A data storage method and system applied to cloud service platform

Abstract

The embodiment of the present application provides a data storage method and system for a cloud service platform. First, labeled and unlabeled cloud service operation data sequences are extracted, and a candidate error estimation network is generated using the labeled data to estimate the network error on the unlabeled data. Through this error parameter, samples are extracted from the unlabeled data for training label configuration, and the labeled data set is expanded. Then, the candidate network and the crash prediction network are cyclically optimized based on the extended data set to generate a target prediction network. Finally, this network is used to predict the crash of any cloud service data, and the critical path data is stored traceably. This method not only improves the accuracy of cloud service crash prediction, but also provides strong data support for cloud service optimization and troubleshooting through the storage of critical path data, thereby enhancing the stability and reliability of cloud services.

Description

Technical Field

[0001] The present application relates to the field of cloud service technology, and more specifically, to a data storage method and system applied to a cloud service platform. Background Art

[0002] With the rapid development and widespread application of cloud computing technology, cloud service platforms have become an important tool for enterprises and individual users to store and process data. However, cloud service platforms may encounter various problems and challenges during operation, among which the stability and reliability of cloud services are particularly critical. The collapse of cloud services will not only lead to data loss, but may also have a serious impact on user experience and business continuity.

[0003] In order to prevent the crash of cloud services, it is necessary to monitor and predict the operating status of cloud services in real time. However, traditional monitoring methods can usually only provide limited operating status information and it is difficult to accurately predict the crash risk of cloud services. In addition, the complexity and large scale of cloud service operation data also bring great challenges to data storage and analysis.

[0004] At present, although there are some methods that can be used to predict cloud service crashes, they are usually based on simple statistical models or machine learning algorithms, and the feature extraction and model training of cloud service operation data are not deep and comprehensive enough. These methods often have difficulty in achieving ideal prediction results when dealing with complex and nonlinear cloud service operation data.

[0005] At the same time, existing cloud service data storage methods usually simply store data in the cloud without in-depth analysis and processing of the data. This method not only takes up a lot of storage space, but also fails to effectively extract key information from the data, making it difficult to accurately monitor and predict the operating status of cloud services. Summary of the invention

[0006] In view of this, the purpose of the present application is to provide a data storage method and system for a cloud service platform.

[0007] In conjunction with the first aspect of the present application, a data storage method applied to a cloud service platform is provided, which is applied to a data storage system applied to a cloud service platform, and the method includes:

[0008] Extracting a first sample cloud service operation data sequence carrying cloud service crash point annotation data and a second sample cloud service operation data sequence not carrying cloud service crash point annotation data;

[0009] Based on the candidate error estimation network generated by the first sample cloud service operation data sequence, estimating the network error parameters of the candidate cloud service crash prediction network corresponding to the candidate error estimation network on the second sample cloud service operation data sequence, and generating a network error parameter estimation result;

[0010] Based on the network error parameter estimation result, extracting sample cloud service operation data from the second sample cloud service operation data sequence for training label configuration to expand the first sample cloud service operation data sequence;

[0011] cyclically optimizing the candidate error estimation network and the candidate cloud service crash prediction network based on the expanded first sample cloud service operation data sequence to generate a target cloud service crash prediction network;

[0012] Based on the target cloud service crash prediction network, cloud service crash prediction is performed on any input target cloud service operation data, a target cloud service crash prediction result corresponding to the target cloud service operation data is generated, and after tracing the critical path data in the target cloud service operation data based on the target cloud service crash prediction result, the critical path data is stored.

[0013] In a possible implementation of the first aspect, the second sample cloud service operation data sequence includes at least unsupervised sample cloud service operation data, and the network error parameter estimation result includes at least a network estimation error corresponding to the unsupervised sample cloud service operation data.

[0014] The step of extracting sample cloud service operation data from the second sample cloud service operation data sequence for training label configuration based on the network error parameter estimation result to expand the first sample cloud service operation data sequence includes:

[0015] Based on the size of each of the network estimation errors, selecting each reference sample cloud service operation data in the second sample cloud service operation data sequence;

[0016] Calibrate the crash points of the reference sample cloud service operation data to generate supervised sample cloud service operation data;

[0017] Each of the supervised sample cloud service operation data is added to the first sample cloud service operation data sequence to expand the first sample cloud service operation data sequence.

[0018] In a possible implementation of the first aspect, the step of calculating a network error parameter estimation result of the unsupervised sample cloud service operation data by using the candidate error estimation network includes:

[0019] Convert the output of each feature encoding unit in the candidate cloud service crash prediction network into an error estimation encoding vector respectively;

[0020] Integrating the error estimation coding vectors to obtain an integrated coding vector;

[0021] The integrated coding vector is fully connected and outputted to generate the network error parameter estimation result.

[0022] In a possible implementation of the first aspect, before the step of extracting the first sample cloud service operation data sequence and the second sample cloud service operation data sequence, the step includes:

[0023] Obtaining an unsupervised sample cloud service operation data sequence for crash point prediction;

[0024] Randomly extracting unsupervised sample cloud service operation data from the unsupervised sample cloud service operation data sequence to perform training label configuration, and dividing the unsupervised sample cloud service operation data sequence into a first sample cloud service operation data sequence and a second sample cloud service operation data sequence;

[0025] Based on the first sample cloud service operation data sequence, a candidate error estimation network and a candidate cloud service crash prediction network are constructed.

[0026] In a possible implementation of the first aspect, the step of cyclically optimizing the candidate error estimation network and the candidate cloud service crash prediction network based on the expanded first sample cloud service operation data sequence to generate a target cloud service crash prediction network includes:

[0027] Calculating a cloud service crash prediction error of the candidate cloud service crash prediction network on the expanded first sample cloud service operation data sequence;

[0028] Calculating, according to the cloud service crash prediction error, an error estimation loss value of the candidate error estimation network on the expanded first sample cloud service operation data sequence;

[0029] The target cloud service crash prediction network is generated according to the cloud service crash prediction error and the error estimation loss value.

[0030] In a possible implementation of the first aspect, the step of calculating, based on the cloud service crash prediction error, an error estimation loss value of the candidate error estimation network on the expanded first sample cloud service operation data sequence includes:

[0031] Based on the candidate error estimation network, a network error parameter of the candidate cloud service crash prediction network on the expanded first sample cloud service operation data sequence is estimated to generate an error estimation loss value.

[0032] In a possible implementation manner of the first aspect, the step of generating the target cloud service crash prediction network according to the cloud service crash prediction error and the error estimation loss value includes:

[0033] Calculating a global error parameter according to the cloud service crash prediction error and the error estimation loss value;

[0034] If the global error parameter does not continue to decrease, taking the candidate cloud service crash prediction network as the target cloud service crash prediction network;

[0035] If the global error parameter continues to decrease, the candidate cloud service crash prediction network is updated according to the network gradient calculated by the global error parameter, and the execution step is returned to: extracting the first sample cloud service operation data sequence and the second sample cloud service operation data sequence.

[0036] In a possible implementation of the first aspect, the step of tracing the critical path data in the target cloud service operation data based on the target cloud service crash prediction result includes:

[0037] Extract all path data from the target cloud service operation data, wherein the path data includes a function call chain, a data transmission path, and / or a resource access path;

[0038] Mapping the target cloud service crash prediction result to the extracted path data, and for each path data, evaluating the impact weight of the path data in causing the cloud service crash according to the crash prediction result corresponding to the path data, to obtain an impact weight evaluation result;

[0039] According to the impact weight evaluation result, path data having a significant impact on the cloud service crash is identified as the critical path data, wherein the impact weight of the critical path data is greater than a preset weight.

[0040] In combination with the second aspect of the present application, a data storage system applied to a cloud service platform is provided, wherein the data storage system applied to the cloud service platform includes a machine-readable storage medium and a processor, wherein the machine-readable storage medium stores machine-executable instructions, and when the processor executes the machine-executable instructions, the data storage system applied to the cloud service platform implements the aforementioned data storage method applied to the cloud service platform.

[0041] In conjunction with the third aspect of the present application, a computer-readable storage medium is provided, in which computer-executable instructions are stored. When the computer-executable instructions are executed, the aforementioned data storage method applied to the cloud service platform is implemented.

[0042] In combination with any of the above aspects, in the present application, by extracting the first sample cloud service operation data sequence carrying cloud service crash point annotation data and the second sample cloud service operation data sequence not carrying cloud service crash point annotation data, the operation status data of the cloud service can be obtained comprehensively and accurately, providing a rich and real data basis for subsequent network training and crash prediction. Secondly, the candidate error estimation network is generated using the first sample cloud service operation data sequence, and the network error parameters of the candidate cloud service crash prediction network on the second sample cloud service operation data sequence are estimated, which not only improves the accuracy of the error estimation, but also provides effective feedback for optimizing the cloud service crash prediction network. Thirdly, based on the network error parameter estimation result, the sample cloud service operation data is extracted from the second sample cloud service operation data sequence for training label configuration to expand the first sample cloud service operation data sequence, effectively increasing the amount of training data and improving the generalization ability and prediction accuracy of the training model. Then, by cyclically optimizing the candidate error estimation network and the candidate cloud service crash prediction network, a target cloud service crash prediction network is generated, so that the network can better adapt to the actual operating environment of the cloud service, and improve the accuracy and stability of the cloud service crash prediction. Finally, based on the target cloud service crash prediction network, cloud service crash prediction is performed on any input target cloud service operation data, and the target cloud service crash prediction results corresponding to the target cloud service operation data are generated. The critical path data in the target cloud service operation data is traced based on the prediction results, and the critical path data is stored. This can not only timely discover and prevent the risk of cloud service crashes, but also provide strong data support for subsequent cloud service optimization and problem troubleshooting by storing critical path data. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained by combining these drawings without paying creative work.

[0044] Figure 1 A flowchart of a data storage method applied to a cloud service platform provided in an embodiment of the present application. DETAILED DESCRIPTION

[0045] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all of the embodiments. According to the embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product or end that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units inherent to these processes, methods, products or ends.

[0047] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0048] Figure 1 The flowchart of the data storage method applied to the cloud service platform provided by the embodiment of the present application is shown. It should be understood that in other embodiments, the order of some steps of the data storage method applied to the cloud service platform of this embodiment can be shared with each other according to actual needs, or some steps can be omitted or maintained. The details of the data storage method applied to the cloud service platform include:

[0049] Step S110: extracting a first sample cloud service operation data sequence carrying cloud service crash point annotation data and a second sample cloud service operation data sequence not carrying cloud service crash point annotation data.

[0050] In this embodiment, the data storage system applied to the cloud service platform acts as a server and can extract data from the cloud service operation database. The data includes two parts: one part is the cloud service operation data with cloud service crash points marked, i.e., the first example cloud service operation data sequence; the other part is the cloud service operation data without cloud service crash points marked, i.e., the second example cloud service operation data sequence.

[0051] For example, the server retrieves 1,000 cloud service operation data from the database, of which 500 have crashed in previous operations and the crash points are accurately marked. These data constitute the first cloud service operation data sequence. The other 500 are normal operation data without crash points marked. These data constitute the second cloud service operation data sequence.

[0052] Step S120, based on the candidate error estimation network generated by the first sample cloud service operation data sequence, estimate the network error parameters of the candidate cloud service crash prediction network corresponding to the candidate error estimation network on the second sample cloud service operation data sequence, and generate a network error parameter estimation result.

[0053] In this embodiment, the server uses the first sample cloud service operation data sequence to generate a candidate error estimation network. Then, the server uses the candidate error estimation network to estimate the network error parameters of the corresponding candidate cloud service crash prediction network on the second sample cloud service operation data sequence.

[0054] For example, the server first trains a candidate error estimation network based on 500 data with crash points marked. Then, the server uses this network to estimate the error parameters of another 500 data without crash points marked, predicting the possible errors of these data if crashes occur, and generating a network error parameter estimation result.

[0055] Step S130: based on the network error parameter estimation result, extract sample cloud service operation data from the second sample cloud service operation data sequence for training label configuration to expand the first sample cloud service operation data sequence.

[0056] In this embodiment, based on the network error parameter estimation results, the server extracts cloud service operation data with larger network error parameters from the second sample cloud service operation data sequence, and configures training labels for these data (i.e., marking crash points), thereby expanding the first sample cloud service operation data sequence.

[0057] For example, based on the network error parameter estimation results, the server found that 100 pieces of data had particularly large error parameters, which means that these data are likely to hide unknown crash points. Therefore, the server configured crash point labels for these 100 pieces of data and added them to the first example cloud service operation data sequence, expanding the first example cloud service operation data sequence from the original 500 to 600.

[0058] Step S140: cyclically optimize the candidate error estimation network and the candidate cloud service crash prediction network based on the expanded first sample cloud service operation data sequence to generate a target cloud service crash prediction network.

[0059] In this embodiment, the server uses the expanded first sample cloud service operation data sequence to cyclically optimize the candidate error estimation network and the candidate cloud service crash prediction network until a target cloud service crash prediction network is generated.

[0060] For example, the server uses 600 cloud service operation data with crash points marked to train and optimize the candidate error estimation network and the candidate cloud service crash prediction network multiple times. After multiple iterations, the server's prediction performance reaches a stable and high level, and the server determines this optimized network as the target cloud service crash prediction network.

[0061] Step S150, based on the target cloud service crash prediction network, a cloud service crash prediction is performed on any input target cloud service operation data, a target cloud service crash prediction result corresponding to the target cloud service operation data is generated, and after tracing the critical path data in the target cloud service operation data based on the target cloud service crash prediction result, the critical path data is stored.

[0062] In this embodiment, the server uses the target cloud service crash prediction network to predict the cloud service crash of any input target cloud service operation data, and traces back to the critical path data in the target cloud service operation data based on the prediction results, and finally stores these critical path data.

[0063] For example, when a new cloud service operation data is input to the server, the server uses the target cloud service crash prediction network to predict the crash of this data. The prediction result shows that there is a critical path in the data that may cause a crash. The server further traces the specific data of this critical path (such as a specific function call chain or data transmission path) and extracts this data separately for storage for subsequent analysis and optimization.

[0064] Based on the above steps, by extracting the first sample cloud service operation data sequence carrying cloud service crash point annotation data and the second sample cloud service operation data sequence without cloud service crash point annotation data, the operation status data of the cloud service can be obtained comprehensively and accurately, providing a rich and real data basis for subsequent network training and crash prediction. Secondly, the candidate error estimation network is generated using the first sample cloud service operation data sequence, and the network error parameters of the candidate cloud service crash prediction network on the second sample cloud service operation data sequence are estimated, which not only improves the accuracy of error estimation, but also provides effective feedback for optimizing the cloud service crash prediction network. Thirdly, based on the network error parameter estimation results, sample cloud service operation data is extracted from the second sample cloud service operation data sequence for training label configuration to expand the first sample cloud service operation data sequence, effectively increasing the amount of training data and improving the generalization ability and prediction accuracy of the training model. Then, by cyclically optimizing the candidate error estimation network and the candidate cloud service crash prediction network, the target cloud service crash prediction network is generated, so that the network can better adapt to the actual operating environment of the cloud service and improve the accuracy and stability of cloud service crash prediction. Finally, based on the target cloud service crash prediction network, cloud service crash prediction is performed on any input target cloud service operation data, and the target cloud service crash prediction results corresponding to the target cloud service operation data are generated. The critical path data in the target cloud service operation data is traced based on the prediction results, and the critical path data is stored. This can not only timely discover and prevent the risk of cloud service crashes, but also provide strong data support for subsequent cloud service optimization and problem troubleshooting by storing critical path data.

[0065] In a possible implementation, the second sample cloud service operation data sequence includes at least unsupervised sample cloud service operation data, and the network error parameter estimation result includes at least a network estimation error corresponding to the unsupervised sample cloud service operation data.

[0066] The step of extracting sample cloud service operation data from the second sample cloud service operation data sequence for training label configuration based on the network error parameter estimation result to expand the first sample cloud service operation data sequence includes:

[0067] Based on the size of each of the network estimation errors, each reference sample cloud service operation data is selected in the second sample cloud service operation data sequence.

[0068] The crash point of each reference sample cloud service operation data is calibrated to generate each supervised sample cloud service operation data.

[0069] Each of the supervised sample cloud service operation data is added to the first sample cloud service operation data sequence to expand the first sample cloud service operation data sequence.

[0070] In a possible implementation, the step of calculating the network error parameter estimation result of the unsupervised sample cloud service operation data by using the candidate error estimation network includes:

[0071] The output of each feature encoding unit in the candidate cloud service crash prediction network is converted into an error estimation encoding vector.

[0072] The error estimation coding vectors are integrated to obtain an integrated coding vector.

[0073] The integrated coding vector is fully connected and outputted to generate the network error parameter estimation result.

[0074] In a possible implementation manner, before the step of extracting the first sample cloud service operation data sequence and the second sample cloud service operation data sequence, the step includes:

[0075] Obtain the unsupervised sample cloud service operation data sequence for crash point prediction.

[0076] Unsupervised sample cloud service operation data are randomly extracted from the unsupervised sample cloud service operation data sequence for training label configuration, and the unsupervised sample cloud service operation data sequence is divided into a first sample cloud service operation data sequence and a second sample cloud service operation data sequence.

[0077] Based on the first sample cloud service operation data sequence, a candidate error estimation network and a candidate cloud service crash prediction network are constructed.

[0078] In a possible implementation, the step of cyclically optimizing the candidate error estimation network and the candidate cloud service crash prediction network based on the expanded first sample cloud service operation data sequence to generate a target cloud service crash prediction network includes:

[0079] The cloud service crash prediction error of the candidate cloud service crash prediction network on the expanded first sample cloud service operation data sequence is calculated.

[0080] According to the cloud service crash prediction error, an error estimation loss value of the candidate error estimation network on the expanded first sample cloud service operation data sequence is calculated.

[0081] The target cloud service crash prediction network is generated according to the cloud service crash prediction error and the error estimation loss value.

[0082] In a possible implementation, the step of calculating the error estimation loss value of the candidate error estimation network on the expanded first sample cloud service operation data sequence according to the cloud service crash prediction error includes:

[0083] Based on the candidate error estimation network, a network error parameter of the candidate cloud service crash prediction network on the expanded first sample cloud service operation data sequence is estimated to generate an error estimation loss value.

[0084] In a possible implementation manner, the step of generating the target cloud service crash prediction network according to the cloud service crash prediction error and the error estimation loss value includes:

[0085] A global error parameter is calculated according to the cloud service crash prediction error and the error estimation loss value.

[0086] If the global error parameter does not continue to decrease, the candidate cloud service crash prediction network is used as the target cloud service crash prediction network.

[0087] If the global error parameter continues to decrease, the candidate cloud service crash prediction network is updated according to the network gradient calculated by the global error parameter, and the execution step is returned to: extracting the first sample cloud service operation data sequence and the second sample cloud service operation data sequence.

[0088] In a possible implementation, the step of tracing the critical path data in the target cloud service operation data based on the target cloud service crash prediction result includes:

[0089] All path data are extracted from the target cloud service operation data, where the path data includes a function call chain, a data transmission path, and / or a resource access path.

[0090] The target cloud service crash prediction result is mapped to the extracted path data. For each path data, the impact weight of the path data in causing the cloud service crash is evaluated according to the crash prediction result corresponding to the path data to obtain an impact weight evaluation result.

[0091] According to the impact weight evaluation result, path data having a significant impact on the cloud service crash is identified as the critical path data, wherein the impact weight of the critical path data is greater than a preset weight.

[0092] In the above embodiments, the data storage system applied to the cloud service platform for executing the above method embodiments has at least one processor, a control module (chip set) coupled to at least one of the (at least one) processors, a memory coupled to the control module, a non-volatile memory (NVM) / storage device coupled to the control module, at least one input / output device coupled to the control module, and a network interface coupled to the control module.

[0093] The processor may include at least one single-core or multi-core processor, and the processor may include any combination of general-purpose processors or special-purpose processors (such as graphics processors, application processors, baseband processors, etc.). For some alternative implementations, the data storage system applied to the cloud service platform can be used as the data storage system device applied to the cloud service platform such as the gateway described in the embodiments of the present application.

[0094] For some alternative embodiments, a data storage system applied to a cloud service platform may include at least one computer-readable medium (e.g., a memory or NVM / storage device) having instructions and at least one processor integrated with the at least one computer-readable medium and configured to execute instructions to implement a module to perform the actions described in the present disclosure.

[0095] For one embodiment, the control module may include any suitable interface controller to provide any suitable interface to at least one of the processor(s) and / or any suitable device or component in communication with the control module.

[0096] The control module may include a memory controller module to provide an interface to the memory. The memory controller module may be a hardware module, a software module, and / or a firmware module.

[0097] The memory may be used, for example, to load and store data and / or instructions for a data storage system applied to a cloud service platform. For one embodiment, the memory may include any suitable volatile memory, for example, a suitable DRAM.

[0098] For one embodiment, the control module may include at least one input / output controller to provide an interface to the NVM / storage device and the (at least one) input / output device.

[0099] For example, the NVM / storage device may be used to store data and / or instructions. The NVM / storage device may include any suitable non-volatile memory (e.g., flash memory) and / or may include any suitable (at least one) non-volatile storage device (e.g., at least one hard disk drive (HDD), at least one compact disk (CD) drive, and / or at least one digital versatile disk (DVD) drive).

[0100] The NVM / storage device may include storage resources that are physically part of a device on which a data storage system for a cloud service platform is installed, or that may be accessible to the device without being part of the device. For example, the NVM / storage device may be accessed via (at least one) input / output device over a network.

[0101] (At least one) input / output device may provide an interface for the data storage system applied to the cloud service platform to communicate with any other appropriate device, and the input / output device may include a communication component, a pinyin component, a sensor component, etc. The network interface may provide an interface for the data storage system applied to the cloud service platform to communicate based on at least one network, and the data storage system applied to the cloud service platform may wirelessly communicate with at least one component of a wireless network based on any standard and / or protocol of at least one wireless network standard and / or protocol, such as accessing a wireless network based on a communication standard.

[0102] For one embodiment, at least one of the (at least one) processor may be loaded together with the logic of at least one controller of the control module (e.g., a memory controller module). For one embodiment, at least one of the (at least one) processor may be loaded together with the logic of at least one controller of the control module to form a system level load. For one embodiment, at least one of the (at least one) processor may be integrated on the same die with the logic of at least one controller of the control module. For one embodiment, at least one of the (at least one) processor may be integrated on the same die with the logic of at least one controller of the control module to form a system on chip (SoC).

[0103] The embodiments of the present application are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of ​​the present application. At the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

[0104] An embodiment of the present invention discloses a computer-readable storage medium storing a computer program for electronic data exchange, wherein the computer program enables a computer to execute the steps of the data storage method applied to a cloud service platform described in the aforementioned embodiment.

[0105] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute the steps in the data storage method applied to a cloud service platform described in the aforementioned embodiment.

[0106] The device embodiments described above are only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative work.

[0107] Through the specific description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solution can be essentially or partly contributed to the prior art in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0108] Finally, it should be noted that what is disclosed above is only a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, it should be understood by those skilled in the art that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A data storage method applied to a cloud service platform, characterized in that: The method comprises: Extracting a first sample cloud service operation data sequence carrying cloud service crash point annotation data and a second sample cloud service operation data sequence not carrying cloud service crash point annotation data; Based on the candidate error estimation network generated by the first sample cloud service operation data sequence, estimating the network error parameters of the candidate cloud service crash prediction network corresponding to the candidate error estimation network on the second sample cloud service operation data sequence, and generating a network error parameter estimation result; Based on the network error parameter estimation result, extracting sample cloud service operation data from the second sample cloud service operation data sequence for training label configuration to expand the first sample cloud service operation data sequence; cyclically optimizing the candidate error estimation network and the candidate cloud service crash prediction network based on the expanded first sample cloud service operation data sequence to generate a target cloud service crash prediction network; Based on the target cloud service crash prediction network, a cloud service crash prediction is performed on any input target cloud service operation data, a target cloud service crash prediction result corresponding to the target cloud service operation data is generated, and after tracing back the critical path data in the target cloud service operation data based on the target cloud service crash prediction result, the critical path data is stored; The second sample cloud service operation data sequence includes at least unsupervised sample cloud service operation data, and the network error parameter estimation result includes at least a network estimation error corresponding to the unsupervised sample cloud service operation data. The step of extracting sample cloud service operation data from the second sample cloud service operation data sequence for training label configuration based on the network error parameter estimation result to expand the first sample cloud service operation data sequence includes: Based on the size of each of the network estimation errors, selecting each reference sample cloud service operation data in the second sample cloud service operation data sequence; Calibrate the crash points of the reference sample cloud service operation data to generate supervised sample cloud service operation data; Adding each of the supervised sample cloud service operation data to the first sample cloud service operation data sequence to expand the first sample cloud service operation data sequence; The step of cyclically optimizing the candidate error estimation network and the candidate cloud service crash prediction network based on the expanded first sample cloud service operation data sequence to generate a target cloud service crash prediction network includes: Calculating a cloud service crash prediction error of the candidate cloud service crash prediction network on the expanded first sample cloud service operation data sequence; Calculating, according to the cloud service crash prediction error, an error estimation loss value of the candidate error estimation network on the expanded first sample cloud service operation data sequence; generating the target cloud service crash prediction network according to the cloud service crash prediction error and the error estimation loss value; The step of generating the target cloud service crash prediction network according to the cloud service crash prediction error and the error estimation loss value comprises: Calculating a global error parameter according to the cloud service crash prediction error and the error estimation loss value; If the global error parameter does not continue to decrease, taking the candidate cloud service crash prediction network as the target cloud service crash prediction network; If the global error parameter continues to decrease, the candidate cloud service crash prediction network is updated according to the network gradient calculated by the global error parameter, and the execution step is returned to: extracting the first sample cloud service operation data sequence and the second sample cloud service operation data sequence.

2. The data storage method applied to the cloud service platform according to claim 1, characterized in that: The steps of calculating the network error parameter estimation results of the unsupervised sample cloud service operation data using the candidate error estimation network include: Convert the output of each feature encoding unit in the candidate cloud service crash prediction network into an error estimation encoding vector respectively; Integrating the error estimation coding vectors to obtain an integrated coding vector; The integrated coding vector is fully connected and outputted to generate the network error parameter estimation result.

3. The data storage method applied to the cloud service platform according to claim 1, characterized in that: The step of extracting the first sample cloud service operation data sequence and the second sample cloud service operation data sequence includes: Obtaining an unsupervised sample cloud service operation data sequence for crash point prediction; Randomly extracting unsupervised sample cloud service operation data from the unsupervised sample cloud service operation data sequence to perform training label configuration, and dividing the unsupervised sample cloud service operation data sequence into a first sample cloud service operation data sequence and a second sample cloud service operation data sequence; Based on the first sample cloud service operation data sequence, a candidate error estimation network and a candidate cloud service crash prediction network are constructed.

4. The data storage method applied to the cloud service platform according to claim 1, characterized in that: The step of calculating the error estimation loss value of the candidate error estimation network on the expanded first sample cloud service operation data sequence according to the cloud service crash prediction error comprises: Based on the candidate error estimation network, a network error parameter of the candidate cloud service crash prediction network on the expanded first sample cloud service operation data sequence is estimated to generate an error estimation loss value.

5. The data storage method applied to the cloud service platform according to claim 1, characterized in that: The step of tracing the critical path data in the target cloud service operation data based on the target cloud service crash prediction result includes: Extract all path data from the target cloud service operation data, wherein the path data includes a function call chain, a data transmission path, and / or a resource access path; Mapping the target cloud service crash prediction result to the extracted path data, and for each path data, evaluating the impact weight of the path data in causing the cloud service crash according to the crash prediction result corresponding to the path data, to obtain an impact weight evaluation result; According to the impact weight evaluation result, path data having a significant impact on the cloud service crash is identified as the critical path data, wherein the impact weight of the critical path data is greater than a preset weight.

6. A data storage system applied to a cloud service platform, characterized in that: It includes a processor and a computer-readable storage medium, wherein the computer-readable storage medium stores machine-executable instructions, and when the machine-executable instructions are executed by a computer, the data storage method applied to a cloud service platform as described in any one of claims 1 to 5 is implemented.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores machine-executable instructions, and when the machine-executable instructions are executed by a computer, the data storage method applied to a cloud service platform described in any one of claims 1 to 5 is implemented.

Images