AI-based hardware resource dynamic evaluation method and system, and electronic device

Abstract

The application relates to the technical field of information systems, and discloses an AI-based hardware resource dynamic evaluation method and system and electronic equipment; the system comprises: a data sensing layer, which is used for collecting original data of hardware resources and pre-processing the collected original data; the original data comprises estimated indexes of different resource types, monitoring indexes, and correlation index data between different resource types; an analysis and decision layer, which is used for performing feature extraction, similarity retrieval, and resource demand prediction on the pre-processed original data based on a hardware resource evaluation model, generating a resource evaluation strategy and optimization suggestion; and a feedback execution layer, which is used for performing resource adjustment operations according to the optimization suggestion, tracking execution effects, and feeding back the execution results to the analysis and decision layer to update the hardware resource evaluation model. The application effectively solves the problems of extensive resource evaluation, dependence on manual monitoring, and difficulty in dynamic adjustment in the prior art.

Description

Technical Field

[0001] This application relates to the field of information system technology, and more specifically, to an AI-based method, system, and electronic device for dynamic evaluation of hardware resources. Background Technology

[0002] Hardware resources account for a large portion of the R&D investment in information systems. With the development of cloud-native technology stacks, information systems are gradually adopting design and deployment methods such as microservices, containerization, and dynamic orchestration, and are striving to achieve on-demand elastic scaling for computing, storage, and network resources.

[0003] However, in some critical business areas, the flexible scaling down of resources often cannot rely entirely on automatic scaling down due to the lack of a complete and accurate resource demand assessment model and the importance of core business systems. The demand for information system hardware resources mainly remains at the level of a one-time static assessment strategy based on experience, supplemented by some campaign-style manual monitoring and scaling down rectification tasks. Summary of the Invention

[0004] In view of the above situation, this application provides an AI-based method, system and electronic device for dynamic evaluation of hardware resources, which aims to solve the above problems or at least partially solve the above problems.

[0005] In a first aspect, embodiments of this application provide an AI-based dynamic hardware resource evaluation system, the system comprising: The data perception layer is used to collect raw data of hardware resources and preprocess the collected raw data; the raw data includes estimated indicators, monitoring indicators, and correlation indicators between different resource types. The analysis and decision-making layer is used to perform feature extraction, similarity retrieval, and resource demand prediction on the preprocessed raw data based on the hardware resource assessment model, and to generate resource assessment strategies and optimization suggestions. The feedback execution layer is used to perform resource adjustment operations based on the optimization suggestions, track the execution effect, and feed the execution results back to the analysis and decision layer to update the hardware resource evaluation model.

[0006] Secondly, embodiments of this application also provide an AI-based method for dynamic evaluation of hardware resources, characterized in that the method includes: Raw data of hardware resources is collected and preprocessed; the raw data includes estimated indicators, monitoring indicators and correlation indicators between different resource types. Based on the hardware resource assessment model, feature extraction, similarity retrieval, and resource demand prediction are performed on the preprocessed raw data to generate resource assessment strategies and optimization suggestions. Based on the optimization suggestions, resource adjustment operations are performed, and the execution effect is tracked. The execution results are then fed back to the analysis and decision-making layer to update the hardware resource evaluation model.

[0007] Thirdly, embodiments of this application also provide an electronic device, including: a processor; and a memory arranged to store computer-executable instructions, which, when executed, cause the processor to perform the steps described in the second aspect.

[0008] Fourthly, embodiments of this application also provide a computer-readable storage medium that stores one or more programs, which, when executed by an electronic device including multiple applications, cause the electronic device to perform the steps described in the second aspect.

[0009] The at least one technical solution adopted in this application embodiment can achieve the following beneficial effects: By constructing a three-layer closed-loop architecture in which the data perception layer, analysis and decision-making layer, and feedback and execution layer work together, comprehensive perception, intelligent analysis, and automated optimization of hardware resources are realized; the data perception layer not only collects estimated indicators and real-time monitoring indicators of different resource types, but also innovatively introduces cross-resource type correlation indicator data, breaking the data silos between resource types such as databases, storage, and containers in traditional monitoring, and providing a richer and more interconnected data foundation for subsequent analysis; the analysis and decision-making layer performs feature extraction, similarity retrieval, and resource demand prediction on the preprocessed data based on the hardware resource evaluation model, and can mine resource configuration schemes under similar scenarios from historical data, dynamically predict future resource demands by combining real-time monitoring data, and generate accurate resource evaluation strategies and optimization suggestions, avoiding This invention overcomes the subjectivity and lag inherent in traditional methods that rely on expert experience and static assessments. The feedback execution layer automatically performs resource adjustments based on optimization suggestions and continuously tracks the execution results, feeding them back in real-time to the analysis and decision-making layer to update the hardware resource assessment model. This forms a closed-loop optimization mechanism of "perception-analysis-execution-feedback," enabling the system to learn and evolve continuously. This application effectively solves the problems of extensive resource assessment, reliance on manual monitoring, and difficulty in dynamic adjustment in existing technologies. It significantly improves the accuracy, dynamism, and automation of hardware resource assessment, achieving collaborative optimization of multiple resources such as databases, storage, and containers. It can accurately identify resource waste or performance bottlenecks, output differentiated optimization strategies, maximize resource utilization and reduce hardware investment costs while ensuring business stability, and provide a scientific basis for information system capacity planning and risk warning. Attached Figure Description

[0010] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 The system architecture diagram of the AI-based dynamic hardware resource evaluation system provided in the embodiments of this application is shown. Figure 2 A flowchart of the AI-based dynamic hardware resource evaluation method provided in an embodiment of this application is shown; Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0012] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such use can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the term "comprising" and its variations should be interpreted as open-ended terms meaning "including but not limited to."

[0013] Figure 1 This diagram illustrates the structure of an AI-based dynamic hardware resource evaluation system according to an embodiment of this application. The AI-based dynamic hardware resource evaluation method provided in this embodiment can, for example, […]. Figure 1 The system implementation, from Figure 1 It can be seen that the AI-based hardware resource dynamic evaluation system includes a data perception layer, an analysis and decision-making layer, and a feedback and execution layer.

[0014] The data perception layer is used to collect raw data from hardware resources and preprocess the collected raw data.

[0015] The raw data includes estimated metrics, monitoring metrics, and correlation metrics between different resource types. Different resource types include at least database hardware resources, file system / object storage hardware resources, and container resources. Specifically, estimated metrics are those pre-estimated based on business needs, historical experience, or industry standards before system launch or during resource planning. These are used to initially determine the scale of resource configuration. Examples include: database hardware resources: number of users, concurrent users, data size, estimated QPS, estimated TPS, estimated connections; file / object storage resources: storage capacity growth rate, file quantity growth rate, and retention period; container resources: core business scenario TPS and response latency. Monitoring metrics are dynamic metrics collected in real-time by monitoring tools during system operation. Examples include: database resources: CPU utilization, memory utilization, disk IOPS, storage throughput, lock wait time; file / object storage resources: file download TPS, file upload TPS, and data access frequency; container resources: CPU utilization, memory utilization, and database latency. Cross-resource type correlation metrics represent the combination, alignment, or derivation of metrics from different resource types to reveal the mutual influence or synergistic relationships between resources. For example, the correlation metrics between container resources and database hardware resources are the number of application Pods and database QPS. Whether the database QPS increases synchronously when the number of container replicas increases is used to determine whether the database needs to be expanded. Another example is the correlation metrics between object storage hardware resources and database hardware resources, which are the QPS of unstructured data access and the QPS of database metadata operations. This indicates whether there is a linkage between GET requests in object storage and metadata queries in the database, and is used to predict storage resource requirements. This application's embodiments use correlation metrics for cross-validation to avoid misjudgments that are easily made by a single metric.

[0016] The analysis and decision-making layer is used to perform feature extraction, similarity retrieval, and resource demand prediction on the preprocessed raw data based on the hardware resource assessment model, and to generate resource assessment strategies and optimization suggestions.

[0017] Feature extraction involves extracting the most representative and discriminative feature vectors for resource assessment from the preprocessed raw data. Similarity retrieval retrieves historical resource demand solutions most similar to the current scenario from a historical database or feature vector library based on the feature vectors. Resource demand prediction predicts resource demand over a future period based on the results of feature extraction and similarity retrieval. The generated resource assessment strategy is an assessment conclusion for the current configuration, such as "The current database CPU peak has reached 85%, posing an overload risk." Optimization suggestions provide specific and actionable recommendations based on the assessment conclusions, such as "It is recommended to immediately upgrade the database instance specification from 2C4G to 4C8G."

[0018] The feedback execution layer is used to perform resource adjustment operations based on optimization suggestions, track the execution effect, and feed the execution results back to the analysis and decision layer to update the hardware resource evaluation model.

[0019] Specifically, resource adjustments are performed based on optimization suggestions. For example, after upgrading the database instance specification from 2C4G to 4C8G, changes in relevant indicators are continuously monitored to assess whether the adjustments have achieved the expected results. The quantitative results of the execution, as well as new data generated during the execution process, are returned to the analysis and decision-making layer for model training and optimization.

[0020] In this embodiment, a three-layer closed-loop architecture consisting of a data perception layer, an analysis and decision-making layer, and a feedback and execution layer is constructed to achieve comprehensive perception, intelligent analysis, and automated optimization of hardware resources. The data perception layer not only collects estimated and real-time monitoring indicators for different resource types but also innovatively introduces cross-resource type correlation indicators, breaking down data silos between databases, storage, containers, and other resource types in traditional monitoring. This provides a richer and more interconnected data foundation for subsequent analysis. The analysis and decision-making layer performs feature extraction, similarity retrieval, and resource demand prediction on the preprocessed data based on a hardware resource assessment model. It can mine resource configuration schemes for similar scenarios from historical data and dynamically predict future resource demands by combining real-time monitoring data, generating accurate resource assessment strategies and optimization suggestions. This avoids the traditional reliance on expert experience and static... This application addresses the subjectivity and lag in dynamic assessment. The feedback execution layer automatically performs resource adjustment operations based on optimization suggestions and continuously tracks the execution effect, feeding the results back in real-time to the analysis and decision-making layer to update the hardware resource assessment model. This forms a closed-loop optimization mechanism of "perception-analysis-execution-feedback," enabling the system to have self-learning and continuous evolution capabilities. This application effectively solves the problems of extensive resource assessment, reliance on manual monitoring, and difficulty in dynamic adjustment in existing technologies. It significantly improves the accuracy, dynamism, and automation level of hardware resource assessment, achieving collaborative optimization of multiple resources such as databases, storage, and containers. It can accurately identify resource waste or performance bottlenecks, output differentiated optimization strategies, maximize resource utilization and reduce hardware investment costs while ensuring business stability, and provide a scientific basis for information system capacity planning and risk warning.

[0021] In some embodiments of this application, the data awareness layer includes a data source access module and a data preprocessing module; The data source access module is used to connect to multiple heterogeneous data sources and collect raw data; The data preprocessing module is used to clean, standardize, tag metadata, and normalize time-series data of raw data; and to establish quantitative dependencies between different hardware resource indicators, analyze the positive correlation between changes in business transaction volume and the utilization rate of various hardware resources, and generate correlation indicator data.

[0022] Specifically, multiple heterogeneous data sources can include a system-level resource monitoring platform, a business requirement assessment system, and a budget management system. The system-level resource monitoring platform collects real-time performance monitoring data, such as CPU utilization, memory usage, IOPS, network throughput, and container Pod metrics. The business requirement assessment system collects business-level requirement parameters, such as the number of users, concurrent users, transaction volume, estimated QPS / TPS, and estimated data size. The business requirement assessment system also collects static information about storage resources, such as total storage capacity, used capacity, storage type, storage tier, and lifecycle policies. The budget management system collects cost data related to hardware resources, such as procurement budgets, actual expenditures, resource unit prices, and billing models. By unifying access to multi-source heterogeneous data, the isolated data sources in traditional monitoring systems are broken down, providing rich data raw materials for subsequent correlation analysis.

[0023] Furthermore, the raw data undergoes preprocessing, such as data cleaning to remove outliers, fill in missing data, and deduplication; standardization to unify data from different units and align formats from different data sources; metadata tagging to label data for subsequent dimensional retrieval and analysis; time-series data normalization to resample and align unaligned time-series data, generating normalized time-series data to prepare for time-series analysis; quantitative dependency identification to quantify the dependencies between different hardware resource indicators; and positive correlation analysis to analyze the correlation between changes in business transaction volume and the utilization rate of various hardware resources, such as calculating the correlation coefficient between transactions per second and database CPU utilization, outputting the correlation indicator data "transaction volume - database CPU correlation coefficient".

[0024] This application embodiment breaks the isolation of data sources in traditional monitoring systems by setting up a unified data source access module to connect to the system-level resource monitoring platform, business demand assessment system, storage facility ledger system, and budget management system. This enables comprehensive collection of performance data, business data, asset data, and cost data, providing a more comprehensive and multi-dimensional data foundation for subsequent resource assessment. The data preprocessing module performs data cleaning, standardization, metadata tagging, and time-series data normalization on the raw data, effectively solving problems such as inconsistent formats, asynchronous timing, and noise in multi-source data, ensuring that the data input to the analysis and decision-making layer is of high quality, high consistency, and... High availability; more importantly, this module generates cross-resource type correlation index data by establishing quantitative dependencies between different hardware resource indicators and analyzing the positive correlation between changes in business transaction volume and the utilization rate of various hardware resources. This enables the system to capture resource correlation patterns in complex scenarios such as the pressure transmission of container expansion to the database and the linkage impact of business peaks on storage IOPS. This provides the analysis and decision-making layer with in-depth insights that traditional isolated indicators cannot support for feature extraction, similarity retrieval, and resource demand prediction, thereby significantly improving the accuracy and scientific nature of resource assessment and laying a solid data foundation for the closed-loop optimization of the subsequent feedback execution layer.

[0025] In some embodiments of this application, the analysis decision layer includes a management layer and a model layer.

[0026] The management layer comprises a specification library management module, a policy and rule base management module, and a feedback and execution engine module. The specification library management module maintains the enterprise's standardized cloud resource specification library, which includes at least database instance specifications, container POD specifications, and storage tier specifications for file systems / object storage. The policy and rule base management module stores and manages standardized expert experience, operational guidelines, and security compliance rules. The feedback and execution engine module is used to orchestrate decision instructions, configure execution modes, and measure and provide feedback on execution results. The orchestration of decision instructions involves transforming the optimization suggestions output by the model layer into an executable sequence of instructions. For example, for a combined suggestion of "database expansion + connection pool adjustment," the orchestration would be to first call the cloud API to expand the database, then connect to the application server to modify the connection pool configuration, and finally restart the application. The execution mode configuration supports configuring different execution methods for optimization suggestions, such as automatic execution, semi-automatic execution, simulated execution, and scheduled execution. Execution result measurement involves calling the monitoring interface of the data awareness layer after execution to obtain a comparison of key indicators before and after execution, and to quantitatively evaluate the execution effect.

[0027] In this embodiment, by setting up a specification library management module, a strategy and rule library management module, and a feedback and execution engine module in the management layer, a three-in-one analysis and decision-making system integrating intelligent decision-making, knowledge management, and closed-loop control is constructed, which significantly improves the intelligence level, compliance assurance capability, and closed-loop optimization efficiency of the hardware resource dynamic evaluation system.

[0028] In some embodiments of this application, the model layer includes at least one of the following evaluation models: a database hardware resource evaluation model, a file system / object storage hardware resource evaluation model, and a container resource evaluation model. Each evaluation model includes an AI feature engineering and vectorization module, a similarity retrieval module, and a resource rationality measurement module.

[0029] The AI ​​feature engineering and vectorization module is used to convert the preprocessed raw data into resource status feature vectors; the similarity retrieval module is used to retrieve similar resource demand scenarios in the historical database based on the feature vectors; and the resource rationality measurement module is used to perform resource demand prediction and resource allocation rationality assessment based on the current feature vectors and similar scenario retrieval results, and generate resource assessment strategies and optimization suggestions.

[0030] Specifically, the AI ​​feature engineering and vectorization module's work includes feature extraction, feature fusion, vectorization transformation, and dimensionality reduction. Feature extraction involves extracting key features from time-series monitoring data, such as the average / peak / fluctuation range of CPU utilization, the growth trend of memory utilization, the periodicity of QPS, and the burst characteristics of IOPS. Feature fusion involves combining features of different resource types within the same time period; for example, combining container CPU features, database QPS features, and business transaction volume features into a multi-dimensional feature vector. Vectorization transformation involves encoding the extracted features according to a unified data structure (such as a fixed-dimensional floating-point array) to generate a resource status feature vector; for example, the system status at a certain moment can be represented as a vector of [0.75, 0.32, 0.91, 0.45, ...]. Dimensionality reduction involves reducing the dimensionality of high-dimensional features to reduce computational complexity while retaining key information.

[0031] The similarity detection module calculates the similarity between the current system's resource state feature vector and historical vectors in the vector library. Common algorithms include cosine similarity, Euclidean distance, and inner product. It returns one or more historical scenarios that are most similar to the current state. Each scenario includes the resource configuration, business load, optimization measures, and execution effect at that time. Based on the degree of difference between the current system and historical scenarios, the recalled solutions are weighted or adjusted. For example, if the current system has a 90% similarity to historical scenario A, then solution A has a higher reference weight.

[0032] The resource rationality measurement module predicts resource demand over a future period based on current feature vectors and future trends of similar historical scenarios. For example, it predicts that "database QPS will increase by 50% and CPU peak will reach 85% in the next 2 hours." It assesses the rationality of current resource configuration, identifying two types of abnormal states: resource waste (resource utilization consistently below the threshold, such as average CPU <10%) and resource starvation (resources frequently approaching the allocated amount, such as peak CPU >85% for 30 minutes). Combining the prediction results and rationality measurement results, it generates specific resource assessment strategies and optimization suggestions, such as: "The average container CPU is only 8%, it is recommended to downgrade the Pod specification from 2C4G to 1C2G"; "The database is predicted to be overloaded in 2 hours, it is recommended to expand capacity in advance and add read-only replicas"; "Object storage access volume has decreased by 70% in the past 30 days, it is recommended to migrate cold data to the archive layer, which is expected to save 40% of storage costs." The generated optimization suggestions are assessed for risk level and benefit, and marked as "immediate optimization items" and "planned optimization items." Real-time tuning identifies low-risk optimization operations that can be performed immediately without affecting the application, while planned tuning generally refers to operations that need to be implemented according to the deployment window, such as system configuration parameters. Tuning can support simulating reasonable resource demand solutions based on changes in demand and performance monitoring data.

[0033] In this embodiment, the collaboration of three modules realizes a complete intelligent link from data to decision-making, which significantly improves the accuracy and intelligence level of resource assessment.

[0034] In some embodiments of this application, the database hardware resource evaluation model takes relational databases as the evaluation object and incorporates evaluation rules for estimated indicators, monitoring indicators, and related indicators. These rules are used to output recommended database server specifications, connection pool size adjustment strategies, performance risk warning results, and comprehensive prediction reports. The estimated indicators include at least the number of users, the number of concurrent users, the query frequency per user, the data size, QPS, TPS, and the number of connections. The monitoring indicators include at least CPU utilization, CPU load, memory utilization, disk IOPS, and storage throughput.

[0035] As shown in Tables 1 and 2, Table 1 shows the data required for database resource assessment, and Table 2 shows the output data of the database hardware resource assessment model.

[0036] Table 1

[0037] Table 2

[0038] In some embodiments of this application, the file system / object storage hardware resource evaluation model incorporates evaluation rules for predicted indicators, monitoring indicators, and related indicators, used to output system capacity space prediction reports and system-level storage tier migration strategies. The predicted indicators include at least storage capacity growth rate, file quantity growth rate, file upload TPS, file download TPS, data retention period, and database metadata access QPS; the monitoring indicators include at least data access frequency.

[0039] As shown in Tables 3 and 4, Table 3 shows the data required for the file system / object storage hardware resource evaluation model, and Table 4 shows the output data of the file system / object storage hardware resource evaluation model.

[0040] Table 3

[0041] Table 4

[0042] In some embodiments of this application, the container resource evaluation model has built-in evaluation rules for estimated indicators, monitoring indicators, and related indicators, which are used to output recommended schemes for container POD resource specifications and POD quantity; wherein, the estimated indicators include at least TPS and response latency in core business scenarios; the monitoring indicators include at least CPU utilization and memory utilization; and the related indicators include at least downstream database latency.

[0043] As shown in Tables 5 and 6, Table 5 shows the data required for the file system / object storage hardware resource evaluation model, and Table 6 shows the output data of the file system / object storage hardware resource evaluation model.

[0044] Table 5

[0045] Table 6

[0046] In this embodiment, differentiated evaluation models are constructed for three core hardware resources: databases, file systems / object storage, and containers. Each model has built-in complete evaluation rules covering prediction indicators, monitoring indicators, and related indicators, and can output specific executable results such as server specification recommendations, connection pool adjustment strategies, capacity prediction reports, storage tier migration strategies, and container resource specification and quantity schemes. By integrating the three models into the same system, a full-coverage evaluation and collaborative optimization of structured data storage, unstructured data storage, and elastic computing resources of cloud-based information systems is achieved. Each model independently undertakes the task of accurate evaluation of specific resource types, and forms a cross-resource type mutual verification and joint decision-making mechanism through related indicators (such as the linkage between downstream database latency and container resources). This significantly improves the accuracy and global consistency of the evaluation results, avoids decision bias caused by evaluation of a single resource type, and provides information systems with a comprehensive, refined, and interconnected dynamic evaluation capability for hardware resources.

[0047] In some embodiments of this application, the AI-based hardware resource dynamic evaluation system also includes an application layer that communicates with the analysis and decision-making layer. The application layer includes a resource evaluation and demand forecasting module, a dynamic monitoring and optimization workbench, a cost analysis and performance dashboard, a capacity planning and scenario simulation module, and a unified alarm and anomaly monitoring module.

[0048] The resource assessment and demand forecasting module is used to output a hardware resource demand assessment report for new systems or new business scenarios. The dynamic monitoring and optimization workbench is used to visualize abnormal resource configuration nodes in the form of an architecture topology diagram, and to output and manage real-time and planned optimization solutions. Cost analysis and performance dashboards are used to display resource usage by system, business department, and project dimensions, and output resource utilization rankings and resource waste rankings. The capacity planning and scenario simulation module is used to output a hardware resource demand plan and a corresponding procurement suggestion list for a preset time based on a business strategy template. The unified alarm and anomaly monitoring module is used to output performance bottleneck warnings and abnormal resource demand alarms based on resource assessment results and performance trend predictions.

[0049] In this embodiment, a user-oriented, visual, and operable application layer is constructed, which transforms the intelligent capabilities of the underlying analysis and decision-making layer into intuitive business functions, thereby democratizing and commercializing complex technologies and enabling the entire hardware resource dynamic evaluation system to move from back-end technical capabilities to front-end business value.

[0050] Figure 2 This illustration shows a flowchart of an AI-based dynamic hardware resource evaluation method according to an embodiment of this application. Figure 2 It can be seen that the method may include steps S101 to S103: Step S101: Collect raw data of hardware resources and preprocess the collected raw data.

[0051] The raw data includes estimated indicators, monitoring indicators, and correlation indicators between different resource types.

[0052] Step S102: Based on the hardware resource assessment model, perform feature extraction, similarity retrieval, and resource demand prediction on the preprocessed raw data to generate resource assessment strategies and optimization suggestions.

[0053] Step S103: Perform resource adjustment operations according to the optimization suggestions, track the execution effect, and feed the execution results back to the analysis and decision-making layer to update the hardware resource evaluation model.

[0054] from Figure 2 As shown in the method, this application innovatively introduces cross-resource type correlation index data by collecting estimated and real-time monitoring indicators for different resource types. This breaks down the data silos between resource types such as databases, storage, and containers in traditional monitoring, providing a richer and more interconnected data foundation for subsequent analysis. Based on a hardware resource assessment model, feature extraction, similarity retrieval, and resource demand prediction are performed on the preprocessed data. This allows for the mining of resource allocation schemes under similar scenarios from historical data, and combined with real-time monitoring data, dynamic prediction of future resource demands. This generates accurate resource assessment strategies and optimization suggestions, avoiding the subjectivity and lag of traditional reliance on expert experience and static assessments. Resource adjustment operations are automatically executed based on the optimization suggestions, and... By continuously tracking the execution results and feeding them back to the analysis and decision-making layer in real time to update the hardware resource assessment model, a closed-loop optimization mechanism of "perception-analysis-execution-feedback" is formed, enabling the system to have self-learning and continuous evolution capabilities. This application effectively solves the problems of extensive resource assessment, reliance on manual monitoring, and difficulty in dynamic adjustment in existing technologies, significantly improving the accuracy, dynamism, and automation level of hardware resource assessment. It achieves collaborative optimization of multiple types of resources such as databases, storage, and containers, accurately identifies resource waste or performance bottlenecks, and outputs differentiated optimization strategies. Under the premise of ensuring business stability, it maximizes resource utilization, reduces hardware investment costs, and provides a scientific basis for information system capacity planning and risk warning.

[0055] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Figure 3As shown, at the hardware level, this electronic device includes a processor, and optionally also includes an internal bus, a network interface, and memory. The memory may include main memory, such as high-speed random-access memory (RAM), or it may include non-volatile memory, such as at least one disk drive. Of course, this electronic device may also include other hardware required for other business operations.

[0056] The processor, network interface, and memory can be interconnected via an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 3 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0057] Memory is used to store programs. Specifically, programs may include program code, which includes computer operation instructions. Memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0058] The processor reads the corresponding computer program from non-volatile memory into main memory and then runs it, forming an AI-based dynamic hardware resource evaluation device at the logical level. The processor executes the program stored in memory and specifically performs the aforementioned methods.

[0059] The processor may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in the memory, and the processor reads the information in the memory and, in conjunction with its hardware, completes the steps of the above method.

[0060] The electronic device can execute the AI-based dynamic hardware resource evaluation method provided in several embodiments of this application, and is implemented as an AI-based dynamic hardware resource evaluation device. Figure 1 The functions of the embodiments shown are not described in detail here.

[0061] This application also proposes a computer-readable storage medium that stores one or more programs, the programs including instructions that, when executed by an electronic device including multiple applications, enable the electronic device to perform the AI-based dynamic hardware resource evaluation method provided in several embodiments of this application.

[0062] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0063] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0064] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0065] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0066] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0067] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0068] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0069] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0070] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0071] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. An AI-based dynamic hardware resource evaluation system, characterized in that, The system includes: The data perception layer is used to collect raw data of hardware resources and preprocess the collected raw data; the raw data includes estimated indicators, monitoring indicators and correlation indicators between different resource types. The analysis and decision-making layer is used to perform feature extraction, similarity retrieval, and resource demand prediction on the preprocessed raw data based on the hardware resource assessment model, and to generate resource assessment strategies and optimization suggestions. The feedback execution layer is used to perform resource adjustment operations based on the optimization suggestions, track the execution effect, and feed the execution results back to the analysis and decision layer to update the hardware resource evaluation model.

2. The system according to claim 1, characterized in that, The data perception layer includes a data source access module and a data preprocessing module; The data source access module is used to connect to multiple heterogeneous data sources and collect the raw data; The data preprocessing module is used to perform data cleaning, standardization, metadata tagging, and time-series data normalization on the raw data. Furthermore, it establishes quantitative dependencies between different hardware resource indicators, analyzes the positive correlation between changes in business transaction volume and the utilization rate of various hardware resources, and generates correlation indicator data.

3. The system according to claim 1, characterized in that, The analysis and decision-making layer includes a management layer and a model layer; wherein, the management layer includes a specification library management module, a strategy and rule library management module, and a feedback and execution engine module; The specification library management module is used to maintain the enterprise's standardized cloud resource specification library; The strategy and rule base management module is used to store and manage standardized expert experience, operation and maintenance specifications, and security compliance rules. The feedback and execution engine module is used to complete the arrangement of decision instructions, the configuration of execution modes, and the measurement and feedback of execution effects.

4. The system according to claim 3, characterized in that, The model layer includes at least one of the following evaluation models: database hardware resource evaluation model, file system / object storage hardware resource evaluation model, and container resource evaluation model. Each evaluation model includes an AI feature engineering and vectorization module, a similarity retrieval module, and a resource rationality measurement module. The AI ​​feature engineering and vectorization module is used to convert the preprocessed raw data into resource status feature vectors; The similarity retrieval module is used to retrieve similar resource demand scenarios in the historical database based on the feature vector; The resource rationality measurement module is used to perform resource demand prediction and resource allocation rationality assessment based on the current feature vector and similar scenario retrieval results, and generate the resource assessment strategy and optimization suggestions.

5. The system according to claim 1, characterized in that, The database hardware resource assessment model takes relational databases as the assessment object and has built-in assessment rules for prediction indicators, monitoring indicators and correlation indicators. It is used to output recommended database server specifications, connection pool size adjustment strategies, performance risk warning results and comprehensive prediction reports. The estimated metrics include at least the number of users, concurrent users, query frequency per user, data size, QPS, TPS, and number of connections; the monitoring metrics include at least CPU utilization, CPU load, memory utilization, disk IOPS, and storage throughput.

6. The system according to claim 1, characterized in that, The file system / object storage hardware resource assessment model has built-in assessment rules for estimated indicators, monitoring indicators, and related indicators, which are used to output system capacity space prediction reports and system-level storage tier migration strategies. The estimated indicators include at least the growth rate of storage capacity, the growth rate of the number of files, file upload TPS, file download TPS, data retention period, and database metadata access QPS; the monitoring indicators include at least the data access frequency.

7. The system according to claim 1, characterized in that, The container resource assessment model has built-in assessment rules for estimated indicators, monitoring indicators, and related indicators, which are used to output recommended solutions for container POD resource specifications and POD quantity. The estimated metrics include at least TPS and response latency for core business scenarios; the monitoring metrics include at least CPU utilization and memory utilization; and the related metrics include at least downstream database latency.

8. The system according to claim 1, characterized in that, The system also includes an application layer that communicates with the analysis and decision-making layer. The application layer includes a resource assessment and demand forecasting module, a dynamic monitoring and optimization workbench, a cost analysis and performance dashboard, a capacity planning and scenario simulation module, and a unified alarm and anomaly monitoring module. The resource assessment and demand forecasting module is used to output a hardware resource demand assessment report for new systems or new business scenarios. The dynamic monitoring and optimization workbench is used to visually display abnormal resource configuration nodes in the form of an architecture topology diagram, and to output and manage real-time optimization and planned optimization solutions. The cost analysis and performance dashboard is used to display resource usage by system, business department and project dimensions, and output resource utilization ranking and resource waste ranking. The capacity planning and scenario simulation module is used to output a hardware resource demand plan and a corresponding procurement suggestion list for a preset time based on the business strategy template. The unified alarm and anomaly monitoring module is used to output performance bottleneck warnings and abnormal resource demand alarms based on resource assessment results and performance trend predictions.

9. A dynamic evaluation method for hardware resources based on AI, characterized in that, The method includes: Raw data of hardware resources is collected and preprocessed; the raw data includes estimated indicators, monitoring indicators and correlation indicators between different resource types. Based on the hardware resource assessment model, feature extraction, similarity retrieval, and resource demand prediction are performed on the preprocessed raw data to generate resource assessment strategies and optimization suggestions. Based on the optimization suggestions, resource adjustment operations are performed, and the execution effect is tracked. The execution results are then fed back to the analysis and decision-making layer to update the hardware resource evaluation model.

10. An electronic device, comprising: processor; as well as A memory configured to store computer-executable instructions, characterized in that, when executed, the executable instructions cause the processor to perform the steps of the AI-based dynamic hardware resource evaluation method as described in claim 9.

Images