Cloud phone fault prediction method, device and equipment

Abstract

The present disclosure provides a cloud phone fault prediction method, device and equipment, which relates to the field of artificial intelligence technology, specifically the field of deep learning technology, and can be applied to cloud phone resource management or operation and maintenance platforms. A specific implementation of the method includes: obtaining a feature vector sequence of the cloud phone in the current time period; inputting the feature vector sequence into a cloud phone fault prediction model to obtain the failure rate of the cloud phone in the future time period. This implementation can provide early warnings for cloud phones that may fail, so that operation and maintenance work can intervene before the cloud phone fails, reduce the failure rate of the cloud phone, and ensure the reliability and stability of the cloud phone service.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, specifically the field of deep learning technology, and can be applied to cloud phone resource management or operation and maintenance platforms. Background Art

[0002] As a new type of application, cloud phones can provide users with efficient and stable services in scenarios such as games, mobile office, advertising, artificial intelligence Internet of Things, and industrial Internet, based on ARM (Advanced RISC Machines) server architecture and virtualization technology, through virtual native mobile phone instances in the cloud. However, with the development of cloud phone business, facing the management of millions of cloud phone resource instances, cloud phone failures have become an inevitable problem in daily life, which not only increases the difficulty of cloud phone operation and maintenance, but also seriously affects service stability and product experience.

[0003] At present, the cloud phone service anomaly detection method is usually that the operation and maintenance personnel or developers define a time window. In this time window, if the service heartbeat of the cloud phone cannot be monitored, it will be warned that the cloud phone has a fault. Summary of the invention

[0004] The embodiments of the present disclosure provide a cloud phone fault prediction method, apparatus, device, storage medium, and program product.

[0005] In the first aspect, an embodiment of the present disclosure proposes a cloud phone fault prediction model training method, including: obtaining a sample feature vector sequence of a sample cloud phone in a first historical time period; obtaining a sample fault label of a sample cloud phone in a second historical time period, wherein the second historical time period is after the first historical time period; taking the sample feature vector sequence as input and the sample fault label as output, training a long short-term memory network LSTM to obtain a cloud phone fault prediction model.

[0006] In the second aspect, the disclosed embodiment proposes a cloud phone fault prediction method, including: obtaining a feature vector sequence of the cloud phone in a current time period; inputting the feature vector sequence into a cloud phone fault prediction model to obtain the failure rate of the cloud phone in a future time period, wherein the cloud phone fault prediction model is trained using the method described in the first aspect.

[0007] In the third aspect, the disclosed embodiment proposes a cloud phone fault prediction model training device, including: a first acquisition module, configured to obtain a sample feature vector sequence of a sample cloud phone within a first historical time period; a second acquisition module, configured to obtain a sample fault label of a sample cloud phone within a second historical time period, wherein the second historical time period is after the first historical time period; a training module, configured to take the sample feature vector sequence as input and the sample fault label as output, train a long short-term memory network LSTM, and obtain a cloud phone fault prediction model.

[0008] In the fourth aspect, an embodiment of the present disclosure proposes a cloud phone fault prediction device, comprising: an acquisition module, configured to obtain a feature vector sequence of the cloud phone in a current time period; a prediction module, configured to input the feature vector sequence into a cloud phone fault prediction model to obtain the failure rate of the cloud phone in a future time period, wherein the cloud phone fault prediction model is trained using the device described in the third aspect.

[0009] In a fifth aspect, an embodiment of the present disclosure proposes an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described in any implementation manner in the first aspect or the method described in any implementation manner in the second aspect.

[0010] In a sixth aspect, an embodiment of the present disclosure proposes a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute a method as described in any implementation manner in the first aspect or a method as described in any implementation manner in the second aspect.

[0011] In a seventh aspect, an embodiment of the present disclosure proposes a computer program product, including a computer program, which, when executed by a processor, implements the method described in any implementation manner in the first aspect or the method described in any implementation manner in the second aspect.

[0012] The cloud phone fault prediction method provided in the embodiment of the present disclosure uses a cloud phone fault prediction model to predict the probability of cloud phone failure in the future in order to improve the operation and maintenance efficiency of cloud phone failures and reduce the impact of cloud phone failures on services. It can provide early warning for cloud phones that may fail so that operation and maintenance work can intervene before the cloud phone fails, thereby reducing the failure rate of the cloud phone and ensuring the reliability and stability of the cloud phone service.

[0013] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Other features, purposes and advantages of the present disclosure will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure. Among them:

[0015] Figure 1 is a flowchart of an embodiment of a cloud phone fault prediction model training method according to the present disclosure;

[0016] Figure 2 It is the structural diagram of the cloud phone failure prediction model;

[0017] Figure 3 is a flowchart of another embodiment of a cloud phone fault prediction model training method according to the present disclosure;

[0018] Figure 4 is a flow chart of an embodiment of a cloud phone fault prediction method according to the present disclosure;

[0019] Figure 5 This is the training prediction flow chart of the cloud phone fault prediction model;

[0020] Figure 6 This is the application architecture diagram of the cloud phone fault prediction model;

[0021] Figure 7 It is a structural schematic diagram of an embodiment of a cloud phone fault prediction model training device according to the present disclosure;

[0022] Figure 8 is a structural schematic diagram of an embodiment of a cloud phone fault prediction device according to the present disclosure;

[0023] Fig. 9 It is a block diagram of an electronic device used to implement the cloud phone fault prediction model training method or cloud phone fault prediction method of the embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0025] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other. The present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0026] Figure 1 A flowchart of an embodiment of a cloud phone fault prediction model training method according to the present disclosure is shown. The cloud phone fault prediction model training method includes the following steps:

[0027] Step 101, obtaining a sample feature vector sequence of a sample cloud phone in a first historical time period.

[0028] In this embodiment, the execution subject of the cloud phone fault prediction model training method can obtain a sample feature vector sequence of the sample cloud phone in the first historical time period.

[0029] Usually, the execution subject is a cloud server, on which a large number of cloud phones are built. A cloud phone is a virtual phone, whose applications run on the cloud server and the data generated is also stored on the cloud server. The sample cloud phone can be at least part of the cloud phone built on the cloud server. Therefore, the execution subject can collect the historical feature data of the sample cloud phone in the first historical time period from the data it stores, and process the historical feature data according to the data feature engineering method to obtain the sample feature vector sequence [X1, X2…X t ]. Among them, the historical feature data may include feature data related to the performance of the cloud phone physical device and feature data of the cloud phone service monitoring. The feature data related to the performance of the cloud phone physical device may include but is not limited to at least one of the following: memory usage C, disk usage D, and network delay N (in milliseconds), etc. The cloud phone service monitoring feature data may include the time S (in seconds) from the service to the last heartbeat.

[0030] Since this application uses LSTM (Long Short-Term Memory) to train the cloud phone fault prediction model, the sample feature vector sequence [X1, X2…X t] depends on the number of input neurons of LSTM. Since the sample feature vector sequence is essentially a time series, and as a special recurrent neural network, LSTM is very suitable for time series prediction. Unlike general feedforward neural networks, LSTM can use time series to analyze inputs. In short, when using a feedforward neural network, the neural network will think that the content input at time T is completely unrelated to the content input at time T+1. However, LSTM can use the input information more reasonably by using the input at time T or before to process the time T+n. In addition, LSTM can also overcome the shortcomings of gradient explosion and attenuation encountered by traditional recurrent neural networks in back propagation. By adding memory units to the hidden layer, the short-term and long-term features of the time series are correlated with each other, and the deletion and storage of valuable information are controlled to form a memory network. Among them, each neural unit of LSTM contains an input gate, an output gate, and a forget gate. The storage unit is at the center of the entire neuron, and is mainly responsible for the transmission and output of neuron state values ​​at different times.

[0031] Step 102, obtaining sample fault labels of sample cloud phones in a second historical time period.

[0032] In this embodiment, the above-mentioned execution entity can obtain the sample fault labels of the sample cloud phones in the second historical time period.

[0033] Generally, if the sample cloud phone fails in the second historical time period, the value of its sample failure label is 1; if the sample cloud phone does not fail in the second historical time period, the value of its sample failure label is 0. The second historical time period is a time period immediately following the first historical time period. For example, the first historical time period is the past six months, and the second historical time period is one hour after the past six months.

[0034] It should be noted that training samples can be generated based on the sample feature vector sequence and the sample fault label, wherein the sample feature vector sequence is the input of the model and the sample fault label is the expected output of the model.

[0035] Step 103, taking the sample feature vector sequence as input and the sample fault label as output, training the long short-term memory network LSTM to obtain a cloud phone fault prediction model.

[0036] In this embodiment, the above-mentioned execution entity can take the sample feature vector sequence as input and the sample fault label as output, train the LSTM, and obtain the cloud phone fault prediction model.

[0037] Usually, the sample feature vector sequence [X1,X2…X t] is input into LSTM, and the sample failure rate can be learned. Based on the difference between the sample failure rate and the sample failure label, the parameters of LSTM are adjusted to make the difference small enough to obtain the cloud phone failure prediction model. Among them, the sample failure rate is the probability of the sample cloud phone predicted by LSTM to fail in the second historical time period.

[0038] Since the sample feature vector sequence is essentially a time series, and as a special recurrent neural network, LSTM is very suitable for time series prediction. It can overcome the shortcomings of gradient explosion and attenuation encountered by traditional recurrent neural networks in back propagation. By adding memory units to the hidden layer, the short-term and long-term features of the time series are correlated with each other, and the deletion and storage of valuable information are controlled to form a memory network. Among them, each neural unit of LSTM contains an input gate, an output gate, and a forget gate. The storage unit is at the center of the entire neuron, and is mainly responsible for the transmission and output of neuron state values ​​at different times.

[0039] Here, based on the LSTM training cloud phone fault prediction model, we take advantage of LSTM's memory advantage for short-term and long-term valuable feature information. On the basis of the cloud phone service heartbeat data features, we further integrate the data features of physical device memory usage, disk usage, and network delay that affect the cloud phone's operating status, and design the following: Figure 2 The structure diagram of the cloud phone fault prediction model shown in the figure includes an input layer, an LSTM network layer, and a classification output layer. Among them, Sigmoid is used as an activation function in the input layer, the LSTM network layer, and the classification output layer to generate a value between 0 and 1. Sigmoid can be used for hidden layer neuron output, with a value range of (0,1). It can map a real number to the interval of (0,1) and can be used for binary classification.

[0040] The cloud phone fault prediction model training method provided in the embodiment of the present disclosure is based on LSTM training of the cloud phone fault prediction model in order to improve the operation and maintenance efficiency of cloud phone faults and reduce the impact of cloud phone faults on services. The cloud phone fault prediction model is used to predict the probability of cloud phones failing in the future. Cloud phones that may fail can be warned in advance so that operation and maintenance work can intervene before the cloud phones fail, thereby reducing the failure rate of cloud phones and ensuring the reliability and stability of cloud phone services.

[0041] Continue to refer Figure 3 , which shows a flow chart of another embodiment of a cloud phone fault prediction model training method according to the present disclosure. The cloud phone fault prediction model training method includes the following steps:

[0042] Step 301, collecting historical feature data of a sample cloud phone in a first historical time period.

[0043] In this embodiment, the execution subject of the cloud phone fault prediction model training method can collect historical feature data of the sample cloud phone in the first historical time period. Among them, the historical feature data may include feature data related to the performance of the cloud phone physical device and feature data of the cloud phone service monitoring. The feature data related to the performance of the cloud phone physical device may include but is not limited to at least one of the following: memory usage C, disk usage D, and network delay N (in milliseconds), etc. The cloud phone service monitoring feature data may include the time S (in seconds) from the service to the last heartbeat.

[0044] Step 302: pre-process the historical feature data at every preset time interval as a sampling point to generate a historical feature vector for each sampling point.

[0045] In this embodiment, the execution subject may preprocess the historical feature data at a preset time interval as a sampling point to generate a historical feature vector for each sampling point. The preprocessing may include but is not limited to at least one of the following: statistics, data cleaning, and normalization.

[0046] For example, every minute is a sampling point, and the historical feature data of the sample cloud phone in the past six months is counted and cleaned, and then each feature vector obtained is normalized by Min-Max standardization, so that a sample feature vector X can be generated every minute. t =(C t ,D t ,N t ,S t ). Among them, C t is the characteristic vector of memory usage C, D t is the feature vector of disk usage D, N t is the characteristic vector of network delay N, S t It is the feature vector of the time S between the service and the last heartbeat.

[0047] Step 303: Generate a sample feature vector sequence based on the historical feature vector of each sampling point.

[0048] In this embodiment, the execution entity may generate a sample feature vector sequence based on the historical feature vector of each sampling point.

[0049] Usually, based on n sample feature vectors X t , and finally generate a sample feature vector sequence [X1,X2…X n ]. By sampling historical feature data for preprocessing, the generated sample feature vector sequence is made more standardized, which is convenient for subsequent data processing.

[0050] Step 304: Obtain sample fault labels of sample cloud phones in the second historical time period.

[0051] In this embodiment, the specific operation of step 304 is already described in Figure 1 In the illustrated embodiment, step 102 is described in detail and will not be repeated here.

[0052] Step 305: input the sample feature vector sequence into LSTM to obtain the sample failure rate of the sample cloud phone in the second historical time period.

[0053] In this embodiment, the execution subject may input the sample feature vector sequence into LSTM to obtain the sample failure rate of the sample cloud phone in the second historical time period, wherein the sample failure rate is the probability of failure of the sample cloud phone in the second historical time period predicted by LSTM.

[0054] Typically, based on the sample feature vector sequence, the LSTM is forward propagated to obtain the sample failure rate of the sample cloud phone in the second historical time period.

[0055] Step 306 , calculating a binary cross entropy loss function based on the sample fault label and the sample failure rate.

[0056] In this embodiment, the execution entity may calculate a binary cross-entropy loss function based on the sample fault label and the sample failure rate.

[0057] Here, the cloud phone failure prediction model can predict whether the cloud phone will fail in the future time period, which is a binary classifier. When training a binary classifier, Binary Cross-Entroy is often used as a loss function. Among them, Binary Cross-Entroy is a commonly used loss function in binary classification problems, and its formula is as follows:

[0058]

[0059] Among them, N is the number of training samples, y i is the sample fault label of the i-th training sample, whose value is 0 or 1, p(y i ) is the sample failure rate predicted using the i-th training sample. Binary Cross-Entroy can be used to judge the quality of the prediction results of the binary classifier. That is, for the case where the label y is 1, if the predicted value p(y) approaches 1, then the value of Binary Cross-Entroy should approach 0. Conversely, if the predicted value p(y) approaches 0, then the value of BinaryCross-Entroy should be very large.

[0060] Step 307, iteratively update the model parameters of LSTM through the adaptive moment estimation optimizer until the binary cross entropy loss function converges to obtain the cloud phone fault prediction model.

[0061] In this embodiment, the above-mentioned execution entity can iteratively update the model parameters of LSTM through the Adam (adaptive moment estimation) optimizer until the binary cross entropy loss function converges to obtain a cloud phone fault prediction model.

[0062] Usually, based on the binary cross entropy loss function, LSTM is back-propagated to obtain the gradient, and the Adam optimizer is used to update the model parameters of LSTM based on the gradient. After multiple rounds of training, a cloud phone fault prediction model with prediction accuracy that meets the requirements is obtained. Among them, Adam is a first-order optimization algorithm that can replace the traditional stochastic gradient descent process. It can iteratively update the weights of the neural network based on the training data. Adam is different from traditional stochastic gradient descent. Traditional stochastic gradient descent maintains a single learning rate to update all weights, and the learning rate does not change during the training process. Adam designs independent adaptive learning rates for different parameters by calculating the first-order moment estimate and the second-order moment estimate of the gradient.

[0063] from Figure 3 It can be seen that Figure 1 Compared with the corresponding embodiments, the flowchart of the cloud phone fault prediction method in this embodiment highlights the sample feature vector sequence generation step and the model training step. Therefore, the solution described in this embodiment integrates more feature attributes such as service heartbeat data, memory usage, disk usage, and network delay, and calculates more reasonable model parameters through Adam optimizer training, so that the short-term prediction results of the final model are more accurate and more timely.

[0064] Further references Figure 4 , which shows a flow chart of an embodiment of a cloud phone fault prediction method according to the present disclosure. The cloud phone fault prediction method comprises the following steps:

[0065] Step 401, obtaining the feature vector sequence of the cloud phone in the current time period.

[0066] In this embodiment, the execution subject of the cloud phone fault prediction method can obtain the feature vector sequence of the cloud phone in the current time period.

[0067] Typically, the feature data of the cloud phone is collected in real time, and the feature data is processed according to the method of data feature engineering to obtain a feature vector sequence. Among them, the feature data may include feature data related to the performance of the cloud phone physical device and feature data of the cloud phone service monitoring. The feature data related to the performance of the cloud phone physical device may include but is not limited to at least one of the following: memory usage C, disk usage D, and network delay N (in milliseconds), etc. The cloud phone service monitoring feature data may include the time S (in seconds) from the service to the last heartbeat.

[0068] Usually, every preset time is a sampling point, and the feature data is preprocessed to generate a feature vector for each sampling point. Based on the feature vector of each sampling point, a feature vector sequence can be generated. For example, every one minute is a sampling point, and the feature data of the cloud phone in the current time period is counted and cleaned, and then each feature vector obtained is normalized by Min-Max standardization, so that a feature vector can be generated every minute. Based on the feature vector of each sampling point, a feature vector sequence is generated.

[0069] Step 402, input the feature vector sequence into the cloud phone failure prediction model to obtain the failure rate of the cloud phone in the future time period.

[0070] In this embodiment, the above-mentioned execution subject can input the feature vector sequence into the cloud phone failure prediction model to obtain the failure rate of the cloud phone in the future time period. Figure 1 or Figure 3 The method shown in the figure is used for training and will not be described here.

[0071] The cloud phone fault prediction method provided in the embodiment of the present disclosure uses a cloud phone fault prediction model to predict the probability of cloud phone failure in the future in order to improve the operation and maintenance efficiency of cloud phone failures and reduce the impact of cloud phone failures on services. It can provide early warning for cloud phones that may fail so that operation and maintenance work can intervene before the cloud phone fails, thereby reducing the failure rate of cloud phones, ensuring the reliability and stability of cloud phone services, and ultimately ensuring a better product experience.

[0072] For ease of understanding, Figure 5 The training and prediction flow chart of the cloud phone fault prediction model is shown. Figure 5 As shown, the historical feature data of the sample cloud phone in the first historical time period is collected as a training data set. The training data set is preprocessed to obtain a processed data set. The processed data set is used for model training and model verification to obtain a prediction model. The real-time data of the cloud phone in the current time period is collected. The real-time data is input into the prediction model to obtain a prediction result.

[0073] For ease of understanding, Figure 6 The application architecture diagram of the cloud phone fault prediction model is shown in FIG. Figure 6 As shown in the figure, the cloud phone is based on the ARM architecture server (IAAS) and provides services to users. The monitoring center of the ARM cloud phone operation and maintenance management platform collects the feature data of the cloud phone in real time, and processes the feature data according to the data feature engineering method to obtain the feature vector sequence. The feature vector sequence is input into the prediction model in the fault prediction service to obtain the failure rate of the cloud phone in the future time period.

[0074] Further references Figure 7 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a cloud phone fault prediction model training device. Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0075] like Figure 7 As shown, the cloud phone fault prediction model training device 700 of this embodiment may include: a first acquisition module 701, a second acquisition module 702 and a training module 703. Among them, the first acquisition module 701 is configured to obtain a sample feature vector sequence of a sample cloud phone in a first historical time period; the second acquisition module 702 is configured to obtain a sample fault label of a sample cloud phone in a second historical time period, wherein the second historical time period is after the first historical time period; the training module 703 is configured to take the sample feature vector sequence as input and the sample fault label as output, train the long short-term memory network LSTM, and obtain a cloud phone fault prediction model.

[0076] In this embodiment, in the cloud phone fault prediction model training device 700: the specific processing of the first acquisition module 701, the second acquisition module 702 and the training module 703 and the technical effects thereof can be referred to respectively. Figure 1 The relevant descriptions of steps 101 - 103 in the corresponding embodiment are not repeated here.

[0077] In some optional implementations of this embodiment, the first acquisition module 701 is further configured to: collect historical feature data of the sample cloud phone within the first historical time period; pre-process the historical feature data at every preset time period as a sampling point to generate a historical feature vector for each sampling point, wherein the preprocessing includes at least one of the following: statistics, data cleaning and normalization; and generate a sample feature vector sequence based on the historical feature vector of each sampling point.

[0078] In some optional implementations of the present embodiment, the training module 703 is further configured to: input the sample feature vector sequence into the LSTM to obtain the sample failure rate of the sample cloud phone in the second historical time period; calculate the binary cross entropy loss function based on the sample fault label and the sample failure rate; iteratively update the model parameters of the LSTM through the adaptive moment estimation optimizer until the binary cross entropy loss function converges to obtain a cloud phone failure prediction model.

[0079] In some optional implementations of this embodiment, the historical feature data includes feature data related to the performance of the cloud phone physical device and feature data for cloud phone service monitoring. The feature data related to the performance of the cloud phone physical device includes at least one of the following: memory usage, disk usage, and network latency. The feature data for cloud phone service monitoring includes the time since the service last heartbeat.

[0080] In some optional implementations of this embodiment, the LSTM includes an input layer, an LSTM network layer, and a classification output layer.

[0081] Further references Figure 8 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a cloud phone fault prediction device. Figure 4 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0082] like Figure 8 As shown, the cloud phone fault prediction device 800 of this embodiment may include: an acquisition module 801 and a prediction module 802. The acquisition module 801 is configured to acquire the feature vector sequence of the cloud phone in the current time period; the prediction module 802 is configured to input the feature vector sequence into the cloud phone fault prediction model to obtain the failure rate of the cloud phone in the future time period, wherein the cloud phone fault prediction model is based on Figure 8 The device shown is trained.

[0083] In this embodiment, in the cloud phone fault prediction device 800, the specific processing of the acquisition module 801 and the prediction module 802 and the technical effects thereof can be referred to respectively. Figure 4 The relevant descriptions of steps 401 - 402 in the corresponding embodiment are not repeated here.

[0084] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0085] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0086] Fig. 9 A schematic block diagram of an example electronic device 900 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0087] like Fig. 9 As shown, the device 900 includes a computing unit 901, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 902 or a computer program loaded from a storage unit 908 into a random access memory (RAM) 903. In the RAM 903, various programs and data required for the operation of the device 900 can also be stored. The computing unit 901, the ROM 902, and the RAM 903 are connected to each other via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0088] A number of components in the device 900 are connected to the I / O interface 905, including: an input unit 906, such as a keyboard, a mouse, etc.; an output unit 907, such as various types of displays, speakers, etc.; a storage unit 908, such as a disk, an optical disk, etc.; and a communication unit 909, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 909 allows the device 900 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0089] The computing unit 901 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 901 performs the various methods and processes described above, such as a cloud phone fault prediction method or a cloud phone fault prediction model training method. For example, in some embodiments, the cloud phone fault prediction method or the cloud phone fault prediction model training method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 908. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 900 via the ROM 902 and / or the communication unit 909. When the computer program is loaded into the RAM 903 and executed by the computing unit 901, one or more steps of the cloud phone fault prediction method or the cloud phone fault prediction model training method described above may be performed. Alternatively, in other embodiments, the computing unit 901 may be configured to execute the cloud phone fault prediction method or the cloud phone fault prediction model training method in any other appropriate manner (for example, by means of firmware).

[0090] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0091] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0092] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0093] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0094] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0095] A computer system may include a client and a server. The client and the server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises through computer programs running on respective computers and having a client-server relationship to each other. The server may be a cloud server, a server in a distributed system, or a server combined with a blockchain.

[0096] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions provided by this disclosure can be achieved, and this document does not limit this.

[0097] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A cloud phone fault prediction model training method, comprising: Obtain a sample feature vector sequence of the sample cloud phone in the first historical time period; Obtaining sample fault labels of the sample cloud phone in a second historical time period, wherein the second historical time period is after the first historical time period; The sample feature vector sequence is input into the long short-term memory network LSTM to obtain the sample failure rate of the sample cloud phone in the second historical time period; based on the sample failure label and the sample failure rate, the binary cross entropy loss function is calculated; the model parameters of the LSTM are iteratively updated through an adaptive moment estimation optimizer until the binary cross entropy loss function converges to obtain a cloud phone failure prediction model, wherein the LSTM includes an input layer, an LSTM network layer and a classification output layer, and Sigmoid is used as an activation function in the input layer, the LSTM network layer and the classification output layer to generate a value between 0 and 1.

2. The method according to claim 1, wherein: The step of obtaining a sample feature vector sequence of a sample cloud phone in a first historical time period includes: Collecting historical feature data of the sample cloud phone in the first historical time period; Preprocessing the historical feature data at every preset time interval as a sampling point to generate a historical feature vector for each sampling point, wherein the preprocessing includes at least one of the following: statistics, data cleaning and normalization; The sample feature vector sequence is generated based on the historical feature vector of each sampling point.

3. The method according to claim 2, wherein: The historical feature data includes feature data related to the performance of the cloud phone physical device and feature data of the cloud phone service monitoring. The feature data related to the performance of the cloud phone physical device includes at least one of the following: memory usage, disk usage and network delay. The feature data of the cloud phone service monitoring includes the time from the service to the last heartbeat.

4. A cloud phone fault prediction method, comprising: Get the feature vector sequence of the cloud phone in the current time period; The feature vector sequence is input into a cloud phone failure prediction model to obtain a failure rate of the cloud phone in a future time period, wherein the cloud phone failure prediction model is trained using the method described in any one of claims 1-3.

5. A cloud phone fault prediction model training device, comprising: A first acquisition module is configured to acquire a sample feature vector sequence of a sample cloud phone in a first historical time period; A second acquisition module is configured to acquire sample fault tags of the sample cloud phone in a second historical time period, wherein the second historical time period is after the first historical time period; The training module is configured to input the sample feature vector sequence into the long short-term memory network LSTM to obtain the sample failure rate of the sample cloud phone in the second historical time period; calculate the binary cross entropy loss function based on the sample failure label and the sample failure rate; iteratively update the model parameters of the LSTM through an adaptive moment estimation optimizer until the binary cross entropy loss function converges to obtain a cloud phone failure prediction model, wherein the LSTM includes an input layer, an LSTM network layer and a classification output layer, and Sigmoid is used as an activation function in the input layer, the LSTM network layer and the classification output layer to generate a value between 0 and 1.

6. The device according to claim 5, wherein: The first acquisition module is further configured to: Collecting historical feature data of the sample cloud phone in the first historical time period; Preprocessing the historical feature data at every preset time interval as a sampling point to generate a historical feature vector for each sampling point, wherein the preprocessing includes at least one of the following: statistics, data cleaning and normalization; The sample feature vector sequence is generated based on the historical feature vector of each sampling point.

7. The device according to claim 6, wherein: The historical feature data includes feature data related to the performance of the cloud phone physical device and feature data of the cloud phone service monitoring. The feature data related to the performance of the cloud phone physical device includes at least one of the following: memory usage, disk usage and network delay. The feature data of the cloud phone service monitoring includes the time from the service to the last heartbeat.

8. A cloud phone fault prediction device, comprising: An acquisition module is configured to acquire a feature vector sequence of the cloud phone in a current time period; The prediction module is configured to input the feature vector sequence into a cloud phone fault prediction model to obtain the failure rate of the cloud phone in a future time period, wherein the cloud phone fault prediction model is trained using the device described in any one of claims 5-7.

9. An electronic device, comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 3 or the method of claim 4.

10. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the method of any one of claims 1 to 3 or the method of claim 4.

11. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the method according to any one of claims 1 to 3 or the method according to claim 4.

Images