Transaction-level SLA fracture positioning method and system

Abstract

The invention discloses a transaction-level SLA fracture positioning method and system, and belongs to the field of network application performance monitoring and service quality assurance. The method comprises the steps of firstly collecting network flow data and constructing a transaction unique identifier ID, generating a structured account book data object according to transaction information, calculating baseline drift and an attribution function of a transaction path segment, calculating the attribution function of each segment to identify a responsibility segment causing SLA fracture when a transaction total time delay is greater than a fixed threshold value, and obtaining a structured account book data object; and finally, calculating a path health index based on the stability index of each segment of the transaction to evaluate the transaction path. According to the method, the transaction path is subjected to segmented modeling, the attribution contribution value of each segment is calculated, so that the SLA anomaly can be accurately positioned to a specific path segment, the anomaly positioning precision is improved, the method can quickly identify a related responsibility segment, the troubleshooting time is shortened, a long-term hotspot segment is identified through cross-transaction statistical analysis, and the service life of the SLA is prolonged. And a basis is provided for system performance optimization and resource configuration.

Description

A Transaction-Level SLA Fracture Location Method and System Technical Field

[0001] This invention belongs to the field of network application performance monitoring and service quality assurance, and relates to a transaction-level SLA break location method and system. Specifically, it relates to a method and system for generating path ledger data objects based on transaction information and performing segmented contribution calculation and cross-transaction statistical analysis of SLA breaks based on multi-factor attribution functions. Background Technology

[0002] In network microservices and multi-layered distributed architectures, a single transaction often involves calls to multiple service segments, making the location of transaction-level SLA (Service Level Agreement) breaches a critical issue. Existing technologies primarily employ distributed tracing (Trace / Span), calculation of response times at each layer, SLA violation detection and alarms, and root cause analysis algorithms to locate SLA breaches. However, these methods suffer from the following shortcomings: they only record runtime call chain relationships, failing to form a transaction-level accountability ledger structure; most methods only statistically analyze latency information without statistically analyzing the contribution of various factors; they only analyze single transactions, lacking cross-transaction clustering and decision-making capabilities; and they lack non-linear attribution mechanisms for multiple influencing factors.

[0003] Therefore, a structured path ledger data object and a corresponding SLA break location method are needed to achieve persistent attribution of transaction-level impact factors, calculate the contribution of multiple impact factors, and thus realize SLA break location and support statistical decision-making across transactions. Summary of the Invention

[0004] The technical problem this invention aims to solve is to address the shortcomings of existing SLA (Service Level Agreement) breakage location methods by proposing a transaction-level SLA breakage location method and system. The method generates transaction-level path ledger data objects and constructs segmented multi-factor attribution functions, enabling cross-transaction path health analysis and accurate attribution of responsibility for SLA breaks. Technical Solution

[0005] To achieve the above objectives, the technical solution provided by the present invention is as follows: a transaction-level SLA breakage location method, comprising the following steps: Step 1, collecting network traffic data, the network traffic data including timestamp, session identifier, request-response direction, TCP status, application layer status code, packet length, and retransmission information.

[0006] Step 2: Construct a unique transaction identifier ID based on the network traffic data collected in Step 1. The transaction ID includes a five-tuple of source IP-source port-destination IP-destination port-protocol type, application layer identifier, time window, and session sequence number.

[0007] Step 3: Differentiate the transaction path segments based on the roles of the participating nodes.

[0008] Step 4: After the transaction is completed, a structured ledger data object is generated based on the transaction information. The data object contains the start time, end time, segment delay, number of reconnections, number of failures, delay variance, and baseline delay of all path segments within the transaction.

[0009] Step 5: Extract the segment delay D of segment i from the transaction path segment information of the structured ledger data object. i Baseline delay B i Calculate the baseline offset E of this segment. i The method is: E i = max(0,D i -B i (where e) i, j w is the j-th extraction factor for segment i. i, j These are the configurable weights for each factor.

[0010] The attribution function is normalized to obtain Normal_Contr. i The method is as follows:

[0011] Step 6: When the total transaction latency exceeds the fixed threshold SLA_threshold, calculate the Normal_Contr for each segment it contains. i The values ​​are sorted, and the primary and secondary responsible segments that caused the SLA breakage are identified based on these values. Step 7: Calculate the path health index Path_Health based on the stability index of each segment of the transaction to evaluate the quality of the transaction path. The method is as follows: Step 7-1, calculate the segment stability index S i The method is V i For the segmented latency variance in the structured ledger data object in step 4, the smaller the segmented latency variance, the more stable the segment; in step 7-2, calculate the path health index Path_Health, using the following method: Where E i The baseline drift of segment i obtained in step 5 is calculated. The larger the Path_Health value, the better the path.

[0012] Furthermore, based on the iNormal_Contr segments obtained in step 5... i The value converts the transaction path into a vector V_path, where V_path = [Normal_Contr1, Normal_Contr2, …, Normal_Contr…]. n The similarity between transactions is obtained by calculating vector distance, and different session types are distinguished by clustering.

[0013] Furthermore, based on the iNormal_Contr segments obtained in step 5... i The system identifies hotspot segments and determines the concentrated areas of SLA breakage responsibility based on the location of these segments. The method involves calculating the average contribution of each segment across all transactions and ranking all segments according to this value. Where m is the total number of transactions, k i This represents the Normal_Contr of the i-th segment in the k-th transaction. i The value identifies the segments with the highest average contribution as hot segments.

[0014] Furthermore, another method for hotspot segment identification is to count the number of times segment i is identified as the main responsible segment in step 6, and rank all segments based on this number, with the segment ranked higher being the hotspot segment.

[0015] Furthermore, in step 5, the attribution function Contr i It can be linear or nonlinear.

[0016] A system for implementing the transaction-level SLA break location method of the present invention includes a traffic acquisition module, a structured data object module, a calculation and analysis module, an SLA determination module, and a decision output module, wherein: the traffic acquisition module implements network traffic data acquisition in the method; the structured data object module implements the generation of structured ledger data objects in the method; the calculation and analysis module implements the calculation steps in the method and analyzes the relevant results; the SLA determination module implements transaction segment location of SLA breaks in the method; and the decision output module implements the evaluation of the quality of transaction paths and hotspot segment identification in the method.

[0017] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed, implements the steps of the above method.

[0018] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method described above.

[0019] Beneficial effects

[0020] The SLA breakage location method of this invention improves the accuracy of SLA location by segmenting and modeling the transaction path and calculating the attribution contribution value of each segment. Furthermore, by generating a transaction-level path ledger and sorting the responsible segments, the method can quickly identify primary and secondary responsible segments, reducing troubleshooting time. Simultaneously, cross-transaction statistical analysis identifies long-term hotspot segments, providing a basis for system performance optimization and resource allocation. Attached Figure Description

[0021] Figure 1 is a flowchart of the method of the present invention; Figure 2 is a system composition diagram for implementing the method of the present invention. Detailed Implementation

[0022] To further understand the content of this invention, it will be described in detail with reference to the accompanying drawings and specific embodiments.

[0023] As shown in Figure 1, the SLA breakage location method of the present invention mainly includes the following steps: Step 1, collect network traffic data. The data is collected through mirror port, split device or bypass collection device. In this embodiment, the collected traffic data includes: timestamp, session identifier, request-response direction, TCP status, application layer status code, packet length and retransmission information.

[0024] The request-response direction refers to the transmission direction of network data packets in transactional communication, used to distinguish between transactional requests and responses; TCP status refers to the state or flag information of a TCP protocol connection during communication, such as SYN connection establishment request, SYN-ACK connection confirmation, ACK confirmation, FIN normal connection closure, RST abnormal connection reset, etc.; application layer status codes refer to the business processing result status returned by the application protocol, such as 200 request success, 404 resource not found, 500 server error, etc. in the HTTP protocol.

[0025] Step 2: Construct a unique transaction ID based on network traffic data. The transaction ID includes the following information: a transaction 5-tuple (source IP / source port / destination IP / destination port / protocol type), application layer identifier, time window, and session sequence number. The time window refers to the time range for identifying and aggregating data packets from the same transaction, used to distinguish different transactions within continuous network traffic.

[0026] Step 3: Distinguish transaction path segments according to the roles of the participating nodes. Transaction path segments include: Client to Front End (FE), Front End (FE) to Middle Tier (Mid), Middle Tier (Mid) to Back End (BE), and Back End (BE) to Database (DB).

[0027] Step 4: After the transaction is completed, a structured ledger data object is generated based on the transaction information. The data object contains the start time, end time, segment delay, number of reconnections, number of failures, delay variance, and baseline delay information for all path segments within the transaction. The number of failures refers to the number of failure events that occurred within the corresponding path segment. The delay variance refers to the degree of fluctuation in the delay data of the corresponding path segment within the statistical period, which can be calculated using delay variance, standard deviation, or the delay fluctuation amplitude within a sliding window. The baseline delay refers to the historical statistical benchmark delay value of the corresponding path segment under normal operating conditions.

[0028] Step 5: Extract the segment delay D of segment i from the transaction path segment information of the structured ledger data object. i Baseline delay B i Information, calculate the baseline offset E of this segment. i The method is as follows:

[0029] Extracting segmentation information from structured ledger data objects to construct attribution functions, in this embodiment, the attribution function is based on the segmentation baseline offset E. i Number of reconnections R i Number of failures F i Time delay variance V i Constructing the attribution function Contr i The method is as follows:

[0030] w1-w4 are the configurable weights of each factor.

[0031] The above formula assumes that all factors have the same influence. When the influence of each factor is different, the attribution function can be expressed in a non-linear manner, such as:

[0032] Normalization process yields Normal_Contr i The method is as follows:

[0033] Step 6: When the total transaction latency exceeds the fixed threshold SLA_threshold, calculate the Normal_Contr for each segment it contains. i Sort the values ​​and output the primary responsibility section and the secondary responsibility section.

[0034] Step 7: Calculate the path health index Path_Health based on the stability index of each segment of the transaction to evaluate the quality of the transaction path. The method is as follows: Calculate the segment stability index S i The method is V i The smaller the segmented delay variance, the more stable the segment is.

[0035] Then, the Path Health index is calculated using the following method:

[0036] Where E i The baseline drift of segment i obtained in step 5 is calculated. The larger the Path_Health value, the better the path.

[0037] Based on the iNormal_Contr segment obtained in step 5 i The value converts the transaction path into a vector V_path, where V_path = [Normal_Contr1, Normal_Contr2, …, Normal_Contr…]. n The similarity between transactions is obtained by calculating vector distance (such as Euclidean distance or cosine similarity), and different session types are distinguished by clustering, such as different path types by k-means clustering of transaction V_path.

[0038] Based on the iNormal_Contr segment obtained in step 5 i Value identification of hotspot segments includes the following steps: (1) Calculate the average contribution of each segment in all transactions and rank all segments according to their values, the method is as follows:

[0039] Where m is the total number of transactions, k i This represents the Normal_Contr of the i-th segment in the k-th transaction. i The higher the average contribution of a segment, the higher its ranking in the hot topic list.

[0040] (2) Count the number of times segment i is identified as the main responsibility segment in step 6, and rank all segments based on the number of times.

[0041] (3) Identify the top-ranked segments as hotspot segments, and then determine the area of ​​concentrated responsibility based on the location of the hotspot segments.

[0042] As shown in Figure 2, the system implementing the SLA breakage location method of the present invention consists of a traffic acquisition module, a structured data object module, a calculation and analysis module, an SLA determination module, and a decision output module. The traffic acquisition module implements network traffic data acquisition in the method; the structured data object module implements the generation of structured ledger data objects in the method; the calculation and analysis module implements the calculation steps in the method and analyzes the relevant results; the SLA determination module implements transaction segmentation location of SLA breaks in the method; and the decision output module implements the evaluation of the quality of transaction paths and hotspot segmentation identification in the method.

[0043] The present invention and its embodiments have been described above illustratively. This description is not restrictive, and the figures shown are only one embodiment of the present invention; the actual structure is not limited thereto. Therefore, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the present invention, such designs should fall within the protection scope of the present invention.

Claims

1. A transaction-level SLA fracture localization method, characterized in that, Includes the following steps: Step S1: Collect network traffic data, which includes timestamps, session identifiers, request-response directions, TCP status, application layer status codes, packet lengths, and retransmission information. Step S2: Construct a unique transaction identifier ID based on the network traffic data collected in step S1. The transaction ID identifier includes a five-tuple of source IP-source port-destination IP-destination port-protocol type, application layer identifier, time window, and session sequence number. Step S3: Differentiate transaction path segments based on the roles of participating nodes. Step S4: After the transaction ends, generate a structured ledger data object based on the transaction information. The data object contains the start time, end time, segment delay, number of reconnections, number of failures, delay variance, and baseline delay for all path segments within the transaction. Step S5: Extract the segment delay D of segment i from the transaction path segment information of the structured ledger data object. i Baseline delay B i Calculate the baseline offset E of this segment. i The method is: E i = max(0,D i -B i Extracting information from segment i from structured ledger data objects to construct a weighted attribution function Contr i The method is as follows: Where e i, j w is the j-th extraction factor for segment i. i, j Configurable factor weights; normalizing the attribution function to obtain Normal_Contr i The method is as follows: Step S6: When the total transaction latency exceeds the fixed threshold SLA_threshold, calculate the Normal_Contr of each segment it contains. i The values ​​are sorted, and the primary and secondary responsibility segments causing SLA breakage are identified based on these values; Step S7, the path health index Path_Health is calculated based on the stability index of each segment of the transaction to evaluate the quality of the transaction path. The method is as follows: Step S7-1, calculate the segment stability index S i The method is V i The segmented latency variance in the structured ledger data object in step S4 is defined as follows: the smaller the segmented latency variance, the more stable the segment; in step S7-2, the path health index Path_Health is calculated using the following method: Where E i The baseline drift of segment i obtained in step S5 is calculated by the Path_Health value. The larger the Path_Health value, the better the path.

2. The transaction-level SLA fracture location method as described in claim 1, characterized in that, Based on the iNormal_Contr segment obtained in step S5 i The value converts the transaction path into a vector V_path, where V_path = [Normal_Contr1, Normal_Contr2, …, Normal_Contr…]. n The similarity between transactions is obtained by calculating vector distance, and different session types are distinguished by clustering.

3. The transaction-level SLA fracture localization method as described in claim 1, characterized in that, Based on the iNormal_Contr segment obtained in step S5 i The system identifies hotspot segments and determines the concentrated areas of SLA breakage responsibility based on the location of these segments. The method involves calculating the average contribution of each segment across all transactions and ranking all segments according to this value. Where m is the total number of transactions, k i This represents the Normal_Contr of the i-th segment in the k-th transaction. i The value identifies the segments with the highest average contribution as hot segments.

4. The transaction-level SLA fracture location method as described in claim 3, characterized in that, Another method for identifying hotspot segments is to count the number of times segment i is identified as the main responsible segment in step S6, and rank all segments based on this number, with the segment ranked higher being the hotspot segment.

5. The transaction-level SLA fracture location method as described in claim 1, characterized in that, In step S5, the attribution function Contr i It can be linear or nonlinear.

6. A system for implementing the transaction-level SLA fracture location method according to any one of claims 1-5, characterized in that, It includes a traffic acquisition module, a structured data object module, a calculation and analysis module, an SLA determination module, and a decision output module. Specifically: the traffic acquisition module implements network traffic data acquisition; the structured data object module implements the generation of structured ledger data objects; the calculation and analysis module implements the calculation steps and analyzes the relevant results; the SLA determination module implements the transaction segmentation location of SLA breaks; and the decision output module implements the evaluation of transaction path quality and hotspot segmentation identification.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed, it implements the steps of the method as described in any one of claims 1-5.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the method as claimed in any one of claims 1-5.

Images