Cloud computer service resource management method and device, equipment, medium and product

Abstract

The invention provides a cloud computer service resource management method and device, equipment, a medium and a product, and the method comprises the steps: constructing a user portrait according to a historical user demand and a historical operation record; performing optimization processing on any user portrait based on a neural network model; performing multi-objective optimization solution on the resource demand based on a Pareto optimization algorithm to determine resource configuration information corresponding to any user portrait; and according to the resource configuration information corresponding to the user portrait, performing management operation on the cloud computer service resources, wherein the management operation comprises at least one of resource scheduling, resource migration, resource capacity expansion and resource capacity reduction. According to the embodiment of the invention, the method can better adapt to the use scenes and use preferences of different users for the service resources of the cloud computer, and is better in robustness and adaptability.

Description

Background Art

[0002] At present, the rapid rise of cloud computers is due to the high development of cloud computing technology. Cloud computing virtualizes a large amount of computing, storage and network resources, providing a basis for flexible resource allocation and on-demand resource scheduling of cloud computers.

[0003] In the related technology, during the interaction between two network devices, the data obtained is composed of cloud computer resource data, specifically including the number of pre-prepared cloud computers, the number of cloud computers that can be prepared, and the ratio of the number of cloud computers that need to be prepared to the number of cloud computers that can be prepared, so as to determine the pre-configuration plan for each type of cloud computer.

[0004] However, the existing cloud computer configuration solutions are all configured or predicted based on the number of cloud computers (needed to be prepared / can be prepared / can be sold), and the data involved is single-sided, which cannot effectively improve the usage feedback of cloud computer service resources.

[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0006] The purpose of the present disclosure is to provide a method, device, equipment, medium and product for managing cloud computer service resources, which are used to overcome the problems of poor user experience of cloud computer service resources caused by the limitations and defects of related technologies at least to a certain extent.

[0007] According to a first aspect of an embodiment of the present disclosure, a method for managing cloud computer service resources is provided, comprising: constructing a user portrait based on historical user demands and historical operation records; optimizing any of the user portraits based on a neural network model; performing multi-objective optimization on the resource demands based on a Pareto optimization algorithm to determine resource configuration information corresponding to any of the user portraits; and performing management operations on the cloud computer service resources based on the resource configuration information corresponding to the user portrait, the management operations comprising at least one of resource scheduling, resource migration, resource expansion and resource reduction.

[0008] In an exemplary embodiment of the present disclosure, constructing a user profile according to historical user needs and historical operation records includes:

[0009] Obtain the historical operation record through the API interface and / or log record associated with the resources required by the cloud computer service;

[0010] Determining demand information corresponding to the historical operation record;

[0011] classifying the demand information according to the attribute information of the historical operation record;

[0012] Generate corresponding user portraits based on the classification results.

[0013] In an exemplary embodiment of the present disclosure, generating a corresponding user portrait according to the classification result includes:

[0014] Classifying the demand information according to the attribute information of the historical operation record, wherein the attribute information includes at least one of the user's usage scenario, time period, price sensitivity, and resource usage requirements;

[0015] Determining the user's usage pattern and / or application preference information of resources required for the cloud computing service based on the historical operation records;

[0016] Identify emotional tendency information in the preference information based on a natural language algorithm;

[0017] Determining the user's demand information for resources required for the cloud computer service based on the emotional tendency information;

[0018] The users are clustered based on a dynamic DBSCAN clustering algorithm to divide the users into multiple categories.

[0019] In an exemplary embodiment of the present disclosure, optimizing any of the user portraits based on a neural network model includes:

[0020] Training the neural network model based on the demand information and the behavior data of multiple categories of users to obtain a user behavior pattern-demand model;

[0021] triggering the user behavior pattern-demand model to predict user needs, and performing at least one of identification, merging, and evaluation on the predicted demand information;

[0022] The user portrait is optimized according to the predicted demand information.

[0023] In an exemplary embodiment of the present disclosure, a multi-objective optimization solution is performed on the resource demand based on the Pareto optimization algorithm to determine the resource configuration information corresponding to any of the user portraits, including:

[0024] Determining the quality of the cloud computing service based on the response time of the application loaded by the cloud computing service and / or the network throughput;

[0025] Determine resource costs required for the operation of the cloud computer service as resource costs, wherein the resource costs include at least one of computing resources, storage resources, and network resources;

[0026] Determine user complaints based on service delays and user satisfaction with the cloud computing service;

[0027] The quality of the cloud computer service and the resource cost are configured as input parameters of the Pareto optimization algorithm, and the minimum value of the user complaints is configured as the target output result of the Pareto optimization algorithm;

[0028] A multi-objective optimization solution is performed on the predicted resource demand according to the configured Pareto optimization algorithm to determine the resource configuration information corresponding to any of the user portraits.

[0029] In an exemplary embodiment of the present disclosure, the formula of the Pareto optimization algorithm includes:

[0030]

[0031] sD i ≤D Response ;

[0032] T i ≤T thr ;

[0033] Q i,1 ≥Q CUP ;

[0034] Q i,2 ≥Q Memory ;

[0035] Q i,3 ≥Q Storyage ;

[0036] Q i,4 ≥Q Network ;

[0037]

[0038] Said The calculation result of the minimum value of user complaints is characterized by the d i represents the service delay of the user performing the i-th operation through the cloud computer service, the S i represents the complaint rate of the i-th operation, a and b both represent impact factors, the value range of a and b is (0,1), and they satisfy the constraint of a+b=1, and D Response Characterize the application response time threshold of the cloud computer service, the T i Characterizes the network throughput required for the i-th operation, the T thr Characterizes the network throughput threshold required for the i-th operation, the Q i,1 , Q i,2 , Qi,3 Q i,4 Respectively represent the CPU, memory, storage, and network bandwidth required for the i-th operation. CPU Q Memory Q Storage Q Network Respectively represent the minimum resource requirements of the cloud computing service, the C i,1 , said C i,2 , said C i,3 , said C i,4 Respectively represent the cost of CPU, memory, storage, and network bandwidth required for the i-th operation. Total Characterizes the total resource cost of the cloud computing service.

[0039] In an exemplary embodiment of the present disclosure, performing the management operation of the cloud computer service resources according to the resource configuration information corresponding to the user portrait includes:

[0040] In the process of scheduling resources required for the cloud computer service, collecting usage feedback from the user, where the usage feedback is positive feedback or negative feedback;

[0041] Determining the validity of the usage feedback according to the resource utilization rate of the resources required by the cloud computer service corresponding to the usage feedback;

[0042] The resources required for the cloud computer service of the user with negative feedback of effectiveness are rescheduled until the usage feedback becomes the positive feedback.

[0043] According to a second aspect of an embodiment of the present disclosure, there is provided a cloud computer service resource management device, comprising:

[0044] A construction module is configured to construct a user profile based on historical user needs and historical operation records;

[0045] An optimization module, configured to optimize any of the user portraits based on a neural network model;

[0046] A calculation module is configured to perform multi-objective optimization on the resource requirements based on a Pareto optimization algorithm to determine resource configuration information corresponding to any of the user portraits;

[0047] The management module is configured to perform management operations on the cloud computer service resources according to the resource configuration information corresponding to the user portrait, and the management operations include at least one of resource scheduling, resource migration, resource expansion and resource reduction.

[0048] According to a third aspect of the present disclosure, there is provided an electronic device, comprising: a memory; and a processor coupled to the memory, wherein the processor is configured to execute any one of the methods described above based on instructions stored in the memory.

[0049] According to a fourth aspect of the present disclosure, there is provided a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements a method for managing cloud computer service resources as described in any one of the above.

[0050] According to a fifth aspect of the present disclosure, there is provided a computer program product, comprising a computer program, characterized in that when the computer program is executed by a processor, the computer program implements a method for managing cloud computer service resources as described in any one of the above.

[0051] The disclosed embodiment constructs a user portrait based on historical user demands and historical operation records, optimizes any of the user portraits based on a neural network model, and performs multi-objective optimization on the resource demands based on a Pareto optimization algorithm to determine resource configuration information corresponding to any of the user portraits. Management operations on the cloud computer service resources are performed based on the resource configuration information corresponding to the user portrait, so that the cloud computer service resources are better adapted to the usage scenarios and usage preferences of different users for the cloud computer service resources, and have better robustness and adaptability.

[0052] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0054] Figure 1 A schematic diagram showing an exemplary system architecture of a cloud computer service resource management solution to which an embodiment of the present invention can be applied;

[0055] Figure 2 is a flow chart of a method for managing cloud computer service resources in an exemplary embodiment of the present disclosure;

[0056] Figure 3 is a flow chart of a method for managing cloud computer service resources in an exemplary embodiment of the present disclosure;

[0057] Figure 4is a flow chart of a method for managing cloud computer service resources in an exemplary embodiment of the present disclosure;

[0058] Figure 5 is a flow chart of a method for managing cloud computer service resources in an exemplary embodiment of the present disclosure;

[0059] Figure 6 is a flow chart of a method for managing cloud computer service resources in an exemplary embodiment of the present disclosure;

[0060] Figure 7 is a flow chart of a method for managing cloud computer service resources in an exemplary embodiment of the present disclosure;

[0061] Figure 8 is a flow chart of a cloud computer service resource management solution in an exemplary embodiment of the present disclosure;

[0062] Fig. 9 It is a schematic diagram of the architecture of a cloud computer service resource management solution in an exemplary embodiment of the present disclosure;

[0063] Fig.10 is a block diagram of a cloud computer service resource management device in an exemplary embodiment of the present disclosure;

[0064] Fig.11 is a block diagram of an electronic device in an exemplary embodiment of the present disclosure. DETAILED DESCRIPTION

[0065] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as being limited to the examples set forth herein; on the contrary, these embodiments are provided so that the present disclosure will be more comprehensive and complete, and the concepts of the example embodiments are fully conveyed to those skilled in the art. The described features, structures, or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known technical solutions are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0066] In addition, the accompanying drawings are only schematic illustrations of the present disclosure, and the same reference numerals in the drawings represent the same or similar parts, so their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities, which do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0067] Figure 1 A schematic diagram of an exemplary system architecture of a cloud computer service resource management solution to which an embodiment of the present invention can be applied is shown.

[0068] like Figure 1 As shown, the system architecture 100 may include one or more of terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or optical fiber cables, etc.

[0069] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. According to the implementation requirements, there may be any number of terminal devices, networks and servers. For example, the server 105 may be a server cluster composed of multiple servers.

[0070] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Terminal devices 101, 102, 103 can be various electronic devices with display screens, including but not limited to smart phones, tablet computers, portable computers, desktop computers, etc.

[0071] In some embodiments, the cloud computer service resource management method provided in the embodiments of the present invention is generally executed by the server 105. Accordingly, the cloud computer service resource management device is generally set in the terminal device 103 (or the terminal device 101 or 102). In other embodiments, some terminals may have functions similar to those of the server device to execute the method.

[0072] The exemplary embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0073] Figure 2 It is a flowchart of a method for managing cloud computer service resources in an exemplary embodiment of the present disclosure.

[0074] refer to Figure 2 , the management method of cloud computer service resources may include:

[0075] Step S202, constructing a user profile based on historical user needs and historical operation records;

[0076] Step S204, optimizing any of the user portraits based on a neural network model;

[0077] Step S206, performing multi-objective optimization on the resource requirements based on the Pareto optimization algorithm to determine the resource configuration information corresponding to any of the user portraits;

[0078] Step S208, performing management operations on the cloud computer service resources according to the resource configuration information corresponding to the user portrait, wherein the management operations include at least one of resource scheduling, resource migration, resource expansion and resource reduction.

[0079] The disclosed embodiment constructs a user portrait based on historical user demands and historical operation records, optimizes any of the user portraits based on a neural network model, and performs multi-objective optimization on the resource demands based on a Pareto optimization algorithm to determine resource configuration information corresponding to any of the user portraits. Management operations on the cloud computer service resources are performed based on the resource configuration information corresponding to the user portrait, so that the cloud computer service resources are better adapted to the usage scenarios and usage preferences of different users for the cloud computer service resources, and have better robustness and adaptability.

[0080] Below, each step of the cloud computer service resource management method is described in detail.

[0081] In an exemplary embodiment of the present disclosure, Figure 3 As shown in the figure, building a user profile based on historical user needs and historical operation records includes:

[0082] Step S302, obtaining the historical operation record through the API interface and / or log record associated with the resources required by the cloud computer service;

[0083] Step S304, determining the demand information corresponding to the historical operation record;

[0084] Step S306, classifying the demand information according to the attribute information of the historical operation record;

[0085] Step S308, generating a corresponding user portrait according to the classification result.

[0086] In an exemplary embodiment of the present disclosure, Figure 4 As shown, the corresponding user portraits generated according to the classification results include:

[0087] Step S402, classifying the demand information according to the attribute information of the historical operation record, wherein the attribute information includes at least one of the user's usage scenario, time period, price sensitivity, and resource usage requirements;

[0088] Step S404, determining the user's usage pattern and / or application preference information of the resources required for the cloud computer service according to the historical operation record;

[0089] Step S406, identifying the emotional tendency information in the preference information based on a natural language algorithm;

[0090] Step S408, determining the user's demand information for resources required by the cloud computer service according to the emotional tendency information;

[0091] Step S410: clustering the users based on a dynamic DBSCAN clustering algorithm to divide the users into multiple categories.

[0092] In an exemplary embodiment of the present disclosure, Figure 5 As shown, optimizing any of the user portraits based on the neural network model includes:

[0093] Step S502, training the neural network model based on the demand information and the behavior data of multiple categories of users to obtain a user behavior pattern-demand model;

[0094] Step S504, triggering the user behavior pattern-demand model to predict user needs, and performing at least one of identification, merging, and evaluation on the predicted demand information;

[0095] Step S506: Optimize the user portrait according to the predicted demand information.

[0096] In an exemplary embodiment of the present disclosure, Figure 6 As shown, the resource requirements are multi-objective optimized based on the Pareto optimization algorithm to determine the resource configuration information corresponding to any of the user portraits, including:

[0097] Step S602, determining the quality of the cloud computer service according to the response time of the application loaded by the cloud computer service and / or the network throughput;

[0098] Step S604, determining the resource overhead required for the operation of the cloud computer service as a resource cost, wherein the resource cost includes at least one of computing resources, storage resources and network resources;

[0099] Step S606, determining user complaints according to the service delay and user satisfaction of the cloud computer service;

[0100] Step S608, configuring the quality of the cloud computer service and the resource cost as input parameters of the Pareto optimization algorithm, and configuring the minimum value of the user complaints as the target output result of the Pareto optimization algorithm;

[0101] Step S610, performing multi-objective optimization on the predicted resource demand according to the configured Pareto optimization algorithm to determine the resource configuration information corresponding to any of the user portraits.

[0102] In an exemplary embodiment of the present disclosure, the formula of the Pareto optimization algorithm includes:

[0103]

[0104] sD i ≤D Response ;

[0105] T i ≤T thr ;

[0106] Q i,1 ≥Q CUP ;

[0107] Q i,2 ≥Q Memory ;

[0108] Q i,3 ≥Q Storyage ;

[0109] Q i,4 ≥Q Network ;

[0110]

[0111] Said The calculation result of the minimum value of user complaints, D i represents the service delay of the user performing the i-th operation through the cloud computer service, the S i represents the complaint rate of the i-th operation, a and b both represent impact factors, the value range of a and b is (0,1), and they satisfy the constraint of a+b=1, and D Response Characterize the application response time threshold of the cloud computer service, the T i Characterizes the network throughput required for the i-th operation, the T thr Characterizes the network throughput threshold required for the i-th operation, the Q i,1 Q i,2 Q i,3 Q i,4Respectively represent the CPU, memory, storage, and network bandwidth required for the i-th operation. CPU Q Memory Q Storage Q Network Respectively represent the minimum resource requirements of the cloud computing service, the C i,1 , said C i,2 , said C i,3 , said C i,4 Respectively represent the cost of CPU, memory, storage, and network bandwidth required for the i-th operation. Total Characterizes the total resource cost of the cloud computing service.

[0112] In an exemplary embodiment of the present disclosure, Figure 7 As shown, the management operation of the cloud computer service resources according to the resource configuration information corresponding to the user portrait includes:

[0113] Step S702, in the process of scheduling resources required for the cloud computer service, collecting the user's usage feedback, the usage feedback being positive feedback or negative feedback;

[0114] Step S704, determining the validity of the usage feedback according to the resource utilization rate of the resources required by the cloud computer service corresponding to the usage feedback;

[0115] Step S706, rescheduling the resources required for the cloud computer service of the user with negative feedback of effectiveness until the usage feedback becomes positive feedback.

[0116] This paper proposes a cloud computer user abnormal behavior detection solution based on a multimodal large model, which specifically includes the following improvements:

[0117] (1) Propose a method system for establishing user-oriented cloud computer resource intelligent allocation and scheduling, including a user behavior pattern recognition method based on machine learning, a resource demand prediction model based on LSTM, a resource intelligent matching scheme based on Pareto optimization, a resource dynamic scheduling strategy and a user feedback mechanism;

[0118] (2) Propose a user behavior pattern recognition method based on machine learning. By analyzing the user's historical behavior data, identify the cloud computer usage patterns and preferences of each user group, establish user portraits, and dynamically maintain the user demand database;

[0119] (3) Build a resource demand prediction model based on the time series of cloud computer usage behavior and use the LSTM model to predict the changing trend of cloud computer resource demand within a set time period;

[0120] (4) Design a resource intelligent matching solution that comprehensively considers factors such as resource cost and user experience, and achieves multi-objective optimization based on Pareto optimization technology;

[0121] (5) A dynamic resource scheduling strategy is proposed based on the resource usage of cloud computers and the resource demand prediction results, and a user feedback mechanism is established to optimize the scheduling strategy and form a closed-loop optimization solution.

[0122] In an exemplary embodiment of the present disclosure, Figure 8 As shown, the main stages of the cloud computer service resource management solution disclosed in the present invention include the cloud computer resource configuration stage and the cloud computer resource scheduling stage. The key steps of each stage include:

[0123] 1. The cloud computer resource configuration phase includes the following steps:

[0124] Step S802, collect user behavior data: obtain user behavior data through API interface or log records (with user consent), and classify the needs according to the user's usage scenario (such as office, entertainment, development, etc.), time period, price sensitivity, and resource usage requirements

[0125] Step S804, behavior analysis and sentiment analysis: analyzing the user's historical behavior data, identifying the user's common cloud computer usage patterns and application preferences, and then using natural language technology to identify the emotional tendencies in the user's feedback information and extract potential needs;

[0126] Step S806, establishing user portraits: extracting features based on the first and second layer analyses, using the improved dynamic DBSCAN clustering algorithm to group users, and then defining labels for each user group to create a detailed user portrait behavior document;

[0127] Step S808, dynamically maintain the user demand library: obtain the user behavior pattern-demand model based on deep reinforcement learning algorithm training, adjust the key parameters of the model, complete model iteration, and identify new demands, merge demands, and evaluate demands based on the model;

[0128] Step S810, training a resource demand prediction model based on LSTM: Based on user portraits and demand libraries, combined with time series and other data, an LSTM network is used for model training to learn the changing trends and rules of cloud computer resource demand over time;

[0129] Step S812, predicting changes in user demand: predicting the resource demand of cloud computer users in the future based on the LSTM training resource demand prediction model;

[0130] Step S814, establishing a multi-objective optimization resource intelligent matching model: quantifying the factors that affect cloud computer resource matching, and maximizing user experience while satisfying service quality constraints;

[0131] Step S816, Pareto optimization solution: For the multi-objective optimization problem of cloud computer resource matching, the Pareto optimization method is used to solve the solution on the Pareto frontier.

[0132] Step S818: Select a better solution as the optimal resource matching solution according to the changing trend of user demand.

[0133] 2. The cloud computer resource scheduling phase includes the following steps:

[0134] Step S820, cloud computer resource monitoring: real-time monitoring of available resources on the cloud service platform and user resource usage;

[0135] Step S822, dynamic scheduling of cloud computer resources: according to the user profile and resource demand prediction model, cloud computer resource matching optimal solution, appropriate resources are allocated to users, or resources are migrated, expanded or reduced in capacity;

[0136] Step S824, user feedback mechanism: collect user experience (response speed, service stability, etc.) during the resource scheduling process, combine resource utilization and other indicators to comprehensively evaluate the effectiveness of the strategy, and continuously adjust the plan until the user is satisfied: if the user is satisfied with the current cloud computer resource dynamic scheduling plan, then it will be used as positive feedback data to act on the subsequent cloud computer resource dynamic scheduling plan formulation module, and the cloud computer resource configuration and scheduling tasks are completed; if the user is not satisfied with the current cloud computer resource dynamic scheduling plan, then a cloud computer resource dynamic scheduling plan will be formulated for the user again, and this plan will be used as negative feedback data for the cloud computer resource dynamic scheduling plan formulation module.

[0137] In an exemplary embodiment of the present disclosure, Fig. 9 As shown, the architecture 900 of the cloud computer resource intelligent configuration and scheduling solution mainly includes a user behavior pattern recognition method based on machine learning, a resource demand prediction model based on LSTM, a resource intelligent matching solution based on Pareto optimization, a resource dynamic scheduling strategy and a user feedback mechanism, but is not limited to these.

[0138] Among them, the user behavior pattern recognition method based on machine learning and the resource intelligent matching solution based on Pareto optimization are the key points of this disclosure. The specific functions are described as follows:

[0139] The user behavior pattern recognition method based on machine learning is a method to build a user profile based on user needs and historical behaviors, thereby dynamically maintaining a user demand library of the same type. For different types of user behaviors, this method can accurately identify and match their personalized needs. This method can be divided into two key steps: building a user profile and dynamically maintaining a user demand library.

[0140] In an exemplary embodiment of the present disclosure, the process of establishing a user portrait specifically includes:

[0141] Data collection: obtaining user behavior data through API interfaces or log records, including but not limited to user registration information, login time, online time, application type, consumed computing resources, network bandwidth usage, user feedback, etc.;

[0142] Demand classification: Classify demands according to user usage scenarios (such as office, entertainment, development, etc.), time periods, price sensitivity, and resource usage requirements;

[0143] The first level of behavior analysis: Analyze the user's historical behavior data to identify the user's common cloud computer usage patterns and application preferences, such as the applications that are frequently used within a time period, the preferred resource configuration, etc.

[0144] The second layer of sentiment analysis: using natural language processing technology to identify the emotional tendencies in user feedback information, understand the user's attitude towards the current cloud computer service, extract potential needs, and add them to the demand classification;

[0145] Feature extraction: Extract key features that reflect users' personalized needs and behavioral habits, such as the above-mentioned usage scenarios, time periods, resource usage, etc., and calculate derived features based on the original data, such as resource usage frequency, average peak usage and distribution;

[0146] User grouping: Select the improved dynamic DBSCAN clustering algorithm to divide users with similar characteristics into different groups, establish the initial clustering structure, and dynamically update the user grouping through the process of incremental update of new data and adjustment of cluster boundaries. Then, perform statistical analysis on the user characteristics in each group to obtain the main characteristics and representative behavior patterns of the group.

[0147] User portrait construction: First, define labels for each user group to describe the main characteristics of the group (such as user occupation, etc.), and then create detailed user portrait behavior documents (such as resource usage, active time, etc.).

[0148] In the actual application of cloud computers, user behavior trends can reflect changes in user needs. Therefore, on the basis of building user portraits, it is also necessary to solve the problem of how to update the user demand library in real time based on user feedback and behavior changes to ensure that the user demand library always reflects the latest user needs. Therefore, in the user behavior pattern recognition method based on machine learning, it is also necessary to dynamically maintain the user demand library, regularly evaluate the effectiveness of user portraits, and optimize user portraits based on user behavior:

[0149] Formation of user demand database: User portraits are used to group users and describe their behaviors. While constructing user portraits, the needs of similar users are summarized, and a user demand database is formed accordingly;

[0150] Model training and iteration: Training is performed based on the user behavior data collected in the early stage. The user behavior pattern-demand model is obtained based on deep reinforcement learning algorithm training. The accuracy index is used to evaluate the model performance. At the same time, the key parameters of the model are adjusted to complete the model iteration;

[0151] New demand identification: Based on the updated user behavior pattern-demand model, new user needs or demand changes are identified based on user behavior;

[0152] Requirements merging: Compare newly identified requirements with entries in the existing requirements library, merge duplicate requirements, and add new requirements points;

[0153] Demand assessment: Re-evaluate each requirement in the demand library based on its importance and urgency, so that high-priority requirements can be given priority when allocating cloud computer resources.

[0154] Dynamically maintaining the user demand library also helps to evaluate the effectiveness of user portraits in order to optimize user portraits.

[0155] In an exemplary embodiment of the present disclosure, the resource intelligent matching scheme implemented based on the Pareto optimization algorithm is actually a multi-objective joint optimization method, which maximizes the user experience while meeting the service quality constraints by quantifying the factors affecting the cloud computer resource matching.

[0156] In an exemplary embodiment of the present disclosure, the quality of cloud computer services is quantified from the application response time and network throughput based on the Pareto optimization algorithm. Resource cost refers to the resource overhead required to provide cloud computer services, including the cost of computing resources (CPU, memory), storage resources and network resources. User experience includes the user's subjective feelings, which can be quantified from service delay and user satisfaction. However, due to different evaluation criteria for user satisfaction, the user's complaint rate is more intuitive. Therefore, user satisfaction can be quantified by the user's complaint rate. A low complaint rate indicates high user satisfaction. Then, the core formulas that can be modeled in this solution include but are not limited to:

[0157]

[0158] sD i ≤D Response ;

[0159] T i ≤T thr ;

[0160] Q i,1 ≥Q CUP ;

[0161] Q i,2 ≥Q Memory ;

[0162] Q i,3 ≥Q Storyage ;

[0163] Q i,4 ≥Q Network ;

[0164]

[0165] Among them, D i represents the service delay of the user's i-th operation, S i represents the complaint rate for the ith operation. The impact factors a and b are constrained to be in the range of (0,1), and a+b=1. Response Indicates the normal application response time threshold. T i represents the network throughput required for the i-th operation, T thr Corresponding to the network throughput threshold. In order to ensure the service quality of cloud computers, Q i,1 , Q i,2 , Q i,3 , Q i,4 They represent the CPU, memory, storage, and network bandwidth required for the i-th operation, respectively. CPU , Q Memory , Q Storage , Q NetworkThey represent the minimum resource requirements of cloud computer users. i,1 , C i,2 , C i,3 , C i,4 They represent the cost of CPU, memory, storage, and network bandwidth required for the i-th operation, respectively. Total is the total resource cost.

[0166] In summary, for the proposed multi-objective optimization problem of cloud computer resource matching, the present disclosure adopts the Pareto optimization algorithm to find a set of solutions in the absence of a single optimal solution, so that the improvement of any objective will not cause other objectives to become worse.

[0167] In combination with the description and processing flow of the above technical solution, the present disclosure also provides the following specific embodiments:

[0168] Embodiment 1:

[0169] First, user behavior data is obtained through the cloud computer service API interface or cloud computer operation log, including but not limited to user registration information (application scenario field), cloud computer login time (high-frequency time points, low-frequency time points), cloud computer online time, application type (office software, entertainment software, development tools, etc.), consumed computing resources (CPU, memory, storage, network bandwidth), user satisfaction scores and suggestions, etc.

[0170] Subsequently, multiple layers of analysis are performed based on the initial behavioral data: the first layer of behavioral analysis identifies the cloud computer usage patterns and application preferences of various types of users, focusing on determining key points such as the demand resource weights for this type of user; the second layer of demand analysis uses natural language technology to identify the emotional tendencies of various types of users in order to determine the users' potential needs.

[0171] After data cleaning, denoising, feature extraction and other steps, the improved dynamic DBSCAN clustering algorithm is selected to divide users with similar characteristics into different groups, establish the initial clustering structure, and then dynamically update the user grouping by using new data incremental updates and cluster boundary adjustments. Subsequently, the user characteristics within each group are statistically analyzed to obtain the main characteristics and representative behavior patterns of each user group.

[0172] Based on this, labels are defined for each user group to describe the main characteristics of the group (such as user occupation, etc.), and then detailed user portrait behavior documents are created (such as resource usage, active time, etc.).

[0173] At the same time, training is carried out based on the processed behavioral data, and a deep reinforcement learning algorithm is used to train the user behavior pattern-demand model. Based on the user's behavior, new user needs or demand changes are identified based on the model, and the user demand library is continuously updated.

[0174] Here, you can identify typical user portraits, such as: Developers (main scenarios: Internet, finance, occupations: front-end, back-end, testing) - Behavior documents (application preferences: Visual Studio Code, Tencent Meeting, active time: afternoon, evening, resource usage: CPU occupancy, storage usage, network bandwidth, etc.).

[0175] Accordingly, the user demand library of this user group includes the requirement that Visual Studio Code is frequently used in the afternoon and has high CPU resource requirements. Based on the user behavior pattern-requirement model, new developer requirements are derived: SublimeText is more actively used in the morning and has little impact on resource changes. Based on this process, the user profiles and requirements of each type of user can be determined.

[0176] Furthermore, based on resource usage, user behavior and time series, a resource demand prediction model is obtained based on LSTM model training to learn the changing trends and laws of cloud computer resource demand over time. Based on the trained model, the resource demand of cloud computer users in the future is predicted.

[0177] Secondly, we select factors that affect cloud computer resource matching, maximize user experience while meeting service quality constraints, and establish a multi-objective optimization resource intelligent matching model. Through the Pareto optimization method, we confirm feasible optimization solutions, and determine a better solution as the optimal resource matching solution after considering the changing trend of user demand.

[0178] After cloud computer resource configuration, we implement a dynamic cloud computer resource scheduling strategy for users: we predict that the computing and network resource demands of developers will gradually increase after 2 p.m., reach a peak and then remain stable, and then fall back after 9 p.m. In the afternoon, we gradually schedule idle CPU resources for the developer group to ensure stable and smooth network connections. Finally, we make policy adjustments based on user feedback.

[0179] Embodiment 2:

[0180] First, user behavior data is obtained through the cloud computer service API interface or cloud computer operation log, including but not limited to user registration information (application scenario field), cloud computer login time (high-frequency time points, low-frequency time points), cloud computer online time, application type (office software, entertainment software, development tools, etc.), consumed computing resources (CPU, memory, storage, network bandwidth), user satisfaction scores and suggestions, etc.

[0181] Subsequently, multiple layers of analysis are performed based on the initial behavioral data: the first layer of behavioral analysis identifies the cloud computer usage patterns and application preferences of various types of users, focusing on determining key points such as the demand resource weights for this type of user; the second layer of demand analysis uses natural language technology to identify the emotional tendencies of various types of users in order to determine the users' potential needs.

[0182] After data cleaning, denoising, feature extraction and other steps, the improved dynamic DBSCAN clustering algorithm is selected to divide users with similar characteristics into different groups, establish the initial clustering structure, and then dynamically update the user grouping by using new data incremental updates and cluster boundary adjustments. Subsequently, the user characteristics within each group are statistically analyzed to obtain the main characteristics and representative behavior patterns of each user group.

[0183] Based on this, labels are defined for each user group to describe the main characteristics of the group (such as user occupation, etc.), and then detailed user portrait behavior documents are created (such as resource usage, active time, etc.).

[0184] At the same time, training is carried out based on the processed behavioral data, and a deep reinforcement learning algorithm is used to train the user behavior pattern-demand model. Based on the user's behavior, new user needs or demand changes are identified based on the model, and the user demand library is continuously updated.

[0185] Here, you can identify typical user portraits, such as: game players (main scenarios: client games, web games, stand-alone games) - behavior documents (application preferences: game clients, network accelerators, active time: weekday evenings, weekends all day, resource usage: CPU occupancy, storage usage, network bandwidth, etc.).

[0186] Correspondingly, the user demand library of this user group includes high CPU and GPU requirements, and has high requirements for graphics processing capabilities and low-latency networks. Based on this process, the user profiles and requirements of each type of user can be determined.

[0187] Furthermore, based on resource usage, user behavior and time series, a resource demand prediction model is obtained based on LSTM model training to learn the changing trends and laws of cloud computer resource demand over time. Based on the trained model, the resource demand of cloud computer users in the future is predicted.

[0188] Secondly, we select factors that affect cloud computer resource matching, maximize user experience while meeting service quality constraints, and establish a multi-objective optimization resource intelligent matching model. Through the Pareto optimization method, we confirm feasible optimization solutions, and determine a better solution as the optimal resource matching solution after considering the changing trend of user demand.

[0189] After cloud computer resource configuration, we implement a dynamic cloud computer resource scheduling strategy for users: between 8pm and 11pm, gamers’ resource demands peak, and during these time periods, we allocate high-performance GPUs and low-latency networks to ensure a smooth gaming experience. Finally, we make policy adjustments based on user feedback.

[0190] Corresponding to the above method embodiments, the present disclosure also provides a cloud computer service resource management device, which can be used to execute the above method embodiments.

[0191] Fig.10 It is a block diagram of a management device for cloud computer service resources in an exemplary embodiment of the present disclosure.

[0192] refer to Fig.10 , the cloud computer service resource management device 1000 may include:

[0193] A construction module 1002 is configured to construct a user portrait based on historical user needs and historical operation records;

[0194] An optimization module 1004 is configured to optimize any of the user portraits based on a neural network model;

[0195] The calculation module 1006 is configured to perform a multi-objective optimization solution on the resource demand based on the Pareto optimization algorithm to determine the resource configuration information corresponding to any of the user portraits;

[0196] The management module 1008 is configured to perform management operations on the cloud computer service resources according to the resource configuration information corresponding to the user portrait, and the management operations include at least one of resource scheduling, resource migration, resource expansion and resource reduction.

[0197] In an exemplary embodiment of the present disclosure, the construction module 1002 is further configured to:

[0198] Obtain the historical operation record through the API interface and / or log record associated with the resources required by the cloud computer service;

[0199] Determining demand information corresponding to the historical operation record;

[0200] classifying the demand information according to the attribute information of the historical operation record;

[0201] Generate corresponding user portraits based on the classification results.

[0202] In an exemplary embodiment of the present disclosure, the construction module 1002 is further configured to:

[0203] Classifying the demand information according to the attribute information of the historical operation record, wherein the attribute information includes at least one of the user's usage scenario, time period, price sensitivity, and resource usage requirements;

[0204] Determining the user's usage pattern and / or application preference information of resources required for the cloud computing service based on the historical operation records;

[0205] Identify emotional tendency information in the preference information based on a natural language algorithm;

[0206] Determining the user's demand information for resources required for the cloud computer service based on the emotional tendency information;

[0207] The users are clustered based on a dynamic DBSCAN clustering algorithm to divide the users into multiple categories.

[0208] In an exemplary embodiment of the present disclosure, the optimization module 1004 is further configured to:

[0209] Training the neural network model based on the demand information and the behavior data of multiple categories of users to obtain a user behavior pattern-demand model;

[0210] triggering the user behavior pattern-demand model to predict user needs, and performing at least one of identification, merging, and evaluation on the predicted demand information;

[0211] The user portrait is optimized according to the predicted demand information.

[0212] In an exemplary embodiment of the present disclosure, the optimization module 1004 is further configured to:

[0213] Determining the quality of the cloud computing service based on the response time of the application loaded by the cloud computing service and / or the network throughput;

[0214] Determine resource costs required for the operation of the cloud computer service as resource costs, wherein the resource costs include at least one of computing resources, storage resources, and network resources;

[0215] Determine user complaints based on service delays and user satisfaction with the cloud computing service;

[0216] The quality of the cloud computer service and the resource cost are configured as input parameters of the Pareto optimization algorithm, and the minimum value of the user complaints is configured as the target output result of the Pareto optimization algorithm;

[0217] A multi-objective optimization solution is performed on the predicted resource demand according to the configured Pareto optimization algorithm to determine the resource configuration information corresponding to any of the user portraits.

[0218] In an exemplary embodiment of the present disclosure, the formula of the Pareto optimization algorithm includes:

[0219]

[0220] sD i ≤D Response ;

[0221] T i ≤T thr ;

[0222] Q i,1 ≥Q CUP ;

[0223] Q i,2 ≥Q Memory ;

[0224] Q i,3 ≥Q Storyage ;

[0225] Q i,4 ≥Q Network ;

[0226]

[0227] Said The calculation result of the minimum value of user complaints, D i represents the service delay of the user performing the i-th operation through the cloud computer service, the S i represents the complaint rate of the i-th operation, a and b both represent impact factors, the value range of a and b is (0,1), and they satisfy the constraint of a+b=1, and D Response Characterize the application response time threshold of the cloud computer service, the T i Characterizes the network throughput required for the i-th operation, the T thr Characterizes the network throughput threshold required for the i-th operation, the Q i,1 Q i,2 Q i,3 Q i,4 Respectively represent the CPU, memory, storage, and network bandwidth required for the i-th operation. CPU Q Memory Q Storage Q Network Respectively represent the minimum resource requirements of the cloud computing service, the C i,1 , said C i,2 , said C i,3 , said Ci,4 Respectively represent the cost of CPU, memory, storage, and network bandwidth required for the i-th operation. Total Characterizes the total resource cost of the cloud computing service.

[0228] In an exemplary embodiment of the present disclosure, the management module 1008 is further configured to:

[0229] In the process of scheduling resources required for the cloud computer service, collecting usage feedback from the user, where the usage feedback is positive feedback or negative feedback;

[0230] Determining the validity of the usage feedback according to the resource utilization rate of the resources required by the cloud computer service corresponding to the usage feedback;

[0231] The resources required for the cloud computer service of the user with negative feedback of effectiveness are rescheduled until the usage feedback becomes the positive feedback.

[0232] Since the functions of the device 1000 have been described in detail in the corresponding method embodiments, the present disclosure will not elaborate on them here.

[0233] This method establishes a user behavior pattern recognition method based on machine learning, conducts demand analysis and sentiment analysis on users' historical behavior data, identifies the cloud computer usage patterns and application preferences of each user group, and thus establishes user portraits and dynamically maintains the user demand library;

[0234] 2. This method designs a resource intelligent matching scheme, comprehensively considers various influencing factors in resource allocation, realizes multi-objective optimization based on Pareto optimization technology, covers the optimal area in the entire feasible solution space, and provides diversified matching schemes;

[0235] 3. This method provides a dynamic resource scheduling strategy based on user portraits, resource demand prediction models, and resource intelligent matching feasible solutions, and optimizes the strategy based on user feedback, thereby forming a closed-loop optimization solution that can adapt to real-time changing user needs and diversified usage scenarios, and has certain robustness and strong adaptability.

[0236] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0237] In an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided.

[0238] It will be appreciated by those skilled in the art that various aspects of the present invention may be implemented as a system, method or program product. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as a "circuit", "module" or "system".

[0239] Refer to the following Fig.11 The electronic device 1100 according to this embodiment of the present invention is described. Fig.11 The electronic device 1100 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0240] like Fig.11 As shown, the electronic device 1100 is in the form of a general computing device. The components of the electronic device 1100 may include but are not limited to: at least one processing unit 1110, at least one storage unit 1120, and a bus 1130 connecting different system components (including the storage unit 1120 and the processing unit 1110).

[0241] The storage unit stores program codes, which can be executed by the processing unit 1110, so that the processing unit 1110 performs the steps according to various exemplary embodiments of the present invention described in the above “Exemplary Method” section of this specification. For example, the processing unit 1110 can perform the method shown in the embodiment of the present disclosure.

[0242] The storage unit 1120 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 11201 and / or a cache storage unit 11202 , and may further include a read-only storage unit (ROM) 11203 .

[0243] The storage unit 1120 may also include a program / utility 11204 having a set (at least one) of program modules 11205, such program modules 11205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0244] The bus 1130 may be a representation of one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0245] The electronic device 1100 may also communicate with one or more external devices 1140 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 1100, and / or communicate with any device that enables the electronic device 1100 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 1150. Furthermore, the electronic device 1100 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter 1160. As shown, the network adapter 1160 communicates with other modules of the electronic device 1100 via a bus 1130. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 1100, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0246] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0247] In an exemplary embodiment of the present disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the above method of the present specification is stored. In some possible implementations, various aspects of the present invention can also be implemented in the form of a program product, which includes a program code, and when the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps according to various exemplary embodiments of the present invention described in the above "Exemplary Method" section of the present specification.

[0248] The program product for implementing the above method according to an embodiment of the present invention can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto, and in this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.

[0249] Among them, the readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0250] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0251] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0252] In an exemplary embodiment of the present disclosure, a computer program product is also provided, which can be loaded or stored in any combination of one or more readable media, and can be written in any combination of one or more programming languages ​​to perform the program code of the present invention, including object-oriented programming languages ​​such as Java, C++, etc., and also conventional procedural programming languages ​​such as "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or completely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0253] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to an exemplary embodiment of the present invention, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0254] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are to be considered exemplary only, and the true scope and concept of the present disclosure are indicated by the claims.

Claims

1. A method for managing cloud computer service resources, characterized in that: include: Build user portraits based on historical user needs and historical operation records; Optimizing any of the user portraits based on a neural network model; Perform multi-objective optimization on the resource requirements based on the Pareto optimization algorithm to determine the resource configuration information corresponding to any of the user portraits; The cloud computer service resources are managed according to the resource configuration information corresponding to the user portrait, and the management operation includes at least one of resource scheduling, resource migration, resource expansion and resource reduction.

2. The cloud computing service resource management method according to claim 1, wherein: Building user portraits based on historical user needs and historical operation records includes: Obtain the historical operation record through the API interface and / or log record associated with the resources required by the cloud computer service; Determining demand information corresponding to the historical operation record; classifying the demand information according to the attribute information of the historical operation record; Generate corresponding user portraits based on the classification results.

3. The cloud computing service resource management method according to claim 2, wherein: The corresponding user portraits generated according to the classification results include: Classifying the demand information according to the attribute information of the historical operation record, wherein the attribute information includes at least one of the user's usage scenario, time period, price sensitivity, and resource usage requirements; Determining the user's usage pattern and / or application preference information of resources required for the cloud computing service based on the historical operation records; Identify emotional tendency information in the preference information based on a natural language algorithm; Determining the user's demand information for resources required for the cloud computer service based on the emotional tendency information; The users are clustered based on a dynamic DBSCAN clustering algorithm to divide the users into multiple categories.

4. The method for managing cloud computing service resources according to claim 1, wherein: Optimizing any of the user portraits based on the neural network model includes: Training the neural network model based on the demand information and the behavior data of multiple categories of users to obtain a user behavior pattern-demand model; triggering the user behavior pattern-demand model to predict user needs, and performing at least one of identification, merging, and evaluation on the predicted demand information; The user portrait is optimized according to the predicted demand information.

5. The cloud computing service resource management method according to claim 1, wherein: The resource requirements are solved by multi-objective optimization based on the Pareto optimization algorithm to determine the resource configuration information corresponding to any of the user portraits, including: Determining the quality of the cloud computing service based on the response time of the application loaded by the cloud computing service and / or the network throughput; Determine resource costs required for the operation of the cloud computer service as resource costs, wherein the resource costs include at least one of computing resources, storage resources, and network resources; Determine user complaints based on service delays and user satisfaction with the cloud computing service; The quality of the cloud computer service and the resource cost are configured as input parameters of the Pareto optimization algorithm, and the minimum value of the user complaints is configured as the target output result of the Pareto optimization algorithm; A multi-objective optimization solution is performed on the predicted resource demand according to the configured Pareto optimization algorithm to determine the resource configuration information corresponding to any of the user portraits.

6. The method for managing cloud computer service resources according to claim 1, wherein: The formula of the Pareto optimization algorithm includes: s.t.D i ≤D Response ; T i ≤T thr ; Q i,1 ≥Q CUP ; Q i,2 ≥Q Memory ; Q i,3 ≥Q Storyage ; Q i,4 ≥Q Network ; Said The calculation result of the minimum value of user complaints, D i represents the service delay of the user performing the i-th operation through the cloud computer service, the S i represents the complaint rate of the i-th operation, a and b both represent impact factors, the value range of a and b is (0,1), and they satisfy the constraint of a+b=1, and D Response Characterize the application response time threshold of the cloud computer service, the T i Characterizes the network throughput required for the i-th operation, the T thr Characterizes the network throughput threshold required for the i-th operation, the Q i,1 , Q i,2 , Q i,3 , Q i,4 Respectively represent the CPU, memory, storage, and network bandwidth required for the i-th operation. CPU , Q Memory , Q Storage , Q Network Respectively represent the minimum resource requirements of the cloud computing service, the C i,1 , said C i,2 , said C i,3 , said C i,4 Respectively represent the cost of CPU, memory, storage, and network bandwidth required for the i-th operation. Total Characterizes the total resource cost of the cloud computing service.

7. The method for managing cloud computer service resources according to claim 1, wherein: The management operation of the cloud computer service resources according to the resource configuration information corresponding to the user portrait includes: In the process of scheduling resources required for the cloud computer service, collecting usage feedback from the user, where the usage feedback is positive feedback or negative feedback; Determining the validity of the usage feedback according to the resource utilization rate of the resources required by the cloud computer service corresponding to the usage feedback; The resources required for the cloud computer service of the user with negative feedback of effectiveness are rescheduled until the usage feedback becomes the positive feedback.

8. A cloud computing service resource management device, characterized in that: include: A construction module is configured to construct a user profile based on historical user needs and historical operation records; An optimization module, configured to optimize any of the user portraits based on a neural network model; A calculation module is configured to perform multi-objective optimization on the resource requirements based on a Pareto optimization algorithm to determine resource configuration information corresponding to any of the user portraits; The management module is configured to perform management operations on the cloud computer service resources according to the resource configuration information corresponding to the user portrait, and the management operations include at least one of resource scheduling, resource migration, resource expansion and resource reduction.

9. An electronic device, characterized in that: include: Memory; as well as A processor coupled to the memory, the processor being configured to execute a method for managing cloud computer service resources as described in any one of claims 1-7 based on instructions stored in the memory.

10. A computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the cloud computer service resource management method as described in any one of claims 1 to 7.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for managing cloud computer service resources as described in any one of claims 1 to 7 is implemented.

Images