CN-122027527-A - Internet infrastructure service index measurement method, device, equipment and medium

CN122027527ACN 122027527 ACN122027527 ACN 122027527ACN-122027527-A

Abstract

The invention provides an Internet infrastructure service index measurement method, device, equipment and medium, which are characterized by sending probe requests of various preset protocols to a target network address, identifying the type of Internet infrastructure service provided by a target according to response messages returned by the target network address for the probe requests, calling a corresponding interactive behavior template library according to the type of Internet infrastructure service, generating a simulated interactive request sequence through reinforcement learning agents, dynamically deciding and executing actions based on state feedback of the target service, acquiring response data of the simulated interactive request, and calculating multi-dimensional service indexes according to the response data, thereby fundamentally solving the problems of single interactive mode and poor service adaptation capability, realizing self-adaption and intelligent multi-state interactive test for different infrastructure services, and realizing deep and implicit threat measurement for service health conditions.

Inventors

MU CHUNLIN
LI HONGTAO
MA YONGZHENG
WEI JUNKAI
SUN JINGYI

Assignees

中国互联网络信息中心

Dates

Publication Date: 20260512
Application Date: 20251223

Claims (10)

1. An internet infrastructure service index measurement method, comprising: sending a probe request of a plurality of preset protocols to a target network address, and identifying the type of Internet infrastructure service provided by a target according to a response message returned by the target network address for the probe request; Calling a corresponding interaction behavior template library according to the service type of the Internet infrastructure; Generating a simulation interaction request sequence through a reinforcement learning agent, wherein the reinforcement learning agent makes a dynamic decision and executes actions based on state feedback of a target service, and the actions comprise a behavior template fusing simulation boundary conditions and error responses and disturbance parameters thereof; And collecting response data of the simulation interaction request, and calculating a multi-dimensional service index according to the response data.
2. The internet infrastructure service indicator measurement method according to claim 1, further comprising: extracting a multidimensional feature vector from the response data, wherein the feature vector at least comprises response time, response packet size and a protocol specific field; Inputting the feature vector into a pre-trained isolated forest model, and calculating an anomaly score, wherein the anomaly score is used for quantifying the degree of deviation of the response data from a normal mode; And comparing the anomaly score with a preset judgment threshold, judging the response as anomaly if the anomaly score exceeds the judgment threshold, and outputting a measurement report containing a conventional index and an anomaly detection result.
3. The method for measuring an internet infrastructure service index according to claim 1, wherein the identifying the type of the internet infrastructure service provided by the target according to the response message returned by the target network address for the probe request comprises: constructing probe requests of a plurality of standard protocols; For each protocol probe request, collecting a response message returned by the target network address, and analyzing the matching degree of the response message and each protocol; And probing the service type corresponding to the protocol with the highest matching degree, and determining the service type of the Internet infrastructure provided by the target.
4. The method for measuring the index of the service of the internet infrastructure according to claim 1, wherein the calling the corresponding interactive behavior template library according to the service type of the internet infrastructure comprises: Selecting a matching interaction behavior template from the interaction behavior template library based on the identified internet infrastructure service type; Wherein the interactive behavior template defines a protocol interactive logic framework for measuring indexes for services.
5. The method of claim 1, wherein the generating a simulated interactive request sequence by a reinforcement learning agent that makes dynamic decisions and performs actions based on state feedback of a target service comprises: the reinforcement learning agent receives a current state vector of a target service; Selecting an action from a preset action space based on the current state vector and an internal strategy function, wherein the action is a combination of an interaction behavior template and a group of disturbance parameters; executing the action, and after receiving the service response, calculating instant rewards according to a preset rewarding function; updating the internal policy function according to the current state vector, the executed action, the instant prize and the next state vector; and re-executing the series of actions based on the updated internal strategy function to generate a simulated interactive request sequence.
6. The method for measuring the service index of the internet infrastructure according to claim 1, wherein the disturbance parameters are obtained by dynamically assigning corresponding disturbance parameter fields in a disturbance field library; the disturbance field library designs disturbance parameter fields aiming at each interactive behavior template, configures a value range for each disturbance parameter field, and comprises at least one of a request concurrency number field, a request content length field, a query interval field and a timeout time field.
7. The method for measuring service indexes of internet infrastructure according to claim 1, wherein the collecting response data of the simulated interactive request and calculating the multi-dimensional service indexes according to the response data comprises: Analyzing response messages corresponding to each simulation interaction request, and extracting original data, wherein the original data comprises at least one of response time, protocol state codes, returned data content and network layer characteristics; And calculating service indexes of different dimensions based on the original data, wherein the service indexes of different dimensions comprise an availability index, a performance index, a stability index and a safety index.
8. An internet infrastructure service index measurement device, comprising: the identification module is used for sending probe requests of various preset protocols to a target network address, and identifying the type of the Internet infrastructure service provided by the target according to a response message returned by the target network address for the probe requests; The calling module is used for calling a corresponding interaction behavior template library according to the service type of the Internet infrastructure; The generation module is used for generating a simulation interaction request sequence through the reinforcement learning agent, wherein the reinforcement learning agent makes a dynamic decision and executes actions based on state feedback of target service, and the actions comprise a behavior template and disturbance parameters thereof fusing simulation boundary conditions and error responses; And the calculating module is used for collecting the response data of the simulation interaction request and calculating a multi-dimensional service index according to the response data.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the internet infrastructure service indicator measurement method of any of claims 1 to 7 when the program is executed.
10. A non-transitory readable storage medium having stored thereon a computer program, which when executed by a processor implements the internet infrastructure service indicator measurement method of any of claims 1 to 7.

Description

Internet infrastructure service index measurement method, device, equipment and medium Technical Field The present invention relates to the field of computers and networks, and in particular, to a method, an apparatus, a device, and a medium for measuring service indexes of an internet infrastructure. Background The internet infrastructure service is a basic guarantee of normal operation of the internet, and mainly refers to a protocol service for supporting a network core function and providing key supporting capability to the outside. Common internet infrastructure services include DNS (domain name system service), NTP (time synchronization service), BGP (border gateway protocol), DHCP (dynamic address assignment), and the like. The DNS (Domain NAME SYSTEM) is used for resolving a Domain name accessed by a user into an actual IP address, is an entrance of all Internet access, and NTP (Network Time Protocol) is used for synchronizing system time of each node and is widely applied to key scenes such as log tracking, transaction systems, security authentication and the like. The current operation condition assessment of the service is mainly completed through timing detection, static script or single request simulation mode, and has various problems such as single interaction mode, lack of behavior change, limitation of index dimension, multi-focusing on whether response exists or not, delay degree, lack of complete index system, poor service adaptation capability, incapability of dynamically switching interaction logic according to the target service type, lack of abnormal behavior simulation capability, and lack of test on boundary conditions and error response. Disclosure of Invention The invention provides a method, a device, equipment and a medium for measuring an Internet infrastructure service index, which are used for solving the defects that the prior art mainly adopts a fixed and single interaction mode when evaluating the Internet infrastructure service, is limited to a basic response state and a delay index, lacks automatic service adaptation capability and systematically simulates and deeply detects an abnormal scene. The invention provides a method for measuring service indexes of internet infrastructure, which comprises the following steps: sending a probe request of a plurality of preset protocols to a target network address, and identifying the type of Internet infrastructure service provided by a target according to a response message returned by the target network address for the probe request; Calling a corresponding interaction behavior template library according to the service type of the Internet infrastructure; Generating a simulation interaction request sequence through a reinforcement learning agent, wherein the reinforcement learning agent makes a dynamic decision and executes actions based on state feedback of a target service, and the actions comprise a behavior template fusing simulation boundary conditions and error responses and disturbance parameters thereof; And collecting response data of the simulation interaction request, and calculating a multi-dimensional service index according to the response data. The method for measuring the service index of the Internet infrastructure, provided by the invention, further comprises the following steps: extracting a multidimensional feature vector from the response data, wherein the feature vector at least comprises response time, response packet size and a protocol specific field; Inputting the feature vector into a pre-trained isolated forest model, and calculating an anomaly score of the feature vector, wherein the anomaly score is used for quantifying the degree of deviation of the response data from a normal mode; and comparing the anomaly score with a preset judgment threshold, judging the response as anomaly if the anomaly score exceeds the judgment threshold, and outputting a measurement report containing a conventional index and an anomaly detection result. According to the method for measuring the service index of the internet infrastructure provided by the invention, the probe requests of a plurality of preset protocols are sent to the target network address, and the service types of the internet infrastructure provided by the target are identified according to the response message returned by the target network address for the probe requests, and the method comprises the following steps: constructing probe requests of a plurality of standard protocols; For each protocol probe request, collecting a response message returned by the target network address, and analyzing the matching degree of the response message and each protocol; And probing the service type corresponding to the protocol with the highest matching degree, and determining the service type of the Internet infrastructure provided by the target network address. According to the method for measuring the service index of the internet infrastructure provided by the