Network analysis service processing method, device, equipment, medium and program product

Abstract

The embodiment of the invention discloses a network analysis service processing method and device, equipment, a medium and a program product, and the method comprises the steps: firstly obtaining to-be-analyzed service data, inputting the to-be-analyzed service data into a service grade branch in a service quality model, the data specification processed by the service level branches corresponds to the service level of the to-be-analyzed service data, and the service quality model comprises a plurality of service level branches, and then obtaining a model analysis result of the service quality model according to the output data of the service level branches, the model analysis result is used for evaluating the quality of the business corresponding to the business data. According to the embodiment of the invention, the related data in the network analysis service process can be analyzed, so that the problems in the network analysis service process can be found and solved in time.

Description

Technical Field

[0001] The present disclosure relates to network resolution service technology and Internet technology, and in particular to a processing method, device, equipment, medium and program product for network resolution service. Background Art

[0002] Network resolution services are used to help users achieve normal Internet functions. For example, network domain name recursive resolution services are an important part of Internet infrastructure, allowing users to access websites or applications through domain names. However, in some cases, network resolution services such as domain name recursive resolution services may have problems, resulting in users being unable to use the Internet normally. Therefore, it is necessary to process the relevant information or data of network resolution services in order to promptly discover and resolve problems in the process of network resolution services. Summary of the invention

[0003] Therefore, the embodiments of the present disclosure provide a method, apparatus, device, medium and program product for processing a network resolution service, which can analyze relevant data in the process of the network resolution service so as to timely discover and solve problems existing in the process of the network resolution service.

[0004] In a first aspect, an embodiment of the present disclosure provides a method for processing a network resolution service, including:

[0005] Obtain business data to be analyzed;

[0006] Inputting the service data to be analyzed into a service level branch in a service quality model; the data specification processed by the service level branch corresponds to the service level of the service data to be analyzed; the service quality model includes multiple service level branches;

[0007] Obtaining a model analysis result of the service quality model according to the output data of the service level branch; the model analysis result is used to evaluate the quality of the service corresponding to the service data;

[0008] The service quality model is obtained by the following steps:

[0009] Obtain sample business data of domain name recursive resolution service;

[0010] Clustering the sample business data to obtain the label category to which the sample business data belongs;

[0011] According to the sample business data of the label category, a model training sample is obtained; the model training sample includes a plurality of the label categories, and the model training sample of each label category includes the sample business data corresponding to a plurality of evaluation labels; the plurality of evaluation labels are pre-set labels for the quality of the network resolution service;

[0012] According to the model training samples, a converged service quality model is obtained;

[0013] Updating the sample business data to obtain updated model training samples;

[0014] The updated model training samples are input into the service quality model to be trained and the converged service quality model, and the converged service quality model is retrained according to the output results to obtain a trained service quality model.

[0015] In a second aspect, an embodiment of the present disclosure provides a processing device for a network resolution service, including:

[0016] A first acquisition module is used to obtain business data to be analyzed;

[0017] A first input module is used to input the service data to be analyzed into a service level branch in a service quality model; the data specification processed by the service level branch corresponds to the service level of the service data to be analyzed; the service quality model includes multiple service level branches;

[0018] A first output module, configured to obtain a model analysis result of the service quality model according to the output data of the service level branch; the model analysis result is used to evaluate the quality of the service corresponding to the service data;

[0019] The second acquisition module is used to obtain sample business data of the domain name recursive resolution service;

[0020] A clustering module, used to cluster the sample business data to obtain the label category to which the sample business data belongs;

[0021] A third acquisition module is used to obtain a model training sample according to the sample business data of the label category; the model training sample includes a plurality of the label categories, and the model training sample of each label category includes the sample business data corresponding to a plurality of evaluation labels; the plurality of evaluation labels are pre-set labels for the quality of the network resolution service;

[0022] A fourth acquisition module is used to obtain a converged service quality model according to the model training samples;

[0023] An updating module, used to update the sample business data to obtain updated model training samples;

[0024] The second input module is used to input the updated model training samples into the service quality model to be trained and the converged service quality model, and retrain the converged service quality model according to the output results to obtain a trained service quality model.

[0025] In a third aspect, an embodiment of the present disclosure provides an electronic device, including:

[0026] a memory for storing a computer program product;

[0027] The processor is used to execute the computer program product stored in the memory, and when the computer program product is executed, the method provided by any embodiment of the present disclosure is implemented.

[0028] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium having computer program instructions stored thereon. When the computer program instructions are executed by a processor, the method provided by any embodiment of the present disclosure is implemented.

[0029] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, including computer program instructions, which, when executed by a processor, implement the method provided by any embodiment of the present disclosure.

[0030] Through the method provided by the embodiment of the present disclosure, the business data generated during the network resolution service can be analyzed to obtain the quality of the business, thereby enabling the quality of the network business to be evaluated, problems and deficiencies in the process of carrying out the network business to be discovered in a timely manner, and the quality of the network business to be improved in a targeted manner.

[0031] The technical solution of the present disclosure is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings, which constitute a part of the specification, illustrate embodiments of the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0033] The present disclosure may be more clearly understood from the following detailed description with reference to the accompanying drawings, in which:

[0034] Figure 1 A flowchart of an embodiment of the method disclosed herein;

[0035] Figure 2 A schematic diagram of a method according to another embodiment of the present disclosure;

[0036] Figure 3 A schematic diagram of service levels in an example of the present disclosure;

[0037] Figure 4 A schematic diagram of intelligent analysis in an example of the present disclosure;

[0038] Figure 5 A schematic diagram of forwarding dependency in an example of the present disclosure;

[0039] Figure 6 A flowchart of a method in an example of the present disclosure;

[0040] Figure 7 A schematic diagram of a service quality model in an example of the present disclosure;

[0041] Figure 8 A schematic diagram of a method implementation framework in an example of the present disclosure;

[0042] Fig. 9 for Figure 8 The schematic diagram of the detection terminal layer shown in;

[0043] Fig.10 for Figure 8 The schematic diagram of the architecture topology shown in;

[0044] Fig.11 A schematic diagram of a method flow in another example of the present disclosure;

[0045] Fig.12 A schematic diagram of a detection management center in an example of the present disclosure;

[0046] Fig.13 A schematic diagram of an electronic device in an example of the present disclosure. DETAILED DESCRIPTION

[0047] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present disclosure more clearly understood, the present disclosure is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present disclosure and are not used to limit the present disclosure.

[0048] In the following description, specific details such as specific system structures and technologies are provided to provide a thorough understanding of the embodiments of the present disclosure. However, it should be clear to those skilled in the art that the present disclosure may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obstructing the description of the present disclosure with unnecessary details.

[0049] It should be understood that when used in this disclosure and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0050] It should also be understood that the term “and / or” used in this disclosure and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0051] As used in this disclosure and the appended claims, the term "if" may be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrases "if it is determined" or "if [described condition or event] is detected" may be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0052] In addition, in the description of the present disclosure and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0053] References to "one embodiment" or "some embodiments" etc. described in this disclosure mean that one or more embodiments of the disclosure include specific features, structures or characteristics described in conjunction with the embodiment. Thus, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0054] Figure 1 A flow chart of a method for processing a network resolution service according to an embodiment of the present disclosure is shown, including: the following steps S11 to S13.

[0055] Step S11: Obtain business data to be analyzed.

[0056] In the disclosed embodiments, the business data to be analyzed may be the original data generated during the implementation of the network resolution service. Alternatively, the business data to be analyzed may be obtained based on the original data generated during the implementation of the network resolution service. For example, the original data generated during the implementation of the network resolution service may be screened and cleaned to obtain the business data to be analyzed.

[0057] Accordingly, the business data to be analyzed includes information about the network resolution service that needs to be processed. For example, when a user uses the network resolution service, business data is generated, and this part of the business data is used as the business data to be analyzed. The business data to be analyzed is used to process the network resolution service in order to obtain an evaluation result of the network resolution service. Further, according to the evaluation result, the problem finding and analysis warning of the network resolution service can be realized.

[0058] In the embodiments of the present disclosure, business data may refer to data generated during the business process. The business data to be analyzed may be data generated during one business process, or may be data generated during multiple business processes within a set time period.

[0059] In the disclosed embodiment, the service data to be analyzed can also be referred to as service data. In the training phase of the service quality model, the service data is also sample service data. After the service quality model is trained, the sample service data is also service data to be analyzed (or service data).

[0060] Step S12: inputting the service data to be analyzed into the service level branch in the service quality model; the data specification processed by the service level branch corresponds to the service level of the service data to be analyzed; the service quality model includes multiple service level branches.

[0061] In a possible implementation, inputting the business data to be analyzed into the service level branch in the business quality model may include: dividing the business data to be analyzed according to service levels, determining the target service level to which the business data to be analyzed belongs; and inputting the business data to be analyzed into the service level branch corresponding to the target service level.

[0062] In another possible implementation, inputting the service data to be analyzed into the service level branch in the service quality model may include: inputting the service data to be analyzed into an input module of the service quality model, and the input module of the service quality model inputting the service data to be analyzed into a corresponding branch.

[0063] The service level branch in the service quality model can be used to process the business data to be processed of the corresponding scale. The data specifications processed by the service level branch and the service level of the business data to be analyzed can be used to describe the scale of the business corresponding to the business data.

[0064] The business quality model is used to analyze the business data to be analyzed, analyze the business completion effect based on the business data to be analyzed, and predict (or evaluate) the quality of the business.

[0065] Step S13: obtaining a model analysis result of the service quality model according to the output data of the service level branch; the model analysis result is used to evaluate the quality of the service corresponding to the service data.

[0066] In the embodiment of the present disclosure, step S13 may include: obtaining output data of the service level branch, and outputting the output data of the service level branch as a model analysis result of the service quality model. Alternatively, step S13 may also include: further processing the output data of the service level branch to obtain a model analysis result of the service quality model.

[0067] In the embodiment of the present disclosure, the service quality model is obtained through the following steps S14 to S19.

[0068] Step S14: Obtain sample business data of the domain name recursive resolution service.

[0069] In the disclosed embodiment, the sample service data may belong to multiple target service levels, and the multiple target service levels may correspond to all service level branches of the service quality model. That is, the data volume of the sample service data is greater than the service data to be processed.

[0070] Recursive domain name resolution means that when performing domain name resolution, the DNS (Domain Name System) server will recursively query multiple DNS servers until the IP address of the target domain name is found. This process involves the collaborative work of multiple DNS servers, with the local DNS server (also called the recursive domain name resolution server) acting as an intermediary, responsible for communicating and querying with other DNS servers.

[0071] The method of acquiring the sample business data may be similar to the above-mentioned business data to be analyzed.

[0072] Step S15: clustering the sample business data to obtain the label category to which the sample business data belongs.

[0073] Clustering the sample business data may be performed by using a certain clustering method and clustering the sample business data according to a pre-set label category. Exemplarily, KMeans (k-means clustering algorithm) and / or DBSCAN algorithm (Density-Based Spatial Clustering of Applications with Noise) may be used to cluster the sample business data. Label categories may be pre-selected to be composed of evaluation labels with a moderate number of categories and small differences in data within each category. Each label category may be scored for quality based on empirical information, and the score may be used as a sample reference answer, and the sample reference answer may be used as part of the first round of supervision data.

[0074] The label category to which the sample business data belongs may refer to the aspect (or dimension) of the business quality involved in the sample business data (or business data). For example, a sample business data includes information related to network security, which can reflect the quality of the business related to network security. Therefore, the label category to which the sample business data belongs may be network security.

[0075] In step S15, for a large amount of data in the sample business data, each piece of data may be classified according to the label category to determine the sample business data included in each label category.

[0076] Step S16: Obtain model training samples based on the sample business data of the label category; the model training samples include multiple label categories, and the model training samples of each label category include the sample business data corresponding to multiple evaluation labels; the multiple evaluation labels are pre-set labels for recursive domain name resolution business record data.

[0077] When initially training the quality analysis model to be trained, the model training samples may be equivalent to the first round of supervision data mentioned above. Use the training data and the first round of supervision data to train the service quality model to be trained, and when the loss function of the model converges, the obtained model is the converged service quality model. In the disclosed embodiment, the service quality model may also be referred to as a domain name recursive resolution service quality model or a rating analysis model. In one possible implementation, the service quality model may use a linear regression method to process input data, and therefore may also be referred to as a linear regression model.

[0078] After determining the label category to which the sample business data belongs, the evaluation label of each sample business data can be determined, and the evaluation label can be added to the sample business data. The model training sample can include the sample business data with the determined label category and the added evaluation label. In other words, the model training sample can include the label category and the sample business data. In addition, when the service quality model to be trained is optimized for multiple rounds, the service quality model to be trained also adjusts the internal parameters of the service quality model according to the sample reference answer.

[0079] In a possible implementation, each tag category includes multiple evaluation tags, each evaluation tag being a sub-dimension of the service quality involved in the sample service data (or service data) under the evaluation category. For example, under the dimension of network security, there are multiple sub-dimensions for evaluating the security of different aspects of the network.

[0080] Step S17: Obtain a converged service quality model based on the model training samples.

[0081] After obtaining the model training samples, the model training samples can be input into the service quality model to be trained, and the service quality model to be trained can obtain the training phase output results of the model training samples, calculate the loss value according to the training phase output results and the sample reference answer and the loss function, and then adjust the parameters of the service quality model to be trained according to the loss value. After so many rounds of parameter adjustment, the loss value calculated by the loss function tends to be minimized.

[0082] In the disclosed embodiments, the converged service quality model may refer to the service quality model obtained when the loss function converges. The service quality model to be trained may refer to the service quality model at any model training stage before the loss function converges, or the service quality model to be trained may refer to the service quality model when the loss function does not converge. Alternatively, the service quality model to be trained may refer to an untrained service quality model.

[0083] Step S18: Update the sample business data to obtain updated model training samples.

[0084] In a possible implementation, updating the sample business data to obtain updated model training samples may include: adding new sample business data to the original sample business data to obtain updated model training samples; performing clustering processing on the updated model training samples to obtain updated model training samples. The method of clustering the updated model training samples may be similar to the method of processing the sample business data in step S15 and step S16.

[0085] Step S19: input the updated model training samples into the service quality model to be trained and the converged service quality model, and retrain the converged service quality model according to the output results to obtain a trained service quality model.

[0086] In a possible implementation, after obtaining the updated model training sample, in step S19, the updated model training sample can be input into the service quality model to be trained, that is, the service quality model to be trained before step S17, to obtain a first output result; and the updated model training sample can be input into the converged service quality model to obtain a second output result. Then, according to the first output result and the second output result, it is judged whether the converged service quality model meets the expected conditions. If it is judged not, the converged service quality model is optimized again according to the first output result and the second output result, and then the process returns to step S18, and step S18 is re-executed to re-judge whether the new first output result and the second output result meet the expected conditions, until the expected conditions are finally met, and the trained service quality model is obtained.

[0087] In a possible implementation, step S19 may further include the following steps S191 to S193.

[0088] Step S191: input the updated model training sample into the service quality model to be trained to obtain a first output result.

[0089] Step S192: input the updated model training samples into the converged service quality model to obtain a second output result.

[0090] Step S193: Based on the difference between the first output result and the second output result, determine whether the difference between the converged service quality model and the service quality model to be trained meets expectations. If so, obtain the trained service quality model; otherwise, return to step S18.

[0091] In step S19, the converged service quality model is used to perform quality assessment on the updated model training samples to obtain the above-mentioned second output result, and the second round of supervision data is obtained according to the second output result, and training is performed again, and so on, to obtain the final service quality assessment model. In this process, quality adjustment instructions can be sent to the converged service quality model at any time; the quality adjustment instructions can be generated by checking the quality of the above-mentioned first output result and / or the second output result and adjusting, and finally the trained service quality model is obtained, which is deployed on the target server side as an available evaluation model for the network resolution service, or sent to the client side, and the above-mentioned steps S11 to S13 are executed.

[0092] In the model training phase, by designing rating indicators and iterative convergence processes, a service quality model that can accurately analyze service quality is formed, and the corresponding indicators of the service scope in the quality indicator system of the network resolution service are designed and modeled. The evaluation phase uses the evaluation model obtained in the training phase, inputs the data item to be evaluated, and obtains the quality score of the data. The service quality model used in the embodiment of the present disclosure can be obtained through weakly supervised learning. In the process of weakly supervised learning, the role of sample reference answers may be weakened, so that the training process of the service quality model is less dependent on the sample reference answers, and the impact of inaccurate sample reference answers on the service quality model can be avoided.

[0093] Through the method provided by the embodiment of the present disclosure, the business data generated during the network resolution service can be analyzed to obtain the quality of the business, thereby enabling the quality of the network business to be evaluated, problems and deficiencies in the process of carrying out the network business to be discovered in a timely manner, and the quality of the network business to be improved in a targeted manner.

[0094] In one embodiment, the method of obtaining a converged service quality model based on model training samples includes: determining the number of sample service data corresponding to each evaluation label for a target label category among the multiple label categories; determining the initial weight of each evaluation label in each service level branch of the service quality model to be trained based on the number of sample service data corresponding to each evaluation label; outputting the training phase model analysis results of the sample service data of the target label category; adjusting the initial weights based on the training phase model analysis results and the sample analysis results to obtain adjusted weights; and obtaining a converged service quality model based on the adjusted weights.

[0095] In the disclosed embodiment, the target tag category may be one of a plurality of tag categories. For the target tag category in the plurality of tag categories, determining the number of sample business data corresponding to each of the evaluation tags may include: for each of the plurality of tag categories, determining the number of sample business data corresponding to the evaluation tags under each tag category one by one.

[0096] By adjusting the initial weights of the evaluation labels in each service level branch of the service quality model to be trained, the model parameters used to process input data inside the model can be changed, thereby optimizing the model.

[0097] In one embodiment, the initial weight of each evaluation tag in each service level branch of the converged service quality model is determined based on the number of sample business data corresponding to each evaluation tag, including: for a single evaluation tag, when the number of sample business data corresponding to the evaluation tag meets the set quantity threshold requirement, the initial weight of the evaluation tag is set according to the default weight value pre-set for the evaluation tag.

[0098] In the disclosed embodiment, the initial weight of the evaluation tag can be used as an internal parameter of the service quality model to be trained when it is not trained. Under each tag type, a default weight value is set corresponding to the evaluation tag. The default weight value is the internal initial parameter of the service quality model to be trained under ideal conditions. Ideally, the evaluation tags under each tag type have sufficient sample business data. In this case, setting the initial weight corresponding to the evaluation tag according to the default weight value can save the initialization time of the service quality model to be trained and improve the efficiency of model optimization.

[0099] In one embodiment, the initial weight of each evaluation tag in each service level branch of the converged service quality model is determined based on the number of sample business data corresponding to each evaluation tag, including: for a single evaluation tag, when the number of sample business data corresponding to the evaluation tag does not meet the set quantity threshold requirement, obtaining a first evaluation tag that meets the set quantity threshold requirement and a second evaluation tag that does not meet the set quantity threshold requirement among the multiple evaluation tags; determining the sum of the default weight values ​​pre-set corresponding to the first evaluation tag and the ratio of the default weight values ​​pre-set corresponding to the second evaluation tag; determining the allocation ratio of the sum on the sample business data corresponding to the second evaluation tag based on the ratio; adjusting the default weight values ​​pre-set corresponding to each evaluation tag based on the allocation ratio to obtain updated weight values ​​corresponding to each evaluation tag; and determining the initial weight of each evaluation tag based on the updated weight value.

[0100] If the quantity of sample business data corresponding to the evaluation label does not meet the set quantity threshold requirement, it can be considered that the sample business data corresponding to the evaluation label is missing.

[0101] When sample business data corresponding to some evaluation labels are missing, the default weight value can be recalculated according to the missing situation. This allows the business quality model to be trained normally when the sample business data does not meet the quantity requirements, thereby improving the utilization rate of the sample business data.

[0102] In one embodiment, determining the proportion of the sum to be allocated to the sample business data corresponding to the second evaluation tag based on the proportion includes: determining the proportion of each second evaluation tag to be allocated to each first evaluation tag based on the product of the proportion of each second evaluation tag and the sum.

[0103] By calculating the allocation ratio of the evaluation labels with missing sample business data, it is possible to train the business quality model to be trained using the existing sample business data when there is no sample business data for some evaluation labels.

[0104] In one embodiment, adjusting the default weight value pre-set for each of the evaluation tags according to the allocation ratio to obtain the updated weight value corresponding to each of the evaluation tags includes: adjusting the default weight value pre-set for each of the second evaluation tags to zero to obtain the updated weight value corresponding to each of the second evaluation tags; for a single first evaluation tag, adding the allocation ratio of each of the second evaluation tags to the first evaluation tag and the default weight value pre-set for the first evaluation tag to obtain the updated weight value corresponding to the first evaluation tag.

[0105] By setting the evaluation label of the missing sample business data to 0 and increasing the weight value of the evaluation label of the remaining non-missing sample business data, the weight value corresponding to each evaluation label can be adjusted to a reasonable range, so that the initial weight value is adapted to the sample business data, thereby making better use of the sample business data in the process of model training.

[0106] In one embodiment, before obtaining the model analysis result of the business quality model based on the output data of the service level branch, it includes: obtaining a sub-module processing result obtained by at least one sub-module in the service level branch processing the business data; each sub-module in the at least one sub-module is used to process the business data in the dimension of at least one label type among the multiple label types; and obtaining the output data based on the sub-module processing result.

[0107] In a possible implementation, different calculation methods may be used within different sub-modules to obtain sub-module processing results.

[0108] For example, in one embodiment, for a part of the submodules, a decision tree algorithm can be used to obtain the submodule output result. The submodule processing result obtained by processing the business data by at least one submodule in the service level branch includes: calculating the score corresponding to each evaluation tag in the submodule according to the weight value corresponding to each evaluation tag in the service level branch and the score corresponding to the business data value of each evaluation tag; calculating the total score corresponding to all the evaluation tags in the submodule according to the score corresponding to each evaluation tag in the submodule, and obtaining the submodule processing result.

[0109] By calculating the total score corresponding to each evaluation label, the quality level of the sample business data under the dimensions corresponding to multiple evaluation labels can be obtained, thereby achieving quality evaluation of the sample business data.

[0110] In one embodiment, obtaining sample business data of a domain name recursive resolution service includes: sending a collection task of sample business data to a business data detection end; the collection task is generated based on an evaluation tag in which sample business data is missing; receiving detection data returned by the detection end for the collection task; determining the consistency of the detection data based on a unified identifier in the detection data; and filtering the detection data based on the consistency to obtain the sample business data.

[0111] Through the detection end, the business data generated during the network resolution service can be efficiently obtained. At the same time, sample business data can be obtained by screening according to the consistency of the detection data, which can effectively reduce the data noise in the detection data and improve the data quality.

[0112] Because the quality requirements and emphasis of domain name recursive resolution services are different at different service levels, for example, the national-level domain name recursive resolution services have higher requirements for service quality, hierarchical dependency and security indicators, the provincial-level domain name recursive resolution services have higher requirements for service quality and hierarchical dependency, and the enterprise-level domain name recursive resolution services need to have basic protocol completeness, service availability, etc. In one embodiment, the service level is determined according to the scope level of the business service object; the service levels include: national level, provincial level and enterprise level.

[0113] The above-mentioned national level may refer to that the business impact scope of the network resolution service is nationwide, and the scope of the domain name recursive resolution service is defined as national recursion, which can be referred to as DG3 level. The above-mentioned provincial level may refer to that the business impact scope of the network resolution service is provincial, and the scope of the domain name recursive resolution service is defined as provincial recursion, which can be referred to as DG2 level. The above-mentioned enterprise level may refer to that the business impact scope of the network resolution service is enterprise. In one example of the present disclosure, the scope of the domain name recursive resolution service that has not been compliantly registered is defined as enterprise-level recursion, which can be referred to as DG1 level. The larger the level number of each of the above service levels, the greater the influence of the domain name recursive resolution service, and the higher the evaluation weight score. Figure 3 As shown, in one example, the evaluation weight of DG3 is [80%-100%], the evaluation weight of DG2 is [60%-80%], and the evaluation weight of DG1 is [40%-60%].

[0114] The weight scores of the national, provincial and enterprise-level indicator evaluations for domain name recursive resolution services are designed in a graded manner. Business data or sample business data of different service levels are processed using different branches within the business quality model, thereby assigning different weight values ​​to data of different service levels and assigning different important information to sample business data of different service levels.

[0115] In one embodiment, the network resolution service includes a domain name recursive resolution service; the business data includes: business log data of the domain name recursive resolution service; the tag categories include: domain name recursive resolution service geographical scope, domain name recursive resolution service quality, recursive hierarchy dependency related and recursive resolution security related.

[0116] At present, the subject of domain name recursive resolution service is complex, difficult to manage, and illegal recursion is rampant. In the traditional classification process of domain name recursive resolution service, whether some network indicators of network resolution service are met plays an important role in the quality evaluation of network resolution service, such as request response time, request response success rate, resolution result accuracy, TTL (time tolive, life cycle) configuration, etc. In the current various evaluation classification standards and specifications, analysis is generally based on the above data indicators, and the analysis dimension is relatively single.

[0117] As an Internet public service, it is difficult to manage and evaluate it through a single resource management method. It is necessary to conduct multi-angle and comprehensive measurements of recursive resolution dependencies, domain name recursive resolution service scope, domain name recursive resolution service quality and security measures. Traditional analysis methods have fewer dimensions and cannot achieve a comprehensive evaluation of domain name recursive resolution services in the network environment. Therefore, it is necessary to redefine the evaluation grading method to adapt to long-term usage needs.

[0118] The disclosed embodiment provides a method for evaluating the quality of network services by sorting out the evaluation contents of network service data classification and service index requirements, analyzing and summarizing the general principles, processes and methods of data classification in the Internet industry, constructing the above-mentioned label types, and implementing a network resolution method using a PDCA (plan, do, check, action) cycle. The method is used to evaluate and analyze the quality of network services. The processing method of the network resolution service provided by the disclosed embodiment converts the processing method of the single closed-loop network resolution service into a long-term continuous model optimization and business data processing, combines the security risks of recursive resolution services that require long-term monitoring and the effectiveness of security protection measures for recursive resolution services with recursive analysis and evaluation, and forms a hierarchical evaluation system for domain name recursive resolution services, ultimately achieving long-term maintenance of the security status of domain name recursive resolution services.

[0119] Because the embodiment of the present disclosure adopts a weak supervision method to train the service quality model, in order to obtain more valuable model training data, the sample service data is clustered according to the label category and the evaluation label. The service quality index of the network resolution service involves many indicators, and the indicators contain both discrete data and continuous data.

[0120] In one example, during the training phase of the service quality model, firstly, based on the batch recursive data obtained by the domain name recursive resolution service detection system, service quality related evaluation tags can be selected to mark sample service data to obtain training data. Before marking the sample service data, Figure 2As shown in the figure, in the training phase, KMeans and DBSCAN algorithms can be used to cluster sample business data, select clustering results with a moderate number of categories and small differences in data within each category, and annotate sample reference answers for each label type. Sample business data, evaluation labels, label types, and sample reference answers are the first round of supervision data. Use the first round of supervision data and other related training data to train the domain name recursive resolution service quality model (linear regression model). After the model converges, as shown in Figure 2 As shown in the figure, in the evaluation phase, the model is used to evaluate the quality of the training data, and the quality evaluation results are used as the second round of supervision data for retraining to obtain the final service quality evaluation model. During this process, it is necessary to manually check the quality of the data generated by the model at any time and follow up adjustments to finally obtain a usable evaluation model.

[0121] The evaluation phase uses the evaluation model obtained in the training phase, inputs the data item to be evaluated, and obtains the quality score of the data.

[0122] Among multiple indicator types, different types of evaluation indicators are independent of each other; evaluation indicators of the same type are not repeated; and a reasonable hierarchical structure is established between evaluation indicator items with dependencies. As an analysis and rating of Internet basic public services, combined with the service model of the domain name recursive resolution service itself, in one example, the label types include domain name recursive resolution service scope (also called domain name recursive resolution service geographical scope), domain name recursive resolution service quality, recursive hierarchical dependency (also called recursive hierarchical dependency correlation) and recursive resolution security (also called recursive resolution security correlation). The valid evaluation labels under each indicator dimension (label type) are as follows.

[0123] Evaluation tags under the domain name recursive resolution service scope may include: node setting IP (Internet Protocol) whitelist, node geographic location, and number of daily active users.

[0124] The evaluation tags under the quality of domain name recursive resolution service may include: response time of user recursive request, service availability SLA (service level ability), user request response success rate, TTL no hijacking, domain name resolution result no hijacking / no clipping, support for intelligent resolution of domain name recursive resolution services at the three major operator levels, DNS protocol support completeness, support for IDN (Internationalized Domain Names) domain name query, support for multiple record rotation and sorting, activation of domain name service privacy leakage protection, and access to DRSS (Domain Name Resolution Supervision Service System).

[0125] Evaluation tags under recursive hierarchical dependencies may include: whether there is replica root access, whether there is forwarding to other recursive DNS services.

[0126] The evaluation tags under recursive resolution security include: whether DNSSEC (Domain Name System Security Extensions) is configured and deployed, DLV (DNSSEC Lookaside Validation), independent public IP only provides TCP (Transmission Control Protocol) and UDP (User Datagram Protocol) port 53, technical means to clear the cache data of the domain name recursive resolution server, Any resolution type security configuration, TXT (Text) resolution type security configuration, whether the resolution entry and exit are separated, whether there is a CNAME (Canonical Name) loop, domain name abuse monitoring, whether to open file tunnel detection, support iterative query Qname mini (queryname mini), and the ability to switch copies. Among them, the monitoring of domain name abuse further includes: pornographic, gambling, and drug domain names, DGA (domain generation algorithm) domain names, and other malicious domain names.

[0127] Finally, some evaluation labels are briefly described as follows.

[0128] Recursive DNS service for intelligent resolution: When receiving a domain name resolution request, traditional DNS resolution does not determine the source of the visitor, but returns a fixed IP address or randomly selects an IP address to the visitor. Intelligent resolution can intelligently determine the source of the visitor (such as the operator to which the visitor's IP belongs) and match an IP address for it, reducing cross-network access and improving resolution efficiency. The intelligent resolution process is as follows: Figure 4 shown.

[0129] Recursive domain name resolution service forwarding dependency: When a recursive domain name resolution server receives a DNS resolution request, if the recursive domain name resolution server does not have the mapping information for the domain name, it will forward the query request to another domain name server. This process is called forwarding dependency. However, in some cases, this forwarding process may loop indefinitely, so when configuring DNS, you need to avoid circular dependencies. The forwarding dependency process is shown below:

[0130] DNSSEC: By adding digital signature information to the data in DNS, the client can determine whether the response data is authoritative and authentic by checking the signature information after receiving the response, providing data source verification and data integrity verification for DNS data, and preventing related attacks against DNS.

[0131] In one example, for each evaluation tag in the above four tag types, the horizontal and vertical span characteristics of the four dimensional indicators can be analyzed by the information weight method, and the business data corresponding to the evaluation tag can be given a reasonable value range and initial weight value in combination with the known domain name recursive resolution service information. When the value of the business data is a continuous value, a value range of the business data corresponds to a default weight value. If the value range of the business data corresponding to the evaluation tag is a discrete value, each discrete value corresponds to a default weight value, and the default weight value can be directly used as the initial weight value or after dynamic adjustment through apportionment to obtain the initial weight value. The score corresponding to each evaluation tag can be a pre-set score, so that when the value of the business data corresponding to the evaluation tag is a continuous value, a value range of the business data corresponds to a pre-set score. When the value of the business data corresponding to the evaluation tag is a discrete value, each discrete value corresponds to a pre-set score.

[0132] When evaluating the domain name recursive resolution services at all levels using a series of evaluation tags provided by the embodiments of the present disclosure, it may be impossible to obtain sample business data of some evaluation tags. In order to more reasonably evaluate the quality of the network resolution services of these recursive domain name resolutions, it is necessary to first dynamically adjust the default weight values ​​when the sample business data of some evaluation tags is missing, and then use the adjusted default weight values ​​to evaluate the remaining sample business data, so that the process of analyzing the input data by the business quality model is more reasonable and accurate.

[0133] In one example, the dynamic adjustment of the parameters within the model is an iterative cycle. When determining the initial weight value of the service quality analysis model to be trained, it is necessary to accurately count all the indicator items involved in each tag type. Figure 5 As shown, the process includes the following steps S51 to S56.

[0134] Step S51: Count the set A of evaluation tags for which no value can be obtained under the tag type.

[0135] The default weight values ​​of the evaluation labels in set A are denoted as a1, a2, ..., am.

[0136] Step S52: Count the indicator item set B that can obtain values ​​within the indicator dimension.

[0137] The default weight values ​​of the evaluation labels in set B are denoted as b1, b2, ..., bn.

[0138] Step S53: Calculate the sum sa of the default weight values ​​of all evaluation tags in set A.

[0139] sa=a1+a2+……+am.

[0140] Step S54: Calculate the ratio brn of the default weight values ​​of all evaluation tags in set B.

[0141] The calculation process is as follows.

[0142] br1=b1 / (b1+b2+...+bn); br2=b2 / (b1+b2+...+bn);...brn=bn / (b1+b2+...+bn).

[0143] Step S55: multiply the sum sa and each proportion brn to obtain the apportionment weight bsn of each indicator item in the set B.

[0144] The calculation process is as follows: bs1=sa×br1; bs2=sa×br2; bsn=sa×brn.

[0145] Step S56: The weights of the indicators in set A are reset to zero, and the original weights of the indicators in set B are added to the apportioned weights to obtain the latest weight values.

[0146] Exemplarily, the calculation results are: a1=0; a2=0; ...; am=0; b1=b1+bs1; ​​b2=b2+bs2; bn=bn+bsn.

[0147] After completing the preliminary design of the evaluation label system, the actual domain name recursive resolution service is evaluated in combination with the service quality model. When the evaluation result does not match the actual domain name recursive resolution service, it is necessary to continue to iterate, evaluate, monitor, and evolve the service quality model for multiple rounds, and finally gradually converge the evaluation label system and the service quality model to match the actual domain name recursive resolution service. Figure 6 As shown, it includes steps S61 to S66.

[0148] Step S61: Obtaining service classification requirement information and quality assessment requirement dimension information.

[0149] The service level and dimension information may be determined according to the received instruction, and each evaluation label may be determined according to the dimension information.

[0150] Step S62: Determine the indicator type and evaluation indicator according to the dimensional information of the quality evaluation requirement.

[0151] In step S62, the evaluation indicators of the domain name recursive resolution service are identified, sorted, and evaluated to obtain the indicator type corresponding to each dimension of information and the evaluation indicators under each indicator type at different service levels.

[0152] Step S63: Determine the service level and the evaluation index corresponding to the service level according to the service classification requirement information.

[0153] Step S64: Determine the evaluation index corresponding to each service level and the default weight value of the evaluation index corresponding to each service level.

[0154] Step S65: Determine the score corresponding to the evaluation indicator.

[0155] Step S66: Evaluate the business data according to the service level, indicator type, evaluation indicator, default weight value and the score corresponding to the evaluation indicator.

[0156] In step S66, the indicator type, service level, evaluation indicator, default weight value and the score corresponding to the evaluation indicator can be integrated into the model parameters of the service quality model to be trained. The service quality model to be trained can be trained by the method provided by any embodiment of the present disclosure, and the service quality model can be used to evaluate the service data.

[0157] Step S67: adjusting at least one of the service level, indicator type, evaluation indicator, default weight value and the score corresponding to the evaluation indicator.

[0158] The service level, indicator type, evaluation indicator, default weight value and the score corresponding to the evaluation indicator constitute the quality evaluation indicator system. During the operation of the system, at least one of the service level, indicator type, evaluation indicator, default weight value and the score corresponding to the evaluation indicator can be adjusted according to the actual situation, so that the quality evaluation indicator system evolves with the display conditions.

[0159] During the evolution and maintenance of the quality assessment indicator system, the following events will trigger the iterative establishment process of the indicator system.

[0160] Event 1: The value range of the domain name recursive resolution service quality evaluation indicator item changes, and the indicator item value range score needs to be adjusted. For example, a new member is added to the protocol family supported by recursion.

[0161] Event 2: The domain name recursive resolution service quality evaluation indicator dimension has not changed, but the indicator items or indicator item weights within the indicator dimension need to be adjusted. For example, if a new evaluation indicator is added, the weight values ​​of all indicator items within this evaluation indicator dimension need to be recalculated and adjusted.

[0162] Event 3: The evaluation indicator dimension of the domain name recursive resolution service has changed. For example, a new evaluation indicator dimension has been added, and the weights of all evaluation indicator dimensions need to be adjusted.

[0163] Event 4: When the existing domain name recursive resolution service grading and quality assessment model is not applicable, it is necessary to re-implement processes such as quality assessment indicator identification and sorting, domain name recursive resolution service grading, domain name recursive resolution service quality assessment indicator system establishment, quality assessment algorithm model design, model convergence monitoring, and algorithm optimization.

[0164] In possible implementations, the quality assessment of domain name recursive resolution services is comprehensively evaluated in four dimensions: domain name recursive resolution service scope, domain name recursive resolution service quality, recursive hierarchical dependency, and recursive resolution security assurance capabilities. Since the above four dimensions differ in the number of indicators and evaluation difficulty, it is considered to set different evaluation methods for each dimension. At the same time, since the scope of domain name recursive resolution services (i.e., service level) has a great impact on various business indicators in domain name recursive resolution services, it is planned to first divide the recursion into multiple categories based on the service scope indicators, and then set corresponding algorithm models according to different categories for evaluation, so as to achieve the purpose of being more in line with the actual situation.

[0165] In one example, at least three different branches are set within the service quality model, corresponding to three different service levels, namely: DG3 level, DG2 level and DG1 level.

[0166] For example, in the service data generated by the domain name recursive resolution service process, the service data related to the service level is discrete data, and the decision tree algorithm can be directly used to classify the service level of the obtained service data. In view of the high feasibility of manual rating of service scope, the manually annotated service level reference answers can be appropriately used as supervision data in the early stage of training the service quality model. In other words, in the training stage of the service quality model, the service level corresponding to the model training data can be determined through manually annotated information.

[0167] After the service level is confirmed, model branches with different focuses are designed according to the service level, such as the national model branch corresponding to the national level, and the model branch of each service level is evaluated according to the evaluation labels of multiple dimensions (indicator types).

[0168] Figure 7An example of a domain name recursive resolution service rating model based on service grading (ie, the service quality model in the aforementioned embodiment) is presented.

[0169] according to Figure 7 As shown, the input data of the service quality model are the service data of three label categories (namely, domain name recursive resolution service quality, recursive hierarchical dependency, and recursive resolution security assurance capability), and the output data are the recursive evaluation results (equivalent to the model analysis results in the aforementioned embodiments).

[0170] In order to reduce the coupling degree of each module in the service quality model, the model can be divided into sub-modules according to the indicator type: domain name recursive resolution service quality part, recursive hierarchical dependency part, recursive resolution security assurance capability part, and total model part. Among them, except for the total model part, the other three parts score the domain name recursive resolution service separately, and after outputting the scoring results, the scores of the three dimensions are used as the input of the total model. In terms of the total model, since the number of indicators is small and the correlation is not strong, it is planned to select a multi-level logistic regression or ordered logistic regression model with strong interpretability and not easy to overfit, and output the final category of the domain name recursive resolution service rating (that is, the model analysis result in the aforementioned embodiment).

[0171] The above domain name recursive resolution service quality part is used to evaluate the domain name recursive resolution service quality dimension of business data. The domain name recursive resolution service quality dimension involves many evaluation indicators, which contain both discrete data and continuous data. There is no supervised data, and the accuracy of direct manual labeling and scoring is low. Based on the above situation, consider using weak supervised learning to divide the formation of the domain name recursive resolution service quality part into training and evaluation stages.

[0172] That is, in one example, the service quality model to be trained may include a domain name recursive resolution service quality part to be trained.

[0173] The process of training the service quality model to be trained includes the process of training the domain name recursive resolution service quality part, which includes a training phase and an evaluation phase. In the training phase, you can first obtain more valuable supervision data (model training data) and train the available domain name recursive resolution service quality part.

[0174] First, based on the business data obtained by the system, select the data corresponding to the evaluation label of the domain name recursive resolution service quality dimension as the training data used to train the domain name recursive resolution service quality part. When the above training data is obtained, the KMeans and DBSCAN algorithms can be used to cluster the business data, and the clustering results with a moderate number of data within the label category and small differences in the data within each label category are manually selected. Each label category is manually scored for quality based on experience as the first round of supervision data for the domain name recursive resolution service quality part. Use the training data and the first round of supervision data to train the business quality model, that is, train the domain name recursive resolution service quality part of the business quality model.

[0175] After obtaining the converged service quality model, the converged service quality model is used to evaluate the quality of the training data and / or updated training data, and the quality evaluation results are used as the second round of supervision data for retraining to obtain the final service quality model. During the training process of the service quality model, the quality of the data generated by the service quality model must be checked at any time and followed up and adjusted, and finally a trained service quality model (including the domain name recursive resolution service quality part) is obtained.

[0176] In the evaluation phase, you can use the trained business quality model, input the data item to be evaluated (equivalent to business data), and get the quality score of the data item.

[0177] In the recursive hierarchical dependency part and the recursive parsing security assurance capability part, the business data or sample business data corresponding to the relevant evaluation labels are all discrete data types. The sample business data corresponding to the evaluation labels can be manually labeled and rated using the decision tree algorithm.

[0178] For example, recursive hierarchical dependency data can be labeled as no dependency, domestic dependency only, and foreign dependency. Recursive parsing security assurance capabilities can be labeled as good, medium, and poor based on the level of assurance capabilities.

[0179] After the business data is labeled for the sample business data of the recursive hierarchical dependency dimension and the sample business data of the recursive parsing security assurance capability dimension, the labeled business data is used as supervision data to train the corresponding decision tree algorithm. After the algorithm converges, a single piece of data is input to output the evaluation and grading results of the data.

[0180] Reference Figure 7The evaluation results of a single submodule (i.e., the processing results of the submodule) need to be converted into numerical scores as the input of the overall model, and the output results of the submodules of the recursive hierarchical dependency part and the recursive parsing security assurance capability part need to be continuous. The non-continuous grading results within the submodule can be converted as follows: the recursive dependency rating results have no dependency, only domestic dependency, and foreign dependency, corresponding to values ​​of 0, 50, and 100, respectively; the recursive security rating results have good, medium, and poor values ​​of 100, 50, and 0, respectively.

[0181] After the service quality model is trained, the model analysis results need to be continuously monitored to ensure that the model analysis results gradually converge to match the actual domain name recursive resolution service. When the following situations occur, it is necessary to perform indicator fine-tuning, algorithm fine-tuning, model optimization, and continuous convergence to optimize the service quality model.

[0182] In case 1, the evaluation tag of the domain name recursive resolution service has not changed, but the default weight value of the evaluation tag needs to be adjusted, such as adjusting the default weight value corresponding to the completeness of protocol support.

[0183] In the second case, the label type of the recursive resolution service quality evaluation has not changed, but the indicator items within the label type (i.e., the evaluation labels) need to be adjusted. For example, if a new evaluation indicator (i.e., the evaluation label) is added, the weight values ​​of all indicator items within the corresponding evaluation indicator dimension need to be recalculated and adjusted.

[0184] Case three: the domain name recursive resolution service evaluation indicator dimension (i.e., label type) changes. For example, a new evaluation indicator dimension is added, and the default weight values ​​corresponding to all evaluation indicator dimensions need to be adjusted.

[0185] Case 4: When the requirements for domain name recursive resolution service companies in relevant regulations change, resulting in the existing domain name recursive resolution service classification and business quality model being no longer applicable, it is necessary to re-implement processes such as domain name recursive resolution service classification (confirming service level), identifying and sorting out quality assessment indicators, evaluating and monitoring the business quality model of domain name recursive resolution services, model convergence, and algorithm optimization.

[0186] Figure 8 An example is shown, showing the various devices involved in the process from sample business data acquisition, business data acquisition, to the creation, training and use of business quality models.

[0187] Figure 8In the multi-device structure shown, a large-scale discovery and analysis rating product framework for domain name recursive resolution services is established in accordance with the principle of "easy to use and simple to operate". The domain name resolution service is fully monitored based on the active and passive monitoring nodes deployed nationwide to obtain the analysis data source (as business data or sample business data). Abnormal resolution behaviors are discovered in real time from the dimensions of recursive resolution dependency, geographical scope of domain name recursive resolution services, quality of domain name recursive resolution services, and deployment of security enhancement measures. Through data analysis on the center side and combined with the recursive analysis rating model, a recursive resolution service quality assessment report is output to achieve long-term monitoring of the security status of domain name recursive resolution services.

[0188] exist Figure 8 The framework shown includes the data and business layer, the central management layer, and the terminal detection layer. The data and business layer is used to support business visualization, including the recursive business module, the domain name detection module, and the visualization detection tool module. Through the data and business layer, recursive business data and domain name detection data can be visualized. Through the visualization detection tool, the visualized information can be detected.

[0189] The central management layer includes an infrastructure management module, a detection management center module, a recursive service quality assessment module, and a recursive grading and classification module. The basic setting management module is used to manage detection nodes, recursive nodes, and detection targets. The recursive node is used to implement the domain name recursive resolution service, and the detection target may include a business object. The detection management center module may include a task bus and a scheduling center. The task bus is used to manage each recursive task or detection task, and the scheduling center is used to schedule resources such as detection nodes. The recursive service quality assessment module is used to run the business quality model and obtain the quality of the business according to the model analysis results of the business quality model. The recursive grading and classification module includes a hierarchical management and an indicator system. The hierarchical management is used to manage the service level, and the indicator system is used to manage the evaluation label. The central management layer is used to issue tasks to the terminal detection layer and set basic information management. The terminal detection layer includes a domain name resolution detection module, and further includes a plurality of detection nodes, and the detection nodes are used to obtain business data or sample business data.

[0190] Fig. 9 for Figure 8 The topology diagram of the framework shown includes Nginx (engine X), page gateway cluster, service API (application program interface), registration discovery (load balancing), nacos (dynamic naming and configuration service) cluster, application server cluster, database, and file server. The application server cluster is used to deploy the business quality model.

[0191] Nginx is the static resource entry of Html (Hypertext Markup Language) and is used for intranet load. The page gateway is used for authentication, access control, and load balancing. The authentication part can be extracted to the gateway layer, and then the microservice system does not need to pay attention to the logic of authentication but only needs to focus on its own business. It can also set whitelists, access times, access frequencies and other settings for some specific interfaces. It can also flexibly formulate load balancing strategies at the gateway layer, and can also uniformly add an additional protection layer at the gateway layer to prevent malicious attacks.

[0192] Service API (web, network) is used as the service interface call entry, mainly to complete API interface call authentication, billing, service driver call, etc.

[0193] Registration discovery can be used to achieve service load balancing.

[0194] Figure 8 The domain name detection and resolution module in the system covers two detection capabilities: general capabilities and basic detection capabilities. The general capabilities mainly include the reception and resolution of tasks, task classification and sharding, (local) consistency comparison and data reporting. The basic detection capabilities support basic detection capabilities such as DNS detection (basic settings configuration), ping (Internet packet explorer) / traceroute, MTR (My Traceroute, footprint), etc., and initiate network dialing requests for domain names or IPs to evaluate and analyze the performance, security and reliability of the DNS system. The functional architecture of the domain name resolution detection module is as follows: Fig.10 shown.

[0195] exist Fig.10 The architecture shown includes business support, detection management center and terminal detection layer. Business support is used to manage resolution accuracy, domain name service quality, recursive configuration security, and recursive classification. The detection management center is used to manage task execution status.

[0196] As the basic framework of detection, the detection terminal can flexibly configure detection tasks, support scanning of the entire network IP address segment, initiate active detection of any IP and port, and make entry and exit judgments on the detection target type (public recursion, operator recursion, etc.).

[0197] The detection terminal receives and analyzes the detection task, identifies the detection task type, detection target, detection cycle, detection source address and other attributes, and performs classification and slicing according to the detection target scale and business type to ensure the efficient operation of the detection system; at the same time, the detection terminal compares and marks the detection results with the expected results. After the data is sorted and classified, the detection data and the preliminary comparison result data are reported to the data center in a specified manner. The detection task processing flow is as follows: Fig.11 As shown, the following steps are included.

[0198] Step S111: The detection management center module issues a task.

[0199] Step S112: The domain name resolution detection module receives and resolves the task.

[0200] Step S113: The domain name resolution detection module resolves the task successfully.

[0201] Step S114: The domain name resolution detection module divides the task into different levels and slices.

[0202] Step S115: The detection node performs the detection task.

[0203] Step S116: collecting business data corresponding to the task.

[0204] Step S117: consistency comparison of business data.

[0205] Step S118: data reporting.

[0206] During the detection task reception and analysis process, the detection terminal center receives and analyzes the detection tasks from the detection management center, and analyzes the detection task's task ID, detection type, detection target (port), detection source address, detection target level (domain name), detection cycle, detection frequency, expected results and other information.

[0207] In the process of grading and slicing, the grading level and slicing method can be determined according to the scale and business characteristics of the detection target. At the same time, communication and collaboration between levels and slices should be maintained to ensure that the entire detection system can operate in a coordinated and efficient manner.

[0208] During the consistency comparison process, the collected detection result data is compared with the expected result data analyzed. The comparison content can be divided into complete match, partial match, detection result included in the expected result, etc. At the same time, the comparison results are marked to facilitate data classification and reporting.

[0209] When reporting data, according to the task type and the timeliness of the task requirements, it supports message queue reporting, SFTP\FTP (Secure File Transfer Protocol\File Transfer Protocol) reporting, one-time interface return reporting, etc. Different reporting data supports different storage modes according to business needs.

[0210] The detection management center consists of two parts: the task bus and the dispatch center. The two parts work together to complete the scheduling and management of tasks. The collaboration between the task bus and the dispatch center can achieve a more flexible and scalable task scheduling mechanism. Dynamically monitor the task status, improve the overall performance and response speed of the system, and better meet various business needs. The functional architecture of the detection management center is as follows: Fig.12 shown.

[0211] The task bus can dynamically add or delete tasks, making the system more flexible and scalable.

[0212] The task bus can also provide information such as task metadata and execution status to the scheduling center.

[0213] The task bus can also execute detection task instructions according to the scheduling strategy.

[0214] The dispatch center can identify the attributes, status, and priority of the task, formulate a more efficient dispatch strategy based on the task's bus information and other system status, and issue dispatch instructions to the task bus for execution.

[0215] The present disclosure also provides a network resolution service processing device, including:

[0216] A first acquisition module is used to obtain business data to be analyzed;

[0217] A first input module is used to input the service data to be analyzed into a service level branch in a service quality model; the data specification processed by the service level branch corresponds to the service level of the service data to be analyzed; the service quality model includes multiple service level branches;

[0218] A first output module, configured to obtain a model analysis result of the service quality model according to the output data of the service level branch; the model analysis result is used to evaluate the quality of the service corresponding to the service data;

[0219] The second acquisition module is used to obtain sample business data of the domain name recursive resolution service;

[0220] A clustering module, used to cluster the sample business data to obtain the label category to which the sample business data belongs;

[0221] A third acquisition module is used to obtain a model training sample according to the sample business data of the label category; the model training sample includes a plurality of the label categories, and the model training sample of each label category includes the sample business data corresponding to a plurality of evaluation labels; the plurality of evaluation labels are pre-set labels for the quality of the network resolution service;

[0222] A fourth acquisition module is used to obtain a converged service quality model based on the model training samples;

[0223] An updating module, used to update the sample business data to obtain updated model training samples;

[0224] The second input module is used to input the updated model training samples into the service quality model to be trained and the converged service quality model, and retrain the converged service quality model according to the output results to obtain a trained service quality model.

[0225] Below, reference Fig.13 The electronic device according to the embodiment of the present disclosure is described. The electronic device may be any one or both of the first device and the second device, or a stand-alone device independent of them, and the stand-alone device may communicate with the first device and the second device to receive the collected input signals from them.

[0226] Fig.13 A block diagram of an electronic device according to an embodiment of the present disclosure is illustrated.

[0227] like Fig.13 As shown, the electronic device includes one or more processors and memory.

[0228] The processor may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.

[0229] The memory may store one or more computer program products, and the memory may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc. One or more computer program products may be stored on the computer-readable storage medium, and the processor may run the computer program product to implement the methods of the various embodiments of the present disclosure described above and / or other desired functions.

[0230] In one example, the electronic device may further include: an input device and an output device, and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0231] In addition, the input device may also include, for example, a keyboard, a mouse, and the like.

[0232] The output device can output various information to the outside, including determined distance information, direction information, etc. The output device can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, and the like.

[0233] Of course, to simplify, Fig.13Only some of the components related to the present disclosure in the electronic device are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application situations, the electronic device may further include any other appropriate components.

[0234] In addition to the above-mentioned methods and devices, an embodiment of the present disclosure may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps in the processing method of the network resolution service according to various embodiments of the present disclosure described in the above part of this specification.

[0235] The computer program product may be written in any combination of one or more programming languages ​​to write program code for performing the operations of the disclosed embodiments, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0236] In addition, an embodiment of the present disclosure may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the steps in the processing method of the network resolution service according to various embodiments of the present disclosure described in the above part of this specification.

[0237] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0238] The basic principles of the present disclosure are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present disclosure are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present disclosure. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, and are not limitations. The above details do not limit the present disclosure to the necessity of adopting the above specific details to be implemented.

[0239] Each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system embodiment, since it basically corresponds to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0240] The block diagrams of the devices, apparatuses, equipment, and systems involved in this disclosure are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including," "comprising," "having," and the like are open words, referring to "including but not limited to," and can be used interchangeably therewith. The words "or" and "and" used herein refer to the words "and / or," and can be used interchangeably therewith, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to," and can be used interchangeably therewith.

[0241] The method and apparatus of the present disclosure may be implemented in many ways. For example, the method and apparatus of the present disclosure may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present disclosure are not limited to the order specifically described above, unless otherwise specifically stated. In addition, in some embodiments, the present disclosure may also be implemented as a program recorded in a recording medium, which includes machine-readable instructions for implementing the method according to the present disclosure. Therefore, the present disclosure also covers a recording medium storing a program for executing the method according to the present disclosure.

[0242] It should also be noted that in the apparatus, device and method of the present disclosure, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present disclosure.

[0243] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0244] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present disclosure to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for processing a network resolution service, characterized in that: include: Obtain business data to be analyzed; Inputting the service data to be analyzed into the service level branch in the service quality model; The data specification processed by the service level branch corresponds to the service level of the service data to be analyzed; the service quality model includes multiple service level branches; Obtaining a model analysis result of the service quality model according to the output data of the service level branch; the model analysis result is used to evaluate the quality of the service corresponding to the service data; The service quality model is obtained by the following steps: Obtain sample business data of domain name recursive resolution service; Clustering the sample business data to obtain the label category to which the sample business data belongs; According to the sample business data of the label category, a model training sample is obtained; the model training sample includes a plurality of the label categories, and the model training sample of each label category includes the sample business data corresponding to a plurality of evaluation labels; the plurality of evaluation labels are pre-set labels for the quality of the network resolution service; According to the model training samples, a converged service quality model is obtained; Updating the sample business data to obtain updated model training samples; The updated model training samples are input into the service quality model to be trained and the converged service quality model, and the converged service quality model is retrained according to the output results to obtain a trained service quality model.

2. The method according to claim 1, characterized in that The step of obtaining a converged service quality model based on the model training samples includes: For a target label category among the multiple label categories, determining the quantity of sample business data corresponding to each of the evaluation labels; Determining, according to the quantity of sample service data corresponding to each evaluation label, the initial weight of each evaluation label in each service level branch of the service quality model to be trained; Outputting the training phase model analysis results of the sample business data of the target label category; According to the model analysis results and sample analysis results of the training phase, the initial weights are adjusted to obtain adjusted weights; According to the adjusted weights, a converged service quality model is obtained.

3. The method according to claim 2, characterized in that The step of determining, according to the quantity of sample service data corresponding to each evaluation label, the initial weight of each evaluation label in each service level branch of the service quality model to be trained comprises: For a single evaluation tag, when the quantity of sample business data corresponding to the evaluation tag meets the set quantity threshold requirement, the initial weight of the evaluation tag is set according to the default weight value pre-set corresponding to the evaluation tag.

4. The method according to claim 2 or 3, characterized in that: The determining, according to the quantity of sample service data corresponding to each evaluation label, the initial weight of each evaluation label in each service level branch of the converged service quality model comprises: For a single evaluation tag, when the quantity of sample business data corresponding to the evaluation tag does not meet the set quantity threshold requirement, obtain a first evaluation tag that meets the set quantity threshold requirement and a second evaluation tag that does not meet the set quantity threshold requirement among the multiple evaluation tags; Determine the sum of the default weight values ​​pre-set corresponding to the first evaluation tag and the ratio of the default weight values ​​pre-set corresponding to the second evaluation tag; According to the ratio, determining the allocation ratio of the sum to the sample business data corresponding to the second evaluation tag; According to the apportionment ratio, the default weight value pre-set corresponding to each of the evaluation tags is adjusted to obtain an updated weight value corresponding to each of the evaluation tags; According to the updated weight value, the initial weight of each evaluation label is determined.

5. The method according to claim 4, characterized in that Determining, according to the ratio, the proportion of the sum to be allocated to the sample business data corresponding to the second evaluation tag includes: The proportion of each second evaluation label on each first evaluation label is determined according to the product of the proportion of each second evaluation label and the sum.

6. The method according to claim 5, characterized in that The step of adjusting the default weight values ​​pre-set for each evaluation tag according to the apportionment ratio to obtain the updated weight values ​​corresponding to each evaluation tag includes: Adjusting the default weight value pre-set corresponding to each second evaluation tag to zero, to obtain an updated weight value corresponding to each second evaluation tag; For a single first evaluation tag, the allocation ratio of each second evaluation tag to the first evaluation tag is added to the default weight value pre-set corresponding to the first evaluation tag to obtain an updated weight value corresponding to the first evaluation tag.

7. The method according to claim 1, characterized in that Before obtaining the model analysis result of the service quality model according to the output data of the service level branch, the method includes: Obtaining a submodule processing result obtained by at least one submodule in the service level branch processing the service data; each submodule in the at least one submodule is used to process the service data in a dimension of at least one tag type among the multiple tag types; The output data is obtained according to the processing results of the submodules.

8. The method according to claim 7, characterized in that The obtaining of the submodule processing result obtained by at least one submodule in the service level branch processing the service data comprises: Calculate the score corresponding to each evaluation tag in the submodule according to the weight value corresponding to each evaluation tag in the service level branch and the score corresponding to the business data value of each evaluation tag; According to the score corresponding to each evaluation tag in the submodule, the total score corresponding to all the evaluation tags in the submodule is calculated to obtain the submodule processing result.

9. The method according to claim 1, characterized in that: The sample business data for obtaining the domain name recursive resolution service includes: Sending a sample business data collection task to the business data detection terminal; the collection task is generated based on the evaluation tag where the sample business data is missing; Receiving detection data returned by the detection end for the collection task; Determining consistency of the detection data according to a unified identifier in the detection data; According to the consistency, the detection data is screened to obtain the sample business data.

10. The method according to claim 1, characterized in that The service level is determined according to the scope level of the business service object; the service levels include: national level, provincial level and enterprise level.

11. The method according to claim 1, characterized in that: The network resolution service includes a domain name recursive resolution service; the business data includes: business log data of the domain name recursive resolution service; the tag categories include: domain name recursive resolution service geographical scope, domain name recursive resolution service quality, recursive hierarchy dependency related and recursive resolution security related.

12. A processing device for a network resolution service, characterized in that: include: A first acquisition module is used to obtain business data to be analyzed; A first input module, used for inputting the service data to be analyzed into the service level branch in the service quality model; The data specification processed by the service level branch corresponds to the service level of the service data to be analyzed; the service quality model includes multiple service level branches; A first output module, configured to obtain a model analysis result of the service quality model according to the output data of the service level branch; the model analysis result is used to evaluate the quality of the service corresponding to the service data; The second acquisition module is used to obtain sample business data of the domain name recursive resolution service; A clustering module, used to cluster the sample business data to obtain the label category to which the sample business data belongs; A third acquisition module is used to obtain a model training sample according to the sample business data of the label category; the model training sample includes a plurality of the label categories, and the model training sample of each label category includes the sample business data corresponding to a plurality of evaluation labels; the plurality of evaluation labels are pre-set labels for the quality of the network resolution service; A fourth acquisition module is used to obtain a converged service quality model according to the model training samples; An updating module, used to update the sample business data to obtain updated model training samples; The second input module is used to input the updated model training samples into the service quality model to be trained and the converged service quality model, and retrain the converged service quality model according to the output results to obtain a trained service quality model.

13. An electronic device, characterized in that: include: a memory for storing a computer program product; A processor is used to execute the computer program product stored in the memory, and when the computer program product is executed, it implements the method described in any one of claims 1 to 11.

14. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method described in any one of claims 1 to 11 is implemented.

15. A computer program product comprising computer program instructions, characterized in that When the computer program instructions are executed by a processor, the method described in any one of claims 1 to 11 is implemented.

Images