CN116578442B

CN116578442B - Application crash analysis method based on artificial intelligence decision and big data storage system

Info

Publication number: CN116578442B
Application number: CN202310497113.6A
Authority: CN
Inventors: 靳艳; 陈丹丹
Original assignee: Beijing Infinite Free Culture Media Co ltd
Current assignee: Beijing Infinite Free Culture Media Co ltd; Beijing Peihong Wangzhi Technology Co ltd
Priority date: 2023-05-05
Filing date: 2023-05-05
Publication date: 2024-02-06
Anticipated expiration: 2043-05-05
Also published as: CN116578442A

Abstract

The embodiment of the application crash analysis method and the big data storage system based on artificial intelligence decision relate to the technical field of digital service and artificial intelligence, and realize error point analysis by carrying out application on the digital service online application correspondence, so that potential relation analysis is carried out on abnormal crash description knowledge chains corresponding to the error points of each application, and error point association information among the error points of each application realization is determined, thereby combining the error point association information among the error points of the application realization, further accurately tracing crash exposure requirements of the digital service online application, avoiding the problem of poor repair update reliability caused by the fact that only a single application is combined to realize repair update at present, and further improving the running stability of the digital service online application.

Description

Application crash analysis method based on artificial intelligence decision and big data storage system

Technical Field

The embodiment of the application relates to the technical field of digital services, in particular to an application crash analysis method based on artificial intelligence decision and a big data storage system.

Background

The digitization wave is coming, and the digitization technologies of mobile interconnection, internet of things, artificial intelligence, big data, cloud architecture, blockchain … … are subverting the traditional industry patterns and market environment, leading to the change of consumer habits of users. Most internet service providers are pushing business innovations with these digitization techniques to achieve the spanned growth of internet service providers. For various digital service online applications, online application crashes occur in the cloud service process, and the crashes interrupt the current workflow of a user, so that data is lost, and the operation of the application in the background is disturbed. For developers of internet service providers, how to accurately trace the breakdown exposure requirement of online application of digital service so as to provide reference basis for subsequent repair and update is a technical problem to be solved. In the related art, the repair updating performed by combining only a single application to implement the error point may result in poor reliability of the repair updating, and the running stability of the on-line application of the digital service cannot be effectively ensured.

Disclosure of Invention

In order to at least overcome the above-mentioned shortcomings in the prior art, an object of an embodiment of the present application is to provide an application crash analysis method and a big data storage system based on artificial intelligence decision.

In a first aspect, an embodiment of the present application provides an application crash analysis method based on artificial intelligence decision, applied to a big data storage system, the method including:

searching each application abnormal breakdown event with abnormal breakdown from an application response log of the digital service online application, and carrying out application implementation error point analysis on each application abnormal breakdown event by combining an application breakdown decision network converged by network weight information to obtain application implementation error point decision data corresponding to the digital service online application, wherein the application breakdown decision network converged by the network weight information is applied to an AI dispatch service center, and the application implementation error point decision data comprises a plurality of application implementation error points and an abnormal breakdown description knowledge chain corresponding to each application implementation error point;

performing potential relation analysis on an abnormal crash description knowledge chain corresponding to each application implementation error point, determining error point association information among the application implementation error points, and determining a target crash exposure requirement meeting the requirement by combining the error point association information among the application implementation error points, wherein the target crash exposure requirement represents at least one crash exposure element in an application service response example;

And repairing and updating the digital service online application based on a cloud crash repairing strategy corresponding to the target crash exposure requirement.

In a second aspect, an embodiment of the present application further provides an application crash analysis system based on an artificial intelligence decision, where the application crash analysis system based on an artificial intelligence decision includes a big data storage system and a plurality of AI operation service nodes communicatively connected to the big data storage system;

the big data storage system is used for:

In a third aspect, embodiments of the present application also provide a big data storage system, the big data storage system including a processor and a machine-readable storage medium having stored therein a computer program loaded and executed in conjunction with the processor to implement the artificial intelligence decision-based application crash analysis method of the first aspect above.

In a fourth aspect, embodiments of the present application provide a computer-readable storage medium storing computer-executable instructions for, when executed by a processor, implementing the artificial intelligence decision-based application crash analysis method of the first aspect above.

In a fifth aspect, embodiments of the present application provide a computer program product comprising a computer program or computer executable instructions which, when executed by a processor, implement the artificial intelligence decision-based application crash analysis method of the first aspect above.

The embodiment of the application has at least the following beneficial effects:

according to any aspect of the method, the analysis of the error points is realized by correspondingly carrying out application on the digital service online application, so that the potential relation analysis is carried out on the abnormal collapse description knowledge chain corresponding to the error points realized by each application, and the error point association information between the error points realized by each application is determined, thereby combining the error point association information between the error points realized by the application, further accurately tracing the collapse exposure requirement of the digital service online application, avoiding the problem of poor reliability of repairing and updating caused by repairing and updating by combining only a single application, and further improving the running stability of the digital service online application.

Drawings

Fig. 1 is a flow chart of an application crash analysis method based on artificial intelligence decision according to an embodiment of the present application.

Description of the embodiments

For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the present application will be described in further detail with reference to the accompanying drawings, and the described embodiments should not be construed as limiting the present application, and all other embodiments obtained by those skilled in the art without making any inventive effort are within the scope of the present application.

In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is to be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments and can be combined with each other on a non-conflicting basis.

In the following description, references to the term "first/second" are merely to distinguish similar virtual character conversational voices and do not represent a particular ordering for objects, it being understood that the "first/second" may be interchanged with a particular order or sequence, as allowed, to enable embodiments of the present application described herein to be implemented in an order other than illustrated or described herein.

Unless defined otherwise, all technical and scientific terms used in the examples of this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the embodiments of the application is for the purpose of describing the embodiments of the application only and is not intended to be limiting of the application.

See fig. 1:

STEP100 searches each application abnormal breakdown event with abnormal breakdown from an application response log of the digital service online application, and applies the application abnormal breakdown event to realize error point analysis by combining an application breakdown decision network converged by network weight information to obtain application realization error point decision data corresponding to the digital service online application.

In an alternative embodiment, the application crash decision network with converged network weight information is applied to an AI dispatch service center, and the application implementation error point decision data includes a plurality of application implementation error points and an abnormal crash description knowledge chain corresponding to each application implementation error point.

In an alternative embodiment, the application abnormal crash event may represent abnormal crash data of the online application of the digital service in the application function response process, and the application abnormal crash event may reflect an application implementation error point of the online application of the digital service in the function response implementation process. Therefore, application implementation error point analysis can be performed on each application abnormal breakdown event through the application breakdown decision network with the converged network weight information, a plurality of application implementation error points corresponding to the digital service online application and an abnormal breakdown description knowledge chain corresponding to each application implementation error point are obtained, wherein the abnormal breakdown description knowledge chain comprises a plurality of abnormal breakdown description knowledge with chain relations and is used for carrying out feature characterization on the corresponding application implementation error points.

The application crash decision network can be combined with training data to perform convergence optimization generation, for example, application crash template data can be collected, the application crash template data comprises application crash learning data and corresponding labeled template application implementation error points and template abnormal crash description knowledge chains, then the application crash learning data is input into an initialization application crash decision network to obtain decision application implementation error points and abnormal crash description knowledge chains, then the decision application implementation error points and abnormal crash description knowledge chains are combined with the labeled template application implementation error points and template abnormal crash description knowledge chains to calculate corresponding loss values, and then the loss values are combined to iteratively update the weight parameter layers of the initialization application crash decision network so as to obtain the application crash decision network. The application crash decision network may be a neural network model composed of a plurality of network parameter layers (representing a network operation node).

STEP200 carries out potential relation analysis on abnormal crash description knowledge chains corresponding to the error points of the application implementation, determines error point association information among the error points of the application implementation, and determines target crash exposure requirements meeting requirements according to the error point association information among the error points of the application implementation.

Wherein the target crash exposure requirement characterizes at least one crash exposure element in an application business response instance.

In an alternative embodiment, the error point association information between the error points implemented by each application can be determined by performing potential relation analysis on the abnormal crash description knowledge chains corresponding to the error points implemented by each application. For example, when three application implementation error points are respectively an application implementation error point 1, an application implementation error point 2 and an application implementation error point 3, through carrying out potential relation analysis on an abnormal collapse description knowledge chain of the application implementation error point 1, the application implementation error point 2 and the application implementation error point 3, it can be determined that the target collapse exposure requirement is finally the application implementation error point 3, but not the application implementation error point 1 and the application implementation error point 2, the application implementation error point 1 and the application implementation error point 2 are only front application implementation error points before the application implementation error point 3, and only the application implementation error point 1 and the application implementation error point 2 can not directly cause the application operation response collapse, so that when the application is subsequently restored and updated, the collapse exposure element is preferentially defined as the collapse exposure requirement corresponding to the application implementation error point 3, and the exposure element of the application implementation error point 1 and the application implementation error point 2 can be determined as the front collapse exposure element of the application implementation error point 3, and the subsequent relation is updated according to the relation.

STEP300, repairing and updating the digital service online application based on the cloud crash repairing strategy corresponding to the target crash exposure requirement.

In this embodiment, cloud crash repair policies corresponding to different crash exposure requirements may be preconfigured in the cloud server, so that the cloud crash repair policies corresponding to the target crash exposure requirements may be downloaded from the cloud server, thereby repairing and updating the online application of the digital service.

According to the technical scheme, the analysis of the error points is realized by carrying out application corresponding on the digital service online application, so that the potential relation analysis is carried out on the abnormal collapse description knowledge chain corresponding to the error points realized by each application, and the error point association information between the error points realized by each application is determined, so that the error point association information between the error points realized by each application is combined, further, the collapse exposure requirement of the digital service online application can be accurately traced, the problem that the repair updating reliability is poor due to the fact that only a single application is combined for realizing the repair updating at present is avoided, and the running stability of the digital service online application is improved.

The embodiment of the application also provides an application implementation error point analysis method based on artificial intelligence, which comprises the following steps.

STEP102, obtaining branch influence values of all decision operation branches in the decision operation topological structure, wherein the branch influence values are calculated by combining all decision operation branch connection information.

The decision operation topological structure refers to a network parameter layer representation diagram corresponding to an application collapse decision network, and the network parameter layer representation diagram refers to a neural network model formed by network parameter layers (representing a network operation node). The network parameter layer may be configured to perform a preset model calculation, such as a convolution operation, a weight summation operation, a full-connection classification operation, and so on. The decision operation topology structure is a topology structure formed by each network parameter layer corresponding to the application crash decision network. The decision operation branches refer to the same number of decision operation branches by applying the operation units corresponding to the network parameter layers in the breakdown decision network and applying how many network parameter layers are in the breakdown decision network. The branch influence value is calculated by combining the decision operation branch connection information and used for representing the span between the decision operation branch and the initial decision operation branch. The decision branch connection information refers to branch connection information corresponding to a decision branch, and the branch connection information refers to other decision branch information on which the decision branch depends when performing a decision operation, that is, decision operation execution data of the first decision branch depends on other decision branches when performing the decision operation, and the other decision branches may be zero, that is, the first decision branch does not depend on decision operation execution data of other decision branches when performing the decision operation, for example, the first decision branch may perform decision operation execution by using input data. The number of the decision branches may be several, that is, the first decision branch needs to depend on the decision execution data of the decision branches when executing the decision.

In an alternative embodiment, the big data storage system may directly obtain, from the network configuration data, a branch impact value (branch depth) of each decision branch in the decision operation topology, where the branch impact value may be calculated by using the connection information of each decision branch.

In an alternative embodiment, the big data storage system may also obtain, from the network configuration data, a decision operation topology structure to be subjected to topology splitting, then obtain, from the network configuration data, branch connection information corresponding to each decision operation branch in the decision operation topology structure, and determine, in combination with the branch connection information corresponding to each decision operation branch, a branch influence value corresponding to each decision operation branch.

STEP104, acquiring AI operation resource service information of each decision operation branch, and carrying out topology splitting on each decision operation branch by combining a branch influence value corresponding to each decision operation branch, branch connection information of each decision operation branch and AI operation resource service information of each decision operation branch to obtain a decision operation topology substructure and a decision operation branch to be processed.

The AI operation resource service information is used for characterizing whether the decision operation branch supports decision operation execution in the performance computing unit (such as a hardware accelerator), when the decision operation branch supports decision operation execution in the performance computing unit, the AI operation resource service information is the performance computing unit service information, and when the decision operation branch does not support decision operation execution in the performance computing unit, the AI operation resource service information is the non-performance computing unit service information. The decision operation branches not adapted to perform decision operation execution in the performance calculation unit may perform decision operation execution in the balance calculation unit. The corresponding AI operation resource service information can be preset by combining the operation attribute of the decision operation branch. The decision operation topology substructure refers to an operation splitting structure formed by decision operation branches obtained after topology splitting, namely the decision operation topology substructure refers to an operation splitting structure obtained by topology splitting. All decision branches in the same decision topology are processed by the same AI computing unit, which may be a performance computing unit or a balance computing unit. The decision operation branches to be processed refer to decision operation branches to be processed in a decision operation topological structure, namely decision operation branches except for the decision operation branches in a decision operation topological sub-structure in the decision operation topological structure are decision operation branches which are not topologically split into the decision operation topological sub-structure. The band structure decision operation branch may include a plurality of.

In an alternative embodiment, the big data storage system may use the branch impact value corresponding to each decision operation branch, the branch connection information of each decision operation branch, and the AI operation resource service information of each decision operation branch to perform topology splitting on each decision operation branch, where when the decision operation branch requiring topology splitting meets the requirement of the branch impact value and meets the requirement of AI operation service node operation service resource matching, the decision operation branch is used as the decision operation branch in the decision operation topology substructure, and when the decision operation branch requiring topology splitting meets the operation coordination requirement and meets the requirement of AI operation service node operation service resource matching, the decision operation branch is used as the decision operation branch in the decision operation topology substructure. The requirement of the branch influence value refers to a preset requirement of continuous branch influence value. The operation service resource matching requirement of the AI operation service node refers to a preset requirement of matching the computing unit of the AI operation service node. The operation coordination requirement refers to a preset unique scheduling requirement. At this time, a decision operation topology sub-structure is obtained by combining all decision operation branches meeting the requirements, and other decision operation branches in the decision operation topology structure are used as decision operation branches to be processed.

STEP106, taking the decision operation topological sub-structure and the decision operation branches to be processed as decision operation branches corresponding to the application breakdown decision network respectively, obtaining an optimized decision operation topological structure, and returning to the STEP of obtaining branch influence values of all the decision operation branches in the decision operation topological structure for iterative execution until all the decision operation branches in the decision operation topological structure complete topological splitting, so as to obtain all the target decision operation topological sub-structures and corresponding sub-structure operation resource service information.

The updated decision operation topology structure is obtained by taking a decision operation topology sub-structure as a decision operation branch of an application crash decision network and associating the decision operation branch with a decision operation branch to be processed. The decision operation branches in the updated decision operation topology are fewer than the decision operation topology. The existence decision operation branches in the updated decision operation topology are decision operation topology substructures. The target decision operation topology substructure refers to a decision operation topology substructure in which final topology splitting is completed. The sub-structure operation resource service information refers to AI operation resource service information corresponding to a target decision operation topology sub-structure, and is used for representing whether the target decision operation topology sub-structure supports decision operation execution in a performance calculation unit. That is, the target decision topology may support decision operations performed in the performance computing unit or in the balance computing unit.

In an alternative embodiment, the big data storage system uses the decision operation topology sub-structure and the decision operation branches to be processed as the decision operation branches corresponding to the application crash decision network respectively, that is, the big data storage system uses all the decision operation branches in the decision operation topology sub-structure together as one decision operation branch, and then combines the decision operation branches and other decision operation branches to be processed in the decision operation topology together to obtain the updated decision operation topology. At this time, the big data storage system returns to the iterative execution of the step of obtaining the branch influence value of each decision operation branch in the decision operation topological structure until each decision operation branch in the decision operation topological structure completes the topological splitting, namely, each decision operation branch in the decision operation topological structure is topologically split into decision operation topological sub-structures, and each target decision operation topological sub-structure is obtained. And meanwhile, combining the AI operation resource service information corresponding to each decision operation branch in the target decision operation topology substructure to obtain the substructure operation resource service information corresponding to the target decision operation topology substructure.

STEP108 stores the corresponding decision signature data of the decision operation topology, the mapping characteristic information between each target decision operation topology sub-structure and the corresponding sub-structure operation resource service information, wherein the mapping characteristic information is used for responding to the application crash decision task, and the application crash decision task is combined with the mapping characteristic information to configure each target decision operation topology sub-structure into the corresponding AI operation service node for topology configuration and application crash decision respectively.

The decision signature data is used for uniquely identifying the decision operation topological structure. The AI operation service node refers to an AI computing unit for performing target decision operation topology substructure operation, and may be a performance computing unit or a balance computing unit. The big data storage system may determine, in combination with the substructure operation resource service information, whether the target decision operation topology substructure is operated in the performance calculation unit or in the balance calculation unit. Different target decision topology substructures perform decision operations in different AI operation service nodes. The operation sequence of each target decision operation topology sub-structure is determined by combining the operation sequence of each decision operation branch in the decision operation topology structure.

In an alternative embodiment, the big data storage system establishes mapping characteristic information with decision signature data corresponding to the decision operation topology structure, each target decision operation topology sub-structure and corresponding sub-structure operation resource service information. The decision signature data is associated with all target decision operation topological substructures corresponding to the decision operation topological structure, and then is associated with the substructure operation resource service information corresponding to the target decision operation topological substructures. The mapping characteristic information is then stored and may be stored in the network configuration data. The big data storage system obtains an application collapse decision task, the application collapse decision task carries decision signature data, the application collapse decision task is responded, the decision signature data is used for combining mapping characteristic information to search all target decision operation topology substructures and corresponding substructures operation resource service information, and then all target decision operation topology substructures are configured into corresponding AI operation service nodes by combining the substructures operation resource service information to respectively carry out topology configuration and application collapse decision.

By adopting the technical scheme, the branch influence value of each decision operation branch in the decision operation topological structure is obtained, then the branch influence value corresponding to each decision operation branch, branch connection information of each decision operation branch and AI operation resource service information of each decision operation branch are used for carrying out topological splitting on each decision operation branch to obtain a decision operation topological sub-structure and a decision operation branch to be processed, then the decision operation topological sub-structure and the decision operation branch to be processed are respectively used as the decision operation branch corresponding to an application breakdown decision network, the optimized decision operation topological structure is obtained and iterative topological splitting is carried out to obtain each target decision operation topological sub-structure and corresponding sub-structure operation resource service information, and because the branch influence value, the branch connection information and the AI operation resource service information of each decision operation branch are used for carrying out topological splitting on each decision operation branch, the network parameter layer topology which is adjacent and can be adapted by an AI operation service node chip is enabled to be divided into the same decision operation topological sub-structure, the accuracy of the decision operation topological sub-structure is improved, the number of each decision operation topological sub-structure which is split is then the decision operation topological sub-structure is used as the decision operation branch corresponding to an application breakdown decision operation network, and the corresponding decision operation topological sub-structure is mapped to the corresponding decision operation topological service information when the corresponding to the application decision operation topological structure is mapped to the corresponding decision operation topological service node, therefore, the speed of dispatching calculation data among different AI operation service nodes can be improved, the decision performance of the application collapse decision network can be improved, and the utilization rate of operation resources of the AI operation service nodes is improved.

In an alternative embodiment, STEP104 performs topology splitting on each decision branch by combining a branch impact value corresponding to each decision branch, branch connection information of each decision branch, and AI operation resource service information of each decision branch, to obtain a decision operation topology sub-structure and a decision operation branch to be processed, and includes:

STEP202 determines a first decision branch from each decision branch, and obtains a second decision branch corresponding to the first decision branch from each decision branch.

The first decision operation branch is a decision operation branch which is not split by topology and meets the matching requirement of the operation service resources of the AI operation service node. The second decision branch refers to a decision branch adjacent to the first decision branch.

In an alternative embodiment, the big data storage system combines the sequence of each decision operation branch in the application crash decision network to select a decision operation branch which is not topologically split from each decision operation branch, and then judges the selected decision operation branch, where it can be judged whether the selected decision operation branch meets the requirement of AI operation service node on matching of operation service resources, where the requirement of AI operation service node on matching of operation service resources refers to support of decision operation execution in the performance calculation unit. And then when the matching requirement of the operation service resources is met, taking the selected decision operation branch which is not split by the topology as a first decision operation branch. And then, combining the sequences of the decision operation branches to acquire neighbor decision operation branches corresponding to the first decision operation branch, and acquiring a second decision operation branch.

STEP204, acquiring target prior topology splitting information of the second decision operation branch, determining a target branch influence value corresponding to the second decision operation branch from branch influence values corresponding to the decision operation branches, determining target AI operation resource service information corresponding to the second decision operation branch from AI operation resource service information of the decision operation branches, and determining target branch connection information corresponding to the second decision operation branch from branch connection information of the decision operation branches.

The target prior topology splitting information refers to prior topology splitting information of the second decision operation branch, and the prior topology splitting information is used for representing whether the second decision operation branch is topologically split. The target branch impact value refers to the depth of the second decision branch in the decision operation topology. The target AI-computing resource service information refers to AI-computing resource service information of the second decision branch. The target branch connection information refers to the branch connection information of the second decision operation branch.

In an alternative embodiment, the big data storage system may set and save the already topologically split feature vector when the decision branch is topologically split. When the topology splitting of the second decision operation branch is needed, it is first needed to find out whether the second decision operation branch has been topologically split. At this time, the big data storage system may acquire target prior topology splitting information of the second decision operation branch from the network configuration data, where the big data storage system may search for a topologically split feature vector corresponding to the second decision operation branch, and when the topologically split feature vector is found, the obtained target prior topology splitting information is that the topologically split is completed, and when the topologically split feature vector is not found, the obtained target prior topology splitting information is that the topologically split is not completed. And simultaneously, the big data storage system searches the branch influence value corresponding to the decision operation branch which is the same as the second decision operation branch from the branch influence values corresponding to the decision operation branches to obtain a target branch influence value. And searching the AI operation resource service information corresponding to the decision operation branch identical to the second decision operation branch from the AI operation resource service information of each decision operation branch to obtain target AI operation resource service information. And searching the branch connection information corresponding to the decision operation branch identical to the second decision operation branch from the branch connection information of each decision operation branch to obtain target branch connection information.

STEP206 performs sub-structure decision branch verification on the second decision branch in combination with the target AI operation resource service information, the target prior topology split information, the target branch impact value, and the target branch connection information.

The verification of the substructure decision operation branch refers to verifying whether the decision operation branch is a decision operation branch in a decision operation topology substructure.

In an alternative embodiment, the big data storage system uses the service information of the target AI operation resource to verify whether the second decision operation branch is a decision operation branch in the decision operation topology sub-structure, so as to obtain a verification result. And verifying whether the second decision operation branch is a decision operation branch in the decision operation topology substructure by using the target branch influence value by the big data storage system to obtain a verification result. And the big data storage system uses the target branch connection information to verify whether the second decision operation branch is the decision operation branch in the decision operation topology sub-structure, so as to obtain a verification result. And the big data storage system uses the target AI operation resource service information to verify whether the second decision operation branch is a decision operation branch in the decision operation topology sub-structure, so as to obtain a verification result. When any one of the verification results indicates that the second decision operation branch is not the decision operation branch in the decision operation topology substructure, the substructure decision operation branch verification is unsuccessful.

STEP208, when the verification of the second decision operation branch is successful, takes the first decision operation branch and the second decision operation branch as updated first decision operation branches, and returns to the STEP of obtaining the second decision operation branch corresponding to the first decision operation branches for execution until the updated first decision operation branches are not changed any more, and obtaining the second decision operation branches with successful verification.

The verification success of the second decision operation branch means that the second decision operation branch is verified to be a decision operation branch in the decision operation topology sub-structure.

In an alternative embodiment, the second decision branch is described as a decision branch in the decision topology sub-structure when the verification of the second decision branch is successful. At this time, the big data storage system takes the first decision operation branch and the second decision operation branch as updated first decision operation branches, wherein, since the first decision operation branch and the target network parameter layer are adjacent decision operation branches and are all decision operation branches in the decision operation topology substructure, the first decision operation branch and the second decision operation branch can be taken as the same decision operation branch at this time, and thus the updated first decision operation branch is obtained. And then performing iterative loop, namely returning to execute the step of acquiring the second decision operation branch corresponding to the first decision operation branch until the updated first decision operation branch is not changed any more, namely, when the updated first decision operation branch is not changed any more, namely, the updated first decision operation branch is the same as the first decision operation branch before updating, and combining the second decision operation branches included in the updated first decision operation branch to obtain each second decision operation branch with successful verification.

STEP210, combining the first decision operation branch and each successfully verified second decision operation branch to obtain a decision operation topology substructure, and combining each decision operation branch, the first decision operation branch and each successfully verified second decision operation branch to obtain a decision operation branch to be processed.

In an alternative embodiment, the big data storage system uses the first decision operation branch and each successfully verified second decision operation branch as decision operation branches in the same decision operation topology sub-structure, so as to obtain the decision operation topology sub-structure. And then taking decision operation branches except the first decision operation branch and the second decision operation branch which are successfully verified in each decision operation branch as decision operation branches to be processed.

By adopting the technical scheme, the first decision operation branch is determined from each decision operation branch, the second decision operation branch corresponding to the first decision operation branch is obtained from each decision operation branch, and then the target AI operation resource service information, the target prior topology splitting information, the target branch influence value and the target branch connection information are used for carrying out substructure decision operation branch verification on the second decision operation branch, so that the verification precision of the second decision operation branch is improved. And when the verification of the second decision operation branch is successful, the first decision operation branch and the second decision operation branch are used as updated first decision operation branches and are iteratively executed until the updated first decision operation branches are not changed any more, so that each verified second decision operation branch is obtained, finally, a decision operation topology sub-structure is obtained by combining the first decision operation branch and each verified second decision operation branch, and the decision operation branch topology belonging to one topology split can be split into the same decision operation topology sub-structure, thereby improving the reliability of the obtained decision operation topology sub-structure and further reducing the number of each decision operation topology sub-structure of the topology split. And then, when decision operation is performed, the performance efficiency is improved.

In an alternative embodiment, STEP202 determines a first decision branch from each decision branch, and obtains a second decision branch corresponding to the first decision branch from each decision branch, including:

STEP302, selecting a targeting decision operation branch from each decision operation branch, acquiring targeting prior topology splitting information of the targeting decision operation branch, and determining targeting operation resource service information corresponding to the targeting decision operation branch from AI operation resource service information of each decision operation branch.

The targeting decision operation branches refer to targeting decision operation branches in the decision operation branches when determining the first decision operation branch. The target prior topology splitting information refers to the prior topology splitting information corresponding to the target decision operation branch. The targeted computing resource service information refers to AI computing resource service information corresponding to a targeted decision computing branch.

In an alternative embodiment, the big data storage system may sequentially select the targeted decision operation branches from the respective decision operation branches in combination with the order of the decision operation branches in the decision operation topology. And then acquiring the target prior topology splitting information of the target decision operation branch from the network configuration data, and searching the AI operation resource service information corresponding to the decision operation branch identical to the target decision operation branch from the AI operation resource service information corresponding to each decision operation branch to obtain the target operation resource service information.

STEP304 performs substructure decision branch verification on the target decision branch in combination with the target prior topology resolution information and the target operation resource service information.

In an alternative embodiment, the big data storage system uses the targeted prior topology split information and the targeted computing resource service information to verify and judge the targeted decision computing branch, wherein the targeted decision computing branch is judged whether the targeted decision computing branch has been topologically split or not in combination with the targeted prior topology split information. And then, judging whether the targeting decision operation branch supports decision operation execution in the performance calculation unit by combining the targeting operation resource service information.

STEP306, when the verification of the targeting decision operation branch is successful, the targeting decision operation branch is taken as a first decision operation branch, a forward decision operation branch and a backward decision operation branch corresponding to the first decision operation branch are obtained from each decision operation branch, and the forward decision operation branch and the backward decision operation branch are respectively taken as a second decision operation branch.

The forward decision operation branch refers to a decision operation branch which is determined by combining the operation sequence of the decision operation branch in the decision operation topological structure and performs decision operation execution before the target decision operation branch. The backward decision operation branch refers to a decision operation branch which is determined by combining the operation sequence of the decision operation branch in the decision operation topological structure and performs decision operation execution after the decision operation branch is targeted.

In an alternative embodiment, when the verification of the targeting decision operation branch is successful, it is indicated that the targeting decision operation branch is not topologically split and the targeting decision operation branch supports the decision operation execution in the performance calculation unit, and the targeting decision operation branch is taken as the first decision operation branch. And then, acquiring a forward decision operation branch and a backward decision operation branch corresponding to the first decision operation branch from each decision operation branch, taking the forward decision operation branch and the backward decision operation branch as second decision operation branches respectively, namely taking the forward decision operation branch as the second decision operation branch, then executing the step of carrying out subsequent processing on the second decision operation branch, and simultaneously taking the backward decision operation branch as the second decision operation branch, and executing the step of carrying out subsequent processing on the second decision operation branch.

STEP308, when the verification of the targeting decision operation branch is unsuccessful, taking the targeting decision operation branch as a decision operation topology substructure, and returning to the STEP of obtaining the branch influence value of each decision operation branch in the decision operation topology structure for iterative execution until each decision operation branch in the decision operation topology structure completes the topology splitting.

In an alternative embodiment, when the verification of the targeting decision operation branch is unsuccessful, it is indicated that the targeting decision operation branch is topologically split, or the targeting decision operation branch does not support the decision operation in the performance calculation unit, or the targeting decision operation branch is topologically split and the targeting decision operation branch does not support the decision operation in the performance calculation unit. At this time, the targeting decision operation branch is directly used as a decision operation topology substructure, i.e. the subsequent decision operation branch is used as a separate decision operation topology substructure. And the independent decision operation topological sub-structure and the decision operation branches to be processed are subjected to iterative execution to obtain updated decision operation topological structures, and the step of obtaining branch influence values of all the decision operation branches in the decision operation topological structures is returned until all the decision operation branches in the decision operation topological structures complete topological splitting.

In an alternative embodiment, when verification of a plurality of consecutive targeting decision branches is unsuccessful, the plurality of consecutive targeting decision branches may be used as decision branches in the same decision topology, so as to obtain an operation splitting structure operated in the balance calculation unit.

By adopting the technical scheme, the target decision operation branches are selected from the decision operation branches, the substructure decision operation branch verification is carried out on the target decision operation branches, and when the verification is successful, the forward decision operation branch and the backward decision operation branch corresponding to the first decision operation branch are obtained from the decision operation branches and are respectively used as the second decision operation branches, so that the accuracy of the obtained second decision operation branches is improved.

In an alternative embodiment, STEP304 performs sub-structure decision operation branch verification on the target decision operation branch in combination with the target previous topology splitting information and the target operation resource service information, and includes the STEPs of:

performing prior topology splitting verification on the target decision operation branch by combining the prior topology splitting information of the target, and performing operation resource verification on the target decision operation branch by combining the service information of the target operation resource; when the prior topology splitting verification of the targeting decision operation branch is successful and the operation resource verification of the targeting decision operation branch is successful, judging that the targeting decision operation branch is successful; and when the prior topology splitting verification of the targeting decision operation branch is unsuccessful or the operation resource verification of the targeting decision operation branch is unsuccessful, judging that the targeting decision operation branch is unsuccessful.

Wherein, the prior topology splitting verification refers to verifying whether the decision operation branch is split by topology. The operation resource verification means verifying whether the decision operation branch supports operation in the performance calculation unit.

In an alternative embodiment, the big data storage system uses the target previous topology splitting information to perform previous topology splitting verification on the target decision operation branch, wherein the big data storage system can verify whether a topological splitting feature vector corresponding to the target decision operation branch exists in the target previous topology splitting information, and when the topological splitting feature vector corresponding to the target decision operation branch does not exist, the previous topology splitting verification of the target decision operation branch is successful. When the topological split feature vector corresponding to the targeting decision operation branch exists, the prior topological split verification of the targeting decision operation branch is unsuccessful. And then the server uses the targeting operation resource service information to perform operation resource verification on the targeting decision operation branch, wherein the big data storage system can verify whether the targeting operation resource service information is that the targeting decision operation branch supports operation in the performance calculation unit. When the target decision operation branch is verified to support the decision operation execution in the performance calculation unit, the success of the operation resource verification of the target decision operation branch is indicated, and when the target decision operation branch is verified to not support the decision operation execution in the performance calculation unit, the unsuccessful of the operation resource verification of the target decision operation branch is indicated.

Therefore, the sub-structure decision operation branch verification is carried out on the target decision operation branch through the target prior topology splitting information and the target operation resource service information, so that the accuracy of the target decision operation branch verification can be ensured.

In an alternative embodiment, STEP206 performs sub-structure decision branch validation on the second decision branch in conjunction with the target AI operation resource service information, the target prior topology resolution information, the target branch impact value, and the target branch connection information, including:

STEP402 performs prior topology splitting verification on the second decision branch in combination with the target prior topology splitting information, performs operation resource verification on the second decision branch in combination with the target AI operation resource service information, and performs continuity verification on the second decision branch in combination with the target branch influence value and the target branch connection information.

The continuity verification is to verify whether the second decision operation branch has continuity between the first decision operation branches, where the continuity may be that a branch impact value corresponding to the second decision operation branch is continuous with a branch impact value corresponding to the first decision operation branch or that the second decision operation branch is uniquely scheduled by the first decision operation branch.

In an alternative embodiment, the big data storage system uses the target previous topology splitting information to perform previous topology splitting verification on the second decision operation branch, wherein the big data storage system can verify whether a topological splitting feature vector corresponding to the second decision operation branch exists in the target previous topology splitting information. And when the topological split feature vector corresponding to the second decision operation branch does not exist, the previous topological split verification of the second decision operation branch is successful. When the topological split feature vector corresponding to the second decision operation branch exists, the former topological split verification of the second decision operation branch is not successful.

In an alternative implementation manner, when the topology splitting is completed, the big data storage system stores the decision operation branch log data of the topology splitting completion, then searches the decision operation branch log data of the topology splitting completion for a second decision operation branch, when the second decision operation branch is not searched, the second decision operation branch is the decision operation branch without the topology splitting, that is, the previous topology splitting verification of the second decision operation branch is successful, and when the second decision operation branch is searched, the second decision operation branch is the decision operation branch with the topology splitting, that is, the previous topology splitting verification of the second decision operation branch is unsuccessful.

And the big data storage system uses the target AI operation resource service information to perform operation resource verification on the second decision operation branch, wherein the big data storage system can verify whether the target AI operation resource service information is that the second decision operation branch supports operation in the performance calculation unit. And when the second decision operation branch is verified to support the decision operation execution in the performance calculation unit, the operation resource verification of the second decision operation branch is proved to be successful, and when the second decision operation branch is verified to not support the decision operation execution in the performance calculation unit, the operation resource verification of the second decision operation branch is proved to be unsuccessful.

And the big data storage system uses the target branch impact value and the target branch connection information to perform continuity verification on the second decision operation branch. The big data storage system can verify whether the target branch influence value is a neighbor value of the branch influence value corresponding to the first decision operation branch or not, and simultaneously verify whether the target branch connection information is that the second decision operation branch is uniquely scheduled by the first decision operation branch or not. And when the target branch influence value is a neighbor value of the branch influence value corresponding to the first decision operation branch or the target branch connection information is that the second decision operation branch is uniquely scheduled by the first decision operation branch, the success of the continuity verification of the second decision operation branch is indicated. When the target branch influence value is not the neighbor value of the branch influence value corresponding to the first decision operation branch and the target branch connection information does not mean that the second decision operation branch is uniquely scheduled by the first decision operation branch, the continuity verification of the second decision operation branch is not successful.

STEP404, when the prior topology split verification of the second decision operation branch, the operation resource verification of the second decision operation branch and the continuity verification of the second decision operation branch all pass, determining that the second decision operation branch verification is successful.

STEP406, when any one of the previous topology split verification of the second decision operation branch, the operation resource verification of the second decision operation branch and the continuity verification of the second decision operation branch is unsuccessful, determining that the second decision operation branch verification is unsuccessful.

In an alternative embodiment, the big data storage system determines that the first topology splitting verification of the second decision operation branch, the verification of the operation resource of the second decision operation branch, and the continuity verification of the second decision operation branch are all passed, and indicates that the second decision operation branch is a decision operation branch in the decision operation topology sub-structure formed by the first decision operation branch, that is, the verification of the second decision operation branch is successful. When any one of the prior topology splitting verification of the second decision operation branch, the operation resource verification of the second decision operation branch and the continuity verification of the second decision operation branch is unsuccessful, the second decision operation branch is not the decision operation branch in the decision operation topology sub-structure formed by the first decision operation branch, namely, the second decision operation branch verification is determined to be unsuccessful.

By adopting the technical scheme, the first decision operation branch is subjected to the prior topology splitting verification by using the target prior topology splitting information, the second decision operation branch is subjected to the operation resource verification by using the target AI operation resource service information, and the second decision operation branch is subjected to the continuity verification by using the target branch influence value and the target branch connection information. And then combining the prior topology splitting verification, the operation resource verification and the continuity verification to determine whether the second decision operation branch is successfully verified, thereby ensuring the accuracy of the second decision operation branch verification.

In an alternative embodiment, after STEP206, i.e. after the sub-structure decision branch verification of the second decision branch by the target prior topology splitting information and the target branch impact value in combination with the target AI operation resource service information, the method further comprises the STEPs of:

when the verification of the second decision operation branch is unsuccessful, combining the first decision operation branch to obtain a decision operation topology sub-structure, and combining each decision operation branch with the first decision operation branch to obtain a decision operation branch to be processed.

In an alternative embodiment, when the verification of the second decision branch is unsuccessful, it is indicated that the second decision branch is not a decision branch in the decision topology formed by the first decision branch. At this time, the first decision operation branch is directly taken as a decision operation topology sub-structure alone. Meanwhile, decision operation branches except the first decision operation branch in each decision operation branch are used as decision operation branches to be processed, so that decision operation branches which can be adapted by a performance calculation unit and decision operation branch topologies which cannot be adapted by the performance calculation unit are prevented from being split into the same operation splitting structure, and the accuracy of the obtained decision operation topology sub-structure is guaranteed.

In an alternative embodiment, STEP102 obtains a branch impact value of each decision branch in the decision topology, where the branch impact value is calculated in combination with the connection information of each decision branch, and includes the STEPs of:

arranging all decision operation branches in the decision operation topological structure to obtain decision operation branch arrangement information; and calculating branch influence values corresponding to the decision operation branches by combining the branch connection information and the decision operation branch arrangement information of the decision operation branches.

The arrangement can sort a network parameter layer topology structure to obtain an ordered linear set. The decision operation branch arrangement information is an ordered linear set obtained by ordering all decision operation branches in the decision operation topological structure, and is used for representing the sequence of the decision operation branches when the decision operation is executed.

In an alternative embodiment, the big data storage system uses an arrangement algorithm to arrange each decision operation branch in the decision operation topological structure to obtain the arrangement information of the decision operation branches. And then the big data storage system uses the branch connection information and the branch arrangement information of each decision operation branch to determine the branch influence value corresponding to each decision operation branch. The method comprises the steps of determining decision operation branches to be calculated according to the sequence of the arrangement information of the decision operation branches, determining associated decision operation branches of the first decision operation branches by using branch connection information of the first decision operation branches, and calculating branch influence values of the first decision operation branches according to branch influence values of the associated decision operation branches. And finally traversing each decision operation branch in the decision operation branch arrangement information to obtain a branch influence value corresponding to each decision operation branch.

According to the technical scheme, the decision operation branches are arranged to obtain the decision operation branch arrangement information, and then the branch connection information and the decision operation branch arrangement information of the decision operation branches are used for calculating the branch influence values corresponding to the decision operation branches, so that the accuracy of the obtained branch influence values is improved.

In an alternative embodiment, calculating the branch impact value corresponding to each decision branch by combining the branch connection information and the decision branch arrangement information of each decision branch includes the steps of:

determining target decision operation branches from all decision operation branches by combining the arrangement order of the decision operation branch arrangement information, and verifying whether corresponding associated decision operation branches exist in the target decision operation branches by combining the branch connection information of all decision operation branches; when the target decision operation branch has a corresponding association decision operation branch, acquiring a branch influence value corresponding to the association decision operation branch, and determining the branch influence value corresponding to the target decision operation branch by combining the branch influence value corresponding to the association decision operation branch.

The target decision operation branch refers to a decision operation branch which needs to calculate a branch influence value at present. The associated decision operation branch refers to a decision operation branch of target decision operation branch scheduling. That is, the target decision operation branch needs the decision operation execution data of the associated decision operation branch to execute the decision operation when executing the decision operation. The branch impact value of the associated decision branch is determined prior to the branch impact value calculation performed by the target decision branch.

In an alternative embodiment, the big data storage system determines a target decision branch from the decision branches in combination with the arrangement order of the decision branch arrangement information, and then verifies whether the target decision branch has a corresponding associated decision branch in combination with the branch connection information of the decision branches. When the target decision operation branch has a corresponding association decision operation branch, the big data storage system directly acquires a branch influence value corresponding to the association decision operation branch from the network configuration data, and then determines the branch influence value corresponding to the target decision operation branch by combining the branch influence value corresponding to the association decision operation branch, which can be the sum of the branch influence value corresponding to the association decision operation branch and the first calibration value, so as to obtain the branch influence value corresponding to the target decision operation branch. The first calibration value is preset and may be one.

By adopting the technical scheme, by verifying whether the target decision operation branch has the corresponding association decision operation branch, when the target decision operation branch has the corresponding association decision operation branch, the branch influence value corresponding to the association decision operation branch is used for determining the branch influence value corresponding to the target decision operation branch, so that the accuracy and the efficiency of calculating the branch influence value are improved.

In an alternative embodiment, there are at least two associated decision branches; acquiring a branch influence value corresponding to the associated decision operation branch, and determining the branch influence value corresponding to the target decision operation branch by combining the branch influence value corresponding to the associated decision operation branch, wherein the method comprises the following steps:

acquiring branch influence values corresponding to at least two associated decision operation branches respectively, and determining the maximum branch influence value from the branch influence values corresponding to the at least two associated decision operation branches respectively; and calculating the sum of the maximum branch influence value and the first calibration value to obtain the branch influence value corresponding to the target decision operation branch.

The first calibration value refers to a preset value used when the branch influence value is required to be calculated, and may be set to be one.

In an alternative embodiment, the big data storage system verifies that there are at least two associated decision branches corresponding to the target decision branch. At this time, a branch influence value corresponding to each associated decision operation branch is obtained from the network configuration data, and then the magnitude of the branch influence value corresponding to each associated decision operation branch is compared, and the maximum branch influence value is determined from the branch influence value corresponding to each associated decision operation branch. And then calculating the sum of the maximum branch influence value and the first calibration value to obtain the branch influence value corresponding to the target decision operation branch.

By adopting the technical scheme, when a plurality of associated decision operation branches exist in the target decision operation branch, the maximum branch influence value is determined from the branch influence values respectively corresponding to at least two associated decision operation branches, and then the sum of the maximum branch influence value and the first calibration value is calculated to obtain the branch influence value corresponding to the target decision operation branch, so that the accuracy of the branch influence value corresponding to the obtained target decision operation branch is ensured.

In an alternative embodiment, after verifying whether the target decision operation branch has a corresponding associated decision operation branch in combination with the branch connection information of each decision operation branch, the method further includes the steps of:

and when the corresponding association decision operation branch does not exist in the target decision operation branch, determining the branch influence value corresponding to the target decision operation branch as a second calibration value.

The second calibration value refers to a value when the preset target decision operation branch is the initial decision operation branch, and may be set to zero.

In an alternative embodiment, when the big data storage system determines that the target decision operation branch does not have a corresponding associated decision operation branch, it indicates that the target decision operation branch is a starting decision operation branch in the decision operation topology structure, that is, the target decision operation branch does not schedule decision operation execution data of other decision operation branches to perform decision operation execution. At this time, the branch influence value corresponding to the target decision operation branch is directly determined to be the second calibration value, so that the accuracy of the branch influence value of the target decision operation branch is ensured.

In an alternative implementation manner, the embodiment of the application further provides an application crash analysis method based on an artificial intelligence model, which comprises the following steps:

STEP502 obtains an application crash decision task, which carries target decision signature data and application abnormal crash events.

The target decision signature data is used for marking a decision operation topological structure which needs to perform decision operation on the application abnormal crash event, namely ID information of the application crash decision network. The application abnormal crash event refers to data which needs to use a decision operation topological structure corresponding to the target decision signature data to perform decision operation execution.

STEP504, obtain each goal decision operation topological substructure and correspondent substructure operation resource service information that the goal decision signature data correlated with in combination with mapping characteristic information, each goal decision operation topological substructure and correspondent substructure operation resource service information are through obtaining the branch influence value of each decision operation branch in the decision operation topological structure, the branch influence value is calculated in combination with each decision operation branch connection information, obtain AI operation resource service information of each decision operation branch, combine branch influence value, branch connection information and AI operation resource service information of each decision operation branch that each decision operation branch corresponds to, carry on the topological split to each decision operation branch, obtain decision operation topological substructure and pending decision operation branch, regard decision operation topological substructure and pending decision operation branch as the decision operation branch that applies the collapse decision network to correspond respectively, obtain the decision operation topological structure after optimizing, and return to obtain the STEP iteration execution of the branch influence value of each decision operation branch in the decision operation topological structure, until each decision operation branch in the decision operation topological structure is all accomplished the time division and obtained.

In an alternative embodiment, the big data storage system searches all target decision operation topology substructures associated with the target decision signature data and the substructures corresponding to each target decision operation topology substructures in combination with application breakdown decision tasks from storage to mapping feature information. The mapping characteristic information may be obtained by the above embodiment. The sub-structure of the decision-making operation topology of each target and the service information of the corresponding sub-structure operation resource obtained by the mapping characteristic information can also be obtained by carrying out the topological resolution of the decision-making operation topology in the above embodiment.

STEP506, combining the application abnormal breakdown event and the sub-structure operation resource service information to configure each target decision operation topology sub-structure to the corresponding AI operation service node to respectively perform topology configuration and application breakdown decision, so as to obtain application implementation error point decision data corresponding to the application abnormal breakdown event.

The application implementation error point decision data refers to a result obtained after the application abnormal crash event is executed by decision operation through a decision operation branch in a decision operation topological structure.

In an alternative implementation manner, the big data storage system determines the operation sequence of each target decision operation topology substructure in combination with the operation sequence of the decision operation branches in the decision operation topology, and then determines the AI operation service node corresponding to each target decision operation topology substructure in combination with the substructure operation resource service information. And then configuring the application abnormal breakdown event and the target decision operation topological sub-structure into corresponding AI operation service nodes by combining the operation sequence of each target decision operation topological sub-structure, and respectively carrying out topology configuration and application breakdown decision, wherein the application abnormal breakdown event and the initial target decision operation topological sub-structure are configured into the initial AI operation service nodes to obtain initial decision operation execution data output by the initial AI operation service nodes, then configuring the initial decision operation execution data and the next target decision operation topological sub-structure into the AI operation service nodes corresponding to the topology splitting to obtain decision operation execution data of the topology splitting, and carrying out decision operation execution until the final target decision operation topological sub-structure carries out decision operation execution through the corresponding AI operation service nodes to obtain application realization error point decision data corresponding to the application abnormal breakdown event.

In an alternative implementation manner, when the big data storage system obtains each decision operation topological sub-structure, each decision operation topological sub-structure can be deployed on a corresponding AI operation service node in advance, when an application crash decision task is received, an application abnormal crash event is sent to the AI operation service node corresponding to the initial decision operation topological sub-structure, decision operation execution is carried out through the decision operation topological sub-structure in the AI operation service node, decision operation execution data is obtained, then the decision operation execution data is input to the next AI operation service node for decision operation execution, and decision operation execution is carried out through the AI operation service node corresponding to the decision operation topological sub-structure in combination with an operation sequence until application realization error point decision data corresponding to the application abnormal crash event is obtained through operation of the decision operation topological sub-structure in all AI operation service nodes.

According to the technical scheme, through obtaining the application collapse decision task, and then combining the mapping characteristic information to obtain each target decision operation topological sub-structure and corresponding sub-structure operation resource service information related to target decision signature data, as each target decision operation topological sub-structure is obtained by obtaining the branch influence value of each decision operation branch in the decision operation topological structure, then carrying out topological resolution on each decision operation branch by using the branch influence value corresponding to each decision operation branch, the branch connection information of each decision operation branch and the AI operation resource service information of each decision operation branch, obtaining a decision operation topological sub-structure and a decision operation branch to be processed, taking the decision operation topological sub-structure and the decision operation branch to be processed as the decision operation branch corresponding to an application collapse decision network respectively, obtaining the optimized decision operation topological structure and carrying out iterative execution until each decision operation branch in the decision operation topological structure is completely topologically split, thereby improving the precision of each obtained target decision operation topological sub-structure, further reducing the quantity of each decision operation topological sub-structure which is topologically split, then carrying out topological resolution on each target operation sub-structure by combining the application of abnormal event and the AI operation resource service information of the decision operation sub-structure to the corresponding decision operation sub-structure to the application collapse decision operation network, thereby improving the application error calculation service node can be applied to the application collapse decision operation network, and can realize the application error calculation of the application error node, and the application of the error calculation can be improved, and the error calculation service node can be improved, and the error can be applied to the application and the failure calculation network can be caused to realize to the corresponding to the false calculation of the failure calculation operation node, and the utilization of the operation resources of the AI operation service node is improved.

In an alternative embodiment, the above method is specifically:

STEP602, arranging all decision operation branches in the decision operation topological structure to obtain decision operation branch arrangement information; and calculating branch influence values corresponding to the decision operation branches by combining the branch connection information and the decision operation branch arrangement information of the decision operation branches. The method comprises the steps of acquiring AI operation resource service information of each decision operation branch, selecting a targeting decision operation branch from each decision operation branch, acquiring targeting prior topology splitting information of the targeting decision operation branch, and determining targeting operation resource service information corresponding to the targeting decision operation branch from the AI operation resource service information of each decision operation branch.

STEP604, performing prior topology splitting verification on the targeting decision operation branch by combining the targeting prior topology splitting information, and performing operation resource verification on the targeting decision operation branch by combining the targeting operation resource service information. When the verification of the target decision operation branch is unsuccessful, the target decision operation branch is used as a decision operation topology substructure, and the step of obtaining the branch influence value of each decision operation branch in the decision operation topology structure is returned to be executed iteratively until each decision operation branch in the decision operation topology structure completes the topology splitting.

STEP606, when the verification of the targeting decision operation branch is successful, the targeting decision operation branch is taken as a first decision operation branch, a forward decision operation branch and a backward decision operation branch corresponding to the first decision operation branch are obtained from each decision operation branch, and the forward decision operation branch and the backward decision operation branch are taken as second decision operation branches respectively.

STEP608 obtains the prior topology splitting information of the target of the second decision branch, determines the target branch influence value corresponding to the second decision branch from the branch influence values corresponding to the decision branches, determines the target AI operation resource service information corresponding to the second decision branch from the AI operation resource service information of the decision branches, and determines the target branch connection information corresponding to the second decision branch from the branch connection information of the decision branches.

STEP610 performs prior topology splitting verification on the second decision branch in combination with the target prior topology splitting information, performs operation resource verification on the second decision branch in combination with the target AI operation resource service information, and performs continuity verification on the second decision branch in combination with the target branch influence value and the target branch connection information.

STEP612, when any one of the prior topology split verification of the second decision operation branch, the operation resource verification of the second decision operation branch and the continuity verification of the second decision operation branch is unsuccessful, determining that the second decision operation branch verification is unsuccessful. And when the prior topology splitting verification of the second decision operation branch, the operation resource verification of the second decision operation branch and the continuity verification of the second decision operation branch are all passed, judging that the second decision operation branch is successful in verification.

STEP614, when the verification of the second decision operation branch is successful, taking the first decision operation branch and the second decision operation branch as updated first decision operation branches, and returning to the STEP of obtaining the second decision operation branch corresponding to the first decision operation branches for execution until the updated first decision operation branches are not changed any more, so as to obtain second decision operation branches with successful verification; combining the first decision operation branch and each second decision operation branch successful in verification to obtain a decision operation topology substructure, and combining each decision operation branch, the first decision operation branch and each second decision operation branch successful in verification to obtain a decision operation branch to be processed.

STEP616, taking the decision operation topological sub-structure and the decision operation branch to be processed as decision operation branches corresponding to the application breakdown decision network respectively, obtaining an optimized decision operation topological structure, and returning to the iterative execution of the STEP of obtaining the branch influence value of each decision operation branch in the decision operation topological structure until each decision operation branch in the decision operation topological structure completes the topological splitting, so as to obtain each target decision operation topological sub-structure and corresponding sub-structure operation resource service information; and establishing and storing the decision signature data corresponding to the decision operation topological structure, and the mapping characteristic information between each target decision operation topological sub-structure and the corresponding sub-structure operation resource service information.

STEP618 obtains an application crash decision task, which carries the target decision signature data and the application exception crash event. And acquiring each target decision operation topology sub-structure and corresponding sub-structure operation resource service information associated with the target decision signature data by combining the mapping characteristic information. And configuring each target decision operation topology sub-structure to a corresponding AI operation service node by combining the application abnormal breakdown event and the sub-structure operation resource service information to respectively perform topology configuration and application breakdown decision to obtain application implementation error point decision data corresponding to the application abnormal breakdown event.

In some design considerations, a big data storage system, which may be a server, is provided that includes a processor, a memory, an Input/Output interface (I/O) and a communication interface. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface. Wherein the processor of the big data storage system is configured to provide computing and control capabilities. The memory of the big data storage system includes a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the big data storage system is used for storing the data related to the method. The model-loaded data/output interface of the big data storage system is used to exchange information between the processor and the external device. The communication interface of the big data storage system is used for communicating with an external terminal through network connection. The computer program, when executed by a processor, implements an artificial intelligence decision-based application crash analysis method.

In some design considerations, a large data storage system is provided, which may be a terminal. The big data storage system includes a processor, a memory, an input/output interface, a communication interface, a display unit, and an input device. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input/output interface. Wherein the processor of the big data storage system is configured to provide computing and control capabilities. The memory of the big data storage system includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The model-loaded data/output interface of the big data storage system is used to exchange information between the processor and the external device. The communication interface of the big data storage system is used for carrying out wired or wireless communication with an external terminal, and the wireless mode can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. The computer program, when executed by a processor, implements an artificial intelligence decision-based application crash analysis method. The display unit of the big data storage system is used to form a visually viewable picture.

In some design considerations, a big data storage system is provided, comprising a memory and a processor, the memory storing a computer program, the processor implementing the steps of the method embodiments described above when executing the computer program.

In some design considerations, a computer readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the method embodiments described above.

In some design considerations, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of the method embodiments described above.

The above examples only represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the present application. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application shall be subject to the appended claims.

Claims

1. An application crash analysis method based on artificial intelligence decision, the method comprising:

Searching each application abnormal breakdown event with abnormal breakdown from an application response log of the digital service online application, carrying out application implementation error point analysis on each application abnormal breakdown event by combining an application breakdown decision network converged by network weight information to obtain application implementation error point decision data corresponding to the digital service online application, wherein the application breakdown decision network converged by the network weight information is applied to an AI dispatch service center, and the application implementation error point decision data comprises a plurality of application implementation error points and an abnormal breakdown description knowledge chain corresponding to each application implementation error point;

2. The method for analyzing the application crash based on the artificial intelligence decision according to claim 1, wherein the step of searching each application abnormal crash event with abnormal crash from the application response log of the digital service online application, and performing application implementation error point analysis on each application abnormal crash event by combining the application crash decision network converged by the network weight information to obtain application implementation error point decision data corresponding to the digital service online application comprises the following steps:

acquiring branch influence values of all decision operation branches in a decision operation topological structure corresponding to an application breakdown decision network, wherein the branch influence values are calculated by combining with connection information of all the decision operation branches;

acquiring AI operation resource service information of each decision operation branch, and carrying out topology splitting on each decision operation branch by combining a branch influence value corresponding to each decision operation branch, branch connection information of each decision operation branch and AI operation resource service information of each decision operation branch to obtain a decision operation topology substructure and a decision operation branch to be processed;

the decision operation topological sub-structure and the decision operation branch to be processed are respectively used as decision operation branches corresponding to an application breakdown decision network, an optimized decision operation topological structure is obtained, and iterative execution of the step of obtaining branch influence values of all decision operation branches in the decision operation topological structure is returned until all decision operation branches in the decision operation topological structure complete topological splitting, and then all target decision operation topological sub-structures and corresponding sub-structure operation resource service information are obtained;

Storing decision signature data corresponding to the decision operation topological structure, and mapping characteristic information between each target decision operation topological sub-structure and corresponding sub-structure operation resource service information;

obtaining an application crash decision task, wherein the application crash decision task carries target decision signature data and an application abnormal crash event;

acquiring each target decision operation topology sub-structure and corresponding sub-structure operation resource service information associated with the target decision signature data by combining the mapping characteristic information;

and configuring each target decision operation topology sub-structure to a corresponding AI operation service node by combining the application abnormal breakdown event and the sub-structure operation resource service information to respectively perform topology configuration and application breakdown decision to obtain application implementation error point decision data corresponding to the application abnormal breakdown event.

3. The artificial intelligence decision-based application crash analysis method as claimed in claim 2, wherein obtaining a branch impact value of each decision operation branch in the decision operation topology, the branch impact value being calculated in combination with each decision operation branch connection information, comprises:

Arranging all decision operation branches in the decision operation topological structure to obtain decision operation branch arrangement information;

determining a target decision operation branch from the decision operation branches according to the arrangement sequence of the decision operation branch arrangement information, and verifying whether the target decision operation branch has a corresponding associated decision operation branch according to the branch connection information of the decision operation branches;

when the target decision operation branch has a corresponding association decision operation branch, acquiring a branch influence value corresponding to the association decision operation branch, and determining the branch influence value corresponding to the target decision operation branch by combining the branch influence value corresponding to the association decision operation branch; and

when the target decision operation branch does not have a corresponding association decision operation branch, determining a branch influence value corresponding to the target decision operation branch as a second calibration value;

wherein, there are at least two said associated decision branches;

the obtaining the branch influence value corresponding to the association decision operation branch, and determining the branch influence value corresponding to the target decision operation branch by combining the branch influence value corresponding to the association decision operation branch includes:

Acquiring branch influence values corresponding to at least two associated decision operation branches respectively, and determining the maximum branch influence value from the branch influence values corresponding to the at least two associated decision operation branches respectively;

and calculating the sum of the maximum branch influence value and the first calibration value to obtain a branch influence value corresponding to the target decision operation branch.

4. The method of claim 2, wherein the performing topology splitting on the decision branches by combining the branch impact values corresponding to the decision branches, the branch connection information of the decision branches, and the AI operation resource service information of the decision branches to obtain a decision operation topology substructure and a decision operation branch to be processed comprises:

determining a first decision operation branch from the decision operation branches, and acquiring a second decision operation branch corresponding to the first decision operation branch from the decision operation branches;

acquiring target prior topology splitting information of the second decision operation branch, determining a target branch influence value corresponding to the second decision operation branch from branch influence values corresponding to the decision operation branches, determining target AI operation resource service information corresponding to the second decision operation branch from AI operation resource service information of the decision operation branches, and determining target branch connection information corresponding to the second decision operation branch from branch connection information of the decision operation branches;

Combining the target AI operation resource service information, and carrying out substructure decision operation branch verification on the second decision operation branch by the target prior topology splitting information, the target branch influence value and the target branch connection information;

when the verification of the second decision operation branch is successful, the first decision operation branch and the second decision operation branch are used as updated first decision operation branches, and the step of obtaining the second decision operation branch corresponding to the first decision operation branch is returned to be executed until the updated first decision operation branch is not changed any more, and each second decision operation branch which is successful in verification is obtained;

combining the first decision operation branch and the second decision operation branches which are successfully verified to obtain a decision operation topology substructure, and combining the decision operation branches, the first decision operation branch and the second decision operation branches which are successfully verified to obtain decision operation branches to be processed.

5. The method for crash analysis of an application based on artificial intelligence decision according to claim 4, wherein determining a first decision branch from the decision branches and obtaining a second decision branch corresponding to the first decision branch from the decision branches comprises:

Selecting a targeting decision operation branch from each decision operation branch, acquiring targeting prior topology splitting information of the targeting decision operation branch, and determining targeting operation resource service information corresponding to the targeting decision operation branch from AI operation resource service information of each decision operation branch;

performing substructure decision operation branch verification on the targeting decision operation branch by combining the targeting prior topology splitting information and the targeting operation resource service information;

when the target decision operation branch is successfully verified, the target decision operation branch is taken as the first decision operation branch, a forward decision operation branch and a backward decision operation branch corresponding to the first decision operation branch are obtained from each decision operation branch, and the forward decision operation branch and the backward decision operation branch are respectively taken as second decision operation branches;

when the verification of the target decision operation branch is unsuccessful, taking the target decision operation branch as a decision operation topology substructure, and returning to the step of obtaining branch influence values of all decision operation branches in the decision operation topology structure for iterative execution until all decision operation branches in the decision operation topology structure complete topology splitting;

The sub-structure decision operation branch verification of the target decision operation branch by combining the target prior topology splitting information and the target operation resource service information comprises the following steps:

performing prior topology splitting verification on the target decision operation branch by combining the target prior topology splitting information, and performing operation resource verification on the target decision operation branch by combining the target operation resource service information;

judging that the target decision operation branch is successfully verified when the previous topology splitting verification of the target decision operation branch is successful and the operation resource verification of the target decision operation branch is successful;

and judging that the target decision operation branch verification is unsuccessful when the previous topology splitting verification of the target decision operation branch is unsuccessful or the operation resource verification of the target decision operation branch is unsuccessful.

6. The artificial intelligence decision based application crash analysis method as claimed in claim 4, wherein said performing sub-structure decision branch validation on the second decision branch by combining the target AI operation resource service information, the target prior topology split information, the target branch impact value, and the target branch connection information comprises:

Performing prior topology splitting verification on the second decision operation branch by combining the target prior topology splitting information, performing operation resource verification on the second decision operation branch by combining the target AI operation resource service information, and performing continuity verification on the second decision operation branch by combining the target branch influence value and the target branch connection information;

judging that the second decision operation branch is successfully verified when the prior topology splitting verification of the second decision operation branch, the operation resource verification of the second decision operation branch and the continuity verification of the second decision operation branch are all passed;

and when any one of the prior topology splitting verification of the second decision operation branch, the operation resource verification of the second decision operation branch and the continuity verification of the second decision operation branch is unsuccessful, judging that the second decision operation branch is unsuccessful.

7. The method for analysis of breakdown of applications based on artificial intelligence decision according to claim 4, wherein after the sub-structural decision branch verification of the second decision branch by combining the target AI operation resource service information, the target previous topology splitting information and the target branch influence value, further comprises:

And when the verification of the second decision operation branch is unsuccessful, combining the first decision operation branch to obtain a decision operation topology substructure, and combining each decision operation branch with the first decision operation branch to obtain a decision operation branch to be processed.

8. A big data storage system, characterized in that the big data storage system comprises a processor and a memory for storing a computer program capable of running on the processor, the processor being adapted to execute the artificial intelligence decision based application crash analysis method according to any of claims 1-7 when the computer program is run.