flow of the archived object is a single-track archiving scenario or a dual-track archiving scenario. When the archiving mode indicates a dual-track archiving scenario, the business data flow of the archived object includes archive data and archived object entity association data.

[0010] Optionally, S2 includes: The business data stream of the archived object is input into FastCDC content in byte order to define a fragmentation algorithm, and a sliding window is set based on the preset minimum fragment length, target fragment length and maximum fragment length, and the rolling hash value is calculated; After the current cumulative shard length reaches the minimum shard length, the rolling hash value is matched with the preset shard boundary determination rule. When the match is successful, the shard boundary is determined and the current shard is generated. When the match is not valid and the current cumulative fragment length reaches the maximum fragment length, the fragment boundary is forcibly determined and the current fragment is generated. The above process is repeated on the archived object business data stream until the data stream ends, so as to generate a fragment sequence, wherein each fragment includes a fragment number, a fragment starting offset in the archived object business data stream, a fragment length, and fragment data content.

[0011] Optionally, S3 includes: BLAKE3 streaming fingerprint calculation is performed sequentially on the fragment sequence in ascending order of fragment number. For each fragment, at least the fragment number, the starting offset of the fragment in the archived object's business data stream, the fragment length, and the fragment data content are input into BLAKE3 in a preset concatenation order to generate the fragment fingerprint corresponding to that fragment. The fragment fingerprints corresponding to each fragment in the fragment sequence are combined into a fragment fingerprint set in ascending order of the fragment number. The fragment fingerprint set is encoded into a fingerprint vector according to a predetermined encoding rule. A vector commitment value is generated for the fingerprint vector based on the common parameters of the KZG vector commitment algorithm, wherein the vector commitment value corresponds one-to-one with the fingerprint vector and is used for subsequent consistency verification.

[0012] Optionally, S4 includes: An anchor message is constructed based on a vector commitment value. The anchor message includes a vector commitment value, a cross-network gateway identifier, a sending-side system identifier, a receiving-side system identifier, an archiving mode, a generation time, and a business serial number used to identify the business data stream of the archived object. The metadata of the business data stream of the archived object, the cross-network gateway identifier, the sending-side system identifier, the receiving-side system identifier, the archiving mode, the generation time, the business serial number, and parameter identifiers used to identify the content definition fragmentation algorithm parameters, the streaming cryptographic digest algorithm type, and the common parameters of the vector commitment algorithm are combined to form the transmission evidence metadata. An evidence digest is calculated on the transmission evidence metadata and associated with the vector commitment value for storage, generating a local evidence record. The local evidence record is written to immutable storage, wherein the immutable storage includes a storage medium that allows for one-time writing and multiple readings or an audit log storage with append-only writing capabilities.

[0013] Optionally, S5 includes: The anchor message is sent to multiple pre-registered consortium blockchain nodes; After receiving the anchor message, each of the consortium blockchain nodes performs format verification and duplicate verification on the vector commitment value, cross-network gateway identifier, sending system identifier, receiving system identifier, archiving mode, generation time, and business serial number in the anchor message. After the verification is passed, the anchor message is used as a consensus proposal to input into the HotStuff Byzantine Fault Tolerant Consensus Algorithm. Based on the HotStuff Byzantine Fault Tolerant Consensus Algorithm, the submission confirmation of the anchoring message is completed among multiple consortium chain nodes, and the anchoring message is written into the blockchain ledger to generate an on-chain anchoring record; The consortium blockchain node that has completed the write outputs an anchoring receipt, which includes the transaction hash and block height corresponding to the on-chain anchoring record; Furthermore, after outputting the anchor receipt, the process further includes: locating the block containing the on-chain anchor record corresponding to the transaction hash based on the anchor receipt; obtaining the block hash of the block after the block has reached finality through Byzantine fault-tolerant consensus; obtaining the commit certificate corresponding to the block; concatenating the transaction hash, the block hash, and the commit certificate according to a preset concatenation rule and inputting them into a preset cryptographic hash function to calculate a randomness source; generating a challenge index set based on the randomness source according to a preset and publicly available deterministic challenge generation rule, wherein each challenge index in the challenge index set falls within the length range of the fingerprint vector; determining the sampled vector position from the fingerprint vector based on the challenge index set, and generating a KZG open proof set corresponding to the sampled vector position; constructing a sampling proof message from the challenge index set, the shard fingerprint corresponding to the sampled position, and the KZG open proof set and writing it into the blockchain ledger to form an on-chain sampling proof record that can be verified by a third party.

[0014] Optionally, S6 includes: The archived object service data stream is transmitted from the sending side to the receiving side through the unidirectional cross-network channel of the cross-network gateway. The receiving side then receives the archived object service data stream completely to form a receiving-side archived object service data stream. Content definition fragmentation processing is performed on the receiving-side archived object service data stream using the same minimum fragment length, target fragment length, maximum fragment length, and fragment boundary determination rules as in S2, to generate a receiving-side fragment sequence. Streaming cryptographic digest calculation and vector commitment calculation are performed on the receiving-side fragment sequence using the same preset splicing order, predetermined encoding rules, and common parameters of the vector commitment algorithm as in S3, to generate a receiving-side vector commitment value. On-chain anchor records in the blockchain ledger are obtained based on anchor receipts, and vector commitment values ​​are read from these on-chain anchor records. The receiving-side vector commitment value is compared with the vector commitment value in the on-chain anchor record for consistency. If they match, a cross-network authentication result indicating that the cross-network transmission has not been tampered with is generated; if they do not match, a cross-network authentication result indicating that the cross-network transmission has been tampered with or that the received data is abnormal is generated.

[0015] Optionally, the S7 includes: The governance model reads the local evidence records written by S4 in a read-only manner and stores them in an immutable manner, and extracts the vector commitment value, fragment fingerprint set, transmission evidence metadata and evidence digest from the local evidence records. Receive cross-network authentication results and anchor receipts, and extract transaction hashes and block heights based on the anchor receipts; The cross-network authentication result, the vector commitment value, the transmission evidence metadata, the evidence digest, the transaction hash, and the block height are combined into model input data according to a preset input template. The model input data is input into the governance big model to generate a risk assessment result, and an explanatory audit report is output at the same time as the risk assessment result is generated. The explanatory audit report includes the cross-network authentication result, the risk assessment result, the evidence summary to support the risk assessment result, and the transaction hash and block height corresponding to the evidence summary. The risk assessment results are processed using an order-preserving confidence prediction algorithm to calculate the uncertainty index.

[0016] Optionally, S8 includes: Obtain uncertainty indicators, risk assessment results, and explanatory audit reports; The uncertainty index is compared with a preset threshold to generate an uncertainty comparison result; The risk assessment results are matched with preset rules for changes in core sovereignty risks to generate rule matching results; When the uncertainty comparison result indicates that the uncertainty index exceeds the preset threshold, or the rule matching result indicates that the sovereign core change risk rule is matched, a human-machine collaborative review task is generated and output. The human-machine collaborative review task includes business serial number, risk judgment result, uncertainty index, transaction hash, block height and evidence digest. When the uncertainty comparison result indicates that the uncertainty index does not exceed the preset threshold, and the rule matching result indicates that no rule for sovereign core change risk is matched, the explanatory audit report is output.

[0017] The beneficial effects of this invention are: 1. Achieve cross-network and cross-regional authentication and tamper-resistant transmission: By defining and fragmenting the business data stream of archived objects and generating fragment fingerprint vectors, and further generating KZG vector commitments and anchoring them on the consortium blockchain, the receiving side can recalculate the commitments and verify consistency with the on-chain records without relying on the original system of the sending side. This effectively detects transmission tampering behaviors such as interception, replacement, insertion or deletion under the conditions of physical isolation and unidirectional channel.

[0018] 2. Enhance the resilience and accountability of evidence retention: By mutually verifying the locally stored, immutable evidence records with the on-chain anchored records of the multi-node Byzantine fault-tolerant consensus, the risk of evidence loss due to physical or logical loss by internal personnel is reduced, forming a traceable evidence chain that can be located to the business serial number, transaction hash and block height, thereby improving cross-network supervision and accountability capabilities.

[0019] 3. Improve the efficiency of penetrating audit and handling: The governance big data model reads the evidence records offline in read-only on the government intranet and generates explanatory audit reports by combining the cross-network authentication results. At the same time, it outputs uncertainty indicators and triggers human-machine collaborative review tasks when the threshold is exceeded or the core sovereign change risk rules are hit. This reduces misjudgments and omissions while ensuring security and compliance, and supports penetrating verification and audit closed loop for key evidence. 4. Through a unified metadata model and dual-track archiving mechanism, it can flexibly adapt to archiving behaviors in different electronic archive fields (such as cross-network association of spatial elements and attribute data in map spatial data, and mapping of physical objects and electronic records in physical archives). While expanding the protection dimension of business archives, it also takes into account the consistency of security protection between business characteristics and archive characteristics, so as to adapt to the multi-domain data archiving needs based on the dual kernel of business and archives. Attached Figure Description

[0020] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1This is a flowchart of a blockchain cross-network anchoring protection method for archive channels based on a governance model proposed in this invention. Detailed Implementation

[0021] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0022] refer to Figure 1 A blockchain cross-network anchoring protection method for archive channels based on a governance model includes: S1. Obtain the archived object business data stream to be transmitted across networks at the physically isolated cross-network gateway; S2. Perform content definition fragmentation on the archived object business data stream to generate a fragmentation sequence; S3. Generate a fragmentation fingerprint set on the fragmentation sequence, construct a fingerprint vector, and generate a vector commitment value; S4. Generate an anchor message based on the vector commitment value. The anchor message includes the vector commitment value and transmission evidence metadata corresponding to the archived object business data stream. Associate the transmission evidence metadata, fragmentation fingerprint set, and vector commitment value to form a local evidence record and write it to immutable storage; S5. Broadcast the anchor message to multiple consortium blockchain nodes, perform Byzantine fault-tolerant consensus on the anchor message, write it to the blockchain ledger, generate an on-chain anchor record, and output an anchor receipt, including transaction hash and block height; S6. Transmit the archived object business data stream through the one-way cross-network channel of the cross-network gateway. The data flow to the receiving side is repeated, and steps S2 and S3 are executed to generate the receiving side vector commitment value. Based on the anchor receipt, the on-chain anchor record in the blockchain ledger is obtained, and the consistency between the receiving side vector commitment value and the vector commitment value in the on-chain anchor record is verified to generate a cross-network authentication result; S7, the governance model is deployed offline on the government intranet, the local evidence record is read in read-only mode, and the cross-network authentication result and anchor receipt are received to generate a risk assessment result and an explanatory audit report, and the uncertainty index corresponding to the risk assessment result is calculated; S8, the uncertainty index is compared with a preset threshold, and the risk assessment result is matched with the preset sovereign core change risk rule. When the uncertainty index exceeds the preset threshold or matches the sovereign core change risk rule, a human-machine collaborative review task is output; otherwise, an explanatory audit report is output.

[0023] In this specific embodiment, S1 includes: A data access component is deployed on the sending side of the cross-network gateway in a physically isolated network. This data access component connects to the output interface of the archived object service system via the service access network port on the sending side of the cross-network gateway. The archived object service system pushes the archived object service data stream to the output interface in a manner where each message corresponds to a single object to be transmitted across the network. The output interface uses TCP-based long-connection transmission and has a fixed listening port of 4433 on the cross-network gateway side. The archived object service data stream is abstracted as a tuple within the cross-network gateway. ,in Represents the business data flow of the archived object. A sequence of bytes representing data content, arranged in byte order. Represents metadata and is a structured collection of key-value pairs; Upon receiving a complete message, the data access component first reads the metadata length field in the message header and extracts the metadata area accordingly. It then parses the metadata area into a JSON object using UTF-8 encoding and performs key whitelist validation, field type validation, and mandatory field validation to obtain metadata that meets the predefined metadata pattern. The metadata schema permanently includes an archive identifier, an archived object identifier, a data category, an archive mode, a generation time, a sending system identifier, and a receiving system identifier. The archive identifier and the archived object identifier both use 36-character UUIDv4 text format and are used to uniquely indicate the corresponding archive and the corresponding archived object (e.g., cultural relic / physical object) entity. The data category is an enumerated value used to indicate the data content. The business type and the generation time use a Unix millisecond timestamp in the UTC time zone to mark the time when the data was generated. The sending system identifier and the receiving system identifier both use system-encoded strings registered in the cross-network gateway and are used to uniquely indicate the sending system and the receiving system. The archiving mode is an enumeration value and the set of values ​​is fixed. A value of 0 indicates a single-track archiving scenario and the data content at this time. Containing only archival data, a value of 1 indicates a dual-track archiving scenario and the data content at this time. Organized into two logical partitions according to a predefined container format, each holding archive data and associated data of archived object entities respectively, the data access component processes the data content after parsing the archive mode. Perform structure validation consistent with the archiving mode to ensure that single-track archiving and dual-track archiving are clearly distinguished before cross-network deployment and form a unified input object that can be reused in subsequent steps. .

[0024] In this specific embodiment, S2 includes: Business data flow for archived objects Data content in Perform content definition fragmentation processing to generate a fragment sequence, the Let be a sequence of bytes arranged in byte order and its total length be denoted as . The content segmentation algorithm is fixed at FastCDC and is input byte-by-byte in a single-threaded sequential scanning manner on the cross-network gateway sending side. The minimum fragment length is fixed at 1. bytes, target fragment length is fixed. bytes, maximum fragment length is fixed. bytes, the sliding window size is fixed. The minimum observation span used to trigger a rolling hash update is 64 bits, and the rolling hash uses a built-in constant table. As a mapping table from bytes to random numbers, the The fragmentation boundary is fixed during cross-network gateway installation and remains completely consistent on both the sending and receiving sides to ensure the determinism of the fragmentation boundary. The algorithm maintains the current fragmentation start offset pointer during runtime. With the current position pointer ,in Indicates the current fragment is in The starting offset and initial value in Indicates the currently scanned byte in The offset in the middle is initialized to 0, while maintaining the cumulative length of the current fragment. With the current rolling hash value ,in Indicates from arrive The cumulative number of bytes, expressed in bytes. This indicates the current fragment's... The 64-bit unsigned integer obtained by iterating through the scanned bytes according to the Gear hash rule; when Updated only and advancement Without boundary determination, when A fragmentation boundary determination is performed once for each byte advanced, and the fragmentation boundary determination rule is fixed as follows: The symbol & represents a bitwise AND operation. Represents the boundary mask and is fixed in value. To make the average slice length and Alignment This represents the rolling hash value corresponding to the current position. When the above condition is met, the current position is... Determine the partition boundary and output the current partition. ,in Indicates the first segment in the sequence of segments Each fragment and Starting from 0 and incrementing, the fragments Includes fragment number , fragments in Starting offset in Slice length and fragmented data content and will Updated to To begin the next segment; When the above judgment is not true and When a forced fragmentation boundary is triggered, the current position is set to the specified value. Directly identify the sharding boundary and output the shards according to the same field. Then proceed to the next segment; When the scan reaches the end of the data content, that is... When, the remaining bytes The output is the last fragment and the process terminates, thus obtaining the result from all fragments. The fragment sequence is formed by arranging the fragment numbers in ascending order.

[0025] In this specific embodiment, S3 includes: For each fragment sequence, generate a fragment fingerprint set in ascending order of fragment number and construct a fingerprint vector and vector commitment value; the cross-network gateway sending side processes each fragment... Perform BLAKE3 streaming fingerprint calculation, the Includes fragment number Fragmentation in data content Starting offset in Slice length and fragmented data content and will and The fingerprint input byte sequence is formed according to a fixed concatenation order, where Uses little-endian encoding for 8-byte unsigned integers. Uses little-endian encoding for 8-byte unsigned integers. The data is encoded using little-endian encoding of 4-byte unsigned integers, followed by appending the fragmented data content. Then, the BLAKE3 streaming interface is called to perform initialization, update, and termination in the above order, and a fixed 32-byte digest is output as the fragment fingerprint of the fragment. All fragment fingerprints are sorted in ascending order by fragment number to form a fragment fingerprint set. ,in This indicates the number of fragments and is determined when the fragmentation ends in step S2. According to the predetermined coding rules Encoded as fingerprint vector ,in The value is fixed at 4096, representing the maximum number of slices supported by the vector commitment. Indicates the first The field elements of the vector positions satisfy the condition that... Time by fragmented fingerprint Determined and when Filling is completed by taking zero elements at a time, as described above. Sure The process is fixed to 32 bytes According to big-endian interpretation, it is a non-negative integer. Then Scalar domain modulus of BLS12-381 curve Obtain by taking the modulus ,in It is a constant and is used to limit It falls into the scalar domain; Generate vector commitment values ​​based on the KZG vector commitment algorithm The common parameters of the KZG vector commitment algorithm are imported once during cross-network gateway deployment and remain consistent on the sending and receiving sides. These common parameters include the BLS12-381 curve group. generator and power sequences ,in This represents a trapdoor scalar generated by trusted initialization and not publicly disclosed, used to generate structured reference strings, and a vector commitment value. Determine by the following formula: ; in Denotes the vector commitment value and is a group. A point on the top, group generator, Represents fingerprint vector In the A field element at each position. A trapdoor scalar representing a structured reference string. Represents the vector position index and its value ranges from 0 to... Indicates the maximum length of the vector. This represents the summation operation of scalars at each position and modulo 1. In the sense of meaning, the above The serialization is performed using the BLS12-381 G1 compression format and is 48 bytes.

[0026] In this specific embodiment, S4 includes: The cross-network gateway sending side obtains the vector commitment value in step S3. With fragmented fingerprint sets And obtain the archived object business data stream from step S1. metadata in ; Cross-network gateway sending side based on Construct anchor message The anchor message It is a structured data object with fixed fields containing vector commitment values. Cross-network gateway identifier Sending side system identifier Receiver system identifier Archive mode Generation time and business serial number ,in This is a unique string identifier registered by the cross-network gateway on the consortium blockchain side. and Retrieved from metadata The transmitting system identifier and the receiving system identifier in the data. Taken from metadata The archiving mode in the file and its set of values ​​is Taken from metadata The generation time in the data is a Unix millisecond timestamp in the UTC time zone. Each archived object business data stream is allocated a 128-bit unsigned integer by the cross-network gateway and serialized in 16-byte big-endian encoding, and monotonically increased within the cross-network gateway to ensure uniqueness; The cross-network gateway sending side further constructs transmission evidence metadata. The transmitted evidence metadata It is a structured key-value collection with fixed fields containing metadata. Cross-network gateway identifier Sending side system identifier Receiver system identifier Archive mode Generation time Business serial number Content definition, fragmentation algorithm parameter identifier Streaming cryptographic digest algorithm type identifier and the common parameter identifier of the vector commitment algorithm ,in The string FASTCDC-Lmin2048-Ltar8192-Lmax16384-w64-mask13 is fixed and used to uniquely indicate step S2. With a set of fragmentation parameters for a boundary mask of 13 bits, The string BLAKE3-256 is fixed and used to uniquely indicate that the BLAKE3 output in step S3 is a 32-byte digest type. The string KZG-BLS12-381-N4096-SRSv1 is fixed and used to uniquely indicate the vector length in step S3. And the corresponding structured reference string version number is The set of KZG common parameters; Cross-network gateway sending side transmits evidence metadata Computational evidence summary The evidence summary According to the formula: ; in This represents a summary of evidence and is a 32-byte result. This indicates that the BLAKE3 cryptographic hash function has a fixed output length of 256 bits. This indicates that a normalized CBOR serialization encoding function is performed on structured objects, and the canonical encoding rule defined in RFC8949 is used to ensure that the sending and receiving sides use the same encoding. A completely identical byte sequence was obtained. This refers to the aforementioned transmitted evidence metadata; The cross-network gateway sends local evidence records. The local evidence storage record It is a structured object with a fixed field that includes the business serial number. Transmitting evidence metadata Evidence Summary Vector commitment value and fragmented fingerprint sets and in Internal establishment from arrive , and The associated indexes are used to ensure that subsequent steps can be performed using... Use the primary key to retrieve the complete set of evidence elements; The cross-network gateway sending side will Write to immutable storage, which is a fixed write-once, read-many object storage system with a retention period locking strategy. Writes are performed using the object key "evidence". storage Furthermore, after the write operation is completed, the disk confirmation interface is called to return the object version number, and this version number is written to the audit log appended locally on the cross-network gateway to form a write loop, thereby ensuring local evidence storage. Anchored messages that cannot be overwritten or deleted during the retention period and can be used in subsequent step S5. Using the same business serial number Achieve one-to-one correspondence.

[0027] In this specific embodiment, S5 includes: The cross-network gateway sending side constructs and obtains the anchor message in step S4. Afterwards, The byte sequence is encoded using normalized CBOR and simultaneously sent to a pre-registered set of consortium blockchain nodes via a TLS two-way authentication channel. ,in Include Each consortium blockchain node stores a cross-network gateway identifier. With system identifier The registration form is used for identity and permission verification. This represents the number of nodes in the consortium blockchain. The Byzantine fault tolerance threshold for the consortium blockchain is fixed at [value missing]. And satisfy , Indicates the number of allowed Byzantine fault nodes; Each consortium blockchain node receives Then, format validation and repeatability validation are performed. The format validation includes validation of the vector commitment value. Execute BLS12-381 curve Point compression encoding length verification and deserialization feasibility verification for cross-network gateway identifiers. Perform an existence check on the registration form and verify the sending system identifier. With the receiver system identifier Perform an existence check on the registration form and check the archiving mode. Execution value set Verification, regarding the generation time Perform a check that ensures the deviation from the node's local UTC time does not exceed 300,000 milliseconds, for the business serial number. The execution length is 16 bytes and can be parsed as an unsigned integer by big-endian. The duplicate check includes checking the local node's ledger status database and the unconfirmed transaction pool using a composite key. Check if an anchoring record already exists to reject duplicate anchoring of the same service serial number on the same cross-network gateway; After the verification is passed, each consortium blockchain node will Encapsulated as a consensus proposal transaction The HotStuff Byzantine Fault Tolerance consensus algorithm is input, which employs a leader rotation mechanism with monotonically increasing view numbers, and the leader of each view is determined by viewmod. It is confirmed that, where "view" represents the view number and is maintained by each node, HotStuff consensus executes three voting phases sequentially on each proposal: prepare, pre-commit, and commit, with each phase requiring no less than [number missing] votes. Each node signs the same proposal block to form a quorum certificate. The quorum certificate uses the BLS12-381 signature scheme to generate an aggregate signature and includes a list of signing node identifiers to form a commit certificate. ,in This indicates that a certificate has been submitted and is used to prove that the corresponding block has achieved finality under the HotStuff rules; When included Once the block reaches finality, the consortium blockchain node that has completed the write returns an anchor receipt to the cross-network gateway. The anchoring receipt Includes transaction hash With block height ,in Indicates to The normalized CBOR byte sequence is calculated using BLAKE3 to obtain a 32-byte result, which is used to uniquely locate the anchor record on the chain. This indicates the height of the block containing the anchored record on the chain, and is a monotonically increasing non-negative integer starting from the genesis block; Output anchoring receipt Afterwards, the consortium blockchain node that has completed the writing is based on Location includes The final block and read the block hash of that block. With submitting certificates ,in This represents the 32-byte result obtained by performing BLAKE3 computation on the normalized CBOR byte sequence of the block header, followed by the calculation of the randomness source. And based on this, a set of challenge indexes is generated. To form a third-party verifiable sample, the randomness source is determined according to the formula: ; in The result represents a random source and a 32-byte output. BLAKE3() indicates the BLAKE3 cryptographic hash function with a fixed output length of 256 bits. This represents the 32-byte sequence corresponding to the transaction hash. This represents the 32-byte sequence corresponding to the block hash. This represents the encoded result of the byte sequence of the submitted certificate. This indicates sequential concatenation operations; Challenge Index Set Fixed generation There are 10 indices, and each index falls within the fingerprint vector. Length range and The The generation process is fixed as follows: As a seed for the BLAKE3 expandable output function, and continuously derive enough bytes to parse a sequence of 32-bit unsigned integers, then process each 32-bit unsigned integer pair... The candidate indices are obtained by taking the modulo and then deduplicated in the order of appearance until exactly the desired result is obtained. Each index is distinct; The consortium blockchain node is based on From fingerprint vectors The sampling vector position is determined and the same common parameter identifier as in step S3 is called. The corresponding KZG open proof algorithm for vector commitment values Generate a set of KZG open proofs ,in and Indicates targeting the index The opening of proof and its use to enable third parties to obtain only Verifiable consistency verification is completed in accordance with the corresponding sampled values; the consortium blockchain node simultaneously retrieves the local evidence record from step S4. Reading from and indexing One-to-one corresponding sampled fragment fingerprint And and Constructing a sampling proof message The sampling proof message Fixed includes business serial number Transaction hash Block height Block hash Submit certificate digest and vector commitment value Challenge Index Set The sampled fragment fingerprint set and the KZG open proof set The submitted certificate digest is a summary of the certificate. The 32-byte result obtained from the BLAKE3 calculation is used to bind the final material without expanding the on-chain storage; The consortium blockchain node will Encapsulated as a sample proof transaction It is then written back into the blockchain ledger via HotStuff consensus, thus forming an on-chain sampling proof record that can be verified by third parties.

[0028] In this specific embodiment, S6 includes: The cross-network gateway transmits archived object business data streams through a one-way cross-network channel. Transmitted from the sending side to the receiving side, the unidirectional cross-network channel uses a hardware unidirectional transmission device to achieve physical layer unidirectionality, and outputs in a framed manner on the sending side and inputs in a reassembly manner on the receiving side. Sending side data content With metadata The encapsulation process generates a frame sequence, with each frame containing a frame header and a frame payload. The frame header fields are fixed and include the business serial number. Frame number Total number of frames Load length With load verification value ,in Anchor message with step S4 The business serial number in the serial number is consistent and used for cross-network reorganization association. This indicates a frame number that monotonically increases from 0. This indicates the total number of frames transmitted across networks and is determined at the start of encapsulation on the sending side. Calculated from the fixed frame payload limit of 1,048,576 bytes, Indicates the first The actual byte length of the frame payload. Indicates the first The CRC32 check result of the frame payload is calculated by byte and used for transport layer error detection; The receiving side receives the complete frame sequence output from the one-way cross-network channel and processes it according to... Aggregate caching, based on each frame Perform sequence rearrangement and apply loads Recalculation for verification Consistency, provided that all frames are received and the set of frame sequence numbers is [value missing]. Then, the payloads of each frame are concatenated and reassembled in ascending order of frame number to form the service data stream of the archived object on the receiving side. ,in This represents the byte sequence of the data content obtained by the receiving side through reconstruction. This represents the set of metadata key-value pairs parsed by the receiving side, and its field set is the same as that in step S1. It is consistent and includes the file identifier, archived object identifier, data category, archive mode, generation time, sending system identifier, and receiving system identifier; Receiving side Perform FastCDC content definition fragmentation processing that is exactly the same as step S2, using the same parameter as the minimum fragment length. bytes, target fragment length bytes, maximum fragment length Bytes, sliding window size Byte, boundary mask bit length is 13 and rolling hash constant table Consistent with the sending side, a receiving-side fragment sequence is generated, ensuring that fragment boundaries are reproducible when the data has not been tampered with. The receiving side performs BLAKE3 streaming fingerprint calculation and KZG vector commitment calculation on the receiving-side fragment sequence, identical to step S3. The fingerprint input is fixed, consisting of fragment number, starting offset, fragment length, and fragment data content encoded in a predetermined concatenation order before being input into BLAKE3, with a fixed digest output length of 32 bytes. The fingerprint vector length is also fixed. And use the same domain element mapping rules and the same public parameter identifiers. The corresponding structured reference string generates the receiver side vector commitment value. ,in Indicates to The recalculated vector commitment value is serialized in BLS12-381 format. The compressed format is 48 bytes. The receiving side is based on the anchoring receipt. Query the on-chain anchoring records from the consortium blockchain ledger, among which Includes transaction hash With block height , This represents a unique location identifier for the anchored message transaction and is a 32-byte sequence. Indicates the block height containing the transaction, which the receiving side accesses via... Read the transaction payload to obtain the on-chain vector commitment value And verify that the transaction occurred at a block height equal to And the business serial number in the transaction field Reassembled with the receiving side Consistent; The receiving side performs a consistency comparison and generates a cross-network authentication result. The According to the formula: ; in This represents the cross-network authentication result and its value set is... This indicates an indicator function that takes the value 1 if the condition within the parentheses is true, otherwise takes the value 0. This represents the receive side vector commitment value. This represents the vector commitment value in the on-chain anchor record. This indicates a byte-by-byte comparison of the two 48-byte sequences; when The system will output a cross-network authentication result indicating that the cross-network transmission has not been tampered with and will... The receiving side appends an audit log entry to create a traceable record. The system outputs cross-network authentication results indicating that cross-network transmission has been tampered with or that received data is abnormal. It also records the verification results of the receiving side fragment quantity and on-chain business serial number in the audit log to support subsequent location.

[0029] In this specific embodiment, S7 includes: Large Model of Governance The system is deployed offline on a government intranet audit server and is physically isolated from the internet. The model weight directory of the audit server is mounted in read-only mode and started by the audit process with a non-privileged account. The governance model... Employs a decoder-based Transformer architecture with a fixed parameter size. The number of network layers is fixed. The hidden dimension is fixed as The number of self-attention heads is fixed. The maximum context length is fixed at 1. Each token is used, and the word segmenter employs BPE with a fixed vocabulary size. The aforementioned governance model In an offline environment, the audit task is adapted and the inference parameters are solidified by fine-tuning the supervision instructions. During inference, the temperature parameter is fixed at 0 and the structured output is generated by deterministic greedy decoding. The audit process writes the object key "evidence" to the immutable storage in read-only mode from step S4. .cbor reads local evidence records ,in Indicate the business serial number and ensure it is consistent with steps S4 and S6, then proceed with... Perform CBOR deserialization and extract the vector commitment value. Fragmented fingerprint set Transmitting evidence metadata and summary of evidence ,in This represents the vector commitment value generated in step S3 and associated and stored in step S4. This represents the set of fragment fingerprints arranged in ascending order of fragment number. Indicates that it contains metadata Metadata of transmission evidence, including fields such as algorithm parameter identifiers and system identifiers. Indicates to The 32-byte evidence digest obtained by executing BLAKE3; The audit process also receives the cross-network authentication result output in step S6. and anchoring receipt and from Extract transaction hash With block height ,in Used to indicate the receive side vector commitment value With on-chain vector commitment value The consistency determination result Used to uniquely locate anchor records on the chain Used to identify the block height containing the on-chain anchored record; The audit process will as well as Enter the preset template Assembled as model input data ,in A structured text template with a fixed field order is used, and binary fields are Base64 encoded for output to ensure consistency across operations. It is fixed to include the fields b_sn, verify, V, E, h_E, h_tx, and H, and the semantics of the fields correspond one-to-one with the above symbols; Will Input Governance Model back, Output risk assessment results With explanatory audit report ,in Fixed includes risk level label Each risk level is categorized as low, medium, and high, and includes a set of evidence supporting those risk levels. Fixed to include cross-network authentication results Risk level label Evidence summary used to support risk assessment and the transaction hash corresponding to the evidence summary. and block height And output in an auditable, field-based format; In generation At the same time, the audit process The classification head output executes an order-preserving confidence prediction algorithm to calculate the uncertainty index. The classification head is for the same input Output the unnormalized score vectors for the three types of risks. And scaled by a fixed temperature factor Perform calibration and calculate the class confidence vector. and uncertainty indicators Defined as the complementary quantity of the maximum confidence level to maintain a monotonically ordered relationship with the confidence level, specifically according to the formula: ; in This represents a function that exponentially and normally distributes the input vector element by element to obtain the probability distribution. This represents the three-dimensional score vector output by the classification head. This represents the temperature scaling factor and is a positive real number. This represents the confidence vector corresponding to the three types of risk. express The highest confidence component in the middle. This indicates an uncertainty index with a value range of [value range missing]. and will and Write them together to the append-written audit log.

[0030] In this specific embodiment, S8 includes: The audit orchestration service is executed on the government intranet and is integrated with the governance model. The reasoning process is decoupled, and the audit orchestration service reads the same business serial number from the audit log in step S7. Corresponding uncertainty indicators Risk assessment results and explanatory audit report and from Extract evidence summary Transaction hash With block height ,in This indicates the unique business sequence number assigned by the cross-network gateway to the business data stream of the archived object. This indicates that the value is calculated according to step S7 and its range is... Uncertainty indicators This indicates the risk assessment result output in step S7, which includes a risk level label. It can also include a set of risk type codes to support rule matching. This refers to the explanatory audit report output in step S7. Indicates the metadata of transmitted evidence The calculated 32-byte evidence digest This represents the 32-byte transaction hash corresponding to the on-chain anchor record. Indicates the block height where the on-chain anchor record is located; The audit orchestration service loads preset thresholds from a local read-only configuration file. ,in This represents the uncertainty threshold and is used to divide model outputs into two categories: those that can be automatically approved and those that require review. Perform stricter-than-comparison to generate uncertain comparison results; The audit orchestration service also loads a pre-defined set of rules for core sovereign change risks. The The matching object is the risk type code, and the code set is fixed as {OWNERSHIP_CHANGE, APPRAISAL_GRADE_CHANGE, LOCATION_CHANGE, COLLECTION_UNIT_CHANGE, EXPORT_RELATED}. Here, OWNERSHIP_CHANGE represents the risk of change in ownership or holder; APPRAISAL_GRADE_CHANGE represents the risk of change in appraisal conclusion or grade; LOCATION_CHANGE represents the risk of change in storage location or warehouse information; COLLECTION_UNIT_CHANGE represents the risk of change in collection management unit; and EXPORT_RELATED represents the risk of cross-border or international transfer. The rule matching process is fixed as follows: Read the risk type code set and determine its relationship with If there is an intersection, the rule matching result is a hit; otherwise, it is a miss. The audit orchestration service generates decision variables based on uncertainty comparison results and rule matching results. The According to the formula: ; in Denotes the decision variable and its set of values ​​as follows: This indicates an indicator function that takes the value 1 if the condition within the parentheses is true, and 0 otherwise. Indicators representing uncertainty Indicates the preset threshold. Represents a logical OR operation. This indicates that the rule matches the indicator and when it hits the sovereign core change risk rule. otherwise ; when The audit orchestration service generates and outputs human-machine collaborative review tasks. The Write to the internal network work order queue and simultaneously append to the audit log, with the business transaction number being a fixed field. Risk level label Uncertainty indicators Transaction hash Block height and summary of evidence To avoid duplicate creation of review tasks, the audit orchestration service uses a composite key. Perform an idempotency check in the work order status database and create a task record if no existing task exists, and set the task status to PENDING. when The audit orchestration service described above will provide interpretive audit reports. Write the audit report to the storage and generate a report index record, which permanently contains the business transaction number. Transaction hash Block height Evidence Summary Risk level label The index record is then appended to the audit log to ensure that the audit loop is traceable.

[0031] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

[0032] This invention addresses cross-network archiving and transmission of archived objects (such as cultural relics / physical objects) and their associated entity information under conditions of high confidentiality, strong sovereignty attributes, and network-physical isolation. It employs an algorithmic link combining content-defined fragmentation, fragment-level streaming fingerprinting, vector commitment anchoring, and consortium blockchain consensus to form a verifiable closed loop of transmission evidence. First, content-defined fragmentation divides the data stream into fragment sequences aligned with content changes, making it difficult for insertions, deletions, and replacements to escape fragment-level consistency checks. Second, for each fragment, the fragment number, starting offset, fragment length, and fragment content are jointly input into the streaming fingerprinting algorithm. Fingerprint calculation ensures that fingerprints are bound to shard locations and boundaries; then, the shard fingerprint set is encoded into a fingerprint vector and a vector commitment value is generated, so that only one commitment value needs to be anchored on the chain to represent the state of all shard fingerprints; finally, the on-chain anchoring record is written through the Byzantine fault-tolerant consensus of the consortium chain nodes, and the receiving side recalculates the vector commitment value and verifies its consistency with the on-chain commitment value. Thus, off-site authentication and tamper-proof auditing are achieved without relying on bidirectional networks and trusted logs on the sending side. At the same time, the local immutable evidence storage and the on-chain anchoring record corroborate each other, reducing the risk of evidence loss caused by internal personnel and enhancing accountability.

[0033] In terms of algorithm structure, this invention also makes scenario-oriented improvements around "penetrating auditing and low-error judgment handling": First, it introduces a sampling mechanism that generates a random source based on on-chain finality materials and deterministically generates challenge indexes. Combined with the open proof of vector commitment, the sampling proof is written on the chain, enabling third parties to verify the consistency of the shard fingerprint of the sampled position without obtaining the full amount of data, reducing the evidence collection burden of cross-network auditing and improving verifiability. Second, the governance big model adopts offline read-only reading of local evidence records and outputs an interpretive audit report in combination with cross-network authentication results. At the same time, it calculates uncertainty indicators and triggers human-machine collaborative review in conjunction with the sovereign core change risk rules. This enables automated risk judgment in highly confidential scenarios to have an interpretable, controllable and retrievable handling path, thus better adapting to the cross-network supervision and security protection needs under physical isolation conditions.

Claims

1. A blockchain cross-network anchoring protection method for archive channels based on a governance model, characterized in that: include: S1. Obtain the archived object business data stream to be transmitted across networks at the physically isolated cross-network gateway; S2. Perform content definition fragmentation on the business data stream of the archived object and generate a fragmentation sequence; S3. Generate a fragmentation fingerprint set on the fragmentation sequence, construct a fingerprint vector and generate a vector commitment value; S4. Generate anchor messages based on vector commitment values. Anchor messages include vector commitment values ​​and transmission evidence metadata corresponding to the business data stream of the archived object. Associate the transmission evidence metadata, fragment fingerprint set and vector commitment values ​​to form local evidence records and write them into tamper-proof storage. S5. Broadcast the anchor message to multiple consortium chain nodes, perform Byzantine fault-tolerant consensus on the anchor message and write it into the blockchain ledger, generate on-chain anchor records and output anchor receipts, including transaction hashes and block heights; S6. Transmit the archived object business data stream to the receiving side through the one-way cross-network channel of the cross-network gateway and repeat steps S2 and S3 to generate the receiving side vector commitment value, obtain the on-chain anchor record in the blockchain ledger based on the anchor receipt, and verify the consistency between the receiving side vector commitment value and the vector commitment value in the on-chain anchor record to generate a cross-network authentication result; S7. Deploy the governance model offline on the government intranet, read local evidence records in read-only mode, receive cross-network authentication results and anchor receipts, generate risk assessment results and explanatory audit reports, and calculate uncertainty indicators corresponding to the risk assessment results. S8. Compare the uncertainty indicators with the preset thresholds and match the risk assessment results with the preset sovereign core change risk rules. When the uncertainty indicators exceed the preset thresholds or match the sovereign core change risk rules, output the human-machine collaborative review task; otherwise, output the explanatory audit report.

2. The method for cross-network anchoring protection of archive channels based on a large governance model according to claim 1, characterized in that, S1 includes: The sending side of the cross-network gateway accesses the output interface of the archive object service system to receive the archive object service data stream; The archived object business data stream is parsed into data content and metadata, the metadata including archive identifier, archived object identifier, data category, archive mode, generation time, sending system identifier and receiving system identifier; The archiving mode is used to indicate whether the business data flow of the archived object is a single-track archiving scenario or a dual-track archiving scenario. When the archiving mode indicates a dual-track archiving scenario, the business data flow of the archived object includes archive data and archived object entity association data.

3. The method for cross-network anchoring protection of archive channels based on a large governance model according to claim 1, characterized in that, S2 include: The business data stream of the archived object is input into FastCDC content in byte order to define a fragmentation algorithm, and a sliding window is set based on the preset minimum fragment length, target fragment length and maximum fragment length, and the rolling hash value is calculated; After the current cumulative shard length reaches the minimum shard length, the rolling hash value is matched with the preset shard boundary determination rule. When the match is successful, the shard boundary is determined and the current shard is generated. When the match is not valid and the current cumulative fragment length reaches the maximum fragment length, the fragment boundary is forcibly determined and the current fragment is generated. The above process is repeated on the archived object business data stream until the data stream ends, so as to generate a fragment sequence, wherein each fragment includes a fragment number, a fragment starting offset in the archived object business data stream, a fragment length, and fragment data content.

4. The method for cross-network anchoring protection of archive channels based on a large governance model according to claim 1, characterized in that, S3 include: BLAKE3 streaming fingerprint calculation is performed sequentially on the fragment sequence in ascending order of fragment number. For each fragment, at least the fragment number, the starting offset of the fragment in the archived object's business data stream, the fragment length, and the fragment data content are input into BLAKE3 in a preset concatenation order to generate the fragment fingerprint corresponding to that fragment. The fragment fingerprints corresponding to each fragment in the fragment sequence are combined into a fragment fingerprint set in ascending order of the fragment number. The fragment fingerprint set is encoded into a fingerprint vector according to a predetermined encoding rule. A vector commitment value is generated for the fingerprint vector based on the common parameters of the KZG vector commitment algorithm, wherein the vector commitment value corresponds one-to-one with the fingerprint vector and is used for subsequent consistency verification.

5. The method for cross-network anchoring protection of archive channels based on a large governance model according to claim 1, characterized in that, S4 include: An anchor message is constructed based on a vector commitment value. The anchor message includes a vector commitment value, a cross-network gateway identifier, a sending-side system identifier, a receiving-side system identifier, an archiving mode, a generation time, and a business serial number used to identify the business data stream of the archived object. The metadata of the business data stream of the archived object, the cross-network gateway identifier, the sending-side system identifier, the receiving-side system identifier, the archiving mode, the generation time, the business serial number, and parameter identifiers used to identify the content definition fragmentation algorithm parameters, the streaming cryptographic digest algorithm type, and the common parameters of the vector commitment algorithm are combined to form the transmission evidence metadata. An evidence digest is calculated on the transmission evidence metadata and associated with the vector commitment value for storage, generating a local evidence record. The local evidence record is written to immutable storage, wherein the immutable storage includes a storage medium that allows for one-time writing and multiple readings or an audit log storage with append-only writing capabilities.

6. The method for cross-network anchoring protection of archive channels based on a large governance model according to claim 1, characterized in that, S5 include: The anchor message is sent to multiple pre-registered consortium blockchain nodes; After receiving the anchor message, each of the consortium blockchain nodes performs format verification and duplicate verification on the vector commitment value, cross-network gateway identifier, sending system identifier, receiving system identifier, archiving mode, generation time, and business serial number in the anchor message. After the verification is passed, the anchor message is used as a consensus proposal to input into the HotStuff Byzantine Fault Tolerant Consensus Algorithm. Based on the HotStuff Byzantine Fault Tolerant Consensus Algorithm, the submission confirmation of the anchoring message is completed among multiple consortium chain nodes, and the anchoring message is written into the blockchain ledger to generate an on-chain anchoring record; The consortium blockchain node that has completed the write outputs an anchor receipt, which includes the transaction hash and block height corresponding to the on-chain anchor record.

7. The method for cross-network anchoring protection of archive channels based on a large governance model according to claim 1, characterized in that, S6 include: The archived object service data stream is transmitted from the sending side to the receiving side through the one-way cross-network channel of the cross-network gateway, and the received archived object service data stream is completely received on the receiving side to form the receiving side archived object service data stream. The receiving-side archived object business data stream is processed using the same minimum fragment length, target fragment length, maximum fragment length, and fragment boundary determination rules as in S2 to generate a receiving-side fragment sequence. The receiving-side fragment sequence is then processed using the same preset splicing order, predetermined encoding rules, and common parameters of the vector commitment algorithm as in S3 to perform streaming cryptographic digest calculation and vector commitment calculation, generating a receiving-side vector commitment value. Based on the anchor receipt, an on-chain anchor record is obtained from the blockchain ledger, and the vector commitment value is read from the on-chain anchor record. The receiving-side vector commitment value is compared with the vector commitment value in the on-chain anchor record for consistency. If they match, a cross-network authentication result indicating that the cross-network transmission has not been tampered with is generated; if they do not match, a cross-network authentication result indicating that the cross-network transmission has been tampered with or that the received data is abnormal is generated.

8. The method for cross-network anchoring protection of archive channels based on a large governance model according to claim 1, characterized in that, S7 includes: The governance model reads the local evidence records written by S4 in a read-only manner and stores them in an immutable manner, and extracts the vector commitment value, fragment fingerprint set, transmission evidence metadata and evidence digest from the local evidence records. Receive cross-network authentication results and anchor receipts, and extract transaction hashes and block heights based on the anchor receipts; The cross-network authentication result, the vector commitment value, the transmission evidence metadata, the evidence digest, the transaction hash, and the block height are combined into model input data according to a preset input template. The model input data is input into the governance big model to generate a risk assessment result, and an explanatory audit report is output at the same time as the risk assessment result is generated. The explanatory audit report includes the cross-network authentication result, the risk assessment result, the evidence summary to support the risk assessment result, and the transaction hash and block height corresponding to the evidence summary. The risk assessment results are processed using an order-preserving confidence prediction algorithm to calculate the uncertainty index.

9. The method for cross-network anchoring protection of archive channels based on a large governance model according to claim 1, characterized in that, S8 includes: Obtain uncertainty indicators, risk assessment results, and explanatory audit reports; The uncertainty index is compared with a preset threshold to generate an uncertainty comparison result; The risk assessment results are matched with preset rules for changes in core sovereignty risks to generate rule matching results; When the uncertainty comparison result indicates that the uncertainty index exceeds the preset threshold, or the rule matching result indicates that the sovereign core change risk rule is matched, a human-machine collaborative review task is generated and output. The human-machine collaborative review task includes business serial number, risk judgment result, uncertainty index, transaction hash, block height and evidence digest. When the uncertainty comparison result indicates that the uncertainty index does not exceed the preset threshold, and the rule matching result indicates that no rule for sovereign core change risk is matched, the explanatory audit report is output.

10. The method for cross-network anchoring protection of archive channels based on a large governance model according to claim 1, characterized in that, After outputting the anchor receipt, the process further includes: locating the block containing the on-chain anchor record corresponding to the transaction hash based on the anchor receipt; obtaining the block hash of the block after the block has reached finality through Byzantine fault-tolerant consensus; obtaining the commit certificate corresponding to the block; concatenating the transaction hash, the block hash, and the commit certificate according to a preset concatenation rule and inputting them into a preset cryptographic hash function to calculate a randomness source; generating a challenge index set based on the randomness source according to a preset and publicly available deterministic challenge generation rule, wherein each challenge index in the challenge index set falls within the length range of the fingerprint vector; determining the sampled vector position from the fingerprint vector based on the challenge index set, and generating a KZG open proof set corresponding to the sampled vector position; constructing a sampling proof message from the challenge index set, the shard fingerprint corresponding to the sampled position, and the KZG open proof set and writing it into the blockchain ledger to form an on-chain sampling proof record that can be verified by a third party.

Images