Flow data tracing and verification system for finance and audit

Abstract

The application relates to the technical field of flow data tracing and verification systems for financial and accounting audits, and discloses a flow data tracing and verification system for financial and accounting audits. The system constructs a layered and progressive security calculation architecture, solves the problem that traditional security multi-party calculation protocols cannot simultaneously achieve real-time response and cryptographic security. Through the cooperation of a trusted execution environment and lightweight homomorphic encryption, real-time consistency comparison of medium and high-risk flows in the ciphertext state is carried out, and the verification response time is shortened under the premise of guaranteeing hardware-level security isolation. For high-risk flows, a strategy combining zero-knowledge proof calculation commitment and asynchronous deep verification in the background is adopted, and the deep verification task is scheduled to be executed during the idle period of system resources, so that the cryptographic security strength is maintained while the real-time business process is avoided from being interrupted.

Description

Technical Field

[0001] This application relates to the field of data identification technology, and in particular to a system for tracing and verifying traffic data for accounting and auditing. Background Technology

[0002] With the deep integration of the digital economy and fintech, the accounting and auditing field is undergoing a paradigm shift from traditional sampling auditing to real-time auditing of full data. In this process, cross-institutional traceability and real-time verification of massive amounts of cash flow data have become key technological aspects of ensuring the authenticity of financial information. Especially in scenarios involving multi-source, heterogeneous data interaction from core bank systems, enterprise ERP systems, third-party payment gateways, and clearing institutions, achieving cross-system data consistency verification without disclosing original transaction details has given rise to distributed flow fingerprint cross-verification systems based on privacy computing technology. This system utilizes secure multi-party computation (MPC), homomorphic encryption, and blockchain notarization to build a technical traceability infrastructure covering the entire cash flow chain, aiming to solve the verification blind spots caused by data silos in traditional auditing.

[0003] However, existing distributed verification architectures based on secure multi-party computation (MPC) face severe real-time bottlenecks in practical applications. Because cryptographic protocols such as SPDZ and BGW require multiple rounds of secret sharing and reconstruction, as well as complex mathematical operations, the cross-system comparison latency for a single transaction is typically as high as hundreds of milliseconds or even seconds. This fundamentally contradicts the second- or even millisecond-level real-time interception requirements for high-risk fund flows (such as large abnormal transfers or suspected money laundering transactions). Current technology is forced to make a trade-off between security and timeliness: either insist on complete MPC verification and adopt a post-event batch auditing model, leading to the risk of attackers exploiting verification gaps to complete fund transfers through "time window arbitrage"; or force real-time MPC comparison for every transaction, resulting in a "cost inversion" phenomenon where computational resource consumption far exceeds the value of the transaction itself, severely restricting the economic feasibility and risk control effectiveness of the system in a market environment. Summary of the Invention

[0004] This application proposes a traffic data tracing and verification system for financial and accounting auditing, which has the advantage of a layered and progressive secure computing architecture to solve the problem that traditional secure multi-party computation protocols cannot simultaneously achieve real-time response and cryptographic security.

[0005] To achieve the above objectives, this application adopts the following technical solution: a traffic data tracing and verification system for accounting and auditing, including a traffic classification module, a real-time verification module, an asynchronous verification module, an automatic recovery module, and an evidence solidification module;

[0006] The traffic classification module extracts local risk features from transaction data through a multi-source collaborative analysis mechanism, generates dynamic risk scores, and classifies risk levels.

[0007] The real-time verification module uses a secure computing hardware environment to verify the consistency of amount and timing in encrypted medium- and high-risk traffic.

[0008] The asynchronous verification module generates cryptographic computation commitments for high-risk traffic and uploads them for evidence storage, and completes deep verification in the background through distributed privacy computing;

[0009] When the automatic recovery module detects inconsistencies in cross-system data, it locks the risk reserve and performs the difference recovery by automatically executing a contract on the blockchain.

[0010] The evidence solidification module aggregates the verification logs, commitment vouchers, and execution records output by each module to generate a multi-level audit evidence package and output a traceability graph.

[0011] Furthermore, the real-time verification module includes a trusted base construction unit and a ciphertext verification unit;

[0012] The Trusted Base Construction Unit initializes a trusted execution environment within a hardware-isolated security domain, verifies memory integrity and establishes a trusted root through a remote proof mechanism, derives a hierarchical key system and distributes it to the isolated memory, performs homomorphic encryption parameter pre-configuration on medium- and high-risk traffic data, and establishes a lightweight ciphertext operation base environment.

[0013] The encrypted verification unit calls upon a lightweight encrypted computation environment to perform homomorphic encrypted computation on the difference in amount and timestamp logic, compare them, and generate real-time verification results. Simultaneously, controllable noise is dynamically injected into the computation process to obfuscate the side-channel information leakage path and block the original data inference channel based on power consumption characteristics or cache mode.

[0014] Furthermore, the trusted base construction unit includes an enclave initialization unit and a key derivation unit;

[0015] The enclave initialization unit instantiates a trusted execution enclave within a hardware-isolated security domain and obtains a memory metric value through a remote proof protocol. It compares the metric value with a benchmark value to establish a trusted root anchor point. Based on the trusted root anchor point, it constructs the enclave memory boundary and enables a real-time encryption mechanism to block unauthorized access paths, establishing a physically-level secure isolation area for cryptographic operations to reside in.

[0016] The key derivation unit executes a hierarchical key derivation algorithm based on the trusted root anchor to generate a hierarchical key structure and establish a binding relationship between the key and the operation domain. It maps homomorphic encryption parameters to an isolated memory space to complete pre-configuration, builds a lightweight ciphertext operation basic environment, establishes a standardized ciphertext operation interface, and forms a path of association with subsequent verification logic.

[0017] Furthermore, the ciphertext verification unit includes a ciphertext processing unit and a side-blocking anti-aliasing unit;

[0018] The ciphertext operation unit calls the standardized interface established by the lightweight ciphertext operation basic environment, performs homomorphic ciphertext operation comparison on the difference in amount and timestamp logic of medium and high risk traffic based on the hierarchical key system, generates real-time consistency verification results and marks anomalies, and passes the verification results and anomalies down to the subsequent processing modules.

[0019] The side-channel anti-aliasing unit dynamically senses power consumption characteristics and cache access patterns during the ciphertext operation execution cycle, injects controllable noise into the operation pipeline to reconstruct the side-channel information distribution pattern, blocks the original data inference path based on power consumption analysis or cache timing, and forms a closed-loop security protection for the entire ciphertext comparison operation process.

[0020] Furthermore, the asynchronous verification module includes a zero-knowledge commitment unit and a background verification unit;

[0021] The Zero-Knowledge Commitment Unit constructs a zero-knowledge proof constraint system for extracting business characteristics of high-risk traffic data. Based on the output of the constraint system, it generates verifiable computational commitments and extracts commitment summaries. The commitment summaries and data location pointers are uploaded to a distributed evidence storage network to establish cryptographic evidence anchor points that are bound to the subsequent smart contract execution logic.

[0022] The background verification unit activates the asynchronous task scheduling queue during system resource idle periods, retrieves the encrypted share from the distributed evidence storage network and executes the secure multi-party computation protocol, performs deep consistency verification on cross-system traffic data to generate deep verification results, and inputs the deep verification results into the automatic recovery module to trigger risk reserve locking and difference recovery operations.

[0023] Furthermore, the zero-knowledge commitment unit includes constraint construction units and commitment anchoring units;

[0024] The constraint construction unit extracts multi-dimensional business features from high-risk traffic data and maps them to the zero-knowledge proof constraint system. After establishing the constraint logic and proof parameters, it generates a circuit structure, forms a basic constraint framework for verifiable computational commitments, and passes it down to the commitment solidification processing stage.

[0025] The commitment anchoring unit extracts the commitment summary and data location pointer and uploads them to the distributed evidence storage network. It establishes a cryptographic evidence anchoring point bound to the smart contract execution logic, establishes an immutable evidence storage structure, and forms a ciphertext share call association path with the backend verification unit.

[0026] Furthermore, the background verification unit includes an asynchronous scheduling unit and a deep verification unit;

[0027] The asynchronous scheduling unit constructs an asynchronous task queue and establishes a task priority sequence during idle periods of system resources. Based on the encrypted share call association path established with the commitment anchoring unit, it extracts secret shared fragments from the distributed evidence storage network and constructs a multi-party computation communication topology. Within the topology, it establishes a secure communication channel between participating nodes and establishes an input association with the deep verification operation.

[0028] The deep verification unit executes a secure multi-party computation protocol within a secure communication channel to perform deep consistency verification operations on the secret-sharing shard, generating deep verification results and establishing a result transmission channel with the automatic recovery module. It maps the verification conclusions to the risk reserve locking and difference recovery operation instruction set, establishes the call association between the instruction set and the smart contract execution logic, and outputs it to the subsequent processing stage.

[0029] Furthermore, the automatic recovery module includes a funds locking unit and a contract execution unit;

[0030] The fund locking unit receives the verification conclusions output by the deep verification unit and parses the risk identifier mapping relationship. Based on the smart contract automatic execution engine, it generates a risk reserve fund locking instruction, establishes a binding relationship between the reserve fund account and the abnormal flow identifier, forms a fund control status, and establishes a call association with the contract execution unit.

[0031] The contract execution unit invokes the fund control status and parses the shortfall recovery instruction set. Through the smart contract automatic execution engine, it completes the reserve release and fund transfer operations, generates execution records, establishes a data association channel with the evidence solidification module, establishes the call association with the subsequent audit evidence generation, and passes it down.

[0032] Furthermore, the fund locking unit includes a risk analysis unit and a locking execution unit;

[0033] The risk analysis unit receives the verification conclusions output by the deep verification unit and analyzes the mapping relationship between risk identifiers and abnormal flow characteristics. It extracts the associated attributes and status information of the risk reserve account, forms locking decision parameters based on the mapping relationship and status information, establishes the input relationship with the instruction generation logic, and passes it down to the locking execution processing stage.

[0034] The execution unit is locked, the locking decision parameters are called to activate the smart contract automatic execution engine to generate a risk reserve locking instruction, establish a binding relationship between the reserve account and the abnormal flow identifier, form a fund control status and establish a call association path with the contract execution unit, and output a standardized locking certificate to the subsequent processing stage.

[0035] Furthermore, the contract execution unit includes an instruction parsing unit and an on-chain execution unit;

[0036] The instruction parsing unit receives the fund control status output by the fund locking unit, parses the operation type and amount parameters in the shortfall recovery instruction set, maps the parsing results to the smart contract execution logic to generate standardized execution instruction packages, establishes an input association with the on-chain execution unit, and passes it down.

[0037] The on-chain execution unit calls the standardized execution instruction package to activate the smart contract automatic execution engine, executes reserve release and fund transfer operations, synchronously generates execution records, establishes a data association channel with the evidence solidification module, establishes a call association with the subsequent audit evidence generation, and outputs it to the processing stage.

[0038] The beneficial effects of this invention are as follows:

[0039] This application provides a traffic data tracing and verification system for accounting and auditing. Through the collaboration of a trusted execution environment and lightweight homomorphic encryption, it performs real-time consistency comparisons of medium- to high-risk traffic in encrypted form, shortening verification response time while ensuring hardware-level security isolation. For high-risk traffic, a strategy combining zero-knowledge proof computation commitments and asynchronous deep verification in the background is adopted. Deep verification tasks are scheduled for execution during periods of system resource idleness, maintaining cryptographic security strength while avoiding interruption of real-time business processes. An on-chain automated processing mechanism for risk reserve locking and shortfall recovery is established through a smart contract automatic execution engine, enabling audit findings to be directly transformed into fund control actions, shortening the risk handling cycle of orphan transactions, and compensating for the lag in fund recovery through traditional post-event manual reconciliation. This layered architecture optimizes computing resource allocation, controls audit verification costs within an economically reasonable range, balances security, real-time performance, and economy, and provides an industrially deployable technical solution for high-frequency trading scenarios, improving the feasibility of cross-system traffic data tracing and verification. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort:

[0041] Figure 1 This is the system flowchart for this application. Detailed Implementation

[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] Example 1: Please refer to Figure 1 A system for tracing and verifying traffic data for accounting and auditing, characterized in that it includes a traffic classification module, a real-time verification module, an asynchronous verification module, an automatic recovery module, and an evidence solidification module;

[0044] The traffic classification module extracts local risk features from transaction data through a multi-source collaborative analysis mechanism, generates dynamic risk scores, and classifies risk levels.

[0045] The real-time verification module uses a secure computing hardware environment to verify the consistency of amount and timing in encrypted medium- and high-risk traffic.

[0046] The asynchronous verification module generates cryptographic computation commitments for high-risk traffic and uploads them for evidence storage, and completes deep verification in the background through distributed privacy computing;

[0047] When the automatic recovery module detects inconsistencies in cross-system data, it locks the risk reserve and performs the difference recovery by automatically executing a contract on the blockchain.

[0048] The evidence solidification module aggregates the verification logs, commitment vouchers, and execution records output by each module to generate a multi-level audit evidence package and output a traceability graph.

[0049] In this embodiment, the traffic classification module extracts risk features and dynamically scores multi-source heterogeneous transaction data through a localized federated learning architecture, classifies traffic into differentiated risk levels and outputs corresponding identifiers, effectively solves the problem of misclassification caused by cross-system clock deviation, achieves synergistic optimization of risk identification accuracy and computing resource allocation efficiency, ensures that high-risk traffic receives priority verification resources while avoiding unnecessary audit overhead for low-risk transactions.

[0050] The real-time verification module constructs a hardware-level secure isolation zone based on a trusted execution environment. It combines lightweight homomorphic encryption technology to compare the consistency of amount and timing in the ciphertext state of medium- and high-risk traffic, and simultaneously injects noise to obfuscate side-channel information. Under the premise of ensuring cryptographic security strength, the verification latency is compressed to the millisecond level, which effectively overcomes the high latency bottleneck of traditional secure multi-party computation protocols and achieves technical compatibility between real-time risk interception and hardware-level security protection.

[0051] The asynchronous verification module generates zero-knowledge proof computation commitments for high-risk traffic and uploads them to the distributed evidence storage network. During periods of system resource idleness, it performs deep consistency verification on the evidence storage data through a secure multi-party computation protocol. This establishes an asynchronous processing mechanism of "allowing first and then verifying," which effectively resolves the fundamental contradiction between deep security verification and business real-time requirements. It ensures that high-risk transactions receive thorough cryptographic verification without blocking normal business processes.

[0052] The automatic recovery module establishes an on-chain automated processing mechanism for risk reserve locking and difference recovery through a smart contract automatic execution engine. When cross-system data inconsistency is detected, risk funds are frozen in a timely manner and difference transfers are executed. This effectively fills the gap in fund recovery under the traditional ex-post audit model, realizes seamless connection from risk discovery to fund control, and significantly improves the timeliness of audit intervention and the reliability of fund preservation.

[0053] The evidence solidification module aggregates technical logs, commitment vouchers, and execution records output from each verification level, constructing a multi-level audit evidence package covering the entire process and visually presenting it in the form of a dynamic lineage diagram. This effectively solves the problems of evidence fragmentation and broken traceability chains in traditional audits, achieving the solidification of the integrity of audit evidence and the presentation of compliance, and providing tamper-proof technical evidence support for regulatory review and judicial evidence presentation.

[0054] This system constructs a layered and progressive secure computing architecture to address the challenge of balancing real-time response and cryptographic security in traditional secure multi-party computation protocols. Through the collaboration of a trusted execution environment and lightweight homomorphic encryption, it performs real-time consistency comparisons of medium- to high-risk traffic in its encrypted state, shortening verification response time while ensuring hardware-level security isolation. For high-risk traffic, a strategy combining zero-knowledge proof computation commitments and asynchronous deep verification in the background is adopted. Deep verification tasks are scheduled for execution during periods of system resource idleness, maintaining cryptographic security strength while avoiding interruption of real-time business processes. An on-chain automated processing mechanism for risk reserve locking and shortfall recovery is established through a smart contract automatic execution engine, directly translating audit findings into fund control actions, shortening the risk handling cycle of orphan transactions, and compensating for the lag in fund recovery through traditional post-event manual reconciliation. This layered architecture optimizes computing resource allocation, keeps audit verification costs within an economically reasonable range, balances security, real-time performance, and economy, and provides an industrially deployable technical solution for high-frequency trading scenarios, improving the feasibility of cross-system traffic data traceability and verification.

[0055] Example 2: Please refer to Figure 1 The real-time verification module includes a trusted base construction unit and a ciphertext verification unit;

[0056] The Trusted Base Construction Unit initializes a trusted execution environment within a hardware-isolated security domain, verifies memory integrity and establishes a trusted root through a remote proof mechanism, derives a hierarchical key system and distributes it to the isolated memory, performs homomorphic encryption parameter pre-configuration on medium- and high-risk traffic data, and establishes a lightweight ciphertext operation base environment.

[0057] The encrypted verification unit calls upon a lightweight encrypted computation environment to perform homomorphic encrypted computation on the difference in amount and timestamp logic, compare them, and generate real-time verification results. Simultaneously, controllable noise is dynamically injected into the computation process to obfuscate the side-channel information leakage path and block the original data inference channel based on power consumption characteristics or cache mode.

[0058] In this embodiment, the trusted base construction unit initializes a hardware-isolated secure computing domain, establishes a trusted execution base, verifies memory integrity, and forms a physical-level security boundary. This unit derives a hierarchical key system and distributes it to the isolated memory space. It pre-configures homomorphic encryption parameters for medium- and high-risk traffic data, constructing a lightweight basic environment that supports ciphertext computation. This achieves rapid deployment of the trusted computing environment and secure key distribution, providing hardware-level security support for subsequent real-time verification while reducing initialization overhead. It effectively avoids the time delays caused by key negotiation and environment construction in traditional secure multi-party computation protocols.

[0059] The ciphertext verification unit invokes the aforementioned lightweight basic environment to perform homomorphic ciphertext operations on the amount difference and timestamp logic for comparison and generates real-time verification results. Simultaneously, this unit dynamically injects controllable noise into the computation process to reconstruct the side-channel information distribution, blocking the original data inference path based on power consumption characteristics or cache timing. This completes the real-time consistency verification and side-channel protection tasks under encrypted conditions, achieving millisecond-level response and preventing information leakage while ensuring cryptographic security strength, effectively solving the side-channel attack risks present in trusted hardware environments.

[0060] Through the collaborative operation of the trusted base construction unit and the ciphertext verification unit, the real-time verification module constructs a technical architecture that integrates hardware isolation and homomorphic encryption. Compared to traditional secure multi-party computation schemes that rely on pure software cryptographic protocols, this architecture simplifies key management complexity by leveraging the hardware isolation characteristics of the trusted execution environment. Combined with lightweight homomorphic encryption technology, it compresses the response time of ciphertext computation, achieving an order-of-magnitude reduction in verification latency while maintaining memory-level security. Simultaneously, a side-channel noise injection mechanism compensates for potential vulnerabilities in the hardware trust assumption, forming a complete security protection chain from environment construction to computation execution. This provides the subsequent asynchronous deep verification module with a pre-screened and security-compliant data foundation, achieving technical compatibility between real-time risk interception capabilities and cryptographic security strength.

[0061] Example 3: Please refer to Figure 1 The trusted base construction unit includes an enclave initialization unit and a key derivation unit;

[0062] The enclave initialization unit instantiates a trusted execution enclave within the hardware isolation security domain and obtains memory measurement values ​​through a remote proof protocol. It compares the measurement values ​​with the benchmark values ​​to establish a trusted root anchor point. Based on the trusted root anchor point, it constructs the enclave memory boundary and enables a real-time encryption mechanism to block unauthorized access paths and establish a physical-level security isolation area for cryptographic operations to reside.

[0063] The key derivation unit executes a hierarchical key derivation algorithm based on the trusted root anchor to generate a hierarchical key structure and establish a binding relationship between the key and the operation domain. It maps homomorphic encryption parameters to an isolated memory space to complete pre-configuration, builds a lightweight ciphertext operation basic environment, establishes a standardized ciphertext operation interface, and forms a path of association with subsequent verification logic.

[0064] In this embodiment, the enclave initialization unit instantiates a trusted execution enclave within a hardware-isolated security domain, uses a remote proof protocol to obtain memory metrics, and compares them with a benchmark value to establish a trusted root anchor. Based on this anchor, the enclave memory boundary is constructed, and a real-time encryption mechanism is enabled to block unauthorized access paths and establish a physically-level secure isolation area. This unit completes the rapid deployment and integrity verification of the trusted execution environment, achieving the goal of providing a hardware-level secure resident environment for cryptographic operations and reducing environment construction overhead, while avoiding the time delays caused by the complex key negotiation and trust establishment processes in traditional software-level secure multi-party computation protocols.

[0065] The key derivation unit executes a hierarchical key derivation algorithm based on a trusted root anchor, generating a hierarchical key structure and establishing a binding relationship between keys and computation domains. It maps homomorphic encryption parameters to isolated memory spaces for pre-configuration. This unit constructs a lightweight ciphertext computation environment, establishes a standardized ciphertext computation interface, and forms a connection path with subsequent verification logic. This unit completes the tasks of secure key system construction and parameterized configuration of the computation environment, achieving the goal of supporting efficient ciphertext computation and seamless integration with subsequent processing modules. It solves the problem of low initialization efficiency caused by the separation of key management and computation environment configuration in traditional schemes.

[0066] Through the collaborative operation of the enclave initialization unit and the key derivation unit, the trusted base construction unit establishes a technical architecture that integrates hardware isolation and hierarchical key management. Compared to traditional secure multi-party computation schemes that rely on pure software cryptographic protocols, this architecture simplifies the establishment process of the trusted environment by leveraging hardware isolation characteristics. It quickly establishes a root of trust through a remote proof mechanism, avoiding initialization delays caused by multiple rounds of communication negotiation. The hierarchical key derivation algorithm achieves tight coupling between the key and the computation domain. Combined with the pre-configuration of homomorphic encryption parameters, it significantly reduces the preparation time for ciphertext computation, constructing a lightweight computational foundation environment. Simultaneously, the establishment of standardized interfaces enables seamless integration with subsequent ciphertext verification units, forming an efficient technical chain from environment construction to computation execution. This provides fundamental support for real-time verification that combines hardware security and computational efficiency, effectively balancing the technical requirements of security strength and initialization efficiency.

[0067] Example 4: Please refer to Figure 1 The ciphertext verification unit includes a ciphertext processing unit and a side-blocking anti-aliasing unit;

[0068] The ciphertext operation unit calls the standardized interface established by the lightweight ciphertext operation basic environment, performs homomorphic ciphertext operation comparison on the difference in amount and timestamp logic of medium and high risk traffic based on the hierarchical key system, generates real-time consistency verification results and marks anomalies, and passes the verification results and anomalies down to the subsequent processing modules.

[0069] The side-channel anti-aliasing unit dynamically senses power consumption characteristics and cache access patterns during the ciphertext operation execution cycle, injects controllable noise into the operation pipeline to reconstruct the side-channel information distribution pattern, blocks the original data inference path based on power consumption analysis or cache timing, and forms a closed-loop security protection for the entire ciphertext comparison operation process.

[0070] In this embodiment, the ciphertext operation unit calls a standardized interface to perform homomorphic ciphertext operations and comparisons on the monetary differences and timestamp logic of medium- and high-risk traffic based on a hierarchical key system. This unit generates a real-time consistency verification result and marks any anomalies, then passes the processing result down to subsequent modules, completing the real-time data consistency verification task under encrypted conditions. This achieves millisecond-level response in a trusted hardware environment, avoiding the high latency issues caused by multiple rounds of communication in traditional secure multi-party computation protocols.

[0071] The side-channel anti-obfuscation unit dynamically senses power consumption characteristics and cache access patterns during the ciphertext operation execution cycle, injecting controllable noise into the computation pipeline to reconstruct the side-channel information distribution. This unit blocks the original data inference path based on power analysis or cache timing, forming a closed-loop security protection system for the entire ciphertext comparison operation process. It completes the side-channel protection task in a trusted hardware environment, achieving the goal of mitigating potential security vulnerabilities in trusted execution environments and resolving the information leakage risks faced by pure hardware isolation solutions.

[0072] Through the collaborative operation of the ciphertext computation unit and the side-channel anti-obfuscation unit, the ciphertext verification unit constructs a technical architecture that combines real-time homomorphic encryption computation with dynamic side-channel protection. Compared to traditional solutions that rely on complex secure multi-party computation protocols or a single trusted hardware environment, this architecture utilizes lightweight homomorphic encryption to complete ciphertext comparison within a hardware-isolated domain, compressing verification response time. Simultaneously, it reconstructs the side-channel information distribution through a dynamic noise injection mechanism, establishing a complete technical chain from computation execution to security protection. This collaborative mechanism achieves technical compatibility between real-time verification response capabilities and cryptographic security strength, providing subsequent asynchronous deep verification modules with traffic data that has undergone preliminary security screening and has clear anomaly identification. While ensuring the efficiency of real-time consistency verification of cross-system traffic data, it establishes a hardware-level security defense against side-channel attacks.

[0073] Example 5: Please refer to Figure 1 The asynchronous verification module includes a zero-knowledge commitment unit and a background verification unit;

[0074] The Zero-Knowledge Commitment Unit constructs a zero-knowledge proof constraint system for extracting business characteristics of high-risk traffic data. Based on the output of the constraint system, it generates verifiable computational commitments and extracts commitment summaries. The commitment summaries and data location pointers are uploaded to a distributed evidence storage network to establish cryptographic evidence anchor points that are bound to the subsequent smart contract execution logic.

[0075] The background verification unit activates the asynchronous task scheduling queue during system resource idle periods, retrieves the encrypted share from the distributed evidence storage network and executes the secure multi-party computation protocol, performs deep consistency verification on cross-system traffic data to generate deep verification results, and inputs the deep verification results into the automatic recovery module to trigger risk reserve locking and difference recovery operations.

[0076] In this embodiment, the zero-knowledge commitment unit extracts business characteristics from high-risk traffic and maps them to the zero-knowledge proof constraint system. Based on the system output, it generates a verifiable computational commitment and a commitment summary. This unit uploads the commitment summary and data location pointer to the distributed evidence storage network, establishing a cryptographic evidence anchor point bound to the smart contract execution logic. This completes the lightweight evidence solidification and asynchronous verification preparatory tasks for high-risk traffic, achieving the goal of establishing immutable verification evidence without blocking real-time business processes and providing a cryptographic evidence foundation for subsequent automatic recourse. This effectively solves the problem of business continuity interference caused by traditional synchronous verification modes.

[0077] During periods of system resource idleness, the background verification unit activates the asynchronous task scheduling queue, retrieves the encrypted share from the distributed evidence storage network, and performs a deep consistency verification operation. This unit generates a deep verification result and inputs it into the automatic recovery module to trigger risk reserve locking and shortfall recovery operations. This completes the task of deep security verification of high-risk traffic and the transformation of verification results into fund control actions. It achieves the goal of optimizing computing resource allocation and driving the subsequent automatic recovery mechanism while ensuring cryptographic security strength, effectively resolving the resource conflict between deep verification requirements and real-time business pressure.

[0078] Existing technologies using secure multi-party computation protocols for cross-system traffic verification typically require all participating parties to complete multiple rounds of interaction and secret sharing reconstruction before business execution. This results in verification delays of hundreds of milliseconds or even seconds, failing to meet the real-time requirements of high-frequency trading scenarios. Furthermore, forced real-time verification consumes extremely high computational resources. This application constructs a technical architecture combining computational commitment and asynchronous deep verification through the collaboration of zero-knowledge commitment units and background verification units. This architecture allows business traffic to pass through after generating computational commitments, scheduling deep security verification tasks to be executed asynchronously during idle system resource periods. This effectively separates the temporal coupling between real-time business flows and deep verification flows while maintaining cryptographic security levels. This asynchronous processing mechanism not only avoids the blockage of real-time business continuity by deep verification but also reduces computational costs through idle resource utilization. Simultaneously, it establishes an immutable verification basis through computational commitments, ensuring that verification results automatically trigger subsequent recourse mechanisms. This achieves technical compatibility between security and real-time performance, as well as reasonable allocation of economic resources, providing an industrially deployable audit verification solution for high-frequency, high-concurrency scenarios.

[0079] Example 6: Please refer to Figure 1 The zero-knowledge commitment unit includes constraint construction units and commitment anchoring units;

[0080] The constraint construction unit extracts multi-dimensional business features from high-risk traffic data and maps them to the zero-knowledge proof constraint system. After establishing the constraint logic and proof parameters, it generates a circuit structure, forms a basic constraint framework for verifiable computational commitments, and passes it down to the commitment solidification processing stage.

[0081] The commitment anchoring unit extracts the commitment summary and data location pointer and uploads them to the distributed evidence storage network. It establishes a cryptographic evidence anchoring point bound to the smart contract execution logic, establishes an immutable evidence storage structure, and forms a ciphertext share call association path with the backend verification unit.

[0082] In this embodiment, the constraint construction unit extracts multi-dimensional business features from high-risk traffic and maps them to a zero-knowledge proof constraint system. After establishing the constraint logic and proof parameters, it generates a circuit structure, forming a basic constraint framework for verifiable computational commitments, which is then passed down. This process completes the task of feature structuring and commitment foundation construction for high-risk traffic, achieving the goal of providing a cryptographic constraint framework for subsequent evidence anchoring without exposing the original business data. This effectively solves the privacy leakage risk caused by plaintext data transmission in traditional verification.

[0083] The commitment anchoring unit extracts the commitment digest and data location pointer and uploads them to the distributed evidence storage network. This establishes a cryptographic evidence anchor point bound to the smart contract execution logic, establishes an immutable evidence storage structure, and forms a ciphertext share call association path with the backend verification unit. This process completes the distributed solidification of the computational commitment and the establishment of the subsequent verification association path, achieving the goal of forming traceable verification evidence while ensuring data privacy and supporting subsequent automated verification calls. This effectively avoids the single point of trust risk faced by traditional centralized evidence storage.

[0084] Existing technologies using secure multi-party computation for cross-system traffic verification typically require all participants to complete complex multi-round interactions and secret-sharing reconstructions before business execution. This results in extremely high verification latency and severely blocks real-time business processes, or forces the sacrifice of security for lightweight verification, failing to provide cryptographic-level evidence. This application constructs a technical architecture for business feature constraint and distributed solidification of computational commitments through the collaboration of constraint construction units and commitment anchoring units. The constraint construction unit maps high-risk traffic business characteristics to a zero-knowledge proof constraint system and generates a circuit structure, building a verifiable commitment foundation without exposing the original data. The commitment anchoring unit uploads the commitment digest and data location pointer to a distributed network, establishing a cryptographic evidence anchor point bound to the smart contract, forming an immutable evidence storage structure and establishing a connection path with subsequent backend verification. This architecture decouples business-first and deep verification in a timely manner, allowing traffic to pass through immediately after commitment generation. It transfers deep security verification tasks to the background for asynchronous execution, eliminating the blockage of real-time business processes while maintaining cryptographic security levels. At the same time, it establishes an immutable verification basis through distributed evidence storage, ensuring that the verification results in the background can automatically trigger the recovery mechanism. This forms a complete closed loop from commitment generation to asynchronous verification and then to fund recovery, effectively balancing the technical contradiction between verification real-time performance and security depth.

[0085] Example 7: Please refer to Figure 1 The background verification unit includes an asynchronous scheduling unit and a deep verification unit;

[0086] The asynchronous scheduling unit constructs an asynchronous task queue and establishes a task priority sequence during idle periods of system resources. Based on the encrypted share call association path established with the commitment anchoring unit, it extracts secret shared fragments from the distributed evidence storage network and constructs a multi-party computation communication topology. Within the topology, it establishes a secure communication channel between participating nodes and establishes an input association with the deep verification operation.

[0087] The deep verification unit executes a secure multi-party computation protocol within a secure communication channel to perform deep consistency verification operations on the secret-sharing shard, generating deep verification results and establishing a result transmission channel with the automatic recovery module. It maps the verification conclusions to the risk reserve locking and difference recovery operation instruction set, establishes the call association between the instruction set and the smart contract execution logic, and outputs it to the subsequent processing stage.

[0088] In this embodiment: the asynchronous scheduling unit constructs an asynchronous task queue and establishes a priority sequence during periods of system resource idleness. Based on the encrypted share call association path established with the commitment anchoring unit, it extracts secret shared fragments from the distributed evidence storage network, constructs a multi-party computation communication topology, and establishes a secure communication channel between participating nodes within the topology. This unit completes the timing optimization and execution environment preparation for the deep verification task, achieving the goal of utilizing idle computing resources without affecting real-time business traffic, and solving the problem of business continuity blocking in traditional synchronous verification modes.

[0089] The deep verification unit executes a secure multi-party computation protocol within a secure communication channel, performs deep consistency verification operations on the secret-sharing fragment, generates a deep verification result, and establishes a result transmission channel with the automatic recovery module. It maps the verification conclusion to the risk reserve locking and shortfall recovery operation instruction set. This unit completes the task of deep security verification of cross-system traffic data and the transformation of the result into fund control behavior, achieving the goal of driving the subsequent automatic recovery mechanism while maintaining cryptographic security strength, thus compensating for the shortcomings of real-time lightweight verification in terms of security depth.

[0090] Existing technologies using secure multi-party computation protocols for cross-system traffic verification typically require all participating parties to complete multiple rounds of interaction and secret sharing reconstruction before business execution, resulting in verification delays of hundreds of milliseconds or even seconds. Furthermore, mandatory real-time verification consumes extremely high computational resources, leading to a cost imbalance in auditing. This application constructs a technical architecture combining computational resource timing optimization and deep security verification through the collaboration of an asynchronous scheduling unit and a deep verification unit. The asynchronous scheduling unit decouples the deep verification task from the real-time business flow, scheduling it for execution during idle system resource periods, fully utilizing idle computing power and avoiding blocking normal business operations. The deep verification unit performs cryptographic-level deep verification within a constructed secure communication channel, ensuring that the verification results can be directly mapped to fund control instructions such as risk reserve fund locking. This collaborative mechanism achieves temporal decoupling between deep security verification and business real-time performance, maintaining cryptographic security levels while keeping computational resource consumption within an economically reasonable range. Simultaneously, it ensures that verification results automatically drive subsequent recovery mechanisms, forming an automated closed loop from deep verification to fund control.

[0091] Example 8: Please refer to Figure 1 The automatic recovery module includes a funds locking unit and a contract execution unit;

[0092] The fund locking unit receives the verification conclusions output by the deep verification unit and parses the risk identifier mapping relationship. Based on the smart contract automatic execution engine, it generates a risk reserve fund locking instruction, establishes a binding relationship between the reserve fund account and the abnormal flow identifier, forms a fund control status, and establishes a call association with the contract execution unit.

[0093] The contract execution unit invokes the fund control status and parses the shortfall recovery instruction set. Through the smart contract automatic execution engine, it completes the reserve release and fund transfer operations, generates execution records, establishes a data association channel with the evidence solidification module, establishes the call association with the subsequent audit evidence generation, and passes it down.

[0094] In this embodiment: the funds locking unit receives the verification conclusion output by the deep verification unit, analyzes the mapping relationship between risk identifiers and abnormal flow, generates a risk reserve fund locking instruction based on smart contract logic, and establishes a binding relationship between the reserve fund account and the abnormal identifier, thereby forming a funds control state and establishing a call association with subsequent execution steps. This process realizes the instant transformation of verification results into funds control behavior, and immediately blocks the outflow of funds when cross-system data inconsistency or unilateral account anomalies are detected, effectively shortening the time delay from risk discovery to control measure initiation under the traditional ex-post manual reconciliation model, and establishing a financial foundation for subsequent difference recovery.

[0095] The contract execution unit invokes the fund control status parsing and deficit recovery instruction set, and completes the reserve release and fund transfer operations through on-chain automatic execution logic. Simultaneously, it generates execution records and establishes a data association channel with the evidence solidification module. This process automates the entire process from audit discovery to fund recovery, directly mapping verification conclusions to executable fund operation instructions. This avoids the time lag and operational uncertainty of traditional manual reconciliation and offline recovery models, ensuring that risk funds are quickly collected or deficits are covered, and simultaneously solidifying the execution trajectory for subsequent audit traceability.

[0096] Existing technologies, upon discovering inconsistencies in cross-system traffic data or orphan transactions, typically rely on manual reconciliation and offline coordination to initiate recovery procedures. This results in a gap of several days or even longer between risk discovery and actual control of funds, during which the risk of fund transfer is extremely high and difficult to reverse. This application constructs a closed-loop technology that automatically transforms verification results into fund control behavior through on-chain collaboration between a fund locking unit and a contract execution unit. The fund locking unit triggers the locking of risk reserves the moment it receives the deep verification conclusion, preventing further outflow of funds; the contract execution unit automatically parses the recovery instruction and completes the fund transfer, compressing the traditional manual approval process into automated on-chain execution. This architecture shortens the recovery response time from several days to hours or even minutes, significantly reducing the risk of fund loss. At the same time, through the instant generation and downward transmission of execution records, it achieves the synchronous solidification of audit evidence, providing a complete fund control trajectory for subsequent regulatory review, effectively filling the gap in the delayed fund recovery of traditional post-audit models.

[0097] Example 9: Please refer to Figure 1 The fund locking unit includes a risk analysis unit and a locking execution unit;

[0098] The risk analysis unit receives the verification conclusions output by the deep verification unit and analyzes the mapping relationship between risk identifiers and abnormal flow characteristics. It extracts the associated attributes and status information of the risk reserve account, forms locking decision parameters based on the mapping relationship and status information, establishes the input relationship with the instruction generation logic, and passes it down to the locking execution processing stage.

[0099] The execution unit is locked, the locking decision parameters are called to activate the smart contract automatic execution engine to generate a risk reserve locking instruction, establish a binding relationship between the reserve account and the abnormal flow identifier, form a fund control status and establish a call association path with the contract execution unit, and output a standardized locking certificate to the subsequent processing stage.

[0100] In this embodiment, the risk analysis unit extracts the associated attributes and status information of the risk reserve account by analyzing the mapping relationship between the verification conclusion and the abnormal flow characteristics, forming locking decision parameters and establishing the input relationship with subsequent execution logic. This completes the structured transformation of verification results into fund control decisions, achieving the goal of accurately identifying risky accounts and generating basic parameters for locking instructions. It solves the problems of delayed and easily overlooked risk identification in the traditional manual review model, providing accurate decision input for automated fund control.

[0101] The locking execution unit calls the locking decision parameters to activate the smart contract's automatic execution engine, establishes a binding relationship between the reserve account and the abnormal flow identifier, forms a fund control state, establishes the call association path with the contract execution unit, and outputs standardized locking credentials. This completes the automated locking and state solidification of the risk reserve, achieving the goal of immediately blocking the outflow of funds and establishing subsequent recovery links. It avoids the time delays and operational uncertainties of traditional manual operation modes, forming an automatically triggered closed-loop fund control system.

[0102] Existing technologies, upon discovering inconsistencies in cross-system data, typically rely on manual identification of risky accounts and offline freezing requests. This results in a significant time delay between the verification conclusion and actual fund control, making it difficult to prevent and trace fund transfers during this period. This application constructs a technical architecture for the automatic transformation of verification conclusions into fund control status through on-chain collaboration between a risk analysis unit and a locking execution unit. The risk analysis unit deeply analyzes the risk identifier mapping relationship in the verification conclusion, extracts account attributes and status information, and forms standardized locking decision parameters, replacing the subjective judgment process of manual review. The locking execution unit instantly activates the automatic execution engine based on the decision parameters, establishes a binding relationship between the account and the anomaly identifier, and solidifies the fund control status, compressing the time delay from verification to locking to minutes or even seconds, significantly reducing the risk of fund loss. Simultaneously, through the generation and downward transmission of standardized locking vouchers, the fund control trajectory is instantly solidified, providing a complete technical evidence chain for subsequent audit tracing, forming a seamless connection from deep verification to fund control to evidence preservation, effectively filling the gap in the lag in fund preservation in traditional post-audit models.

[0103] Example 10: Please refer to Figure 1 The contract execution unit includes an instruction parsing unit and an on-chain execution unit;

[0104] The instruction parsing unit receives the fund control status output by the fund locking unit, parses the operation type and amount parameters in the shortfall recovery instruction set, maps the parsing results to the smart contract execution logic to generate standardized execution instruction packages, establishes an input association with the on-chain execution unit, and passes it down.

[0105] The on-chain execution unit calls the standardized execution instruction package to activate the smart contract automatic execution engine, executes reserve release and fund transfer operations, synchronously generates execution records, establishes a data association channel with the evidence solidification module, establishes a call association with the subsequent audit evidence generation, and outputs it to the processing stage.

[0106] In this embodiment: the instruction parsing unit receives the fund control status output by the fund locking unit, performs structured parsing of the operation type and amount parameters in the difference recovery instruction set, maps the parsing results to the smart contract execution logic to generate standardized execution instruction packages, establishes an input association with the on-chain execution unit, and passes it down. This unit completes the structured processing and standardized encapsulation of recovery instructions, achieving the goal of transforming verification conclusions into automatically executable technical instructions. This solves the problem of delayed and error-prone instruction parsing in the traditional manual approval mode, providing a standardized input foundation for on-chain automatic execution.

[0107] The on-chain execution unit invokes standardized execution instruction packages to activate the smart contract's automatic execution engine, executing reserve release and fund transfer operations. Simultaneously, it generates execution records and establishes a data link with the evidence solidification module, confirming its connection to subsequent audit evidence generation. This unit completes the on-chain transformation of verification conclusions into fund control actions, achieving automated execution of recovery operations and synchronous solidification of execution trajectories. It solves the time delay and operational uncertainty problems of traditional offline recovery models, forming an auditable closed-loop fund control system.

[0108] Existing technologies rely on manual approval and offline bank operations in the recovery and enforcement phase, resulting in delays of several days from instruction generation to actual fund transfer. Furthermore, the lack of real-time recording of the process makes auditing and traceability difficult. This application constructs a closed-loop technology that automatically transforms verification conclusions into fund control actions through the collaboration of an instruction parsing unit and an on-chain execution unit. The instruction parsing unit structurally parses and maps the fund control status into standardized instruction packages, eliminating the subjective delays of manual parsing. The on-chain execution unit, through an automated execution engine, instantly completes fund transfers and simultaneously generates execution records. This architecture compresses the recovery response time from several days to minutes, avoiding the uncertainty risks of manual operations. Simultaneously, the instant generation and downward transmission of execution records achieves the synchronous solidification of audit evidence, forming a complete automated chain from verification conclusions to fund control and evidence preservation, effectively filling the execution lag defects of traditional post-event recovery models.

[0109] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A system for tracing and verifying flow data used in accounting and auditing, characterized by: It includes a traffic classification module, a real-time verification module, an asynchronous verification module, an automatic compensation module, and an evidence solidification module; The traffic classification module extracts local risk features from transaction data through a multi-source collaborative analysis mechanism, generates dynamic risk scores, and classifies risk levels. The real-time verification module uses a secure computing hardware environment to verify the consistency of amount and timing in encrypted medium- and high-risk traffic. The asynchronous verification module generates cryptographic computation commitments for high-risk traffic and uploads them for evidence storage, and completes deep verification in the background through distributed privacy computing; When the automatic recovery module detects inconsistencies in cross-system data, it locks the risk reserve and performs the difference recovery by automatically executing a contract on the blockchain. The evidence solidification module aggregates the verification logs, commitment vouchers, and execution records output by each module to generate a multi-level audit evidence package and output a traceability graph.

2. The accounting and auditing flow data traceability and verification system according to claim 1, characterized in that, The real-time verification module includes a trusted base construction unit and a ciphertext verification unit; The Trusted Base Construction Unit initializes a trusted execution environment within a hardware-isolated security domain, verifies memory integrity and establishes a trusted root through a remote proof mechanism, derives a hierarchical key system and distributes it to the isolated memory, performs homomorphic encryption parameter pre-configuration on medium- and high-risk traffic data, and establishes a lightweight ciphertext operation base environment. The encrypted verification unit calls upon a lightweight encrypted computation environment to perform homomorphic encrypted computation on the difference in amount and timestamp logic, compare them, and generate real-time verification results. Simultaneously, controllable noise is dynamically injected into the computation process to obfuscate the side-channel information leakage path and block the original data inference channel based on power consumption characteristics or cache mode.

3. The accounting and auditing flow data traceability and verification system according to claim 1, characterized in that, The trusted base construction unit includes an enclave initialization unit and a key derivation unit; The enclave initialization unit instantiates a trusted execution enclave within a hardware-isolated security domain and obtains a memory metric value through a remote proof protocol. It compares the metric value with a benchmark value to establish a trusted root anchor point. Based on the trusted root anchor point, it constructs the enclave memory boundary and enables a real-time encryption mechanism to block unauthorized access paths, establishing a physically-level secure isolation area for cryptographic operations to reside in. The key derivation unit executes a hierarchical key derivation algorithm based on the trusted root anchor to generate a hierarchical key structure and establish a binding relationship between the key and the operation domain. It maps homomorphic encryption parameters to an isolated memory space to complete pre-configuration, builds a lightweight ciphertext operation basic environment, establishes a standardized ciphertext operation interface, and forms a path of association with subsequent verification logic.

4. The accounting and auditing flow data traceability and verification system according to claim 1, characterized in that, The ciphertext verification unit includes a ciphertext processing unit and a side-blocking anti-aliasing unit; The ciphertext operation unit calls the standardized interface established by the lightweight ciphertext operation basic environment, performs homomorphic ciphertext operation comparison on the difference in amount and timestamp logic of medium and high risk traffic based on the hierarchical key system, generates real-time consistency verification results and marks anomalies, and passes the verification results and anomalies down to the subsequent processing modules. The side-channel anti-aliasing unit dynamically senses power consumption characteristics and cache access patterns during the ciphertext operation execution cycle, injects controllable noise into the operation pipeline to reconstruct the side-channel information distribution pattern, blocks the original data inference path based on power consumption analysis or cache timing, and forms a closed-loop security protection for the entire ciphertext comparison operation process.

5. The accounting and auditing data traceability and verification system according to claim 1, characterized in that, The asynchronous verification module includes a zero-knowledge commitment unit and a background verification unit; The Zero-Knowledge Commitment Unit constructs a zero-knowledge proof constraint system for extracting business characteristics of high-risk traffic data. Based on the output of the constraint system, it generates verifiable computational commitments and extracts commitment summaries. The commitment summaries and data location pointers are uploaded to a distributed evidence storage network to establish cryptographic evidence anchor points that are bound to the subsequent smart contract execution logic. The background verification unit activates the asynchronous task scheduling queue during system resource idle periods, retrieves the encrypted share from the distributed evidence storage network and executes the secure multi-party computation protocol, performs deep consistency verification on cross-system traffic data to generate deep verification results, and inputs the deep verification results into the automatic recovery module to trigger risk reserve locking and difference recovery operations.

6. The accounting and auditing flow data traceability and verification system according to claim 1, characterized in that, Zero-knowledge commitment units include constraint construction units and commitment anchoring units; The constraint construction unit extracts multi-dimensional business features from high-risk traffic data and maps them to the zero-knowledge proof constraint system. After establishing the constraint logic and proof parameters, it generates a circuit structure, forms a basic constraint framework for verifiable computational commitments, and passes it down to the commitment solidification processing stage. The commitment anchoring unit extracts the commitment summary and data location pointer and uploads them to the distributed evidence storage network. It establishes a cryptographic evidence anchoring point bound to the smart contract execution logic, establishes an immutable evidence storage structure, and forms a ciphertext share call association path with the backend verification unit.

7. The accounting and auditing data traceability and verification system according to claim 1, characterized in that, The background verification unit includes an asynchronous scheduling unit and a deep verification unit; The asynchronous scheduling unit constructs an asynchronous task queue and establishes a task priority sequence during idle periods of system resources. Based on the encrypted share call association path established with the commitment anchoring unit, it extracts secret shared fragments from the distributed evidence storage network and constructs a multi-party computation communication topology. Within the topology, it establishes a secure communication channel between participating nodes and establishes an input association with the deep verification operation. The deep verification unit executes a secure multi-party computation protocol within a secure communication channel to perform deep consistency verification operations on the secret-sharing shard, generating deep verification results and establishing a result transmission channel with the automatic recovery module. It maps the verification conclusions to the risk reserve locking and difference recovery operation instruction set, establishes the call association between the instruction set and the smart contract execution logic, and outputs it to the subsequent processing stage.

8. The accounting and auditing flow data traceability and verification system according to claim 1, characterized in that, The automatic recovery module includes a funds locking unit and a contract execution unit; The fund locking unit receives the verification conclusions output by the deep verification unit and parses the risk identifier mapping relationship. Based on the smart contract automatic execution engine, it generates a risk reserve fund locking instruction, establishes a binding relationship between the reserve fund account and the abnormal flow identifier, forms a fund control status, and establishes a call association with the contract execution unit. The contract execution unit invokes the fund control status and parses the shortfall recovery instruction set. Through the smart contract automatic execution engine, it completes the reserve release and fund transfer operations, generates execution records, establishes a data association channel with the evidence solidification module, establishes the call association with the subsequent audit evidence generation, and passes it down.

9. The accounting and auditing flow data traceability and verification system according to claim 1, characterized in that, The fund locking unit includes a risk analysis unit and a locking execution unit; The risk analysis unit receives the verification conclusions output by the deep verification unit and analyzes the mapping relationship between risk identifiers and abnormal flow characteristics. It extracts the associated attributes and status information of the risk reserve account, forms locking decision parameters based on the mapping relationship and status information, establishes the input relationship with the instruction generation logic, and passes it down to the locking execution processing stage. The execution unit is locked, the locking decision parameters are called to activate the smart contract automatic execution engine to generate a risk reserve locking instruction, establish a binding relationship between the reserve account and the abnormal flow identifier, form a fund control status and establish a call association path with the contract execution unit, and output a standardized locking certificate to the subsequent processing stage.

10. The accounting and auditing flow data traceability and verification system according to claim 1, characterized in that, The contract execution unit includes an instruction parsing unit and an on-chain execution unit; The instruction parsing unit receives the fund control status output by the fund locking unit, parses the operation type and amount parameters in the difference recovery instruction set, maps the parsing results to the smart contract execution logic to generate a standardized execution instruction package, establishes an input association with the on-chain execution unit, and passes it down; The on-chain execution unit calls the standardized execution instruction package to activate the smart contract automatic execution engine, executes reserve release and fund transfer operations, synchronously generates execution records, establishes a data association channel with the evidence solidification module, establishes a call association with the subsequent audit evidence generation, and outputs it to the processing stage.

Images