A data security sharing method and device of a cross-network switching platform

Abstract

The application relates to the technical field of information tracing, in particular to a data security sharing method and equipment of a cross-network exchange platform, which comprises the following steps: determining a first abnormal characteristic value based on the distribution of deviated characteristic values of external requests of all Span data under any service node in each time period; determining a second abnormal characteristic value based on the change trend of the calling duration of all Span data under any service node in each time period, and combining the first abnormal characteristic value to determine a sampling adjustment coefficient of any service node in each time period, so as to adjust the sampling rate of Span data under any service node in the next time period of each time period. The application solves the problem that abnormal data is easy to be missed in the traditional method by dynamically adjusting the link tracking sampling rate, and improves the efficiency and accuracy of fault positioning of the cross-network exchange platform.

Description

Technical Field

[0001] This application relates to the field of information traceability technology, specifically to a data security sharing method and device for a cross-network exchange platform. Background Technology

[0002] A cross-network exchange platform is a system platform for exchanging data between different networks. By leveraging the security isolation features of network gateway products and their capabilities for business data approval and data security detection, it forms a cross-network and cross-domain data security exchange solution. It is mainly used to solve problems such as data sharing, business access, and boundary security between departments. It can be widely used in various industries with cross-network data security exchange needs, such as government agencies at all levels, public security, procuratorates, courts, judicial departments, and natural resources departments, to achieve secure information exchange and provide strong security guarantees for the informatization construction of various industries.

[0003] In the operation and maintenance of cross-network exchange platforms, real-time tracking and monitoring of data exchange and user operations across the entire chain is crucial to ensuring system security and reliability. While the current mainstream fixed sampling rate strategy can reduce storage costs, it ignores the uneven distribution of Span data value. This can easily lead to the over-collection of a large amount of normal Span data, while critical abnormal Span data caused by system failures or performance anomalies is missed. This affects the accuracy and efficiency of troubleshooting, and weakens the cross-network exchange platform's ability to promptly detect and respond to security risks. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a method and device for secure data sharing across network exchange platforms, the specific technical solution of which is as follows:

[0005] In a first aspect, embodiments of this application provide a method for secure data sharing across a network exchange platform, the method comprising the following steps:

[0006] On the cross-network exchange platform, Span data and its call duration under each service node are obtained in real time, and all Span data under each external request within each preset time period are obtained.

[0007] Based on the correlation between all Span data under each external request within each time period, a service call path graph is constructed; all normal service call path graphs under various types of external requests are obtained; based on the differences between the service call path graph of each external request within each time period and the normal service call path graphs under all types of external requests, the service difference value between the service call path graph of each external request within each time period and the normal service call path graphs is determined, so as to determine the deviation characteristic value of each external request within each time period; based on the distribution of the deviation characteristic values ​​of the external requests to which all Span data under any service node belong within each time period, the first abnormal characteristic value of any service node within each time period is determined.

[0008] Based on the changing trend of call duration of all Span data under any service node in each time period, the second abnormal feature value of any service node in each time period is determined, and combined with the first abnormal feature value, the sampling adjustment coefficient of any service node in each time period is determined, so as to adjust the sampling rate of Span data under any service node in the next time period of each time period.

[0009] Preferably, the method for constructing the service call path graph is as follows:

[0010] All Span data under each external request within each time period are treated as nodes in a directed graph, and the connections between nodes with call relationships are treated as edges in the directed graph, with the direction of the edges pointing from the caller to the callee. The resulting directed graph is denoted as the service call path graph.

[0011] Preferably, the various external requests are data queries, browsing, and data exchanges.

[0012] Preferably, the method for determining the service difference value between the service call path diagram of each external request and each normal service call path diagram within each time period is as follows:

[0013] Obtain the service call path graph for each external request and the adjacency matrix of all normal service call graphs for all classes of external requests;

[0014] The difference in the adjacency matrix between the service call path graph of each external request and the normal service call path graph under all types of external requests within each time period is used as the service difference value between the service call path graph of each external request and the normal service call path graph within each time period.

[0015] Preferably, the deviation characteristic value of each external request in each time period is the minimum value among the service difference values ​​between the service call path map of each external request and the total normal service call path map in each time period.

[0016] Preferably, the first abnormal feature value of any service node in each time period is the result of averaging the deviation feature values ​​of all external requests to which Span data belongs under any service node in each time period.

[0017] Preferably, the method for determining the second abnormal feature value of any service node within each time period is as follows:

[0018] The call duration of all Span data under any service node in each time period is fitted according to the time sequence of all Span data, and the slope of the fitted line is used as the second abnormal feature value of any service node in each time period.

[0019] Preferably, the sampling adjustment coefficient of any service node in each time period is the result of the positive fusion of the normalized value of the first abnormal feature value and the normalized value of the second abnormal feature value of any service node in each time period.

[0020] Preferably, adjusting the sampling rate of Span data at any service node in the next time period of each time period includes:

[0021] The sampling rate of Span data at service node i in the next time period after time period T. The expression is: In the formula, This represents the sampling adjustment coefficient for service node i within the time period T; , These represent a first preset value and a second preset value, respectively, where the first preset value is less than the second preset value; norm() represents a normalization function; and round[] represents a rounding function. Secondly, embodiments of this application also provide a data security sharing device for a cross-network exchange platform, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described data security sharing methods for a cross-network exchange platform.

[0022] This application has at least the following beneficial effects:

[0023] This application constructs a service call path graph and compares it with a normal service call path graph to quantify the deviation feature value of each external request. This allows for the assessment of the prevalence of path deviations at each service node, forming a first abnormal feature value. This method can accurately identify abnormal call chains triggered by platform failures, helping to improve the data accuracy and fault location efficiency of the link tracing system in complex fault-tolerant scenarios. Furthermore, compared to traditional link tracing methods with fixed sampling rates, this application constructs a sampling adjustment coefficient by comprehensively considering the path deviation and performance degradation trends of service nodes. By dynamically adjusting the sampling rate, it effectively avoids the problem of reduced operational efficiency and accuracy of cross-network exchange platforms due to the link tracing system missing too much abnormal Span data caused by platform system failures or performance anomalies. This achieves accurate identification of abnormal nodes and intelligent control of the sampling rate, ensuring comprehensive capture of key abnormal data and improving the efficiency and accuracy of fault location in cross-network exchange platforms. Attached Figure Description

[0024] To more clearly illustrate the technical solutions and advantages in the embodiments of this application 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A flowchart illustrating the steps of a data security sharing method across a network exchange platform, as provided in one embodiment of this application;

[0026] Figure 2 This is a schematic diagram of the sampling adjustment coefficient extraction process provided in one embodiment of this application. Detailed Implementation

[0027] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a data security sharing method and device for a cross-network exchange platform proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0029] The following description, in conjunction with the accompanying drawings, details the specific scheme of the data security sharing method and equipment for a cross-network exchange platform provided in this application.

[0030] Please see Figure 1 The diagram illustrates a flowchart of a data security sharing method for a cross-network exchange platform according to an embodiment of this application. The method includes the following steps:

[0031] Step S1: On the cross-network exchange platform, obtain the Span data and its call duration under each service node in real time, and obtain all Span data under each external request within each preset time period.

[0032] This embodiment uses the distributed link tracing (Zipkin) system to monitor the cross-network exchange platform in real time. The Zipkin system is mainly divided into a client and a server. Through injection, a client is deployed on each service node in the cross-network exchange platform. The client collects Span data from the log data of the service node in the form of a data stream. It uses multi-threaded streaming data processing and data synchronization technology based on time windows to realize the collection and transmission of Span data on different service nodes in the cross-network exchange platform. The Span data collected by the client is uploaded to the server for storage.

[0033] This embodiment adds a sampling adjustment module to the Zipkin system to set appropriate sampling rates for the Span data acquisition terminals of clients deployed on each service node in the cross-network switching platform at different time periods. This enables dynamic adjustment of the Span data sampling rate at each service node, thereby avoiding over-collection of normal Span data and omission of abnormal Span data at each service node.

[0034] On the cross-network exchange platform, Span data and its call duration under each service node are acquired in real time, and all Span data under each external request within each preset time period are acquired. The data acquisition frequency is f. In this embodiment, the values ​​of data acquisition frequency f and the length of each preset time period are set manually. In this embodiment, the value of data acquisition frequency f is 1Hz and the length of the time period is 2min. In actual application, as other implementation methods, implementers can also set them according to specific circumstances. This embodiment does not impose any special restrictions.

[0035] In addition, in this embodiment, each Span includes a corresponding traceid, Spanid, ParentSpanid, call start time and end time, and the name of the service node to which it belongs. The traceid represents the globally unique link identifier of the service call chain in which the Span is located. The Spanid and ParentSpanid represent the hierarchical identifier of a service call event in the service call chain in which the Span is located, which is used to restore the hierarchical call relationship of the Span in the service call chain in which it is located.

[0036] The start and end times of each Span are converted into timestamps, and the converted end time of each Span is subtracted from the converted start time of each Span to obtain the call duration of each Span. This duration is used to evaluate the time consumed when the service call event corresponding to each Span is invoked. The conversion of timestamps is a well-known technique, and the specific process will not be described in detail.

[0037] Step S2: Construct a service call path graph based on the association between all Span data under each external request in each time period; obtain all normal service call path graphs under various types of external requests; determine the service difference value between the service call path graph of each external request and each normal service call path graph under all types of external requests in each time period based on the difference between the service call path graph of each external request and each normal service call path graph in each time period, so as to determine the deviation characteristic value of each external request in each time period; determine the first abnormal characteristic value of any service node in each time period based on the distribution of the deviation characteristic values ​​of the external requests to which all Span data under any service node belong in each time period.

[0038] In general, when a cross-network exchange platform processes external requests for data query, browsing, or cross-webpage data exchange, if a node in the service call path fails, such as a service crash or network interruption, the fault tolerance mechanism of the cross-network exchange platform, such as service retry or error degradation, will be triggered. These mechanisms will generate additional service calls outside the original normal call path, thereby generating abnormal Span data. This abnormal data will interfere with the accuracy of link tracing, making it unable to accurately and truthfully reflect the preset business processing flow.

[0039] Therefore, to ensure the effectiveness of link tracing, this embodiment constructs a service call path graph based on the association between all Span data under each external request within each time period; obtains all normal service call path graphs under various types of external requests; and determines the service difference value between the service call path graph of each external request and each normal service call path graph under all types of external requests within each time period based on the difference between them, thereby determining the deviation characteristic value of each external request within each time period; and determines the first abnormal characteristic value of any service node within each time period based on the distribution of the deviation characteristic values ​​of the external requests to which all Span data under any service node belong within each time period, in order to process the abnormal Span data caused by the cross-network exchange platform failure. The specific process is as follows:

[0040] First, in this embodiment, a service call path graph is constructed based on the association between all Span data under each external request within each time period, specifically as follows:

[0041] In this embodiment, all Span data under each external request within each time period are taken as nodes in the directed graph, and the connection between nodes with a calling relationship is taken as edges in the directed graph. The direction of the edges is from the caller to the callee. The constructed directed graph is denoted as the service call path graph.

[0042] Furthermore, under the stable state of the cross-network exchange platform without faults or performance anomalies, the service call path diagrams for all types of external requests are obtained according to the aforementioned method. Since these are obtained under the stable state of the cross-network exchange platform without faults or performance anomalies, the service call path diagrams for all types of external requests are recorded as normal service call path diagrams. Among these, all types of external requests include: data query, browsing, and data exchange.

[0043] Furthermore, this embodiment determines the service difference value between the service call path diagram of each external request and the normal service call path diagram under all types of external requests within each time period based on the difference between the service call path diagram of each external request within each time period and the normal service call path diagram within each time period. Specifically:

[0044] In this embodiment, the adjacency matrix of all nodes in the service call path graph of each external request is obtained, and the adjacency matrix of the same nodes in the service call path graph of each service request is obtained in the normal service call graph. The row and column indices of the adjacency matrix are the nodes in the service call path graph. The elements in the adjacency matrix are: if there is a call relationship between nodes A and B, the element in the adjacency matrix corresponding to node A and B is set to 1; otherwise, it is set to 0. For easier understanding, the following example is given:

[0045] Suppose that the service call path graph of an external request Q contains nodes A, B, C, and D, where there are call relationships between nodes A and B, B and D, and D and C. Then the corresponding adjacency matrix is: Let U be the adjacency matrix, where the rows are labeled A, B, C, and D from left to right, and the columns are labeled A, B, C, and D from top to bottom. Assume that the normal service call path graph for a certain type of external request contains nodes A, B, C, and E, with call relationships between nodes A and C, C and B, and B and E. Then, within the normal service call graph, obtain the adjacency matrix of the nodes that are identical to those in the external request Q service call path graph. Let H be the adjacency matrix, where the rows are first, then the columns. Each row from left to right contains nodes A, B, C, and E, and each column from top to bottom contains nodes A, B, C, and E.

[0046] Furthermore, the difference in the adjacency matrix between the service call path graph of each external request and the normal service call path graph under all types of external requests within each time period is used as the service difference value between the service call path graph of each external request and the normal service call path graph within each time period. The specific calculation process is illustrated using the above two adjacency matrices as examples:

[0047] Let adjacency matrices U and H contain nodes A, B, and C. In adjacency matrix U, the value between nodes A and B is 1, the value between nodes A and C is 0, and the value between nodes B and C is 0. In adjacency matrix H, the value between nodes A and B is 0, the value between nodes A and C is 1, and the value between nodes B and C is 1. Calculate the absolute difference between the corresponding values ​​of the same nodes in adjacency matrices U and H, i.e., the difference between the corresponding values ​​of nodes A and B is... Similarly, calculate the average of all absolute differences between nodes A and C, and between nodes B and C, and use the average of these differences as the service difference value.

[0048] It should be noted that the calling relationship can be obtained through Spanid and ParentSpanid in step S1, which will not be elaborated further.

[0049] Furthermore, this embodiment determines the deviation characteristic value of each external request within each time period based on the service difference value, specifically:

[0050] In this embodiment, the minimum value among the service difference values ​​between the service call path graph of each external request and the total normal service call path graph within each time period is used as the deviation feature value of each external request within each time period.

[0051] Based on the deviation feature value of each external request within each time period, it can be understood that the deviation feature value is used to characterize the degree of difference between the actual service call path of a single external request and any normal service call path among all known types of external requests in the cross-network exchange platform. It reflects whether the processing flow of the external request follows the expected healthy business logic. If the deviation feature value of the current external request is larger, it means that the matching degree between the actual service call path of the current external request and the normal service call path is extremely low, reflecting that the current external request is very likely to have deviated from the path during the processing. For example, due to the failure of a certain service node, a retry, degradation, circuit breaker or error handling process is triggered, thereby adding unexpected service nodes or call relationships in the call connection. In this case, the Span data under the current external request is very likely to contain abnormal Span data generated by the failure of the cross-network exchange platform, which has important fault diagnosis value.

[0052] Conversely, the smaller the deviation characteristic value of the current external request, the higher the degree of matching between the actual service call path of the current external request and a known normal service call path. This reflects that the current external request strictly follows the expected and healthy business logic during processing, and its service call chain is complete without any path deviation. In this case, the Span data under the current external request belongs to normal business processing data and is used for routine performance monitoring and business processes.

[0053] Furthermore, this embodiment determines the first abnormal feature value of any service node in each time period based on the distribution of deviation feature values ​​of all external requests to which Span data belongs under any service node in each time period, specifically as follows:

[0054] In this embodiment, the average deviation feature value of all external requests belonging to Span data under any service node within each time period is taken as the first abnormal feature value of any service node within each time period.

[0055] Based on the first abnormal feature value of any service node within each time period, it can be understood that the first abnormal feature value is used to characterize the prevalence of path deviation among all requests processed by the service node within a time period. It reflects the possibility of structural failure of the service node itself or its downstream nodes. The magnitude of the first abnormal feature value is determined by the deviation feature value of all external requests flowing through the service node. That is, if the average deviation feature value of all external requests belonging to the current service node is larger, it indicates that a large number of requests passing through the current service node have experienced serious path deviation. This reflects that the current service node itself or its closely dependent downstream nodes are very likely to be in a fault state, causing a large number of requests to be unable to complete normally and forced to take the wrong processing path. Correspondingly, the larger the first abnormal feature value, the more attention should be paid to the current service node.

[0056] Conversely, the smaller the first abnormal characteristic value of the current service node, the lower the deviation characteristic value of most requests flowing through the current service node. This means that the actual service call paths of these requests can be highly matched with a certain type of normal service call path. This reflects that the current service node and its closely dependent downstream nodes are running stably, and there are no faults that cause structural changes in the request process. The service call link is smooth and meets expectations. Therefore, the smaller the corresponding first abnormal characteristic value, the healthier the operating status of the service node in this time period.

[0057] Thus, this embodiment quantifies the deviation feature value of each external request by constructing a service call path graph and comparing it with the topological difference of the normal service call path graph. In turn, it assesses the prevalence of path deviation of each service node and forms the first abnormal feature value. This method can accurately identify abnormal call links triggered by platform failures, which helps to improve the data accuracy and fault location efficiency of the link tracing system in complex fault-tolerant scenarios.

[0058] Step S3: Based on the changing trend of the call duration of all Span data under any service node in each time period, determine the second abnormal feature value of any service node in each time period, and combine it with the first abnormal feature value to determine the sampling adjustment coefficient of any service node in each time period, so as to adjust the sampling rate of Span data under any service node in the next time period of each time period.

[0059] When a cross-network exchange platform is processing external requests, if the system experiences performance abnormalities, such as connection blockage caused by CPU or memory exhaustion or network latency, it will directly affect the processing efficiency of relevant nodes in the service call chain. Specifically, the response time of the service node will increase significantly, resulting in an increasing trend in the processing time of the Span data.

[0060] Therefore, to prevent the link tracing system from missing critical Span data caused by performance anomalies due to the use of a fixed sampling rate, this embodiment determines the second abnormal feature value of any service node in each time period based on the changing trend of the call duration of all Span data under any service node in each time period. Combined with the first abnormal feature value, a sampling adjustment coefficient for any service node in each time period is determined to adjust the sampling rate of Span data under any service node in the next time period. The specific process is as follows:

[0061] In this embodiment, firstly, based on the changing trend of the call duration of all Span data under any service node within each time period, the second abnormal feature value of any service node within each time period is determined, specifically:

[0062] In this embodiment, the call duration of all Span data under any service node in each time period is fitted according to the time sequence of all Span data, and the slope of the fitted line is used as the second abnormal feature value of any service node in each time period.

[0063] It should be noted that there are many commonly used fitting methods. In this embodiment, the least squares fitting method is used to fit the call duration. In practical applications, as other implementation methods, implementers may also choose other fitting algorithms such as polynomial function fitting according to specific circumstances. This embodiment does not impose any special restrictions on the selection of fitting methods.

[0064] The least squares method is a well-known technique, and the specific process of using it to fit the call duration will not be elaborated here.

[0065] Specifically, if the slope of the fitted line is negative, its second abnormal feature value is assigned to 0.

[0066] Based on the second abnormal feature value of any service node within each time period, it can be understood that the second abnormal feature value reflects the degradation trend of the service node's processing performance within the time period. It is used to assess whether the service node has experienced performance anomalies, such as resource exhaustion, connection blocking, etc., which cause the service node to respond slowly. If the slope of the fitted line obtained by fitting the call duration of all Span data under the current service node according to the time sequence of all Span data is larger, it means that the call duration of the current service node shows a significant and continuous growth trend. This directly reflects that the current service node is facing a serious performance bottleneck, which leads to a continuous decline in the processing capacity of the current service node and a slower and slower response. If not intervened in time, it may eventually lead to service timeout or crash. Therefore, the larger the corresponding second abnormal feature value, the better.

[0067] Conversely, if the slope of the fitted line obtained by fitting the call duration of all Span data under the current service node according to the time sequence of all Span data is smaller, or even close to zero or negative, it means that the call duration of the current service node remains stable or shows a downward trend. This directly reflects that the processing performance of the current service node is in a healthy and stable state, with sufficient resources, no continuous deterioration in response time, and the ability to process external requests efficiently and reliably. Therefore, the smaller the corresponding second abnormal characteristic value.

[0068] Furthermore, in this embodiment, based on the first abnormal feature value and the second abnormal feature value, the sampling adjustment coefficient is determined as follows:

[0069] In this embodiment, the result of positively fusing the normalized value of the first abnormal feature value and the normalized value of the second abnormal feature value of any service node in each time period is used as the sampling adjustment coefficient of any service node in each time period.

[0070] Preferably, the schematic diagram of the sampling adjustment coefficient extraction process provided in this embodiment is as follows: Figure 2 As shown.

[0071] It should be understood that positive fusion refers to combining two or more indicators through addition or multiplication to obtain a comprehensive indicator, thereby more comprehensively and accurately assessing a phenomenon or problem. This fusion method is not limited to simple arithmetic operations, but can also include more complex statistical models and analytical methods. Implementers can choose according to specific circumstances, and this embodiment does not impose any special restrictions.

[0072] Preferably, as one implementation method, in this embodiment, the average of the normalized values ​​of the first abnormal feature value and the normalized values ​​of the second abnormal feature value of any service node in each time period is used as the sampling adjustment coefficient of any service node in each time period. In actual application, as other implementation methods, implementers may also adopt other positive fusion methods such as multiplication or addition according to specific circumstances. This embodiment does not impose any special restrictions.

[0073] Based on the sampling adjustment coefficient of any service node within each time period, it can be understood that the sampling adjustment coefficient reflects the sampling priority that the service node should be assigned within the time period. It comprehensively reflects the overall anomaly degree of the service node simultaneously experiencing structural failure risk and performance degradation risk within the time period, and is used to assess and determine whether it is necessary to increase the sampling rate of the service node's Span data. If the first anomaly characteristic value of the current service node is larger within the current time period, it indicates that there is a widespread path deviation problem within the current service node within the current time period, and the current service node is very likely a source of failure. Therefore, it is necessary to increase its sampling rate to capture more abnormal call chains caused by failure, thereby providing sufficient data support for fault location. Therefore, the corresponding sampling adjustment coefficient is larger. At the same time, if the second anomaly characteristic value of the current service node is larger within the current time period, it indicates that the performance of the current service node is continuously deteriorating within the current period. It is necessary to increase its sampling rate to capture more Span data with continuously increasing response time. Therefore, the corresponding sampling adjustment coefficient is larger, thereby providing key basis for performance bottleneck analysis and capacity planning.

[0074] Conversely, if the first abnormal characteristic value of the current service node is smaller within the current time period, it indicates that the requests flowing through the current service node within the current time period generally follow normal call paths, and the service node itself and its downstream nodes have not experienced structural failures. At the same time, if the second abnormal characteristic value of the current service node is smaller within the current time period, it indicates that the call duration of the current service node remains stable within the current period, there are no signs of continuous performance degradation, and the service node is in a healthy operating state. Therefore, when both the first and second abnormal characteristic values ​​are small, the overall abnormality of the current service node within the time period is low, and there is no need to increase the sampling rate of its Span data; the corresponding sampling adjustment coefficient should be smaller.

[0075] Furthermore, based on the aforementioned sampling adjustment coefficient, this embodiment adjusts the sampling rate of the Span data under any service node in the next time period of each time period. Specifically:

[0076] As one implementation method, in this embodiment, the sampling rate of the Span data under service node i in the next time period of time period T is... The expression is: In the formula, This represents the sampling adjustment coefficient for service node i within the time period T; , represents the first preset value and the second preset value, respectively, where the first preset value is less than the second preset value; norm() represents the normalization function; round[] represents the rounding function.

[0077] It should be noted that the values ​​of the first preset value and the second preset value are both set manually. In this embodiment, the value of the first preset value is 0.01 and the value of the second preset value is 1. In actual application, as other implementation methods, implementers can also set them according to specific circumstances. This embodiment does not impose any special restrictions.

[0078] To clarify, the sampling rate is the proportion of tracking data collected in a distributed tracing system out of all requests. It is a floating-point number between 0 and 1.0, representing the proportion of requests collected.

[0079] Based on the sampling rate, it can be understood that if the sampling adjustment coefficient of the current service node is larger in the current time period, it indicates a higher overall anomaly risk for the current service node in the current time period. The distributed tracing system believes that the current service node is more likely to generate abnormal Span data in the next time period. Therefore, it will increase the sampling rate of Span data in the next time period to ensure that enough Span data can be captured, thus providing a key basis for performance bottleneck analysis and capacity planning. Conversely, if the sampling adjustment coefficient of the current service node is smaller in the current time period, it indicates a lower overall anomaly risk for the current service node in the current time period. The distributed tracing system believes that the current service node is less likely to generate abnormal Span data in the next time period. Therefore, it will reduce or maintain the sampling rate of Span data in the next time period to reduce system resource consumption while ensuring that the monitoring of normal business processes remains at a reasonable level.

[0080] The sampling rate adjustment value of each service node in the cross-network exchange platform system is calculated in the next time period of time period T. The sampling rate adjustment value is used as the sampling rate of the Span data acquisition terminal of the client deployed under each service node in the next time period. The distributed link tracing system is then used to perform real-time tracking and monitoring of the entire process of data flow, business flow and information flow in the cross-network exchange platform.

[0081] Thus, this embodiment constructs a dynamic sampling adjustment mechanism by comprehensively considering the path deviation and performance degradation trends of service nodes. This enables accurate identification of abnormal nodes and intelligent control of the sampling rate, ensuring comprehensive capture of key abnormal data and improving the efficiency and accuracy of fault location in cross-network exchange platforms.

[0082] Based on the same inventive concept as the above method, this application embodiment also provides a data security sharing device for a cross-network exchange platform, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described data security sharing methods for a cross-network exchange platform.

[0083] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, specific embodiments of this specification have been described above. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0084] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0085] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A method for secure data sharing across a cross-network exchange platform, characterized in that, The method comprises the following steps: On the cross-network exchange platform, real-time acquisition of Span data and its calling duration under each service node, and acquisition of all Span data under each external request in each time period; Based on the association between all Span data under each external request in each time period, a service calling path graph is constructed; all normal service calling path graphs under various external requests are acquired, based on the difference between the service calling path graph of each external request in each time period and each normal service calling path graph under various external requests, a service difference value between the service calling path graph of each external request in each time period and each normal service calling path graph is determined to determine the deviation characteristic value of each external request in each time period; based on the distribution of the deviation characteristic value of the external request to which all Span data under any service node in each time period belongs, a first abnormal characteristic value of any service node in each time period is determined; Based on the change trend of the calling duration of all Span data under any service node in each time period, a second abnormal characteristic value of any service node in each time period is determined, and in combination with the first abnormal characteristic value, a sampling adjustment coefficient of any service node in each time period is determined to adjust the sampling rate of Span data under any service node in the next time period of each time period.

2. The method of claim 1, wherein the method further comprises: The construction method of the service calling path graph is: All Span data under each external request in each time period are taken as nodes in a directed graph, the connection line between nodes with calling relationship is taken as an edge in the directed graph, the direction of the edge is from the caller to the callee, and the constructed directed graph is denoted as a service calling path graph.

3. The method of claim 1, wherein the method further comprises: The various external requests are data query, browsing and data exchange.

4. The method of claim 1, wherein the method further comprises: The determination method of the service difference value between the service calling path graph of each external request in each time period and each normal service calling path graph is: The adjacency matrix of each external request service calling path graph and all normal service calling graphs of all external requests is acquired; The difference between the adjacency matrix of each external request service calling path graph and each normal service calling path graph under all external requests is taken as the service difference value between the service calling path graph of each external request in each time period and each normal service calling path graph.

5. The method of claim 1, wherein the method further comprises: The deviation characteristic value of each external request in each time period is the minimum value in the service difference value between the service calling path graph of each external request in each time period and all normal service calling path graphs.

6. The method of claim 1, wherein the method further comprises: The first abnormal characteristic value of any service node in each time period is the result of taking the mean value of the deviation characteristic value of the external request to which all Span data under any service node in each time period belongs.

7. The data security sharing method for a cross-network exchange platform as described in claim 1, characterized in that, The determination method of the second abnormal characteristic value of any service node in each time period is: The calling duration of all Span data under any service node in each time period is fitted in the time sequence order of all Span data, and the slope of the fitted straight line is taken as the second abnormal characteristic value of any service node in each time period.

8. The data security sharing method for a cross-network exchange platform as described in claim 1, characterized in that, The sampling adjustment coefficient of any service node in each time period is the positive fusion result of the first abnormal feature value normalized value and the second abnormal feature value normalized value of any service node in each time period, wherein the positive fusion means combining two or more indexes together by addition or multiplication.

9. A data security sharing method for a cross-network exchange platform as described in claim 1, characterized in that, The adjusting the sampling rate of the Span data of any service node in the next time period of each time period comprises: a sampling rate of the span data under the service node i in a next time period of the time period T The expression is: ; wherein, represents a sampling adjustment coefficient under the service node i in the time period T; , respectively represent a preset first value and a preset second value, wherein the preset first value is less than the preset second value; norm() represents a normalization function; and round[] represents a rounding function.

10. A data security sharing device of a cross-network switching platform, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The processor executes the computer program to realize the steps of the data security sharing method of the cross-network switching platform according to any one of claims 1-9.

Images