Data transmission route control method and device, electronic equipment and storage medium

Abstract

The application relates to the technical field of data security and compliance management, and discloses a data transmission route control method and device, electronic equipment and a storage medium, wherein the method comprises the following steps: generating a lightweight path fingerprint containing a source area, an intermediate area list, a target area and a behavior type on a data stream transmission path; sending the fingerprint and the data stream to an intermediate node; performing a matching query in a local cache based on the fingerprint by the intermediate node and generating a local compliance judgment result; if a hit is obtained, constructing a local compliance feature vector and comparing the local compliance feature vector with a benchmark vector to determine whether to return an enhanced signal; and an export node fuses the enhanced signal to generate a comprehensive route decision and controls a route action. The application has the beneficial technical effects that the perception and distributed collaborative detection of multi-hop data streams under a complex geopolitical path are realized, and the accuracy and real-time performance of compliance judgment are improved.

Description

Technical Field

[0001] This invention relates to the field of data transmission routing control technology, and in particular to a data transmission routing control method, apparatus, electronic device, and storage medium. Background Technology

[0002] With the accelerated development of the global digital economy, multinational corporations, cloud service providers, and cross-border business platforms continuously generate and transmit massive amounts of multi-source, heterogeneous data in their daily operations, covering personal information, trade secrets, financial transactions, and health and medical information. When this data flows between different jurisdictions, it must simultaneously meet multi-layered and dynamically evolving compliance requirements. Currently, cross-border data flow management practices generally rely on static policy configurations, pre-set rule engines, and manual review mechanisms. Compliance determination of cross-border data flows typically requires a comprehensive assessment based on multi-dimensional information, including the data subject's identity, processing purpose, the recipient's jurisdiction, and the specific business context of the transmission scenario.

[0003] Existing cross-border data compliance control systems have long faced the contradiction of lacking multi-hop path awareness capabilities and insufficient accuracy in global decision-making when dealing with complex geographical paths and dynamic regulatory environments. The updating of global detection knowledge relies on centralized push, and there is a lack of an effective collaborative evolution framework among nodes, which restricts the system's ability to continuously adapt to new compliance threats. Summary of the Invention

[0004] Therefore, it is necessary to propose a data transmission routing control method, device, electronic device, and storage medium to address the existing data transmission routing control problem.

[0005] A data transmission routing control method, the method comprising: Along the data stream transmission path, a lightweight path fingerprint corresponding to the transmission path is generated; wherein, the lightweight path fingerprint includes at least a source region, a list of intermediate regions, a target region, and a behavior type; Based on the transmission path, the lightweight path fingerprint and the data stream are sequentially sent to the intermediate node; Each of the intermediate nodes receives the data stream and the lightweight path fingerprint, performs a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint, and generates a local compliance judgment result; If the local compliance determination result indicates that the matching query has matched the corresponding compliance event record, then the intermediate node constructs a local compliance feature vector. The local compliance feature vector is compared with the baseline feature vector in the local compliance event fingerprint cache, and based on the comparison result, it is determined whether to generate and send an enhanced signal back to the exit node. The exit node receives the enhanced signals returned by each intermediate node, merges the enhanced signals to generate a comprehensive routing decision, and controls the routing action of the data flow on the transmission path according to the comprehensive routing decision.

[0006] Further, the step of generating a lightweight path fingerprint corresponding to the transmission path of the data stream includes: Obtain the source region identifier of the data stream, the list of intermediate regions the data stream is expected to pass through, the target region identifier, and the behavior type of the data stream; The source region identifier, the intermediate region list, the target region identifier, and the behavior type are assembled into a path fingerprint tuple according to a predetermined format. The path fingerprint tuple is hashed and a digital signature is appended to generate the lightweight path fingerprint.

[0007] Further, the step of performing a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint includes: Extract the source region, the intermediate region list, and the target region from the lightweight path fingerprint; Multiple retrieval strategies are generated based on the source region, the intermediate region list, and the target region; wherein, the multiple retrieval strategies are used to retrieve at least one of the following combinations in the compliance event record: the combination between the source region and the current intermediate node, the combination between the current intermediate node and the target region, and the combination from the source region to the target region via the current intermediate node.

[0008] Further, the step of constructing a local compliance feature vector by the intermediate node if the local compliance determination result indicates that the matching query has matched the corresponding compliance event record includes: If the local compliance determination result indicates that the matching query has matched the corresponding compliance event record, then the intermediate node extracts the path topology features, data type features, and transmission destination features corresponding to the behavior type of the data stream. Obtain the credibility evaluation parameters of the intermediate node from the preset regional authentication mapping table; The path topology features, data type features, transmission destination features, and credibility evaluation parameters are fused into the local compliance feature vector.

[0009] Further, the step of controlling the routing action of the data flow on the transmission path based on the comprehensive routing decision includes: The exit node parses the comprehensive routing decision to obtain the corresponding routing action; wherein, the routing action includes blocking the data stream from continuing to transmit, limiting the transmission rate, or allowing the transmission. The exit node sends control commands corresponding to the routing action to each intermediate node on the transmission path through the service mesh sidecar proxy mode.

[0010] Furthermore, after the step of the exit node issuing control commands corresponding to the routing action to each intermediate node on the transmission path through the service mesh sidecar proxy mode, the method further includes: If the routing action is to block transmission, the exit node generates a standardized error code and feeds it back to the source application of the data stream; If the routing action is rate-limited transmission, the exit node synchronously sends rate-limiting parameters to each intermediate node through which the data stream passes.

[0011] Furthermore, the step of generating a comprehensive routing decision by fusing the enhanced signals returned by each intermediate node at the exit node includes: The exit node collects the enhanced signals transmitted back from all intermediate nodes along the transmission path; The matching strength of each enhanced signal is weighted and averaged according to the credibility evaluation parameters of the corresponding intermediate nodes to obtain the global compliance risk score; The global compliance risk score is compared with preset blocking thresholds and rate limiting thresholds to generate the comprehensive routing decision.

[0012] A data transmission routing control device, the device comprising: A generation module is used to generate a lightweight path fingerprint corresponding to the transmission path of the data stream; wherein the lightweight path fingerprint includes at least a source region, a list of intermediate regions, a target region, and a behavior type. The sending module is used to sequentially send the lightweight path fingerprint and the data stream to the intermediate nodes based on the transmission path; The receiving module is used for each of the intermediate nodes to receive the data stream and the lightweight path fingerprint, perform a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint, and generate a local compliance judgment result. A construction module is used to construct a local compliance feature vector by the intermediate node if the local compliance determination result indicates that the matching query matches the corresponding compliance event record. The comparison module is used to compare the local compliance feature vector with the baseline feature vector in the local compliance event fingerprint cache, and determine whether to generate and send the enhancement signal back to the exit node based on the comparison result; The control module is used to receive the enhanced signals returned by each intermediate node at the exit node, fuse the enhanced signals to generate a comprehensive routing decision, and control the routing action of the data flow on the transmission path according to the comprehensive routing decision.

[0013] An electronic device includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the following steps: Along the data stream transmission path, a lightweight path fingerprint corresponding to the transmission path is generated; wherein, the lightweight path fingerprint includes at least a source region, a list of intermediate regions, a target region, and a behavior type; Based on the transmission path, the lightweight path fingerprint and the data stream are sequentially sent to the intermediate node; Each of the intermediate nodes receives the data stream and the lightweight path fingerprint, performs a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint, and generates a local compliance judgment result; If the local compliance determination result indicates that the matching query has matched the corresponding compliance event record, then the intermediate node constructs a local compliance feature vector. The local compliance feature vector is compared with the baseline feature vector in the local compliance event fingerprint cache, and based on the comparison result, it is determined whether to generate and send an enhanced signal back to the exit node. The exit node receives the enhanced signals returned by each intermediate node, merges the enhanced signals to generate a comprehensive routing decision, and controls the routing action of the data flow on the transmission path according to the comprehensive routing decision.

[0014] A computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the following steps: Along the data stream transmission path, a lightweight path fingerprint corresponding to the transmission path is generated; wherein, the lightweight path fingerprint includes at least a source region, a list of intermediate regions, a target region, and a behavior type; Based on the transmission path, the lightweight path fingerprint and the data stream are sequentially sent to the intermediate node; Each of the intermediate nodes receives the data stream and the lightweight path fingerprint, performs a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint, and generates a local compliance judgment result; If the local compliance determination result indicates that the matching query has matched the corresponding compliance event record, then the intermediate node constructs a local compliance feature vector. The local compliance feature vector is compared with the baseline feature vector in the local compliance event fingerprint cache, and based on the comparison result, it is determined whether to generate and send an enhanced signal back to the exit node. The exit node receives the enhanced signals returned by each intermediate node, merges the enhanced signals to generate a comprehensive routing decision, and controls the routing action of the data flow on the transmission path according to the comprehensive routing decision.

[0015] The beneficial effects of this invention are as follows: By generating a lightweight path fingerprint containing a list of source regions, intermediate regions, target regions, and behavior types along the data flow transmission path, and sending data flows to intermediate nodes based on this fingerprint, each intermediate node can utilize its local compliance event fingerprint cache to perform matching queries and generate local compliance judgment results. Furthermore, when a matching query hits a compliance event record, the intermediate node constructs a local compliance feature vector by extracting path topology features, data type features, transmission destination features, and credibility evaluation parameters. This vector is then compared with a baseline feature vector to determine whether to send back an enhanced signal. This allows the exit node to receive and fuse the enhanced signals from each intermediate node to generate a comprehensive routing decision. This process achieves explicit awareness and distributed collaborative detection across the entire path from the source to the target. By introducing the credibility evaluation parameters of relay nodes into the construction of local feature vectors and global signal fusion, the differences in compliance risks among different transit regions are effectively quantified. This effectively solves the technical problem that existing cross-border data compliance control systems have long faced in dealing with complex geographical paths and dynamic regulatory environments, which is the lack of multi-hop path awareness and insufficient accuracy of global decision-making. Therefore, it avoids the situation of delayed response to policy changes and difficulty in audit traceability caused by the inability to identify intermediate jurisdictions, and improves the accuracy of compliance judgment, real-time system response and verifiability of evidence throughout the cross-border data flow process. Attached Figure Description

[0016] 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 some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] in: Figure 1 This is an application environment diagram of the data transmission routing control method in one embodiment; Figure 2 This is a flowchart of a data transmission routing control method in one embodiment; Figure 3 This is a structural block diagram of a data transmission routing control device in one embodiment; Figure 4 This is a structural block diagram of an electronic device in one embodiment. Detailed Implementation

[0018] 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.

[0019] Figure 1 This is a diagram illustrating a data transmission routing control application environment in one embodiment. (Refer to...) Figure 1 This data transmission routing control method is applied to a data transmission routing control system. The data transmission routing control system includes a terminal 110 and a server 120. The terminal 110 and server 120 are connected via a network. The terminal 110 can be a desktop terminal or a mobile terminal; the mobile terminal can be at least one of a mobile phone, tablet, or laptop. The server 120 can be a standalone server or a server cluster consisting of multiple servers. The terminal 110 is used to generate a lightweight path fingerprint corresponding to the transmission path, and the server 120 is used to store the data transmission evidence chain.

[0020] like Figure 2 As shown, in one embodiment, a data transmission routing control method is provided. This method can be applied to both terminals and servers; this embodiment uses terminal application as an example. The data transmission routing control method specifically includes the following steps: S1: Generate a lightweight path fingerprint corresponding to the data flow transmission path, wherein the lightweight path fingerprint includes at least the source region, the intermediate region list, the target region, and the behavior type.

[0021] The lightweight path fingerprint, serving as a lightweight signaling carrier accompanying data flow transmission, is used to explicitly declare the compliance context of the flow path between distributed network nodes. This fingerprint employs a compact binary or text-encoded format, and its internal structure consists of multiple fields assembled in a predetermined order. Specifically, the lightweight path fingerprint includes a source region identifier, an intermediate region list, a destination region identifier, and a behavior type. The source region identifier indicates the initiating jurisdiction or network region of the data flow; the intermediate region list records the ordered sequence of regions belonging to all transit nodes the data flow is expected to traverse; the destination region identifier indicates the final receiving region of the data flow; and the behavior type characterizes the business intent of the data transmission (such as cross-border data storage, remote computing, backup archiving, etc.). This allows each hop node to parse complete path topology information from the fingerprint, providing a contextual basis for subsequent localized compliance determinations. Furthermore, to ensure the integrity and non-repudiation of fingerprints during transmission, lightweight path fingerprints are typically appended with digital signatures or hash checksums, forming a compliant fingerprint watermark chain data structure to prevent malicious tampering or forgery. It's important to note that "lightweight" refers to the path fingerprint using a compact binary encoding format, with a data size far smaller than the data stream itself. The appended digital signature occupies only a small number of bytes and does not substantially impact data transmission efficiency. The fingerprint generation process is usually completed at the source exit node or gateway of the data stream. By extracting data packet metadata or stream identification information, the aforementioned fields are encapsulated into a path fingerprint tuple according to a predetermined format, and a secure signature is appended before being injected into the data stream header or extended fields. The compliant fingerprint watermark chain is a data structure formed by linking the path fingerprint tuple, hash checksum, and digital signature in a predetermined order to ensure the integrity, immutability, and non-repudiation of the path fingerprint during transmission. This lightweight path marking method allows the semantic information required for compliance control to be non-intrusively embedded into the network transport layer, providing necessary data input for collaborative detection by intermediate nodes. The update of the local compliance event fingerprint cache can adopt a federated learning approach: each intermediate node updates its local fingerprint database based on newly emerging compliance events, and encrypts and uploads the updated fingerprint features (excluding the original data) to the central server; the central server aggregates the updates from each node to generate a global fingerprint model, and then distributes it to each node. This method achieves collaborative updating of the fingerprint database while protecting the data privacy of each node.

[0022] In one embodiment, step S1, which generates a lightweight path fingerprint corresponding to the transmission path of the data stream, includes: S101: Obtain the source region identifier of the data stream, the list of intermediate regions the data stream is expected to pass through, the target region identifier, and the behavior type of the data stream; S102: Assemble the source region identifier, the intermediate region list, the target region identifier, and the behavior type into a path fingerprint tuple according to a predetermined format; S103: Perform a hash operation on the path fingerprint tuple and attach a digital signature to generate the lightweight path fingerprint.

[0023] As described in steps S101-S103 above, during the data flow initiation of a transmission request phase, key elements required for constructing a path fingerprint are extracted from the traffic context or business request message. The source region identifier is used to characterize the jurisdiction or geographical location of the data initiator, and its form can be, for example, a standardized region code; the destination region identifier is used to characterize the jurisdiction of the final data recipient. The intermediate region list records all transit nodes or transit regions that the data flow is expected to pass through during transmission. This list contains one or more region identifiers to fully characterize the multi-hop transmission path. The behavior type defines the type of operation performed by the data flow on the current path, such as 'export', 'transit', or 'import', etc. This parameter serves as key contextual information for determining whether a specific compliance rule applies. By systematically acquiring the above four core elements, a data foundation is laid for subsequently generating path tuples with legal semantics.

[0024] The process involves assembling the acquired basic elements into structured path fingerprint tuples based on predefined serialization rules. This predefined format specifies the order and data structure of each element within the tuple, for example, using a tuple structure like (source region identifier, [intermediate region list], target region identifier, behavior type). The path fingerprint tuple is not simply a string concatenation, but a structured data sequence representing the semantics of the data flow path. It clearly defines the starting point, waypoints, ending point, and flow purpose, enabling the subsequent compliance matching engine to parse a precise compliance determination path based on this tuple. This assembly process can involve converting each element into a string and concatenating them, or encapsulating them using binary encoding. The resulting data entity is stored in a memory buffer, awaiting integrity protection processing.

[0025] To ensure the integrity and immutability of the path fingerprint during transmission, hash operations and digital signatures are performed on the assembled path fingerprint tuple. A one-way hash algorithm maps path fingerprint tuples of arbitrary length to a fixed-length hash digest. This digest uniquely represents the tuple content; any subtle modification to the tuple will result in a significant change in the hash value, thus verifying content integrity. Subsequently, the private key held by the exit node is used to encrypt either the hash digest or the tuple itself, generating a signature. This signature is strongly bound to the identity of the exit node, allowing the receiver to verify the authenticity of the signature using the corresponding public key, thereby confirming that the path fingerprint was indeed sent from a legitimate source and has not been tampered with by intermediate nodes. Through the dual protection of hash operations and digital signatures, the generated lightweight path fingerprint possesses extremely high reliability and can be directly carried in network protocol headers (such as HTTP custom header fields) for transmission, providing verifiable path determination criteria for distributed intermediate nodes.

[0026] S2: Send lightweight path fingerprints and data streams sequentially to intermediate nodes based on the transmission path.

[0027] After generating the lightweight path fingerprint, the system encapsulates it along with the original data stream to form a message or packet to be transmitted. The transmission process strictly follows the network topology logic of the transmission path, forwarding the encapsulated data unit sequentially to each intermediate node along the path. During this process, the lightweight path fingerprint acts as control signaling, flowing synchronously with the data stream to ensure that each intermediate node receives complete path context information. The transmission action can be implemented based on existing network protocols (such as TCP / IP (Transmission Control Protocol / Internet Protocol), HTTP / 2 (Hypertext Transfer Protocol Version 2), etc.) through header field extensions or dedicated encapsulation protocols, ensuring that the fingerprint information is tightly bound to and inseparable from the business data. By synchronously distributing the path fingerprint along with the data stream, subsequent compliance detection mechanisms can be triggered in a distributed manner at key nodes along the path, achieving a deep integration of compliance control and data transmission.

[0028] S3: Each intermediate node receives the data stream and lightweight path fingerprint, performs a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint, and generates a local compliance judgment result.

[0029] When data flows to an intermediate node, the node's proxy module or processing unit parses a lightweight path fingerprint from the data packet. Subsequently, the node uses key information extracted from the fingerprint, such as the source region, intermediate region list, and target region, to retrieve data from a locally maintained local compliance event fingerprint cache. The local compliance event fingerprint cache is a pre-built and continuously updated high-speed storage structure used to store compliance rules, historical event records, and risk fingerprints related to the node's jurisdiction. This cache is typically built during node initialization or offline phases, aggregating local regulatory texts, compliance rule announcements, and historical compliance precedents, transforming them into searchable fingerprint vectors or hash indexes. At runtime, this cache supports millisecond-level fast queries to meet the real-time processing demands of high-speed network pipelines. To address dynamic regulatory changes, the cache's update mechanism can, for example, employ a lightweight fingerprint federated learning framework to collaboratively train and distribute the latest compliance fingerprint features while protecting the data privacy of each node.

[0030] In one embodiment, step S3, which involves performing a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint, includes: S301: Extract the source region, the intermediate region list, and the target region from the lightweight path fingerprint; S302: Generate multiple retrieval strategies based on the source region, the intermediate region list, and the target region; wherein, the multiple retrieval strategies are used to retrieve at least one of the following combinations in the compliance event record: the combination between the source region and the current intermediate node, the combination between the current intermediate node and the target region, and the combination from the source region to the target region via the current intermediate node.

[0031] As described in steps S301-S302 above, during the matching query process, the system generates various retrieval strategies based on path fingerprints. These strategies include combinations between the source region and the current intermediate node, combinations between the current intermediate node and the target region, or end-to-end combinations from the source region through the current intermediate node to the target region. This comprehensive assessment of the compliance status of the data flow at the current node is crucial. If the query matches a specific compliance event record stored in the cache (such as prohibited cross-border transmission paths or restricted data type combinations), the data flow is deemed to have compliance risks at the current node, resulting in a positive local compliance determination. If no match is found, the data flow is deemed compliant. Furthermore, during the matching query, the system can introduce a dual-axis model of event fingerprint spatiotemporal decay. This model dynamically adjusts the matching weight based on the event fingerprint's temporal decay (e.g., the validity period of regulations) and geographical decay (e.g., the scope of geographical application), preventing the accidental triggering of expired or invalid strategies. This real-time matching mechanism based on local caching enables intermediate nodes to possess independent compliance awareness capabilities, allowing them to complete preliminary risk screening without relying on centralized services.

[0032] The key topological elements of the transmission path are structured and parsed to build a retrieval strategy based on multi-dimensional regional relationships. Intermediate nodes receive and parse a lightweight path fingerprint carrying metadata describing the entire data flow. This parsing process essentially deconstructs the path context into the smallest logical units that can be independently used for compliance determination. The source region identifies the starting jurisdiction or physical location of the data flow, the target region identifies the final destination of the data flow, and the intermediate region list contains the regions to which all forwarding nodes along the path belong. For example, in a scenario where data is transmitted from the source region through intermediate regions to the target region, these three elements correspond to the respective geographical or jurisdictional entities. In practice, intermediate nodes obtain the encapsulated path fingerprint tuple by reading HTTP header fields (such as X-Route-Fingerprint) and decompose it using predefined parsing rules. This parsing process can be based on key-value pairs or regular expression matching with specific delimiters, extracting 'SRC', 'TGT', and the intermediate region array from the tuple and mapping them to memory variables one by one. The extracted list of intermediate regions, as an ordered sequence, not only reflects the spatial trajectory of the data flow but also provides an index for determining the specific location of the current intermediate node within the entire path. These extraction results serve as necessary input parameters for generating a multi-dimensional retrieval strategy, enabling compliance assessments to cover complex topologies including transit regions, overcoming the limitations of traditional reliance solely on source-destination binary relationships.

[0033] This involves constructing a multi-dimensional compliance event matching space to cover complex cross-border data flow scenarios. Based on the regional elements extracted in the preceding steps, retrieval strategies are dynamically generated to adapt to different legal relationship perspectives. These multiple retrieval strategies constitute a three-dimensional query logic set, enabling the retrieval of the local compliance event fingerprint cache (EFC) to penetrate the limitations of a single dimension and accurately locate specific compliance records related to the current path segment. The various retrieval strategies generated based on the source region, intermediate region list, and target region mainly include the following three key combination logics: The first is the combination between the source region and the current intermediate node. This retrieval strategy focuses on the legality of data traveling from the source to the current transit node. For example, when the intermediate node is located in a region with a special trade agreement, this combination strategy is used to search for whether there are specific exemptions or whitelist records that allow data to flow from the source region into the intermediate region. The second is the combination between the current intermediate node and the target region. This retrieval strategy focuses on assessing the risk of continuing downstream transmission from the current node, searching for whether the intermediate node is qualified to forward data to a specific target region, or whether there are records of sanctions or restrictive events targeting this downstream path. The third is the combination of the source region via the current intermediate node to the target region. This retrieval strategy employs a global, end-to-end inspection, treating the source region, current intermediate node, and target region as a whole for retrieval. It aims to capture complex compliance events that only trigger under specific path sequences, such as those involving transit or regulations jointly governed by multiple regions. In engineering implementation, generating these retrieval strategies involves mapping the aforementioned three types of combinations to query keys or filter conditions in a local caching system. For example, wildcard matching syntax can be used to generate flexible query patterns such as 'SRC→Current→TGT' and 'SRC→*→TGT' (where * represents a wildcard), combined with the Approximate Nearest Neighbor (ANN) algorithm for rapid scanning in the local compliance event fingerprint cache. These strategies do not necessarily all need to be executed every time; instead, they can be selectively triggered or executed in parallel based on preset priorities or actual network load conditions. Through the generation and application of this multi-dimensional retrieval strategy, fine-grained matching of regulatory constraints under different scenarios can be achieved, effectively avoiding compliance omissions caused by changes in the legal status of relay nodes, thereby improving policy response speed and accuracy in multi-hop path environments.

[0034] S4: If the local compliance determination result indicates that the matching query has hit the corresponding compliance event record, then the intermediate node constructs a local compliance feature vector.

[0035] When an intermediate node confirms that its local compliance assessment result matches a risk event, it will conduct further analysis to construct a local compliance feature vector. This local compliance feature vector is a multi-dimensional quantitative description of the current data flow's compliance status, used to fine-grainedly characterize risk features. The construction process involves feature extraction and fusion across multiple dimensions: First, extracting the data flow's path topology features, such as the number of hops and the sensitivity level distribution of the traversed areas; second, extracting data type features to identify the categories of data content (e.g., personal identification information, financial data, medical records); third, extracting the transmission purpose features corresponding to the behavior type to analyze the business motivation behind the transmission (e.g., business cooperation, public services, personal use); and finally, obtaining the intermediate node's own credibility evaluation parameters, which reflect the node's accuracy, reliability, or authority level in historical compliance audits (e.g., whether it has passed certification by a designated institution). These features are combined into a high-dimensional vector, with each dimension corresponding to the extracted parameter values. Furthermore, if the local compliance assessment result indicates that the matching query did not match the corresponding compliance event record, the intermediate node will forward the data flow normally, i.e., transmit it according to normal transmission procedures.

[0036] In one embodiment, step S4, where the local compliance determination result indicates that the matching query matches the corresponding compliance event record, involves the intermediate node constructing a local compliance feature vector, includes: S401: If the local compliance determination result indicates that the matching query has matched the corresponding compliance event record, then the intermediate node extracts the path topology features, data type features, and transmission destination features corresponding to the behavior type of the data stream; S402: Obtain the credibility evaluation parameters of the intermediate node from the preset regional authentication mapping table; S403: The path topology features, the data type features, the transmission destination features, and the credibility evaluation parameters are fused into the local compliance feature vector.

[0037] As described in steps S401-S403 above, when the lightweight path fingerprint matches a record in the local compliance event fingerprint cache, the intermediate node initiates a multi-dimensional feature extraction operation on the data flow to capture business semantics and network state information supporting compliance risk assessment. Path topology features characterize the structured attributes of the data flow's transmission trajectory in cyberspace, reflecting the hierarchical relationship and hop count information of data flowing from the source region through the current intermediate node to the target region, used to characterize the path's complexity and geographical span. Data type features describe the sensitive attribute classification of the information carried by the data flow, such as personal identification information, financial transaction data, and health records. This feature determines the applicable compliance benchmarks for data in different legal jurisdictions. Transmission purpose features refer to the data processing intent and business scenario background mapped by the behavior type, such as distinguishing whether the data is used for business analysis, advertising, or disease risk prediction. The compliance obligations triggered by the same type of data under different transmission purposes differ significantly. The above feature extraction process provides the original semantic input for subsequent generation of high-dimensional vectors, enabling the intermediate node to shift from simple identity matching to deep understanding based on context.

[0038] The process involves a quantitative assessment of the compliance qualifications and trustworthiness of intermediate nodes to obtain a trustworthiness evaluation parameter. This parameter is not a general network reputation score, but a numerical indicator determined based on the certification status of the intermediate node's region within international data flow legal frameworks or regional agreements. Input sources include, but are not limited to, international compliance whitelist status such as cross-border privacy rule system certification and data privacy framework certification, or regional data security protection level certification results. By consulting a pre-set regional certification mapping table, the physical location or jurisdiction of the intermediate node is mapped to a specific numerical range, thereby generating this parameter. For example, for regional nodes with the highest level of certification, the trustworthiness evaluation parameter is set to 1.0; for regions with basic certification but additional restrictions, the parameter is set between 0.6 and 0.9. Introducing this parameter integrates a legal-level trustworthiness dimension into local compliance judgments, avoiding treating all transit regions as having the same risk level, and enabling the feature vector to more accurately reflect the compliance and trustworthiness of data flowing through the current node.

[0039] After extracting and acquiring multi-dimensional features, a feature fusion operation is performed to combine discrete business features, topological features, and credibility parameters into a unified mathematical expression for subsequent similarity calculations. The local compliance feature vector adopts a fixed-length vector structure, such as a 128-dimensional high-dimensional vector space. The fusion logic maps path topological features, data type features, and transmission destination features into sub-vector components in the vector space, and then embeds the credibility evaluation parameter as an independent weighted dimension. This fusion process is achieved through vector concatenation or weighted summation, generating a vector containing complete semantic information about path context, data attributes, business intent, and node credibility. By jointly encoding the above four types of features, this local compliance feature vector represents the comprehensive compliance status of the current data flow at a specific intermediate node, providing a standardized data foundation for subsequent vector space distance calculations with the benchmark feature vector in the local compliance event fingerprint cache, thereby realizing the technological evolution from rule matching to semantic similarity measurement.

[0040] S5: Compare the local compliance feature vector with the baseline feature vector in the local compliance event fingerprint cache, and determine whether to generate and send the enhanced signal back to the egress node based on the comparison result.

[0041] After obtaining the local compliance feature vector, the intermediate node compares it with the baseline feature vector pre-stored in the local compliance event fingerprint cache. The baseline feature vector represents a standard feature pattern that meets specific compliance requirements, or a known typical violation feature pattern. This baseline feature vector, pre-stored in the local compliance event fingerprint cache, is generated based on historical compliance event data through feature extraction and cluster analysis, representing a standard feature pattern that meets specific compliance requirements or a known typical violation feature pattern. The comparison process is achieved by calculating the similarity or distance between vectors (such as cosine similarity or Euclidean distance), thereby quantifying the degree of deviation between the current data stream features and the baseline pattern. If the comparison result shows a deviation exceeding a preset threshold, indicating that the compliance features of the current data stream significantly deviate from the standard pattern or match a high-risk pattern, a global attention threshold is triggered. At this time, the intermediate node generates an enhancement signal, which contains information such as the local compliance determination result, a key summary of the local compliance feature vector, or a matching strength index. After the enhancement signal is generated, the intermediate node encrypts and sends it to the exit node of the path (usually the exit gateway of the source region) through a dedicated backhaul channel or in-band signaling mechanism. Conversely, if the comparison result is within the allowable range, the intermediate node can choose not to generate an enhanced signal and only maintain normal forwarding. This backhaul triggering mechanism based on feature vector comparison ensures that only data flows with significant compliance risks or requiring close monitoring will consume network resources to report to the egress node, thereby reducing signaling overhead while ensuring risk control. An "enhanced signal" refers to a feedback data unit generated by the intermediate node when it determines that a data flow has a compliance risk. It contains the local compliance determination result, a key summary of the local compliance feature vector, and matching strength index information.

[0042] S6: The exit node receives the enhanced signals returned by each intermediate node, merges the enhanced signals to generate a comprehensive routing decision, and controls the routing action of the data flow on the transmission path according to the comprehensive routing decision.

[0043] The exit node, acting as the convergence point for path control, is responsible for collecting enhanced signals transmitted back from all intermediate nodes along the transmission path. Upon receiving these signals, the exit node fuses the matching strength, risk score, and credibility evaluation parameters contained within them. The fusion process can, for example, employ a weighted average algorithm, which weights the matching strength transmitted back by each intermediate node based on the credibility evaluation parameters provided, calculating a global compliance risk score representing the overall compliance level of the path. Subsequently, the global compliance risk score is compared with preset blocking and rate-limiting thresholds: if the score exceeds the blocking threshold, a comprehensive routing decision to block transmission is generated; if the score is between the blocking and rate-limiting thresholds, a decision to limit transmission is generated; if the score is below the rate-limiting threshold, a decision to allow transmission is generated. The blocking and rate-limiting thresholds are set as follows: historical transmission samples and their corresponding actual handling results are collected, a Receiver Operational Characteristics (ROC) curve is plotted, and the optimal cutoff point on the ROC curve is selected as the blocking and rate-limiting thresholds based on the tolerance of the business scenario for false alarm and false negative rates.

[0044] In one embodiment, step S6, which controls the routing action of the data flow on the transmission path based on the comprehensive routing decision, includes: S601: The exit node parses the comprehensive routing decision to obtain the corresponding routing action; wherein, the routing action includes blocking the data stream from continuing to transmit, limiting the transmission rate, or allowing the transmission; S602: The exit node sends control commands corresponding to the routing action to each intermediate node on the transmission path through the service mesh sidecar proxy mode.

[0045] As described in steps S601-S602 above, the egress node executes specific routing actions based on the comprehensive routing decision, such as issuing control commands to block the continued transmission of data streams, adjusting bandwidth for rate limiting, or maintaining the original path for transmission. Furthermore, while generating decisions, the egress node can associate and store metadata such as enhanced signals, decision-making basis, and original regulatory links from each intermediate node, forming an intermediate region collaborative audit graph. This provides a structured chain of evidence for subsequent compliance audits and reverse tracing. The intermediate region collaborative audit graph refers to the structured chain of evidence formed by the egress node associating and storing metadata such as enhanced signals, decision-making basis, and original regulatory links from each intermediate node, used for subsequent compliance audits and reverse tracing. Through this mechanism combining distributed perception and centralized decision-making, refined and dynamic routing control of complex cross-domain data streams can be achieved, ensuring that data transmission meets multi-dimensional compliance requirements throughout the entire link.

[0046] As the central hub for data flow routing decisions, the exit node establishes communication monitoring channels with each intermediate node along the transmission path to capture and aggregate the enhanced signals independently transmitted by each node. These enhanced signals are feedback data generated by intermediate nodes after a local compliance determination, based on a matching calculation between a lightweight path fingerprint and a local compliance event fingerprint cache. They not only contain feature information indicating compliance status but also encapsulate data payloads for quantitative assessment, such as match strength or local compliance distance scores. By collecting all these enhanced signals, a complete multi-dimensional input is provided for subsequent global risk fusion, ensuring that decision-making covers the compliance status of each hop along the path.

[0047] In one embodiment, step S6, where the exit node fuses the enhanced signals returned from each intermediate node to generate a comprehensive routing decision, includes: S611: The exit node collects the enhanced signals transmitted back by all intermediate nodes on the transmission path; S612: The matching strength in each enhanced signal is weighted and averaged according to the credibility evaluation parameters of the corresponding intermediate nodes to obtain the global compliance risk score; S613: Compare the global compliance risk score with the preset blocking threshold and rate limiting threshold to generate the comprehensive routing decision.

[0048] As described in steps S611-S613 above, after acquiring the enhanced signals from all intermediate nodes, these local judgment results are fused to form a unified understanding of the compliance risks of the entire transmission path. This process introduces a credibility weighting mechanism instead of a simple arithmetic average. The matching strength characterizes the degree to which the data flow violates compliance rules at a certain intermediate node or its similarity to a specific compliance event; the credibility evaluation parameter reflects the reliability of the intermediate node in terms of legal status, security certification level, and regional policy stability. For each intermediate node, the matching strength value in its returned enhanced signal is extracted and weighted in conjunction with the credibility evaluation parameter corresponding to that node. This weighted averaging process aims to give greater weight to the judgment results of high-credibility nodes while suppressing the excessive influence of low-credibility nodes on the final decision. This calculation process can be formally represented as: In the formula, This represents the calculated global compliance risk score; This indicates the total number of intermediate nodes along the transmission path that transmit the enhanced signal back. Indicates the first The matching strength returned by each intermediate node; Indicates the first The credibility evaluation parameters correspond to each intermediate node. In the above formula, the matching strength... For example, it could be a value obtained by calculating the cosine similarity or distance metric between the local compliant feature vector and the baseline feature vector of an intermediate node. This value is typically normalized (e.g., between 0 and 1). Credibility evaluation parameters. The construction process is not a static assignment in a single dimension, but rather a dynamic parameter determined based on multi-source information. Its formation process can be as follows: During system initialization or policy configuration, firstly, the international compliance certification levels and historical compliance audit records of the regions where each intermediate node is located are collected; secondly, combining a spatiotemporal dual-axis decay model, a time-based decay factor is calculated based on the relative relationship between the current time point and the regional policy's effective window; finally, the quantified certification level value and the time decay factor are aggregated to form the node's credibility evaluation parameter at the current moment. After this parameter is constructed, it is usually stored in the exit node's in-memory database or cache in the form of key-value pairs or a structured configuration list for real-time weighted calculation. When regional policies are updated or node certification status changes, this parameter supports a dynamic update mechanism to ensure the timeliness and accuracy of the global score. Through this weighted fusion method, the exit node can synthesize the local compliance evidence scattered across various nodes in the path into a global score with legal interpretability and engineering robustness.

[0049] Based on the global compliance risk score calculated in the preceding steps, the final threshold determination logic is executed to determine the specific data flow routing action. According to preset blocking and rate-limiting thresholds, continuous risk scores are mapped to discrete decision instructions. The preset blocking threshold is the critical value for determining whether a data flow has a serious compliance risk and must therefore terminate transmission; the preset rate-limiting threshold is the critical value for determining whether a data flow has potential risks but is allowed to continue transmission under controlled conditions. The determination process for these thresholds can be, for example, as follows: based on historical compliance data statistics and business risk preference analysis, a large number of historical transmission samples and their corresponding actual handling results are collected; by analyzing the false positive and false negative rate distribution curves under different risk scores, combined with the different trade-offs between data security and transmission continuity requirements in business scenarios, a blocking threshold (e.g., set to 0.85) is selected to ensure strict interception of high-risk traffic, while a rate-limiting threshold (e.g., set to 0.62) is selected to achieve flexible control of medium-risk traffic. These thresholds can be stored in the policy configuration center after formation and can be dynamically adjusted according to changes in the regulatory environment. The specific determination logic is as follows: First, the global compliance risk score is... Compared with the blocking threshold, if If the data flow is greater than or equal to the blocking threshold, it is determined that the compliance risk is extremely high, and a comprehensive routing decision to block the transmission is generated; if the blocking threshold is not reached, the transmission will continue to be blocked. Compare with the speed limit threshold; if If the data stream is greater than or equal to the rate-limiting threshold, it is determined that the data stream poses a certain risk, and a comprehensive routing decision for rate-limited transmission is generated; if If the data flow is below the rate limit threshold, the compliance risk is determined to be within an acceptable range, and a comprehensive routing decision to allow transmission is generated. After the comprehensive routing decision is generated, it will be directly transmitted to the routing execution module to drive the issuance of subsequent control commands, thereby achieving precise control over data flow behavior.

[0050] The egress node, as the final execution anchor of the routing control policy, is configured to parse and process the comprehensive routing decision generated in the preceding steps. This parsing process aims to map the abstract comprehensive decision result into concrete, executable network layer instructions. The egress node reads the control fields in the comprehensive routing decision and, based on preset action mapping logic, determines the specific routing action category corresponding to the decision. When the determined routing action is to block the continued transmission of the data flow, the egress node immediately terminates the forwarding process of the current data flow and releases relevant network resources. For example, it can forcibly disconnect the connection by configuring firewall rules or directly dropping subsequent data packets using a scalable Berkeley packet filter. When the determined routing action is to rate-limit transmission, the egress node dynamically limits the transmission rate of the data flow, controlling its bandwidth usage within a preset threshold range. For example, it can use a token bucket algorithm to shape the traffic, ensuring that high-priority services are not affected while maintaining the overall stability of the network. When the determined routing action is to allow transmission, the egress node maintains the normal transmission path of the data flow, without imposing any rate limits or blocking operations, allowing the data flow to continue transmitting to the destination according to existing routing rules.

[0051] After determining the specific routing action, the egress node executes the control command distribution operation. This process employs a service mesh sidecar proxy pattern to achieve distributed policy execution. The service mesh sidecar proxy pattern involves deploying a dedicated lightweight network proxy alongside each service instance (such as an intermediate node). This proxy shares the network namespace with the application container and is responsible for handling all inbound and outbound traffic for that instance. The egress node, through its internal control plane component, converts the parsed routing action into standard format control commands that the sidecar proxy can recognize. Subsequently, using the service mesh's secure channels (such as long-lived connections based on gRPC (Google Remote Procedure Call) or discovery services based on the xDS (Discovery Service) protocol), the aforementioned control commands are distributed to the corresponding sidecar proxy instances of each intermediate node along the transmission path. For different action types, the sidecar proxy instance executes the corresponding local policy configuration: if the action type is blocking, the sidecar proxy instance immediately configures local traffic filtering rules upon receiving the instruction to block subsequent data packets; if the action type is rate limiting, the sidecar proxy instance loads the rate limiting configuration to monitor the rate of passing traffic; if the action type is allowing, the sidecar proxy instance maintains the existing forwarding policy or resets the previous temporary restrictions. Through this sidecar proxy mode, routing control logic is pushed down to the edge nodes of the data plane, enabling each intermediate node to independently and quickly respond to the decisions of the egress node without the intervention of a centralized gateway, thereby achieving policy execution with sub-second effectiveness and high availability.

[0052] In one embodiment, after step S602, where the egress node issues control commands corresponding to the routing action to each intermediate node on the transmission path via the service mesh sidecar proxy mode, the method further includes: S6031: If the routing action is to block transmission, the egress node generates a standardized error code and feeds it back to the source application of the data stream; S6032: If the routing action is rate-limited transmission, the exit node synchronously sends rate-limiting parameters to each intermediate node through which the data flow passes.

[0053] As described in steps S6031-S6032 above, when the comprehensive routing decision determines that the final control action is to block transmission, the egress node immediately triggers the interruption process and constructs a standardized error code. This standardized error code is not a simple HTTP status code, but a structured data object constructed according to specifications. Its specific construction process is as follows: Extract the unique Trace-ID of this violation, the compliance event record ID that triggered the violation, the legal jurisdiction involved in the violation (such as RGN-EU), and the specific type of violation reason from the Intermediate Area Collaborative Audit Graph (ICAG) or Compliance Fingerprint Watermark Chain (CWC); assemble the above key audit information into JSON format and attach a digital signature to generate a standardized error code that can be automatically parsed by the source application. This standardized error code provides the source application with a detailed error context, enabling the application layer to automatically execute degradation processing, retry logic, or alert the administrator based on the error type. This mechanism effectively solves the problems of ambiguous feedback information and difficult troubleshooting in traditional interception mechanisms, significantly enhancing the system's observability and traceability.

[0054] In cases where the comprehensive routing decision determines the final control action to be rate-limited transmission, the egress node synchronously distributes rate-limiting parameters to each intermediate node along the transmission path via a service mesh sidecar proxy. These rate-limiting parameters are dynamically calculated by the egress node based on the compliance risk level of the current data flow and network load status. The construction process mainly includes three dimensions: first, determining the rate threshold, i.e., the maximum transmission rate per second allowed for the data flow, such as setting it to 50% of the original rate before compliance determination or a specific value (e.g., 10MB / s); second, setting the burst capacity, defining the amount of data allowed to exceed the rate threshold under short-term burst traffic, such as setting it to the rate threshold multiplied by a 0.1-second time window length; and finally, configuring the sliding window time granularity for precise traffic rate statistics, such as setting it to 1 second or 100 milliseconds. These rate-limiting parameters are encapsulated as a standard xDS (Discovery Service) protocol configuration and distributed to the sidecar proxies of each intermediate node along the path via gRPC long connections. Upon receiving the rate-limiting parameters, the sidecar proxy loads them into its local token bucket / leaky bucket algorithm to perform precise traffic shaping on the data flow. By synchronously sending rate-limiting parameters to each intermediate node, the QoS (Quality of Service) policy is kept consistent throughout the entire path, preventing data backlog or compliance risks caused by single-point policy drift, thus ensuring the continuity of data transmission while meeting compliance constraints.

[0055] Reference Figure 3 The present invention also provides a data transmission routing control device, the device comprising: The generation module 902 is used to generate a lightweight path fingerprint corresponding to the transmission path of the data stream; wherein the lightweight path fingerprint includes at least a source region, a list of intermediate regions, a target region, and a behavior type. Sending module 904 is used to send the lightweight path fingerprint and the data stream sequentially to the intermediate node based on the transmission path; The receiving module 906 is used for each of the intermediate nodes to receive the data stream and the lightweight path fingerprint, perform a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint, and generate a local compliance judgment result; The construction module 908 is used to construct a local compliance feature vector by the intermediate node if the local compliance determination result indicates that the matching query hits the corresponding compliance event record. The comparison module 910 is used to compare the local compliance feature vector with the baseline feature vector in the local compliance event fingerprint cache, and determine whether to generate and send the enhancement signal back to the exit node based on the comparison result; The control module 912 is used to receive the enhanced signals returned by each intermediate node at the exit node, fuse the enhanced signals to generate a comprehensive routing decision, and control the routing action of the data flow on the transmission path according to the comprehensive routing decision.

[0056] In one embodiment, the generation module 902 includes: The information acquisition submodule is used to acquire the source region identifier of the data stream, the list of intermediate regions that the data stream is expected to pass through, the target region identifier, and the behavior type of the data stream; The information assembly submodule is used to assemble the source region identifier, the intermediate region list, the target region identifier, and the behavior type into a path fingerprint tuple according to a predetermined format. The path fingerprint tuple is hashed and a digital signature is appended to generate the lightweight path fingerprint.

[0057] In one embodiment, the receiving module 906 includes: The region extraction submodule is used to extract the source region, the intermediate region list, and the target region from the lightweight path fingerprint. The retrieval strategy generation submodule is used to generate multiple retrieval strategies based on the source region, the intermediate region list, and the target region; wherein, the multiple retrieval strategies are used to retrieve at least one of the following combinations in the compliance event records: the combination between the source region and the current intermediate node, the combination between the current intermediate node and the target region, and the combination from the source region to the target region via the current intermediate node.

[0058] In one embodiment, the construction module 908 includes: The feature extraction submodule is used to extract the path topology features, data type features, and transmission destination features corresponding to the behavior type of the data stream by the intermediate node if the local compliance determination result indicates that the matching query hits the corresponding compliance event record. The parameter acquisition submodule is used to obtain the credibility evaluation parameters of the intermediate node from the preset regional authentication mapping table; The fusion submodule is used to fuse the path topology features, the data type features, the transmission destination features, and the credibility evaluation parameters into the local compliance feature vector.

[0059] In one embodiment, the control module 912 includes: The routing action acquisition submodule is used by the exit node to parse the comprehensive routing decision to obtain the corresponding routing action; wherein, the routing action includes blocking the data stream from continuing to transmit, limiting the transmission rate, or allowing the transmission. The control command sending submodule is used by the exit node to send control commands corresponding to the routing action to each intermediate node on the transmission path through the service mesh sidecar proxy mode.

[0060] In one embodiment, the control module 912 further includes: The first execution submodule is used to generate a standardized error code and feed it back to the source application of the data stream if the routing action is to block transmission. The second execution submodule is used to synchronously send rate limiting parameters to each intermediate node through which the data flow passes if the routing action is rate-limited transmission.

[0061] In one embodiment, the control module 912 includes: An enhanced signal collection submodule is used by the exit node to collect the enhanced signals transmitted back from all intermediate nodes along the transmission path; The weighted average submodule is used to perform a weighted average of the matching strength in each enhanced signal according to the credibility evaluation parameters of the corresponding intermediate node to obtain a global compliance risk score. The comparison submodule is used to compare the global compliance risk score with preset blocking thresholds and rate limiting thresholds to generate the comprehensive routing decision.

[0062] Figure 4 An internal structural diagram of an electronic device in one embodiment is shown. This electronic device can specifically be a terminal or a server, and more specifically, a computer device. Figure 4As shown, the electronic device includes a processor, a memory, and a network interface connected via a system bus. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and may also store a computer program. When executed by the processor, this computer program enables the processor to implement a data transmission routing control method. The internal memory may also store a computer program, which, when executed by the processor, enables the processor to implement a data transmission routing control method. Those skilled in the art will understand that… Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0063] In one embodiment, an electronic device is provided, including a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the following steps: Along the data stream transmission path, a lightweight path fingerprint corresponding to the transmission path is generated; wherein, the lightweight path fingerprint includes at least a source region, a list of intermediate regions, a target region, and a behavior type; Based on the transmission path, the lightweight path fingerprint and the data stream are sequentially sent to the intermediate node; Each of the intermediate nodes receives the data stream and the lightweight path fingerprint, performs a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint, and generates a local compliance judgment result; If the local compliance determination result indicates that the matching query has matched the corresponding compliance event record, then the intermediate node constructs a local compliance feature vector. The local compliance feature vector is compared with the baseline feature vector in the local compliance event fingerprint cache, and based on the comparison result, it is determined whether to generate and send an enhanced signal back to the exit node. The exit node receives the enhanced signals returned by each intermediate node, merges the enhanced signals to generate a comprehensive routing decision, and controls the routing action of the data flow on the transmission path according to the comprehensive routing decision.

[0064] It effectively solves the technical problem that existing cross-border data compliance control systems have long faced with the contradiction of lack of multi-hop path awareness and insufficient accuracy of global decision-making when dealing with complex geographical paths and dynamic regulatory environments. Therefore, it avoids the situation of delayed response to policy changes and difficulty in audit traceability due to the inability to identify intermediate jurisdictions, and improves the accuracy of compliance judgment, real-time system response and verifiability of evidence throughout the cross-border data flow process.

[0065] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, causes the processor to perform the following steps: Along the data stream transmission path, a lightweight path fingerprint corresponding to the transmission path is generated; wherein, the lightweight path fingerprint includes at least a source region, a list of intermediate regions, a target region, and a behavior type; Based on the transmission path, the lightweight path fingerprint and the data stream are sequentially sent to the intermediate node; Each of the intermediate nodes receives the data stream and the lightweight path fingerprint, performs a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint, and generates a local compliance judgment result; If the local compliance determination result indicates that the matching query has matched the corresponding compliance event record, then the intermediate node constructs a local compliance feature vector. The local compliance feature vector is compared with the baseline feature vector in the local compliance event fingerprint cache, and based on the comparison result, it is determined whether to generate and send an enhanced signal back to the exit node. The exit node receives the enhanced signals returned by each intermediate node, merges the enhanced signals to generate a comprehensive routing decision, and controls the routing action of the data flow on the transmission path according to the comprehensive routing decision.

[0066] It effectively solves the technical problem that existing cross-border data compliance control systems have long faced with the contradiction of lack of multi-hop path awareness and insufficient accuracy of global decision-making when dealing with complex geographical paths and dynamic regulatory environments. Therefore, it avoids the situation of delayed response to policy changes and difficulty in audit traceability due to the inability to identify intermediate jurisdictions, and improves the accuracy of compliance judgment, real-time system response and verifiability of evidence throughout the cross-border data flow process.

[0067] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0068] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0069] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A data transmission routing control method, characterized in that, The method includes: Along the data stream transmission path, a lightweight path fingerprint corresponding to the transmission path is generated; wherein, the lightweight path fingerprint includes at least a source region, a list of intermediate regions, a target region, and a behavior type; Based on the transmission path, the lightweight path fingerprint and the data stream are sequentially sent to the intermediate node; Each of the intermediate nodes receives the data stream and the lightweight path fingerprint, performs a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint, and generates a local compliance judgment result; If the local compliance determination result indicates that the matching query has matched the corresponding compliance event record, then the intermediate node constructs a local compliance feature vector. The local compliance feature vector is compared with the baseline feature vector in the local compliance event fingerprint cache, and based on the comparison result, it is determined whether to generate and send an enhanced signal back to the exit node. The exit node receives the enhanced signals returned by each intermediate node, merges the enhanced signals to generate a comprehensive routing decision, and controls the routing action of the data flow on the transmission path according to the comprehensive routing decision.

2. The data transmission routing control method according to claim 1, characterized in that, The step of generating a lightweight path fingerprint corresponding to the transmission path of the data stream includes: Obtain the source region identifier of the data stream, the list of intermediate regions the data stream is expected to pass through, the target region identifier, and the behavior type of the data stream; The source region identifier, the intermediate region list, the target region identifier, and the behavior type are assembled into a path fingerprint tuple according to a predetermined format. The path fingerprint tuple is hashed and a digital signature is appended to generate the lightweight path fingerprint.

3. The data transmission routing control method according to claim 1, characterized in that, The step of performing a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint includes: Extract the source region, the intermediate region list, and the target region from the lightweight path fingerprint; Multiple retrieval strategies are generated based on the source region, the intermediate region list, and the target region; wherein, the multiple retrieval strategies are used to retrieve at least one of the following combinations in the compliance event record: the combination between the source region and the current intermediate node, the combination between the current intermediate node and the target region, and the combination from the source region to the target region via the current intermediate node.

4. The data transmission routing control method according to claim 1, characterized in that, The step of constructing a local compliance feature vector by the intermediate node if the local compliance determination result indicates that the matching query has matched the corresponding compliance event record includes: If the local compliance determination result indicates that the matching query has matched the corresponding compliance event record, then the intermediate node extracts the path topology features, data type features, and transmission destination features corresponding to the behavior type of the data stream. Obtain the credibility evaluation parameters of the intermediate node from the preset regional authentication mapping table; The path topology features, data type features, transmission destination features, and credibility evaluation parameters are fused into the local compliance feature vector.

5. The data transmission routing control method according to claim 1, characterized in that, The step of controlling the routing action of the data flow on the transmission path based on the comprehensive routing decision includes: The exit node parses the comprehensive routing decision to obtain the corresponding routing action; wherein, the routing action includes blocking the data stream from continuing to transmit, limiting the transmission rate, or allowing the transmission. The exit node sends control commands corresponding to the routing action to each intermediate node on the transmission path through the service mesh sidecar proxy mode.

6. The data transmission routing control method according to claim 5, characterized in that, After the step of the exit node issuing control commands corresponding to the routing action to each intermediate node on the transmission path through the service mesh sidecar proxy mode, the method further includes: If the routing action is to block transmission, the exit node generates a standardized error code and feeds it back to the source application of the data stream; If the routing action is rate-limited transmission, the exit node synchronously sends rate-limiting parameters to each intermediate node through which the data stream passes.

7. The data transmission routing control method according to claim 1, characterized in that, The step of generating a comprehensive routing decision by fusing the enhanced signals returned by each intermediate node at the exit node includes: The exit node collects the enhanced signals transmitted back from all intermediate nodes along the transmission path; The matching strength of each enhanced signal is weighted and averaged according to the credibility evaluation parameters of the corresponding intermediate nodes to obtain the global compliance risk score; The global compliance risk score is compared with preset blocking thresholds and rate limiting thresholds to generate the comprehensive routing decision.

8. A data transmission routing control device, characterized in that, The device includes: A generation module is used to generate a lightweight path fingerprint corresponding to the transmission path of the data stream; wherein the lightweight path fingerprint includes at least a source region, a list of intermediate regions, a target region, and a behavior type. The sending module is used to sequentially send the lightweight path fingerprint and the data stream to the intermediate nodes based on the transmission path; The receiving module is used for each of the intermediate nodes to receive the data stream and the lightweight path fingerprint, perform a matching query in the local compliance event fingerprint cache based on the lightweight path fingerprint, and generate a local compliance judgment result. A construction module is used to construct a local compliance feature vector by the intermediate node if the local compliance determination result indicates that the matching query matches the corresponding compliance event record. The comparison module is used to compare the local compliance feature vector with the baseline feature vector in the local compliance event fingerprint cache, and determine whether to generate and send the enhancement signal back to the exit node based on the comparison result; The control module is used to receive the enhanced signals returned by each intermediate node at the exit node, fuse the enhanced signals to generate a comprehensive routing decision, and control the routing action of the data flow on the transmission path according to the comprehensive routing decision.

9. A computer-readable storage medium, characterized in that, The system contains a computer program that, when executed by a processor, causes the processor to perform the steps of the data transmission routing control method as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, The device includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the data transmission routing control method as described in any one of claims 1 to 7.

Images