Network performance testing method and apparatus, and storage medium

Abstract

The application discloses a network performance test method and device and a storage medium, relates to the field of communication, and is used for improving the efficiency of network performance test, and comprises the following steps: acquiring a topology graph of a target network; the topology graph comprises network instances and connection relationships between the network instances; determining a first network instance and a second network instance connected with the first network instance according to the topology graph; the first network instance and the second network instance belong to the target network; the first network instance is any network instance in the target network; acquiring a network performance parameter of the first network instance in the case that the second network instance sends a preset bandwidth of traffic to the first network instance; determining a network performance analysis report of the target network based on the network performance parameter of the first network instance and a target large language model; the target large language model is used for predicting network performance according to the network performance parameter and outputting an analysis report of the network performance. The application is applied to the process of network performance test.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of communication, and in particular to a network performance testing method and device and storage medium. BACKGROUND

[0002] With the popularity of cloud computing and the increasing complexity and scale of cloud environment, higher requirements are put forward for network performance management. As an open source cloud platform, OpenStack has a large user group and a rich ecosystem. In order to ensure the stability of the OpenStack network, it is necessary to accurately analyze the OpenStack network.

[0003] In the related art, technical personnel analyze the OpenStack network performance through OpenStack management tools and plug-ins. However, the manual analysis method is difficult to cope with the surge in data volume and complexity, and cannot quickly and accurately determine the network performance. Therefore, how to improve the efficiency of network performance testing is still a technical problem to be solved. SUMMARY

[0004] The present application provides a network performance testing method, device and storage medium, which can improve the efficiency of network performance testing.

[0005] In a first aspect, the present application provides a network performance testing method, comprising: obtaining a topology graph of a target network; the topology graph comprising network instances and connection relationships between the network instances; determining a first network instance and a second network instance connected to the first network instance according to the topology graph; the first network instance and the second network instance belong to the target network; the first network instance is any network instance in the target network; obtaining a network performance parameter of the first network instance in the case that the second network instance sends a preset bandwidth of traffic to the first network instance; determining a network performance analysis report of the target network based on the network performance parameter of the first network instance and a target large language model; the target large language model is used to predict network performance according to the network performance parameter and output an analysis report of the network performance.

[0006] In a possible implementation, a training set is obtained; the training set comprises: historical network performance parameters of multiple historical periods and actual network performance analysis reports corresponding to the historical network performance parameters; the historical network performance parameters in the training set are input into an initial prediction model to determine an initial prediction result; based on the initial prediction result and the actual network performance analysis report, a loss value of the initial prediction model is determined, and in the case that the loss value does not satisfy a preset condition, the model parameters of the initial prediction model are adjusted based on the loss value, and iterative training based on the training set is continued until a model convergence condition is satisfied, and a target large language model is obtained.

[0007] In a possible implementation, the iperf tool in the second network instance is instructed to send traffic of a preset bandwidth to the first network instance; and the iperf tool is installed in the second network instance through a script.

[0008] In a possible implementation, the network quality parameter received by the first network instance is obtained through the iperf tool in the first network instance; the device state information of the first network instance is obtained through the sar command in the first network instance; and the network performance parameter of the first network instance is determined according to the network quality parameter received by the first network instance and the device state information of the first network instance.

[0009] In a possible implementation, the network quality parameter includes at least one of the following: throughput, jitter, packet loss rate, and maximum transmission unit; and the device state information includes at least one of the following: CPU occupation information, memory occupation information, and disk input / output (I / O) information.

[0010] In a possible implementation, the device state information of the second network instance is obtained through the sar command in the second network instance; and the network performance parameter of the first network instance and the device state information of the second network instance are input into the target large language model to obtain the network performance analysis report of the target network.

[0011] In a possible implementation, the network performance parameter includes at least one of the following: expected bandwidth, actual sending bandwidth, actual receiving bandwidth, jitter, packet loss rate, CPU utilization rate, and memory utilization rate.

[0012] In a second aspect, the present application provides a network performance testing device, comprising: an obtaining unit configured to obtain a topology graph of a target network; the topology graph includes network instances and connection relationships between the network instances; a processing unit configured to determine a first network instance and a second network instance connected to the first network instance according to the topology graph; the first network instance and the second network instance belong to the target network; the first network instance is any network instance in the target network; the processing unit is further configured to instruct the obtaining unit to obtain a network performance parameter of the first network instance in a case where the second network instance sends traffic of a preset bandwidth to the first network instance; and the processing unit is further configured to determine a network performance analysis report of the target network based on the network performance parameter of the first network instance and a target large language model; the target large language model is configured to predict network performance according to the network performance parameter and output an analysis report of the network performance.

[0013] In a possible implementation, the acquisition unit is further configured to acquire a training set, the training set including historical network performance parameters of a plurality of historical periods and actual network performance analysis reports corresponding to the historical network performance parameters; the processing unit is further configured to input the historical network performance parameters in the training set into an initial prediction model to determine an initial prediction result; the processing unit is further configured to determine a loss value of the initial prediction model based on the initial prediction result and the actual network performance analysis reports, and adjust model parameters of the initial prediction model based on the loss value in a case where the loss value does not satisfy a preset condition, and continue iterative training based on the training set until a model convergence condition is satisfied to obtain a target large language model.

[0014] In a possible implementation, the processing unit is further configured to instruct an iperf tool in the second network instance to send traffic of a preset bandwidth to the first network instance; and the iperf tool is installed in the second network instance through a script.

[0015] In a possible implementation, the acquisition unit is further configured to acquire, by an iperf tool in the first network instance, network quality parameters received by the first network instance; the acquisition unit is further configured to acquire, by a sar command in the first network instance, device state information of the first network instance; and the processing unit is further configured to determine network performance parameters of the first network instance according to the network quality parameters received by the first network instance and the device state information of the first network instance.

[0016] In a possible implementation, the network quality parameters include at least one of the following: throughput, jitter, packet loss rate, and maximum transmission unit; and the device state information includes at least one of the following: CPU occupation information, memory occupation information, and disk input / output (I / O) information.

[0017] In a possible implementation, the acquisition unit is further configured to acquire, by a sar command in the second network instance, device state information of the second network instance; and the processing unit is further configured to input the network performance parameters of the first network instance and the device state information of the second network instance into the target large language model to obtain a network performance analysis report of the target network.

[0018] In a possible implementation, the network performance parameters include at least one of the following: expected bandwidth, actual sending bandwidth, actual receiving bandwidth, jitter, packet loss rate, CPU utilization rate, and memory utilization rate.

[0019] In a third aspect, the present application provides a computer readable storage medium storing one or more programs, the one or more programs including instructions, which when executed by the electronic device of the present application, cause the electronic device to perform the network performance testing method as described in the first aspect and any possible implementation of the first aspect.

[0020] In a fourth aspect, the present application provides an electronic device, comprising: a processor and a memory; wherein the memory is configured to store one or more programs, the one or more programs comprising computer-executable instructions; and when the electronic device is running, the processor executes the computer-executable instructions stored in the memory, so that the electronic device executes the network performance testing method as described in the first aspect and any possible implementation manner of the first aspect.

[0021] In a fifth aspect, the present application provides a computer program product comprising instructions which, when executed on a computer, cause the electronic device of the present application to perform the network performance testing method as described in the first aspect and any possible implementation manner of the first aspect.

[0022] In a sixth aspect, the present application provides a chip system applied to a network performance testing device; the chip system comprises one or more interface circuits and one or more processors. The interface circuit and the processor are interconnected through a circuit; the interface circuit is configured to receive a signal from a memory of the network performance testing device and send a signal to the processor, the signal comprising computer instructions stored in the memory. When the processor executes the computer instructions, the network performance testing device executes the network performance testing method as described in the first aspect and any possible implementation manner thereof.

[0023] In the present application, the name of the network performance testing device does not constitute a limitation on the device or functional unit itself, and in actual implementation, these devices or functional units can appear with other names. As long as the functions of each device or functional unit are similar to those of the present application, they are within the scope of the claims of the present application and equivalent technologies.

[0024] Based on the above technical solution, in the present application, the topology graph comprises network instances and connection relationships between the network instances, and the network performance testing device determines a first network instance and a second network instance connected to the first network instance according to the topology graph. Since the first network instance and the second network instance belong to the target network, the network performance testing device can send traffic of a preset bandwidth to the first network instance through the second network instance, and obtain the network performance parameter of the first network instance at the first network instance. In this way, the network quality between the first network instance and the second network instance can be determined. Further, based on the network quality between the first network instance and the second network instance and the target language model, the network performance analysis report of the target network can be quickly and accurately determined. Compared with the prior art, the technical personnel do not need to analyze the network performance of OpenStack through OpenStack management tools and plug-ins to obtain the network performance analysis report. In the present application, the network data does not need to be analyzed by technical personnel in sequence, which saves the labor cost and improves the efficiency of network performance measurement. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 A network performance test system architecture schematic diagram is provided for the embodiments of the present application.

[0026] Figure 2 A network performance test device structure schematic diagram is provided for the embodiments of the present application.

[0027] Figure 3 A network performance test method flowchart is provided for the embodiments of the present application.

[0028] Figure 4 A network topology diagram schematic diagram is provided for the embodiments of the present application.

[0029] Figure 5 A training set acquisition method flowchart is provided for the embodiments of the present application.

[0030] Figure 6 Another network performance test method flowchart is provided for the embodiments of the present application.

[0031] Figure 7 Another network performance test method flowchart is provided for the embodiments of the present application.

[0032] Figure 8 Another network performance test method flowchart is provided for the embodiments of the present application.

[0033] Figure 9 Another network performance test method flowchart is provided for the embodiments of the present application.

[0034] Figure 10 Another network performance test method flowchart is provided for the embodiments of the present application.

[0035] Figure 11 Another network performance test method flowchart is provided for the embodiments of the present application.

[0036] Figure 12 Another network performance test device structure schematic diagram is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0037] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0038] The character " / " herein generally represents an "or" relationship between the front and rear associated objects. For example, A / B can be understood as A or B.

[0039] The terms "first" and "second" in the description and claims of the present application are used to distinguish different objects, rather than to describe a specific order of the objects. For example, the first edge service node and the second edge service node are used to distinguish different edge service nodes, rather than to describe the order of the characteristics of the edge service nodes.

[0040] In addition, the terms "include" and "have" and any variations thereof mentioned in the description of the present application are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but can optionally include other steps or units not listed, or can optionally include other steps or units inherent to the process, method, product or device.

[0041] In addition, in the embodiments of the present application, the words "exemplarily" or "for example" are used to represent as an example, illustration or description. Any embodiment or design scheme described as "exemplarily" or "for example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the words "exemplarily" or "for example" are intended to present the concept in a specific manner.

[0042] With the popularity of cloud computing and the increasing complexity and scale of cloud environment, higher requirements are put forward for network performance management. As an open source cloud platform, OpenStack has a large user group and a rich ecosystem. In order to ensure the stability of OpenStack network, accurate analysis of OpenStack network is needed.

[0043] In the related art, technical personnel analyze OpenStack network performance through OpenStack management tools and plug-ins. However, the manual analysis method is difficult to cope with the surge in data volume and complexity, and cannot quickly and accurately determine the network performance. Therefore, how to improve the efficiency of network performance testing is still a technical problem to be solved.

[0044] Based on the above technical scheme, the present application considers that the topology graph includes network instances and connection relationships between the network instances, and the network performance testing device determines a first network instance and a second network instance connected with the first network instance according to the topology graph. Since the first network instance and the second network instance belong to the target network, the network performance testing device can send traffic of a preset bandwidth to the first network instance through the second network instance, and acquire the network performance parameter of the first network instance at the first network instance. In this way, the network quality between the first network instance and the second network instance can be determined. Further, based on the network quality between the first network instance and the second network instance and the target language model, the network performance analysis report of the target network can be quickly and accurately determined. Compared with the prior art, the related technical personnel need to analyze the OpenStack network performance through the OpenStack management tool and the plug-in to obtain the network performance analysis report. In the present application, the technical personnel do not need to analyze the network data in sequence, which saves the labor cost and improves the efficiency of network performance measurement.

[0045] Exemplarily, as shown in Figure 1 , Figure 1 An architecture schematic diagram of a network performance testing system provided by the present application is provided, and the network performance testing system 10 includes a data acquisition device 101 and a network performance testing device 102.

[0046] The data acquisition device 101 acquires a requirement document and a network topology graph.

[0047] The network performance testing device 102 acquires the topology graph of the target network through the data acquisition device 101; the topology graph includes network instances and connection relationships between the network instances; according to the topology graph, a first network instance and a second network instance connected with the first network instance are determined; the first network instance and the second network instance belong to the target network; the first network instance is any network instance in the target network; in the case that the second network instance sends traffic of a preset bandwidth to the first network instance, the network performance parameter of the first network instance is acquired; based on the network performance parameter of the first network instance and the target large language model, the network performance analysis report of the target network is determined.

[0048] Optionally, the target large language model is used to predict the network performance according to the network performance parameter and output the analysis report of the network performance.

[0049] The target large language model can be set in the network performance testing device 102, or can be set in a processing device connected with the network performance testing device 102.

[0050] Optionally, the entity device of the data acquisition device 101 is a terminal, and the entity device of the network performance testing device 102 is a server.

[0051] Optionally, the terminal described above can refer to a device that provides voice and / or data connectivity for a user, a handheld device having wireless connection capability, or other processing device connected to a wireless modem. The wireless terminal can communicate with one or more core networks through a radio access network (RAN). The wireless terminal can be a mobile terminal, such as a mobile telephone (also known as a "cellular" telephone) and a computer having a mobile terminal, which communicates voice and / or data with a radio access network. The wireless terminal can be portable, pocket, hand-held, computer-included, or vehicle-mounted, which communicates voice and / or data with a radio access network.

[0052] Optionally, the server described above can be one of the servers in a server cluster (composed of multiple servers), can also be a chip in the server, can also be a system on chip in the server, and can also be implemented through a virtual machine (VM) deployed on a physical machine, and the embodiments of the present application do not make any limitation in this regard.

[0053] The embodiments of the present application provide a network performance testing device for executing the network performance testing system provided by the embodiments of the present application, Figure 2 A structural schematic diagram of the network performance testing device provided by the embodiments of the present application is shown in FIG. 2. As shown in the figure, the network performance testing device 200 includes at least one processor 201, a communication line 202, and at least one communication interface 204, and can further include a memory 203. The processor 201, the memory 203, and the communication interface 204 can be connected through the communication line 202. Figure 2

[0054] The processor 201 can be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement one or more embodiments of the present application, such as one or more digital signal processors (DSPs), or one or more field programmable gate arrays (FPGAs).

[0055] The communication line 202 can include a path for transmitting information between the components described above.

[0056] ​The communication interface 204 is used to communicate with other devices or communication networks, and can use any transceiver-like device, such as Ethernet, radio access network (RAN), WLAN, etc.

[0057] The memory 203 may be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including a compact disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to include or store desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to these.

[0058] In one possible design, the memory 203 can exist independently of the processor 201, that is, the memory 203 can be a memory external to the processor 201. In this case, the memory 203 can be connected to the processor 201 via the communication line 202, and is used to store execution instructions or application code, and the execution is controlled by the processor 201 to implement the network performance testing method provided in the following embodiment of this application. In another possible design, the memory 203 can also be integrated with the processor 201, that is, the memory 203 can be the internal memory of the processor 201. For example, the memory 203 is a cache that can be used to temporarily store some data and instruction information.

[0059] As an implementation method, the processor 201 may include one or more CPUs, such as Figure 2 As another implementation, the network performance testing device 200 may include multiple processors, such as Figure 2 As another implementation, the network performance testing apparatus 200 may further include an output device 205 and an input device 206.

[0060] The following, combined with the Figure 3 The network performance testing method provided in the embodiment of the present application is described in detail. Figure 3 As shown, the network performance testing method includes S301-S304.

[0061] S301, the network performance testing device acquires a topology graph of a target network.

[0062] The topology graph includes network instances and connection relationships between the network instances.

[0063] Optionally, the network instances include virtual servers.

[0064] For example, as shown in the figure, Figure 4 The topology graph includes network instance 1, network instance 2, network instance 3, network instance 4, network instance 5, and network instance 6. Network instance 1 is connected to network instance 2 and network instance 6, network instance 2 is connected to network instance 1 and network instance 3, network instance 3 is connected to network instance 2 and network instance 4, network instance 4 is connected to network instance 3 and network instance 5, network instance 5 is connected to network instance 4 and network instance 6, and network instance 6 is connected to network instance 5 and network instance 1.

[0065] S302, the network performance testing device determines a first network instance and a second network instance connected to the first network instance according to the topology graph.

[0066] The first network instance and the second network instance belong to the target network, and the first network instance is any network instance in the target network.

[0067] It should be explained that the network performance testing device determines each network instance and the network instances connected to each network instance in the topology graph according to the topology graph. In this way, the network performance testing device can understand the network quality of each network instance.

[0068] S303, the network performance testing device acquires a network performance parameter of the first network instance in a case where the second network instance sends a preset bandwidth of traffic to the first network instance.

[0069] The network performance parameter includes at least one of the following: expected bandwidth, actual sending bandwidth, actual receiving bandwidth, jitter, packet loss rate, CPU utilization, and memory utilization.

[0070] In a possible implementation, the network performance testing device connects to the first network instance in the openstack network through an SSH client to acquire the network performance parameter of the first network instance.

[0071] It should be explained that the network performance testing device uses an ssh command and the IP address, username (and optionally port number and key file path) of the first network instance to establish a connection with the first network instance, and then acquires the network performance parameter of the first network instance.

[0072] S304, the network performance testing device determines the network performance analysis report of the target network based on the network performance parameter of the first network instance and the target large language model.

[0073] The target large language model is used to predict network performance according to the network performance parameter and output an analysis report of the network performance.

[0074] In a possible implementation, the network performance testing device inputs the network performance parameter of the first network instance into the target large language model to obtain the network performance analysis report of the target network.

[0075] Optionally, the target language model includes a chatGLM model and a baichuan model.

[0076] The above scheme at least brings the following beneficial effects: In the present application, the topology graph includes network instances and connection relationships between network instances, and the network performance testing device determines the first network instance and the second network instance connected with the first network instance according to the topology graph. Since the first network instance and the second network instance belong to the target network, the network performance testing device can send traffic with a preset bandwidth to the first network instance through the second network instance, and obtain the network performance parameter of the first network instance at the first network instance. In this way, the network quality between the first network instance and the second network instance can be determined. Further, based on the network quality between the first network instance and the second network instance and the target language model, the network performance analysis report of the target network can be quickly and accurately determined. Compared with the prior art, the related technical personnel need not analyze the network performance of OpenStack through OpenStack management tools and plug-ins to obtain the network performance analysis report. In the present application, the technical personnel do not need to analyze the network data one by one, which saves the labor cost and improves the efficiency of network performance measurement.

[0077] In combination Figure 3 As shown in Figure 5 The target large language model can be constructed through an electronic device, as shown in Figure 5 The target large language model is trained in the following way:

[0078] S501, the network performance testing device obtains a training set.

[0079] The training set includes historical network performance parameters of multiple historical periods and actual network performance analysis reports corresponding to the historical network performance parameters.

[0080] Optionally, the training set further includes device information, historical test data, traffic scenario data, and fault scenario data. The traffic scenario data includes normal traffic, burst traffic, and abnormal traffic. The fault scenario data includes link jitter, packet loss, and delay.

[0081] It can be understood that the training set data is the original data used for training the model, and through the training set data, the large language model learns the characteristics and rules of the data.

[0082] In an implementable manner, the target large language model can be constructed by an electronic device. The electronic device needs to train an initial prediction model based on the training set, and continuously adjust the model parameters of the initial prediction model according to the initial prediction result and the training set data to obtain the target large language model. Therefore, the electronic device needs to obtain the training set including the historical network performance parameters of multiple historical periods and the actual network performance analysis report corresponding to the historical network performance parameters.

[0083] In a possible implementation manner, the network performance testing apparatus compiles a python script, collects network data through the python script, and automatically cleans the network performance data through the first script.

[0084] Optionally, the network data includes at least one of the following: bandwidth, throughput, CPU utilization, memory utilization, bandwidth utilization, time delay, jitter, packet loss, network congestion, device performance, signal interference, application load, and the like.

[0085] Optionally, the data format after cleaning is

[0086]

[0087]

[0088] It needs to be explained that the network performance testing apparatus trains the first script to distinguish the source of the input data through the user id and the assistant id, from: human is used to represent the user, value is used to represent the user's question content, from: assistant is used to represent the target language model, and value is used to represent the reply content of the target language model.

[0089] S502, the network performance testing apparatus inputs the historical network performance parameters in the training set into the initial prediction model to determine an initial prediction result.

[0090] In an implementable manner, the electronic device can input the service information in the training set into the initial prediction model and determine an initial prediction result, so that the electronic device adjusts the model parameters of the initial prediction model based on the initial prediction result and the actual network performance analysis report in the training set to obtain the target large language model.

[0091] Optionally, the initial prediction model includes baichuan2-13b-chat.

[0092] S503, the network performance testing device determines a loss value of the initial prediction model based on the initial prediction result and the actual network performance analysis report, and adjusts the model parameters of the initial prediction model based on the loss value in a case where the loss value does not satisfy a preset condition, and continues to perform iterative training based on the training set until a model convergence condition is satisfied, and a target large language model is obtained.

[0093] In an implementable manner, the electronic device adjusts the model parameters of the initial prediction model based on the initial prediction result and the actual network performance analysis report to obtain the target large language model.

[0094] Optionally, the electronic device can construct the target large language model based on a neural network.

[0095] Optionally, the neural network is a computational model that simulates the connection and information processing mode between biological neurons.

[0096] Optionally, the neural network is composed of a large number of interconnected nodes (also known as neurons), which are arranged in a certain hierarchical structure (input layer, hidden layer, output layer). Information is transmitted and processed through the connection between neurons. Each neuron receives input signals from other neurons and calculates the input through an activation function to generate an output signal to pass to subsequent neurons.

[0097] Optionally, the electronic device can construct a loss function according to a plurality of loss values.

[0098] Optionally, the loss function includes at least one of a mean square error function, a mean absolute error function, a cross-entropy loss function, a binary cross-entropy loss function, a hinge loss function, a log-likelihood loss function, and a Kullback-Leibler (KL) divergence function.

[0099] Optionally, as shown in Figure 6 The network performance testing device downloads the initial prediction model and prepares a training script. The network performance testing device collects and cleans the network performance data, and then trains the initial prediction model based on the cleaned network performance data to obtain a target large language model. The network performance testing device uses the target language model to answer the performance data, and if the answer result does not meet the demand, the initial prediction model is retrained based on the cleaned network performance data, and if the answer result meets the demand, the performance indicators of the OpenStack instance are collected and formatted, the performance indicators of the OpenStack instance are analyzed using the target language model, a performance analysis report and icon are generated, and an email is sent.

[0100] It needs to be explained that the network performance testing device uses the lora method to fine-tune the target language model, and the main parameters of the training are:

[0101]

[0102] Among them, the Lora training principle: LoRA uses the data corresponding to the downstream task and adapts to the downstream task only by training some newly added parameters.

[0103] It is understandable that after training the new parameters, the network performance test device uses a re-parameterization method to merge the new parameters with the old model parameters. This can achieve the effect of fine-tuning the entire model for the new task without increasing the inference time during inference, thus reducing the consumption of server resources.

[0104] Optionally, the network performance test device uses the trained large model fine_tune_baichuan2 to perform performance analysis on the test set data. The test set data content is as follows:

[0105]

[0106]

[0107] It's important to explain that the network performance tester checks the analysis results. If they meet the requirements, it collects OpenStack instance performance metrics. If not, it optimizes the training data, adjusts the learning rate and number of learning cycles, and continues training the large model until the trained large model meets user requirements. The target language model is manually trained through repeated question-and-answer sessions, exposing the model to questions of varying difficulty. Based on user feedback, the model learns more complex logical relationships and decision-making strategies, further improving the accuracy of the target language model.

[0108] The above scheme brings at least the following beneficial effects: In the embodiment of the present application, the network performance testing device trains the initial prediction model. During the training process of the initial prediction model, the efficiency of obtaining the target language model can be improved by adjusting the number of training times and the learning rate of the initial prediction model.

[0109] In one possible implementation, combining Figure 3 ,like Figure 7 As shown, before the network performance test apparatus obtains the network performance parameters of the first network instance in S303, the network performance test apparatus instructs the second network instance to send traffic to the first network instance. The process of the network performance test apparatus instructing the second network instance to send traffic to the first network instance can be specifically implemented by the following S701.

[0110] S701: The network performance testing apparatus instructs the iperf tool in the second network instance to send traffic of a preset bandwidth to the first network instance.

[0111] The iperf tool is installed in the second network instance through a script.

[0112] In a possible implementation, the network performance testing device installs an iperf tool in the second network instance by running a script, and instructs the iperf tool in the second network instance to send traffic of a preset bandwidth to the first network instance.

[0113] It is understandable that the network performance testing device determines whether the iperf tool is installed in the second network instance, and if the iperf tool does not exist in the second network instance, installs the iperf tool in the second network instance through a script.

[0114] It should be explained that iperf is a client / server-based network performance testing tool that can test the bandwidth quality of the Transmission Control Protocol (TCP), User Datagram Protocol (UDP) and Stream Control Transmission Protocol (SCTP). It can provide network throughput information, as well as statistical information such as jitter, packet loss rate, maximum segment and maximum transmission unit size, thereby testing network performance.

[0115] For example, Figure 8 As shown, the network performance testing device is connected to the first network instance, determines whether the first network instance has the iperf tool and the sar command, and if the first network instance does not have the iperf tool and the sar command, installs the iperf tool and the sar command on the first network instance. If the first network instance has the iperf tool and the sar command, starts the iperf server in the first network instance. The network performance testing device obtains the second network instance connected to the first network instance, searches the second network instance, and determines whether the second network instance has the iperf tool and the sar command. If the second network instance does not have the iperf tool and the sar command, installs the iperf tool and the sar command on the second network instance. If the second network instance has the iperf tool and the sar command, starts the iperf client and the sar command in the second network instance, and collects performance indicators of the second network instance. The network performance testing device instructs the iperf client to send traffic to the iperf client server.

[0116] The above scheme brings at least the following beneficial effects: In an embodiment of the present application, the network performance testing device instructs the iperf tool in the second network instance to send traffic of a preset bandwidth to the first network instance. In this way, the network performance of the network to which the second network instance and the first network instance belong can be determined based on the data sent by the iperf tool in the second network instance and the data received by the iperf tool in the first network instance.

[0117] In one possible implementation, combining Figure 3 ,like Figure 9 As shown, in S303, the process in which the network performance testing apparatus obtains the network performance parameters of the first network instance can be specifically implemented through the following S901-S903.

[0118] S901. A network performance testing apparatus obtains network quality parameters received by a first network instance through an iperf tool in the first network instance.

[0119] The network quality parameters include at least one of the following: throughput, jitter, packet loss rate, and maximum transmission unit.

[0120] Optionally, the iperf tool includes an iperf client and an iperf server. The iperf client is used to send traffic, and the iperf server is used to receive traffic.

[0121] In one possible implementation, the network performance testing device sends a first request message to the iperf server in the first network instance, receives a first request response message from the iperf server in the first network instance, and obtains the network quality parameters received by the first network instance based on the first request response message.

[0122] Optionally, the first request information is used to request traffic information, and the first request response information includes traffic information.

[0123] It should be explained that after the iperf server in the first network instance receives the traffic information, it will save the traffic information to the log file. When the iperf server in the first network instance receives the first request information, it generates the first request response information based on the traffic information in the log file.

[0124] Exemplarily, after receiving the instruction information from the network performance test device, the iperf client in the second network instance sends 1M, 10M, 100M, and 1G of traffic to the iperf server in the first network instance. The iperf server records the network performance parameters corresponding to the 1M, 10M, 100M, and 1G of traffic, and sends the network performance parameters to the network performance test device.

[0125] S902, the network performance testing device obtains device state information of the first network instance through a sar command in the first network instance.

[0126] The device state information includes at least one of CPU occupation information, memory occupation information, and disk input / output (I / O) information.

[0127] It should be explained that sar can collect performance data through a system activity data collector (SADC). The SADC is a background process that periodically collects system performance data and writes it to system log files. These data include CPU usage, memory usage, disk I / O, network activity, etc. The SADC can collect data according to the configured time interval and iteration number, thereby providing real-time or near real-time system performance information.

[0128] S903, the network performance testing device determines the network performance parameter of the first network instance according to the network quality parameter received by the first network instance and the device state information of the first network instance.

[0129] In a possible implementation, the network performance testing device obtains the network quality parameter sent by the second network instance, and determines the network performance parameter of the first network instance according to the network quality parameter received by the first network instance, the device state information of the first network instance, and the network quality parameter sent by the second network instance.

[0130] In another possible implementation, after the network performance testing device determines the network performance parameter of the first network instance, the agent tool is invoked to convert the generated test case into an excel file and send an email.

[0131] It should be explained that the agent tool combines the characteristics of large language models, planning capabilities, memory, and tool usage, and can autonomously understand, plan decisions, and perform complex tasks.

[0132] The above scheme at least brings the following beneficial effects: in the embodiment of the application, the network performance testing device can determine the communication quality between the first network instance and the second network instance according to the network quality parameter received by the first network instance and the network quality parameter sent by the second network instance. Further, the network performance testing device can accurately determine the network performance parameter of the first network instance based on the communication quality between the first network instance and the second network instance and the device state information of the first network instance.

[0133] In a possible implementation, the network performance testing device obtains the network quality parameter sent by the second network instance, and determines the network performance parameter of the first network instance according to the network quality parameter received by the first network instance, the device state information of the first network instance, and the network quality parameter sent by the second network instance. Figure 3 For example, Figure 10As shown, the process in which the network performance testing device determines the network performance analysis report of the target network based on the network performance parameters of the first network instance and the target large language model in S304 can be implemented through S1001-S1002.

[0134] S1001, the network performance testing device obtains the device state information of the second network instance through the sar command in the second network instance.

[0135] In a possible implementation manner, the network performance testing device sends second request information to the sar command, receives second request response information from the sar command, and obtains the device state information of the second network instance based on the second request response information.

[0136] Optionally, the second request response information is used to obtain the device state information of the second network instance, and the second request response information includes the device state information of the second network instance.

[0137] It should be explained that the sar command is a very comprehensive system performance analysis tool under Linux and an important tool in UNIX system. The sar command is used to collect, report and save system activity information, and helps users understand the performance of the system.

[0138] S1002, the network performance testing device inputs the network performance parameters of the first network instance and the device state information of the second network instance into the target large language model to obtain the network performance analysis report of the target network.

[0139] In a possible implementation manner, the network performance testing device constructs a prompt word template, determines the first performance test index and the first content format of the output. The network performance testing device obtains input information meeting the performance test requirements from the network performance parameters of the first network instance and the device state information of the second network instance based on the prompt word template, the first performance test index, and inputs the input information into the target large language model to obtain the network performance analysis report of the target network.

[0140] Optionally, the format of the network performance analysis report is the first content format.

[0141] Illustratively, the network performance testing device determines the bandwidth performance according to the target ratio of the actual transmission bandwidth and the theoretical bandwidth, and determines that the network condition is good when the target ratio is less than or equal to 0.15, and determines that the network condition is poor when the target ratio is greater than 0.15. The target large language model provides possible influencing factors as the direction of adjusting the network condition.

[0142] Illustratively, as Figure 11As shown, the network performance testing device constructs a prompt word template, specifies a performance testing index and a report format, uploads a performance index file, and judges whether the uploading of the performance index file is successful. In the case that the uploading of the performance index file fails, the performance index file is re-uploaded, and in the case that the uploading of the performance index file succeeds, the network performance testing device automatically reads data in the performance index file and stores the performance data into a variable. The network performance testing device analyzes the performance index according to the prompt word template, generates an analysis report, and stores the analysis report into an output folder. The network performance testing device customizes a tool and automatically sends a report email.

[0143] It needs to be explained that the network performance testing device uses an agent, customizes a report content tool, saves generated charts and texts into a csv file, customizes a sending email tool, sets sender information, recipient information and sent texts, and automatically sends a performance analysis report by email.

[0144] The above scheme at least has the following beneficial effects: in the embodiment of the application, the network performance testing device inputs the network performance parameter of the first network instance and the device state information of the second network instance into the target large language model. In this way, the network performance testing device can determine the communication quality between the first network instance and the second network instance. Further, the network performance testing device can accurately determine the network performance parameter of the first network instance based on the communication quality between the first network instance and the second network instance and the device state information of the first network instance and the device state information of the second network instance.

[0145] The embodiment of the application can divide the network performance testing device into functional modules or functional units according to the above method examples, for example, each functional module or functional unit can be divided according to each function, or two or more functions can be integrated into one processing module. The above integrated module can be realized in the form of hardware or in the form of a software functional module or functional unit. In the embodiment of the application, the division of the module or unit is illustrative, and is only a logical functional division. In actual implementation, there can be another division mode.

[0146] Exemplarily, as shown in Figure 12 Fig. 1 is a possible structural schematic diagram of a network performance testing device according to an embodiment of the application. The network performance testing device 120 includes an acquisition unit 1201 and a processing unit 1202.

[0147] The acquisition unit 1201 is configured to acquire a topology graph of a target network; the topology graph comprises network instances and connection relationships between the network instances; the processing unit 1202 is configured to determine, according to the topology graph, a first network instance and a second network instance connected to the first network instance; the first network instance and the second network instance belong to the target network; the first network instance is any network instance in the target network; the processing unit 1202 is further configured to instruct the acquisition unit 1201 to acquire a network performance parameter of the first network instance in a case where the second network instance sends traffic of a preset bandwidth to the first network instance; the processing unit 1202 is further configured to determine, based on the network performance parameter of the first network instance and a target large language model, a network performance analysis report of the target network; the target large language model is used to predict network performance according to the network performance parameter and output an analysis report of the network performance.

[0148] Optionally, the acquisition unit 1201 is further configured to acquire a training set; the training set comprises historical network performance parameters of a plurality of historical periods and actual network performance analysis reports corresponding to the historical network performance parameters; the processing unit 1202 is further configured to input the historical network performance parameters in the training set into an initial prediction model to determine an initial prediction result; the processing unit 1202 is further configured to determine a loss value of the initial prediction model based on the initial prediction result and the actual network performance analysis report, and adjust model parameters of the initial prediction model based on the loss value in a case where the loss value does not satisfy a preset condition, and continue iterative training based on the training set until a model convergence condition is satisfied, to obtain the target large language model.

[0149] Optionally, the processing unit 1202 is further configured to instruct an iperf tool in the second network instance to send traffic of a preset bandwidth to the first network instance; the iperf tool is installed in the second network instance through a script.

[0150] Optionally, the acquisition unit 1201 is further configured to acquire, through an iperf tool in the first network instance, a network quality parameter received by the first network instance; the acquisition unit 1201 is further configured to acquire, through a sar command in the first network instance, device state information of the first network instance; the processing unit 1202 is further configured to determine the network performance parameter of the first network instance according to the network quality parameter received by the first network instance and the device state information of the first network instance.

[0151] Optionally, the network quality parameter comprises at least one of the following: throughput, jitter, packet loss rate, and maximum transmission unit; the device state information comprises at least one of the following: CPU occupation information, memory occupation information, and disk input / output (I / O) information.

[0152] Optionally, the acquisition unit 1201 is further configured to acquire device state information of the second network instance through a sar command in the second network instance; and the processing unit 1202 is further configured to input the network performance parameter of the first network instance and the device state information of the second network instance into the target large language model to obtain the network performance analysis report of the target network.

[0153] Optionally, the network performance parameter comprises at least one of the following: expected bandwidth, actual sending bandwidth, actual receiving bandwidth, jitter, packet loss rate, CPU utilization, and memory utilization.

[0154] The embodiment of the application further provides a chip, which comprises a processor and a communication interface, the communication interface is coupled with the processor, and the processor is configured to run a computer program or instructions to implement the network performance testing method in the above method embodiment.

[0155] The embodiment of the application provides a computer program product comprising instructions which, when executed on a computer, cause the computer to perform the switching method of the industrial terminal in the above method embodiment.

[0156] The computer readable storage medium, for example, can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or apparatus, or any combination of the above. More specific examples (a non-exhaustive list) of the computer readable storage medium include an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a register, a hard disk, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any other suitable combination of the above, or any other form of computer readable storage medium. An exemplary storage medium is coupled to the processor such that the processor can read information from, and write information to, the storage medium. Of course, the storage medium can be a component of the processor. The processor and the storage medium can be located in an application-specific integrated circuit (ASIC). In the embodiments of the application, the computer readable storage medium can be any tangible medium that contains or stores a program, which can be used by or in connection with an instruction execution system, apparatus, or device.

[0157] Since the device, the equipment, the computer readable storage medium, the computer program product in the embodiment of the application can be applied to the above method, the technical effects that can be obtained can also be referred to the above method embodiment, and the embodiment of the application will not be repeated here.

[0158] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, any change or replacement within the technical scope disclosed in the present application should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A network performance testing method, characterized in that: The method comprises: Obtaining a topology map of the target network; the topology map includes network instances and connection relationships between the network instances; Determine, according to the topology map, a first network instance and a second network instance connected to the first network instance; the first network instance and the second network instance belong to the target network; and the first network instance is any network instance in the target network; When the second network instance sends traffic of a preset bandwidth to the first network instance, obtaining network performance parameters of the first network instance; Based on the network performance parameters of the first network instance and a target large language model, a network performance analysis report of the target network is determined; the target large language model is used to predict network performance according to the network performance parameters and output an analysis report of the network performance.

2. The method according to claim 1, characterized in that The method further comprises: Acquire a training set; the training set includes: historical network performance parameters of multiple historical periods and actual network performance analysis reports corresponding to the historical network performance parameters; Inputting the historical network performance parameters in the training set into an initial prediction model to determine an initial prediction result; Based on the initial prediction result and the actual network performance analysis report, a loss value of the initial prediction model is determined. If the loss value does not meet a preset condition, the model parameters of the initial prediction model are adjusted based on the loss value, and iterative training is continued based on the training set until a model convergence condition is met, thereby obtaining the target large language model.

3. The method according to claim 1, characterized in that Before obtaining the network performance parameters of the first network instance, the method further includes: Instructing the iperf tool in the second network instance to send traffic of the preset bandwidth to the first network instance; the iperf tool is installed in the second network instance through a script.

4. The method according to claim 3, characterized in that The obtaining of the network performance parameters of the first network instance includes: Obtaining, by using the iperf tool in the first network instance, network quality parameters received by the first network instance; Obtaining device status information of the first network instance through the sar command in the first network instance; Determine a network performance parameter of the first network instance according to the network quality parameter received by the first network instance and the device status information of the first network instance.

5. The method according to claim 4, characterized in that The network quality parameters include at least one of the following: throughput, jitter, packet loss rate, and maximum transmission unit; the device status information includes at least one of the following: CPU occupancy information, memory occupancy information, and disk input / output (I / O) information.

6. The method according to claim 1, characterized in that The determining, based on the network performance parameters of the first network instance and the target large language model, a network performance analysis report of the target network includes: Obtaining device status information of the second network instance by using the sar command in the second network instance; The network performance parameters of the first network instance and the device status information of the second network instance are input into the target large language model to obtain a network performance analysis report of the target network.

7. The method according to any one of claims 1 to 6, characterized in that The network performance parameter includes at least one of the following: expected bandwidth, actual sending bandwidth, actual receiving bandwidth, jitter, packet loss rate, CPU utilization, and memory utilization.

8. A network performance testing device, characterized in that: include: Acquisition unit and processing unit; The acquisition unit is used to acquire a topology map of the target network; The topology diagram includes network instances and connection relationships between the network instances; The processing unit is configured to determine, based on the topology map, a first network instance and a second network instance connected to the first network instance; the first network instance and the second network instance belong to the target network; and the first network instance is any network instance in the target network; The processing unit is configured to obtain the network performance parameters of the first network instance when the second network instance sends traffic of a preset bandwidth to the first network instance; The processing unit is used to determine a network performance analysis report of the target network based on the network performance parameters of the first network instance and a target large language model; the target large language model is used to predict network performance according to the network performance parameters and output an analysis report of the network performance.

9. A network performance testing device, characterized in that: include: A processor and a memory; wherein the memory is used to store computer-executable instructions, and when the network performance testing device is running, the processor executes the computer-executable instructions stored in the memory, so that the network performance testing device executes the network performance testing method described in any one of claims 1-6.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium includes instructions, which, when executed by a network performance testing device, enable the computer to perform the network performance testing method according to any one of claims 1 to 6.

Images