Cloud rendering resource intelligent management method and device, computer equipment and storage medium
By collecting data by monitoring agents and using AI models to build a global resource view for comprehensive scoring, the problem of low management efficiency of cloud rendering nodes is solved, efficient rendering task execution and resource optimization are achieved, and system performance and user experience are improved.
Patent Information
- Application Number
- CN202411046963.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-01
- Publication Date
- 2024-12-06
AI Technical Summary
The existing cloud rendering node management efficiency is low, and the efficiency of the automatic scheduling method decreases after the node type is complicated, making it difficult to achieve efficient rendering task execution and resource optimization.
The monitoring agent collects resource usage data of cloud rendering nodes, combines the AI model to build a global resource view and performs comprehensive scoring, realizes task scheduling and optimizes resource allocation.
It realizes efficient execution and resource optimization of cloud rendering tasks under the optimal node conditions, reduces latency and waiting time, and improves rendering quality and system efficiency.
Smart Images

Figure CN119088513A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a resource management method, and more specifically to a cloud rendering resource intelligent management method, device, computer equipment and storage medium. Background Art
[0002] In the application of digital twin technology, cloud rendering technology is the key to simulating and visualizing complex industrial, architectural or other business scenarios. At the current stage, the management of cloud rendering nodes has mainly gone through two stages: manual scheduling and automatic scheduling.
[0003] Manual scheduling is usually inefficient in the context of rapid business expansion and high labor costs. It relies on maintenance personnel to manually specify the nodes and priorities of rendering tasks, which is easily limited by human factors and task priority management. Automatic scheduling reduces repetitive maintenance work, reduces labor costs, and improves maintenance efficiency through preset rule scripts. However, with the complexity and diversity of cloud rendering node types, the automatic scheduling method based on manually specified rules has gradually become incapable and its efficiency has declined.
[0004] Therefore, it is necessary to design a new method to solve the problems of efficient execution and resource optimization of cloud rendering tasks, to realize automatic management and optimization of cloud nodes, and to enable rendering tasks to be completed efficiently under the best node conditions. Summary of the invention
[0005] The purpose of the present invention is to overcome the defects of the prior art and provide a cloud rendering resource intelligent management method, device, computer equipment and storage medium.
[0006] To achieve the above objectives, the present invention adopts the following technical solution: a cloud rendering resource intelligent management method, comprising:
[0007] Obtain resource usage data of the cloud rendering node collected by a monitoring agent deployed on the cloud rendering node;
[0008] Determine node resource allocation decisions by combining the AI model and the resource usage data;
[0009] Building a global view of resource usage of each cloud rendering node based on the resource usage data and performing a comprehensive score;
[0010] Task scheduling is performed using the comprehensive score combined with the node resource allocation decision.
[0011] Its further technical solution is: the resource usage data includes CPU utilization, GPU utilization, memory occupancy, and network throughput.
[0012] Its further technical solution is: the AI model is obtained by training the machine learning network using the historical data of the cloud rendering nodes collected by the monitoring agent as a sample set, and the AI model uses cross-validation and A / B testing methods to verify the performance and accuracy of the model.
[0013] A further technical solution is: constructing a global view of resource usage of each cloud rendering node based on the resource usage data and performing a comprehensive score, including:
[0014] Building a global view of resource usage of each cloud rendering node based on the resource usage data;
[0015] The available resources of each cloud rendering node are evaluated in real time to obtain a comprehensive score.
[0016] A further technical solution is: the available resources of each cloud rendering node are evaluated in real time to obtain a comprehensive score, including:
[0017] The available resources of each cloud rendering node are comprehensively evaluated from the comparison of total resources and used resources, historical trends of resource usage, weights of different resource types, hardware configuration of the node and performance indicators to obtain a comprehensive score.
[0018] A further technical solution is: the task scheduling is performed by combining the comprehensive score with the node resource allocation decision, including:
[0019] The node resource allocation decision is converted into a specific task distribution instruction according to the comprehensive score, and sent to the cloud platform API so that the task can be scheduled and executed among the nodes.
[0020] The present invention also provides a cloud rendering resource intelligent management device, including:
[0021] A data acquisition unit, used to acquire resource usage data of a cloud rendering node collected by a monitoring agent program deployed on the cloud rendering node;
[0022] A decision determination unit, used to determine a node resource allocation decision in combination with the AI model and the resource usage data;
[0023] A scoring unit, used to construct a global view of resource usage of each cloud rendering node based on the resource usage data, and perform a comprehensive score;
[0024] A scheduling unit is used to schedule tasks by utilizing the comprehensive score in combination with the node resource allocation decision.
[0025] Its further technical solution is: the scoring unit includes:
[0026] A construction subunit, configured to construct a global view of resource usage of each cloud rendering node based on the resource usage data;
[0027] The real-time evaluation subunit is used to evaluate the available resources of each cloud rendering node in real time to obtain a comprehensive score.
[0028] The present invention further provides a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor implements the above method when executing the computer program.
[0029] The present invention also provides a storage medium, wherein the storage medium stores a computer program, and the computer program implements the above method when executed by a processor.
[0030] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention collects resource usage data of cloud rendering nodes through a monitoring agent, determines the node resource allocation strategy in combination with an AI model, builds a global resource view and comprehensively scores, thereby realizing task scheduling, solving the problems of efficient execution and resource optimization of cloud rendering tasks, and realizing automated management and optimization of cloud nodes, so that rendering tasks can be efficiently completed under optimal node conditions.
[0031] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.
[0033] Figure 1 A schematic diagram of an application scenario of the cloud rendering resource intelligent management method provided by an embodiment of the present invention;
[0034] Figure 2 A schematic diagram of a process flow of a cloud rendering resource intelligent management method provided by an embodiment of the present invention;
[0035] Figure 3 A schematic diagram of a sub-process of a method for intelligent management of cloud rendering resources provided by an embodiment of the present invention;
[0036] Figure 4 A schematic block diagram of a cloud rendering resource intelligent management device provided by an embodiment of the present invention;
[0037] Figure 5 A schematic block diagram of a scoring unit of a cloud rendering resource intelligent management device provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0040] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.
[0041] It should also be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.
[0042] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.
[0043] See also Figure 1 , Figure 1 A schematic flow chart of a cloud-based rendering resource intelligent management method provided in an embodiment of the present invention. The cloud-based rendering resource intelligent management method is applied in a server. The server interacts with cloud-based rendering nodes, and in a cloud environment, based on a distributed architecture, efficiently and dynamically schedules node resources of a large-scale rendering cluster, significantly improving the processing capacity of rendering tasks, and while ensuring excellent rendering quality, greatly reducing latency and waiting time, providing users with a smoother and more efficient service experience, solving the problems of efficient execution of cloud-based rendering tasks and resource optimization, and realizing automated management and optimization of cloud-based nodes, so that rendering tasks can be efficiently completed under optimal node conditions.
[0044] Figure 1 FIG. 1 is a flow chart of a method for intelligently managing cloud rendering resources provided by an embodiment of the present invention. Figure 1 As shown, the method includes the following steps S110 to S140.
[0045] S110: Obtain resource usage data of the cloud rendering node collected by a monitoring agent program deployed on the cloud rendering node.
[0046] In this embodiment, the resource usage data includes CPU utilization, GPU utilization, memory occupancy, and network throughput.
[0047] CPU utilization refers to the degree to which the workload of the processor (CPU) is used relative to its total capacity. It is usually expressed as a percentage, indicating the workload of the CPU in a specific period of time.
[0048] GPU utilization refers to the extent to which the workload of the graphics processing unit (GPU) is used relative to its total capacity. Similar to the CPU, GPU utilization is usually expressed as a percentage to evaluate the efficiency of the GPU in processing graphics calculations or other parallel computing tasks.
[0049] Memory usage refers to the amount of memory in the system that is occupied at any time. This includes the amount of physical memory that is already in use by applications and system services, as well as the portion of memory that is reserved but not in use.
[0050] Network throughput refers to the amount of data that passes through a network. Network throughput can be described as the amount of data that passes through a network device (such as a network card or router) per unit time, usually in bits or bytes per second (bps or Bps).
[0051] Specifically, a lightweight monitoring agent is deployed on the node, and a stable long link can be established with the resource monitoring center. The agent is responsible for periodically collecting the node's resource usage data, including CPU utilization, GPU utilization, memory usage, network throughput and other indicators.
[0052] In this embodiment, a dedicated monitoring agent is deployed on each cloud rendering node to collect key node resource utilization and operating status indicators in real time. These monitoring agents are lightweight and are designed to minimize the impact on node performance. They can be flexibly deployed on each node and support automated deployment of the cloud platform, thereby significantly improving deployment efficiency.
[0053] Data collection and reporting are completed by these monitoring agents. They continuously collect key data such as node resource utilization and task execution. These data are regularly transmitted to the central resource monitoring center through a secure and reliable communication protocol. This process ensures the integrity and confidentiality of data transmission, while enabling the monitoring center to grasp the status of global nodes in real time, providing accurate data support for subsequent resource allocation and optimization decisions.
[0054] S120. Determine a node resource allocation decision by combining the AI model and the resource usage data.
[0055] The AI model is obtained by training a machine learning network using historical data of cloud rendering nodes collected by a monitoring agent as a sample set, and the AI model uses cross-validation and A / B testing methods to verify the performance and accuracy of the model.
[0056] In this embodiment, the resource status of the current node and the predicted resource demand are analyzed based on the trained AI model. According to the analysis results, a specific node resource allocation strategy is formulated, such as increasing or decreasing the computing resources and storage resources of the node. These allocation decisions are implemented through automated scripts to ensure the timeliness and efficiency of scheduling tasks.
[0057] Continuously collect the latest monitoring data and periodically retrain the AI model. Optimize scheduling strategies to ensure that the system can adapt to dynamically changing workloads and resource requirements.
[0058] Specifically, using historical data and AI models, the monitoring center dynamically extracts scheduling rules suitable for the current node cluster. These rules are used for real-time node resource allocation decisions and execute scheduling tasks through automated scripts.
[0059] Based on historical data, use machine learning algorithms (such as linear regression, random forest, neural network, etc.) to predict node resource requirements. Training data includes node resource utilization, task execution, and historical scheduling records. Continuously collect the latest monitoring data, regularly retrain AI models, and optimize scheduling strategies. Use methods such as cross-validation and A / B testing to verify the performance and accuracy of the model.
[0060] This method of combining historical data and AI models enables the monitoring center to efficiently manage and optimize the allocation of node resources to meet the ever-changing rendering task requirements and improve the performance and efficiency of the overall system.
[0061] S130: construct a global view of resource usage of each cloud rendering node based on the resource usage data, and perform a comprehensive score.
[0062] In one embodiment, see Figure 2 , the above-mentioned step S130 may include steps S131 to S132.
[0063] S131. Construct a global view of resource usage of each cloud rendering node according to the resource usage data.
[0064] In this embodiment, the collected data such as resource utilization, storage usage, network bandwidth consumption, etc. of each node are integrated. Through data visualization tools or customized data analysis tools, the resource usage of each node is visualized as a global view. This can be in the form of a dashboard, chart, or report, which makes it easy for administrators or automated systems to quickly understand the current status of each node.
[0065] Specifically, a lightweight monitoring agent is deployed on each cloud rendering node to collect the node's resource usage data in real time.
[0066] The monitoring agent collects the following key metrics: CPU usage: overall and per-core usage; GPU usage: GPU utilization; Memory usage and usage: used memory and total memory; Disk I / O rate: read and write rate; Network bandwidth usage: upload and download bandwidth; Number of rendering tasks currently running.
[0067] Use message queues as the middleware for data transmission to improve the reliability and scalability of data transmission. Set a reasonable reporting frequency (every 30 seconds) to avoid excessive network overhead.
[0068] The monitoring center deploys a cloud data receiving platform to decompress, decrypt, and format-verify the received data, and store the data. It also provides a visual monitoring panel display, showing the node resource usage heat map and the cumulative resource usage trend map.
[0069] Through the above implementation steps, a powerful and reliable resource monitoring center is realized, which can realize real-time monitoring of distributed rendering nodes and build a global resource view, providing important data support for intelligent task scheduling and resource optimization.
[0070] S132. Evaluate the available resources of each cloud rendering node in real time to obtain a comprehensive score.
[0071] In this embodiment, the available resources of each cloud rendering node are comprehensively evaluated from the comparison of the total resources and the used resources, the historical trend of resource usage, the weights of different resource types, the hardware configuration and performance indicators of the node to obtain a comprehensive score.
[0072] Specifically, resource demand prediction and scheduling decisions are made based on the resource usage data of the cloud rendering nodes; a comprehensive score for each node is given by comprehensively considering the node hardware configuration, resource usage trends, etc.
[0073] Define indicators for evaluating node resource conditions, such as the comparison between total resources and used resources, historical resource usage trends, weights of different resource types (such as computing resources, storage resources, network bandwidth, etc.), hardware configuration and performance indicators of nodes (such as number of CPU cores, memory capacity, network speed, etc.). Based on the defined evaluation indicators, the system will calculate and evaluate the available resources of each node in real time. This may involve mathematical models or rule engines to calculate a comprehensive score based on the importance and weight of each indicator. The comprehensive score can serve as an important basis for node resource allocation decisions. Nodes with high comprehensive scores are generally considered to have high resource utilization efficiency and are suitable for processing more tasks or tasks that require high performance.
[0074] Through real-time evaluation, administrators or systems can quickly understand the resource status of each node, supporting rapid decision-making and adjustment; through comprehensive scoring, the system can optimize resource allocation to ensure that the resources of each node are optimally utilized, thereby improving overall system efficiency; through historical trend analysis, changes in resource demand can be predicted, and expansion or reduction can be made in a timely manner to avoid resource shortages or waste; the global view and comprehensive scoring provide transparency and monitoring capabilities for the entire cloud rendering node group, helping administrators to promptly identify and resolve potential problems.
[0075] Specifically, in order to establish a comprehensive scoring model to evaluate the available resources of each cloud rendering node, a weighted scoring formula was used: S = Wcpu*C + Wgpu*G + Wmem*M + Wbw*B;
[0076] Where: S: comprehensive score of the node (range of 0-100 or 0-1); C: CPU performance score (0-100); G: GPU performance score (0-100); M: memory performance score (0-100); B: network bandwidth score (0-100); Wcpu, Wgpu, Wmem, Wbw: weights of each indicator, satisfying the condition Wgpu+Wgpu+Wmem+Wbw=1.
[0077] Calculation of each score:
[0078] CPU performance score C: C = (U ÷ T) × 100%; U: current CPU utilization (%); T: total CPU performance (theoretical maximum performance).
[0079] GPU performance score G: G = (U ÷ T) × 100%; U: current GPU utilization (%); T: total performance of GPU (theoretical maximum performance).
[0080] Memory performance score M: M = (U ÷ T) × 100%; U: current memory utilization (%); T: total memory capacity;
[0081] Network bandwidth score B: B = (U ÷ T) × 100%; U: current network bandwidth utilization (%); T: total network bandwidth.
[0082] Weight setting: The weights Wcpu, Wgpu, Wmem, and Wbw should be adjusted according to the specific application scenario and task requirements. For example, in a graphics rendering task, a higher weight may be given to the GPU.
[0083] The above formula can be used to comprehensively evaluate the available resources of each cloud rendering node, thus providing an effective basis for task scheduling. This model can be adjusted and optimized according to actual conditions to adapt to different rendering requirements and resource configurations.
[0084] S140: Perform task scheduling by utilizing the comprehensive score in combination with the node resource allocation decision.
[0085] In this embodiment, the node resource allocation decision is converted into a specific task distribution instruction according to the comprehensive score, and sent to the cloud platform API so that the task can be scheduled and executed between nodes.
[0086] Specifically, resource scheduling decision: In the resource monitoring center, task scheduling decisions are generated based on real-time collected node resource usage data and comprehensive scores. The specific steps are as follows:
[0087] Score calculation: Obtain node score: Use the comprehensive score calculation formula mentioned above to obtain the comprehensive score (S) of each node in the resource monitoring center;
[0088] Determine task requirements: Analyze the resource requirements of the task to be performed, including the required CPU, GPU, memory, and network bandwidth;
[0089] Task allocation strategy:
[0090] Task classification: Classify tasks according to their resource requirements (CPU intensive, GPU intensive); Node selection: Match task requirements with node scores and select the node with the highest score that meets resource requirements.
[0091] Generate task distribution instructions, the instruction content includes: task ID; target node ID; required resources (number of CPU cores, number of GPUs, memory size, etc.); task execution command; call the cloud platform API; send an HTTP request to send the task distribution instruction to the task scheduling interface of the cloud platform API.
[0092] A lightweight monitoring agent is deployed on each cloud rendering node to collect key indicator data such as the node's resource utilization and operating status in real time.
[0093] The method of this embodiment uses historical monitoring data for learning modeling, dynamically optimizes node resource scheduling and task allocation strategies; achieves adaptability and real-time performance, and is able to cope with dynamically changing rendering needs; automatically adjusts the entry and exit of nodes based on real-time monitoring data to achieve flexible expansion and contraction capabilities; and automatically allocates rendering tasks to maximize the utilization efficiency of node resources.
[0094] Design lightweight, low-overhead monitoring agents to minimize the impact on node performance and ensure the stability of core business services; support flexible monitoring agent deployment methods, automate deployment and seamlessly integrate with cloud platform tools to improve deployment efficiency; ensure the reliability and timeliness of monitoring data, have high fault tolerance, and ensure data accuracy even under high load conditions; continuously optimize scheduling algorithms based on big data and machine learning to achieve intelligent resource management and task allocation, and improve the overall efficiency and performance of the system.
[0095] Specifically, through the real-time resource monitoring and evaluation mechanism, the system can instantly understand the resource usage of each cloud rendering node. This precise resource understanding enables the scheduling system to allocate tasks more effectively, avoid resource waste and node idleness, and thus maximize resource utilization.
[0096] Based on real-time monitoring and evaluation data, the scheduling system can intelligently allocate tasks according to node resource scores. This approach helps avoid over-allocation of tasks to nodes with limited resources, thereby improving the efficiency and performance of overall task processing.
[0097] The monitoring system adopts a distributed architecture, which can distribute monitoring tasks to multiple sub-nodes and increase the overall throughput of the system. This architectural design enables the system to easily cope with the increase in the number of cloud nodes and has better scalability and elasticity.
[0098] Through an efficient monitoring data collection and processing mechanism, the monitoring system can timely acquire and process the resource usage data of the nodes. The denoising and smoothing of this data improves the quality of the data and helps to accurately assess the resource status of the nodes. As a result, the scheduling system can make more reliable and accurate scheduling decisions to respond to real-time changes and demands.
[0099] This link is the data collection stage of monitoring each node. The data reported by the node usually has abnormal jitter, missing values, and abnormal values due to network jitter or accidental peak interference. The main function of this link is to avoid jitter of the scheduling algorithm due to unreasonable data. The common practice is to fill missing values, remove abnormal values, and process inconsistent data. And standardize or normalize the data. So as to meet the input requirements of the scheduling algorithm.
[0100] These optimization strategies work together on the cloud rendering system to improve its efficiency, performance, and reliability, thereby providing users with a better and more efficient service experience.
[0101] The above-mentioned intelligent management method for cloud rendering resources collects resource usage data of cloud rendering nodes through a monitoring agent program, determines the node resource allocation strategy in combination with an AI model, builds a global resource view and performs a comprehensive score, thereby realizing task scheduling, solving the problems of efficient execution and resource optimization of cloud rendering tasks, and realizing automated management and optimization of cloud nodes, so that rendering tasks can be efficiently completed under the best node conditions.
[0102] Figure 3 is a schematic block diagram of a cloud rendering resource intelligent management device 300 provided in an embodiment of the present invention. Figure 3 As shown, corresponding to the above cloud rendering resource intelligent management method, the present invention also provides a cloud rendering resource intelligent management device 300. The cloud rendering resource intelligent management device 300 includes a unit for executing the above cloud rendering resource intelligent management method, and the device can be configured in a server. Figure 3 The cloud rendering resource intelligent management device 300 includes a data acquisition unit 301, a decision determination unit 302, a scoring unit 303 and a scheduling unit 304.
[0103] The data acquisition unit 301 is used to acquire the resource usage data of the cloud rendering node collected by the monitoring agent deployed on the cloud rendering node; the decision determination unit 302 is used to determine the node resource allocation decision in combination with the AI model and the resource usage data; the scoring unit 303 is used to build a global view of the resource usage of each cloud rendering node according to the resource usage data, and perform a comprehensive score; the scheduling unit 304 is used to perform task scheduling by using the comprehensive score in combination with the node resource allocation decision.
[0104] In one embodiment, if Figure 4 As shown, the scoring unit 303 includes a construction subunit 3031 and a real-time evaluation subunit 3032 .
[0105] The construction subunit 3031 is used to construct a global view of the resource usage of each cloud rendering node based on the resource usage data; the real-time evaluation subunit 3032 is used to evaluate the available resources of each cloud rendering node in real time to obtain a comprehensive score.
[0106] In one embodiment, the real-time evaluation subunit 3032 is used to comprehensively evaluate the available resources of each cloud rendering node from the comparison of the total resources and the used resources, the historical trend of resource usage, the weights of different resource types, the hardware configuration of the node and the performance indicators to obtain a comprehensive score.
[0107] In one embodiment, the scheduling unit 304 is used to convert the node resource allocation decision into a specific task distribution instruction according to the comprehensive score, and send it to the cloud platform API so that the task can be scheduled and executed between nodes.
[0108] It should be noted that technicians in the relevant field can clearly understand that the specific implementation process of the above-mentioned cloud rendering resource intelligent management device 300 and each unit can refer to the corresponding description in the aforementioned method embodiment, and for the convenience and brevity of description, it will not be repeated here.
[0109] The cloud rendering resource intelligent management device 300 can be implemented in the form of a computer program. Figure 5 Runs on the computer device shown.
[0110] See also Figure 5 , Figure 5 5 is a schematic block diagram of a computer device provided in an embodiment of the present application. The computer device 500 may be a server, wherein the server may be an independent server or a server cluster composed of multiple servers.
[0111] See also Figure 5 The computer device 500 includes a processor 502 , a memory and a network interface 505 connected via a system bus 501 , wherein the memory may include a non-volatile storage medium 503 and an internal memory 504 .
[0112] The non-volatile storage medium 503 can store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions, and when the program instructions are executed, the processor 502 can execute a cloud rendering resource intelligent management method.
[0113] The processor 502 is used to provide computing and control capabilities to support the operation of the entire computer device 500 .
[0114] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute a cloud rendering resource intelligent management method.
[0115] The network interface 505 is used to communicate with other devices over the network. Figure 5The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present application, and does not constitute a limitation on the computer device 500 to which the solution of the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0116] The processor 502 is used to run the computer program 5032 stored in the memory to implement the following steps:
[0117] Obtain resource usage data of cloud rendering nodes collected by a monitoring agent deployed on the cloud rendering nodes; determine node resource allocation decisions in combination with an AI model and the resource usage data; construct a global view of resource usage of each cloud rendering node based on the resource usage data and perform a comprehensive score; and perform task scheduling using the comprehensive score in combination with the node resource allocation decision.
[0118] The resource usage data includes CPU utilization, GPU utilization, memory usage, and network throughput.
[0119] The AI model is obtained by training a machine learning network using historical data of cloud rendering nodes collected by a monitoring agent as a sample set, and the AI model uses cross-validation and A / B testing methods to verify the performance and accuracy of the model.
[0120] In one embodiment, when the processor 502 implements the step of building a global view of resource usage of each cloud rendering node according to the resource usage data and performing a comprehensive scoring step, the processor 502 specifically implements the following steps:
[0121] A global view of resource usage of each cloud rendering node is constructed based on the resource usage data; and the available resources of each cloud rendering node are evaluated in real time to obtain a comprehensive score.
[0122] In one embodiment, when the processor 502 implements the step of evaluating the available resources of each cloud rendering node in real time to obtain a comprehensive score, the processor 502 specifically implements the following steps:
[0123] The available resources of each cloud rendering node are comprehensively evaluated from the comparison of total resources and used resources, historical trends of resource usage, weights of different resource types, hardware configuration of the node and performance indicators to obtain a comprehensive score.
[0124] In one embodiment, when the processor 502 implements the step of performing task scheduling by using the comprehensive score in combination with the node resource allocation decision, the processor 502 specifically implements the following steps:
[0125] The node resource allocation decision is converted into a specific task distribution instruction according to the comprehensive score, and sent to the cloud platform API so that the task can be scheduled and executed among the nodes.
[0126] It should be understood that in the embodiment of the present application, the processor 502 may be a central processing unit (CPU), and the processor 502 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.
[0127] It can be understood by those skilled in the art that all or part of the processes in the method for implementing the above embodiment can be completed by instructing the relevant hardware through a computer program. The computer program includes program instructions, and the computer program can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiment of the above method.
[0128] Therefore, the present invention also provides a storage medium. The storage medium may be a computer-readable storage medium. The storage medium stores a computer program, wherein when the computer program is executed by a processor, the processor executes the following steps:
[0129] Obtain resource usage data of cloud rendering nodes collected by a monitoring agent deployed on the cloud rendering nodes; determine node resource allocation decisions in combination with an AI model and the resource usage data; construct a global view of resource usage of each cloud rendering node based on the resource usage data and perform a comprehensive score; and perform task scheduling using the comprehensive score in combination with the node resource allocation decision.
[0130] The resource usage data includes CPU utilization, GPU utilization, memory usage, and network throughput.
[0131] The AI model is obtained by training a machine learning network using historical data of cloud rendering nodes collected by a monitoring agent as a sample set, and the AI model uses cross-validation and A / B testing methods to verify the performance and accuracy of the model.
[0132] In one embodiment, when the processor executes the computer program to implement the steps of building a global view of resource usage of each cloud rendering node according to the resource usage data and performing a comprehensive scoring step, the processor specifically implements the following steps:
[0133] A global view of resource usage of each cloud rendering node is constructed based on the resource usage data; and the available resources of each cloud rendering node are evaluated in real time to obtain a comprehensive score.
[0134] In one embodiment, when the processor executes the computer program to implement the step of evaluating the available resources of each cloud rendering node in real time to obtain a comprehensive score, the processor specifically implements the following steps:
[0135] The available resources of each cloud rendering node are comprehensively evaluated from the comparison of total resources and used resources, historical trends of resource usage, weights of different resource types, hardware configuration of the node and performance indicators to obtain a comprehensive score.
[0136] In one embodiment, when the processor executes the computer program to implement the step of performing task scheduling by using the comprehensive score in combination with the node resource allocation decision, the processor specifically implements the following steps:
[0137] The node resource allocation decision is converted into a specific task distribution instruction according to the comprehensive score, and sent to the cloud platform API so that the task can be scheduled and executed among the nodes.
[0138] The storage medium may be a USB flash drive, a mobile hard disk, a read-only memory (ROM), a magnetic disk, or an optical disk, etc., which are computer-readable storage media that can store program codes.
[0139] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.
[0140] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of each unit is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.
[0141] The steps in the method of the embodiment of the present invention can be adjusted in order, combined and deleted according to actual needs. The units in the device of the embodiment of the present invention can be combined, divided and deleted according to actual needs. In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0142] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, terminal, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention.
[0143] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.
Claims
1. Intelligent management method of cloud rendering resources, It is characterized in that include: Obtain resource usage data of the cloud rendering node collected by a monitoring agent deployed on the cloud rendering node; Determine node resource allocation decisions by combining the AI model and the resource usage data; Building a global view of resource usage of each cloud rendering node based on the resource usage data and performing a comprehensive score; Task scheduling is performed using the comprehensive score combined with the node resource allocation decision.
2. The method for intelligent management of cloud rendering resources according to claim 1, It is characterized in that The resource usage data includes CPU utilization, GPU utilization, memory usage, and network throughput.
3. The method for intelligent management of cloud rendering resources according to claim 2, It is characterized in that The AI model is obtained by training a machine learning network using historical data of cloud rendering nodes collected by a monitoring agent as a sample set, and the AI model uses cross-validation and A / B testing methods to verify the performance and accuracy of the model.
4. The method for intelligent management of cloud rendering resources according to claim 1, It is characterized in that The global view of resource usage of each cloud rendering node is constructed based on the resource usage data, and a comprehensive score is performed, including: Building a global view of resource usage of each cloud rendering node based on the resource usage data; The available resources of each cloud rendering node are evaluated in real time to obtain a comprehensive score.
5. The method for intelligent management of cloud rendering resources according to claim 4, It is characterized in that The available resources of each cloud rendering node are evaluated in real time to obtain a comprehensive score, including: The available resources of each cloud rendering node are comprehensively evaluated from the comparison of total resources and used resources, historical trends of resource usage, weights of different resource types, hardware configuration of the node and performance indicators to obtain a comprehensive score.
6. The method for intelligent management of cloud rendering resources according to claim 1, It is characterized in that The utilizing the comprehensive score in combination with the node resource allocation decision to perform task scheduling includes: The node resource allocation decision is converted into a specific task distribution instruction according to the comprehensive score, and sent to the cloud platform API so that the task can be scheduled and executed among the nodes.
7. Cloud rendering resource intelligent management device, It is characterized in that include: A data acquisition unit, used to acquire resource usage data of a cloud rendering node collected by a monitoring agent program deployed on the cloud rendering node; A decision determination unit, used to determine a node resource allocation decision in combination with the AI model and the resource usage data; A scoring unit, used to construct a global view of resource usage of each cloud rendering node based on the resource usage data, and perform a comprehensive score; A scheduling unit is used to schedule tasks by utilizing the comprehensive score in combination with the node resource allocation decision.
8. The cloud rendering resource intelligent management device according to claim 7, It is characterized in that The scoring unit comprises: A construction subunit, configured to construct a global view of resource usage of each cloud rendering node based on the resource usage data; The real-time evaluation subunit is used to evaluate the available resources of each cloud rendering node in real time to obtain a comprehensive score.
9. A computer device, It is characterized in that The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 6 when executing the computer program.
10. A storage medium, It is characterized in that The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.
Citation Information
Cited By
Cloud rendering method based on middleware and middleware system
CN121330139A
