A resource intelligent agent service undertaking method for a computing power network environment

Abstract

This invention relates to a resource intelligent agent service acceptance method for computing power network environments, belonging to the field of computing power network technology. This method uses resource intelligent agents as autonomous resource service units in computing power networks. By constructing a mapping relationship between resource capabilities and service requirements, it introduces service capability profiling, service adaptation analysis, service-specific reputation assessment, lifecycle state awareness, and acceptance capability declaration generation mechanism. This enables resource intelligent agents to transform from expressing resource availability to expressing service acceptability, allowing resource intelligent agents to proactively describe their service capabilities, fulfillment capabilities, and acceptance willingness in specific business scenarios, providing decision-making basis for service scheduling, resource orchestration, and intelligent agent collaboration.

Description

Technical Field

[0001] This invention belongs to the field of computing power network technology and relates to a resource intelligent agent service undertaking method for computing power network environment. Background Technology

[0002] With the rapid development of 5G, cloud computing, edge computing, big data models, and artificial intelligence technologies, the types of services carried by networks are constantly diversifying, and service demands are increasingly characterized by low latency, high reliability, high bandwidth, and intelligence. The traditional resource supply model centered on centralized data centers is insufficient to meet the real-time acquisition needs of massive heterogeneous services for computing resources. Computing, storage, and network resources are gradually evolving towards cloud-edge-device collaboration and ubiquitous distributed deployment. To achieve unified connection, management, and scheduling of distributed computing resources, computing networks have emerged, with their core objective being the deep integration and on-demand service of computing, storage, and network resources.

[0003] In recent years, with the development of generative artificial intelligence and intelligent agent technology, network architecture is gradually evolving from traditional connection networks to intelligent agent networks. Domestic and international research institutions and industries have successively proposed new network architectures such as self-intelligent networks and intelligent agent networks. In these systems, various functional entities in the network are gradually transforming from traditional passive execution units into intelligent agents with autonomous perception, autonomous decision-making, and autonomous execution capabilities. Resource objects no longer exist merely as scheduled resources but are gradually acquiring intelligent characteristics such as capability expression, state awareness, autonomous coordination, and service provision.

[0004] However, existing resource management and scheduling mechanisms in computing networks still primarily employ centralized or semi-centralized architectures. Resource nodes typically report basic resource metrics such as CPU utilization, GPU utilization, storage capacity, and network bandwidth to the scheduling platform periodically. The scheduling system then selects resources and allocates tasks based on resource availability, load conditions, or comprehensive scoring results. In this process, resource entities mainly act as resource providers, lacking the proactive ability to express themselves in relation to business scenarios and the capacity for autonomous decision-making.

[0005] As business types become increasingly complex, relying solely on resource availability or a uniform scoring method is no longer sufficient to accurately reflect the actual capacity of resources to handle specific business needs. For example, while a resource node may have ample computing resources, it might be better suited for batch processing and analysis tasks than for low-latency real-time services. Similarly, some resource nodes, although currently having sufficient resource availability, may be undergoing migration, scaling up or down, fault recovery, or service execution, and are not actually suitable for handling new business tasks. Furthermore, existing resource evaluation mechanisms typically use uniform reputation values ​​or uniform capability scores, which fail to reflect the differentiated performance capabilities of resources across different business types and accurately reflect the adaptability of resources to specific business scenarios.

[0006] On the other hand, with the development trend of resource intelligence, resource entities are gradually gaining autonomous decision-making capabilities. However, existing technologies still lack a resource intelligence capability expression and business acceptance organization mechanism oriented towards business acceptance scenarios. Resource intelligence struggles to form service capability expression results for specific business scenarios based on business needs, its own service capabilities, historical performance, and current operating status, and it is also difficult to output structured business acceptance capability information to the subsequent scheduling system.

[0007] Therefore, how to construct a service acceptance mechanism for resource intelligent agents in computing power networks, taking into account factors such as resource service capabilities, service adaptability, historical fulfillment capabilities, and operational status, to realize the service capability expression, service adaptability analysis, acceptance mode generation, and acceptance capability declaration output of resource intelligent agents for business scenarios, and to form service acceptance information that can be used for subsequent scheduling or task orchestration, has become a technical problem that urgently needs to be solved in the current computing power network field. Summary of the Invention

[0008] In view of this, the purpose of this invention is to solve the problems in existing computing power networks where resource nodes mainly rely on basic resource indicators and unified scoring methods to participate in service scheduling, which makes it difficult to accurately express the service capabilities of resources for different service scenarios, fails to reflect the performance differences of resources under different service types, and lacks a service undertaking mechanism for resource intelligent agents oriented to service scenarios. The invention provides a method and apparatus for service undertaking of computing power network resource intelligent agents.

[0009] To achieve the above objectives, the present invention provides the following technical solution: A method for resource-based intelligent agents to undertake services in a computing power network environment includes the following steps: S1: The node coordinating agent receives business task requests initiated by business users, business orchestration systems, or task scheduling systems, performs structured parsing of the business tasks, extracts business requirement information, and forms a standardized business requirement description. S2: The node coordinating agent sends business requirement information to the candidate resource agent; the candidate resource agent obtains its own basic resource status information, static resource information, and lifecycle status information, and obtains historical service performance information and business reputation information from the shared knowledge plane; the candidate resource agent generates a service capability profile based on its own basic resource status information, static resource information, and historical service performance information. S3: The resource intelligence agent conducts business adaptation analysis based on the correspondence between service capability profiles and business service semantic requirements to form business adaptation information for the current business scenario; S4: The resource intelligence agent combines the basic resource status information, business adaptation information, lifecycle status information and sub-business reputation information to conduct a comprehensive analysis of the current business scenario and generate a business acceptance mode. S5: The node coordinating agent receives the acceptance capability declaration output by the resource agent and provides the acceptance capability declaration to the subsequent business scheduling, task allocation or business orchestration module. S6: After the business is completed, the resource agent's performance in this business service process is recorded and updated. The performance feedback information is written into the shared knowledge plane for storage and maintenance, forming a historical service record under the corresponding business type. At the same time, the resource agent updates its own service capability profile and lifecycle status information according to the business execution results.

[0010] Furthermore, the business requirement information mentioned in step S1 includes business type information, basic resource requirement information, and service semantic requirement information; by uniformly modeling the business requirements, a business requirement description model that can be understood and matched by the resource intelligent agent is formed.

[0011] Furthermore, the basic resource status information mentioned in step S2 is used to describe the computing, storage, and network resource supply capabilities that the resource agent can currently provide; the static resource information is used to describe the long-term stable deployment attributes and hardware characteristics of the resource agent; and the lifecycle status information is used to describe the current operating stage and resource service status of the resource agent.

[0012] Furthermore, in step S2, the shared knowledge plane establishes reputation records corresponding to different business types based on the resource agent's historical business execution results, performance feedback information, and service records. After a business execution is completed, the resource agent's performance result for the corresponding business type is written into the shared knowledge plane and used to update the reputation record for that business type. The historical service performance information and sub-business reputation information are maintained and provided by the shared knowledge plane. The historical service performance information describes the resource agent's historical service experience and performance in different business scenarios. The sub-business reputation information describes the resource agent's historical performance in different business scenarios.

[0013] Furthermore, the service capability profile described in step S2 is used to describe the business service capability characteristics formed by the resource agent based on its own resource conditions and historical service performance. The service capability profile includes the following capability dimensions: AI intelligent computing service capabilities are used to reflect the level of resource intelligent agents' ability to perform artificial intelligence training and inference tasks; Real-time business processing capability is used to reflect the resource intelligence agent's ability to support low-latency, high-real-time business. Data processing service capability reflects the resource intelligence agent's ability to handle large-scale data analysis and data computing tasks. Storage access service capabilities reflect the resource agent's service capabilities in data storage, data retrieval, and data exchange.

[0014] Furthermore, in step S4, when the resource intelligence agent has sufficient current resource supply capacity, high adaptability between service capability profile and business needs, and its life cycle state is in an idle state or a state that can stably provide services, the resource intelligence agent forms a standard business service mode, indicating that the resource intelligence agent can provide business services in a normal service manner. When the resource agent is currently reserving or adjusting resources, or waiting for resources to be released, a delayed business service mode is formed, indicating that the resource agent needs to adjust or reserve resources or wait for business to provide services. When the resource agent has limited resource capabilities, but can still support business by reducing the service level, a degraded business service mode is formed, which means that the resource agent's current service capabilities are limited and business services are provided in a restricted manner. When a resource intelligence agent is currently unable to form an effective business service method, a service mode that cannot be undertaken temporarily is formed, indicating that the resource intelligence agent is currently unable to form an effective business service method.

[0015] Furthermore, the capability declaration mentioned in step S5 is used to provide the subsequent system with the service capability characteristics, business adaptation information, historical service performance and business acceptance mode of the resource agent in the current business scenario, so that the subsequent business scheduling process can obtain the current business service capability expression result of the resource agent.

[0016] The beneficial effects of this invention are as follows: Compared with existing computing networks that mainly rely on resource reserves, unified resource scoring, or centralized scheduling methods for resource selection, this invention has the following advantages: (1) It realizes the ability expression and adaptation of resource intelligence agents to business scenarios.

[0017] This invention constructs a service capability profile and combines it with a business adaptation analysis mechanism to map the basic resource capabilities of a resource intelligence agent into service capability features oriented towards business scenarios. At the same time, it establishes a correlation between service capabilities and business requirements, enabling the resource intelligence agent to express its capability advantages, applicable scope and service characteristics in different business scenarios, thereby realizing the transformation of resource capabilities into business service capabilities and differentiated adaptation to business requirements.

[0018] (2) A reliable assessment of the business undertaking capacity of resource intelligent agents has been achieved.

[0019] This invention introduces a business-specific reputation mechanism and a lifecycle status mechanism, incorporating the historical service performance, contract fulfillment status, and current operational status of resource agents in different business scenarios into the business acceptance capability assessment process. This ensures that business acceptance information not only reflects resource supply capacity but also embodies the service credibility and operational reliability of resource agents, thereby improving the reference value of business acceptance decisions.

[0020] (3) A structured declaration of the business acceptance capability of the resource intelligence agent has been implemented.

[0021] This invention uses a capability declaration mechanism to uniformly organize and structure the service capability profile, business adaptation information, business reputation information, lifecycle status information, and business acceptance mode, forming a standardized description of business acceptance capabilities. This provides a unified information foundation for subsequent business scheduling, task orchestration, resource collaboration, and cross-domain resource selection.

[0022] (4) A continuous evolution mechanism for the business undertaking capability of resource intelligence has been formed.

[0023] This invention uses a performance feedback update mechanism to continuously write performance feedback information generated during business execution into a shared knowledge plane, and uses it to update service capability profiles, business reputation, and related business acceptance information. This enables resource intelligence agents to continuously improve their capability expression based on historical business practices, thereby achieving dynamic updates and continuous optimization of business acceptance capabilities.

[0024] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0025] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein: Figure 1 A method for undertaking resource intelligent agent services in a computing power network environment. Detailed Implementation

[0026] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0027] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0028] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0029] Example 1: This invention provides a resource intelligent agent service acceptance method for computing power network environments. This method uses resource intelligent agents as autonomous resource service units in the computing power network. By constructing a mapping relationship between resource capabilities and service requirements, it introduces service capability profiling, service adaptation analysis, service-specific reputation assessment, lifecycle state awareness, and a service acceptance capability declaration generation mechanism. This enables resource intelligent agents to transform from expressing resource availability to expressing service acceptability, allowing them to proactively describe their service capabilities, fulfillment capabilities, and acceptance willingness in specific service scenarios. This provides a decision-making basis for service scheduling, resource orchestration, and agent collaboration. The method includes the following steps: Step 1: Business Requirements Analysis It receives business task requests initiated by business users, business orchestration systems, or task scheduling systems, performs structured parsing of the business tasks, extracts business requirement information, and forms standardized business requirement descriptions.

[0030] Business requirements information includes: Business type information, such as artificial intelligence inference business, edge low-latency interaction business, data-intensive analysis business, storage access business, etc.; Basic resource requirements, including computing resource requirements, storage resource requirements, and network bandwidth requirements; Service semantic requirements information, including the intensity of AI reasoning requirements, the intensity of real-time requirements, the intensity of data processing requirements, and the intensity of storage access requirements.

[0031] By uniformly modeling business requirements, a business requirement description model can be formed that can be understood and matched by resource intelligent agents.

[0032] Step 2: Resource Agent Information Acquisition After receiving the business requirement description, the node coordinating agent sends the business requirement information to the candidate resource agents. The candidate resource agents obtain their own basic resource status information, static resource information, and lifecycle status information, and retrieve historical service performance information and business-specific reputation information from the shared knowledge plane.

[0033] (1) Generation of resource intelligent agent service capability profile Basic resource status information reflects the current computing, storage, and network resource supply capabilities that the resource agent can provide; static resource information reflects the long-term stable deployment attributes and hardware characteristics of the resource agent; historical service performance information reflects the resource agent's historical service experience and performance in different business scenarios. Based on the above information, the resource agent generates a service capability profile, which is then used for subsequent business adaptation analysis, business acceptance mode generation, and acceptance capability declaration generation.

[0034] Basic resource status information describes the current resource supply capacity of a resource agent, reflecting the real-time resource status of the agent in terms of computing, storage, and networking. Basic resource status information includes, but is not limited to, available CPU capacity, available GPU capacity, available memory capacity, available storage capacity, storage access performance, storage bandwidth, network bandwidth, network latency, packet loss rate, and network jitter.

[0035] Static resource information of a resource agent describes its fundamental attribute information that remains stable over a long period. Static resource information includes the resource agent's deployment location, hardware configuration characteristics, and the management domain it belongs to.

[0036] Service capability profiles are used to describe the business service capability characteristics formed by a resource agent based on its own resource conditions and historical service performance. Unlike basic resource status information, which directly reflects resource supply capacity, service capability profiles further describe the resource agent's capability advantages, applicable scope, and service characteristics in different business scenarios, and are used to realize the mapping expression of resource capabilities to business service capabilities.

[0037] Service capability profiles are formed through comprehensive analysis of the resource agent's basic resource status information, static resource profiles, and historical service performance. Specifically: basic resource status information reflects the resource agent's current resource supply capacity; static resource profiles reflect the resource agent's long-term stable resource attributes and deployment characteristics; and historical service performance reflects the resource agent's actual service experience and fulfillment in different business scenarios. By integrating and analyzing the above information, a service capability profile tailored to specific business scenarios is formed. The service capability profile includes, but is not limited to, the following capability dimensions: AI intelligent computing service capabilities are used to reflect the level of resource intelligent agents' ability to perform tasks such as artificial intelligence training and inference. Real-time business processing capability is used to reflect the resource intelligence agent's ability to support low-latency, high-real-time business. Data processing service capability reflects the resource intelligence agent's ability to handle large-scale data analysis, data computation, and other business operations. Storage access service capabilities are used to reflect the service capabilities of resource agents in areas such as data storage, data retrieval, and data exchange.

[0038] Service capability profiles are used to describe the capabilities and service characteristics of resource intelligence agents in different business scenarios, providing support for subsequent business adaptation analysis, business acceptance mode generation, and acceptance capability declaration generation.

[0039] (2) Credit information by business segment Business-specific reputation information is used to describe the historical performance of resource agents in different business scenarios. Unlike traditional resource management systems that use a single reputation value or a unified capability score, this method maintains corresponding reputation information for different business types to reflect the actual service capabilities and historical service quality of resource agents in specific business scenarios.

[0040] Because different business types have varying resource requirements, the performance of the same resource agent may differ significantly across different business scenarios. For example, one resource agent may have a good historical performance in AI training but perform relatively poorly in low-latency real-time scenarios; another resource agent, while smaller in overall resource size, has consistently handled edge real-time services with high service quality. Therefore, using only a uniform reputation value is insufficient to accurately reflect the actual service capabilities of resource agents in different business scenarios.

[0041] The business-specific reputation information is maintained by the shared knowledge plane. Based on the historical business execution results, performance feedback information, and service records of the resource agent, the shared knowledge plane establishes reputation records corresponding to different business types.

[0042] The performance feedback information includes, but is not limited to, task completion status, service stability status, service level agreement compliance status, resource operation status, user feedback status, and business execution results.

[0043] Once the business is completed, the performance result of the resource agent under the corresponding business type will be written into the shared knowledge plane and used to update the reputation record corresponding to that business type.

[0044] Business-specific reputation information is used to reflect the historical service experience and performance capabilities of resource agents in specific business scenarios, and serves as an important reference for subsequent business adaptation analysis, business acceptance mode generation, and acceptance capability declaration generation.

[0045] (3) Lifecycle status information Lifecycle status information is used to describe the current operational stage and resource service status of the resource agent.

[0046] Unlike basic resource status information, which reflects the current resource supply capacity of a resource agent, lifecycle status information is mainly used to reflect the current operation process and service stage of a resource agent, describing whether the resource agent is in a state of idle, execution, adjustment, or restriction.

[0047] The lifecycle status information is maintained in real time by the resource agent based on its own operating status and is dynamically updated as the operating status changes.

[0048] In this method, the lifecycle states of the resource agent include, but are not limited to, idle state, locked state, executing state, adjustment and migration state, degraded isolation state, and deregistered state.

[0049] Among them: Idle state indicates that the resource agent is currently idle and can participate in new business acceptance processes; Locked state indicates that the resource agent has generated a business acceptance capability declaration and is waiting for subsequent business orchestration or resource scheduling; Execution state indicates that the resource agent is undertaking business tasks and continuously providing services; Adjustment and migration state indicates that the resource agent is performing resource adjustments, scaling up or down, or business migration operations; Degraded isolation state indicates that some of the resource agent's current resource capabilities are restricted, and its service capabilities have decreased; Deregistered state indicates that the resource agent has exited the current service process and can no longer provide business services.

[0050] Lifecycle status information is used to reflect the current operating stage and service status of the resource agent, and serves as an important reference for subsequent business adaptation analysis, business acceptance mode generation, and acceptance capability declaration generation.

[0051] When the lifecycle state of the resource agent changes, the resource agent will update the corresponding state information synchronously, thereby ensuring that the business acceptance process can reflect the current real operating state of the resource agent.

[0052] Step 3: Business Compatibility Analysis Business compatibility analysis is used to describe the compatibility relationship between the resource intelligence agent's service capabilities and business requirements.

[0053] After generating a service capability profile, the resource intelligence agent conducts business adaptation analysis based on current business needs to form business adaptation information for the current business scenario.

[0054] Business adaptation analysis primarily analyzes the correspondence between the resource intelligence agent's service capability profile and the semantic requirements of business services. This analysis describes the resource intelligence agent's capabilities, applicable scope, and service characteristics in the current business scenario. For example: AI training businesses focus more on the resource intelligence agent's AI computing service capabilities; edge real-time businesses focus more on the resource intelligence agent's real-time business processing capabilities; data analysis businesses focus more on the resource intelligence agent's data processing service capabilities; and data storage and access businesses focus more on the resource intelligence agent's storage access service capabilities.

[0055] Based on the service semantic features corresponding to business needs, the resource intelligence agent performs correlation analysis with its own service capability profile to form business adaptation information, which is used to characterize the degree of adaptation between the resource intelligence agent and the current business scenario.

[0056] Business adaptation information is used to reflect the service advantages and applicable characteristics of resource intelligence agents in the current business scenario, and serves as an important input for the generation of subsequent business acceptance models.

[0057] Step 4: Generation of Acceptance Pattern The acceptance mode generates a business service approach for forming a resource intelligence agent oriented towards the current business scenario.

[0058] After completing the business adaptation analysis, the resource intelligence agent combines basic resource status information, business adaptation information, lifecycle status information, and business reputation information to conduct a comprehensive analysis of the current business scenario and generate the corresponding business acceptance mode.

[0059] (1) Resource capability correlation Resource capability association is used to establish the relationship between business resource requirements and the resource supply capabilities of the resource intelligence agent, and to reflect the resource support characteristics of the resource intelligence agent in the current business scenario. Based on basic resource status information, the resource intelligence agent analyzes computing resources, storage resources, network resources, and other business-related resources, and combines this with the basic business resource requirements to form resource capability association results.

[0060] Resource capability association results do not simply describe the quantity or scale of resources, but rather characterize the correspondence between the current resource system of the resource agent and business requirements. For example, different business scenarios have different emphases on computing power, storage capacity, network transmission capacity, and resource coordination capabilities. Through resource capability association, the resource agent can identify the adaptation characteristics between its own resource structure and the business requirement structure.

[0061] Therefore, resource capability association is used to describe the support capability of resource intelligence agents for the current business scenario from the perspective of resource supply, to provide a resource foundation for the formation of subsequent business service methods, and to serve as important reference information for the generation of acceptance models.

[0062] (2) Business compatibility analysis Business adaptation information is used to reflect the fit between the resource intelligent agent service capability profile and the business service semantic requirements, and influences the formation of business acceptance mode.

[0063] The resource agent generates business adaptation information based on service capability profiles and business requirement characteristics. When the service capabilities of the resource agent are highly consistent with the business requirements, it indicates that the resource agent can match the business requirements well; when the consistency between the business requirements and the service capabilities of the resource agent is low, the service capabilities of the resource agent cannot match the business resource requirements well, thus affecting the generation of the acceptance mode.

[0064] Therefore, business adaptation information is used to reflect the degree of business compatibility between the resource intelligence agent and the current business scenario, and to provide a basis for the generation of subsequent business acceptance models. (3) Lifecycle collaborative processing Lifecycle status information is used to reflect the current operating stage of the resource agent and affects the stability and continuity of business undertaking mode and service organization method.

[0065] The resource agent analyzes the sustainable supply of resource service capabilities based on its current lifecycle state. Different lifecycle stages correspond to different operational characteristics. For example, when in the idle or execution state, the resource agent usually has good continuous service conditions; when in the adjustment / migration or degradation / isolation state, the service organization of the resource agent may be affected to some extent, requiring coordination of business services in conjunction with the current operational state; when in the deregistration state, its business service capabilities will gradually withdraw from the business operation system.

[0066] Therefore, lifecycle collaborative processing does not directly determine the business service outcome, but rather serves to ensure that the generated business acceptance model can truly reflect the service characteristics and service conditions of the resource agent in its current operating state.

[0067] (4) Integration of business-specific credit Business-specific reputation information is used to reflect the historical service performance of resource intelligence agents under the current business type and to provide empirical reference for the formation of business acceptance models.

[0068] The resource agent acquires the business reputation information for the corresponding business type and analyzes it in conjunction with the performance, service quality, and business collaboration in the historical service process. Resource agents with rich service experience and stable performance records can usually demonstrate stronger business service capabilities when forming business service methods; while factors such as service fluctuations, collaboration limitations, or performance risks reflected in historical service performance can also provide a reference for the generation of business undertaking models.

[0069] Therefore, the business-specific reputation fusion is used to supplement the service characteristic description of resource agents from the perspective of historical business practices, so that the business acceptance mode can not only reflect the current capability status, but also reflect its long-term service performance in the corresponding business scenarios.

[0070] (5) Generation of the receiving mode The resource intelligence agent integrates the results of resource capability association, business adaptation information, lifecycle status information, and sub-business reputation information to form the business acceptance mode corresponding to the current business scenario.

[0071] Among them, the resource capability correlation results reflect the support of the resource side for business needs; business adaptation information reflects the degree of fit between the resource agent's service capabilities and business needs; lifecycle status information reflects the impact of the current operational stage on the service mode; and business-specific reputation information reflects the service performance in historical business practices. All of these information collectively influence the formation process of the business service mode, enabling the generated service model to comprehensively reflect the current service capability characteristics of the resource agent.

[0072] In this method, the business acceptance modes include, but are not limited to, standard business service mode, delayed business service mode, downgraded business service mode, and temporarily unacceptable service mode.

[0073] Among them, the standard business service mode means that the resource intelligence agent can provide business services in a normal way; the delayed business service mode means that the resource intelligence agent needs to adjust resources, reserve resources, or wait for business to provide services; the degraded business service mode means that some of the service capabilities of the resource intelligence agent are currently limited, and business services are provided in a limited way; and the temporarily unacceptable service mode means that the resource intelligence agent is currently unable to form an effective business service mode.

[0074] The generated business acceptance pattern is used to describe the business service methods that the resource agent can currently provide, and serves as an important basis for the generation of subsequent acceptance capability declarations.

[0075] Step 5: Declaration of Resource Agent's Output Capability The capability declaration is used to structurally express the service capability characteristics and business service methods of the resource intelligence agent in the current business scenario.

[0076] After generating the business acceptance mode, the resource agent forms a corresponding acceptance capability declaration based on the current business scenario and outputs the acceptance capability declaration to the node coordination agent.

[0077] The capability declaration describes the business service capabilities and corresponding service methods that the resource agent can currently provide, providing a reference for subsequent business scheduling, task orchestration, and resource collaboration.

[0078] In this method, the capability declaration includes, but is not limited to, resource agent identification information, current business task identification information, service capability profile information, business adaptation information, sub-business reputation information, lifecycle status information, business acceptance mode information, and declaration validity period information.

[0079] The resource agent identification information identifies the current resource agent; the service capability profile information describes the resource agent's capabilities and service characteristics in different business scenarios; the business adaptation information describes the adaptation relationship between the resource agent and the current business scenario; the business-specific reputation information describes the resource agent's historical service performance under the corresponding business type; the lifecycle status information describes the current operating stage of the resource agent; the business acceptance mode information describes the business service methods that the resource agent can currently provide; and the declaration validity period information describes the applicable time range of the current acceptance capability declaration. After the acceptance capability declaration is generated, it is sent by the resource agent to the node coordinating agent. The node coordinating agent receives the acceptance capability declaration output by the resource agent and provides it to subsequent business scheduling, task orchestration, or resource collaboration modules.

[0080] The capability declaration is not directly used as the result of business scheduling, but is used to express the service capability characteristics and business service methods of the resource intelligence agent in the current business scenario, providing support for subsequent business organization and resource collaboration.

[0081] Step 6: Application of Acceptance Capability Declaration The node coordinating agent receives the acceptance capability declaration output by the resource agent and provides the acceptance capability declaration to the subsequent business scheduling, task allocation or business orchestration modules.

[0082] The capability declaration is used to provide subsequent systems with information such as the service capability characteristics, business adaptation information, historical service performance, and business acceptance mode of the resource agent in the current business scenario, so that the subsequent business scheduling process can obtain the current business service capability expression result of the resource agent.

[0083] The subsequent business scheduling or task allocation module carries out business orchestration, resource selection, task distribution and resource coordination processes based on the capacity declaration, so as to realize the organization and coordination between business needs and resource service capabilities.

[0084] This step does not involve specific scheduling strategies or orchestration algorithms; its main function is to enable the transmission and application of capacity declarations to subsequent business organization processes.

[0085] Step 7: Update Performance Feedback After the business operation is completed, the performance of the resource agent in this business service process will be recorded and updated. The performance feedback information is used to reflect the actual service performance of the resource agent during the business operation, including but not limited to task completion, service stability, service quality, business execution results, user feedback, resource operation, and service collaboration.

[0086] After the business is completed, the relevant performance feedback information is written into the shared knowledge plane for storage and maintenance, forming historical service records under the corresponding business type.

[0087] Based on the newly added performance feedback information, the shared knowledge plane updates the historical service performance and sub-business reputation information of the resource agent under the corresponding business type, so that the sub-business reputation can continuously reflect the long-term service capability and performance level of the resource agent in different business scenarios.

[0088] Meanwhile, the resource intelligence agent updates and adjusts the service capability profile based on business execution results and performance feedback information, so that the service capability profile can continuously reflect the changes in the actual business service capabilities of the resource intelligence agent.

[0089] When resource adjustments, business migrations, service recovery, or service exits occur during business execution, the resource agent synchronously updates its lifecycle status information to ensure that the lifecycle status accurately reflects the current operating stage of the resource agent.

[0090] Through the performance feedback and update mechanism, the continuous accumulation of historical service experience, the dynamic evolution of service capability profiles, and the continuous updating of business reputation are achieved, thereby forming a closed-loop optimization process for the business undertaking capacity of the resource intelligence agent.

[0091] Example 2: This embodiment uses AI inference service as an example to illustrate the intelligent agent service undertaking method of computing power network resources described in this invention. The overall process of the embodiment is as follows: Figure 1 As shown.

[0092] In a certain computing network scenario, a user-side business system submits an AI inference service request to a service node. This service is used to execute model inference tasks and has certain requirements on computing resources, model loading speed, data reading capabilities, and service response latency.

[0093] 1. Receiving and parsing business requests After receiving the business request, the node coordinating agent in the service node calls the business requirement parsing module to parse the business request and obtain the business requirement description.

[0094] The business requirements description includes business type, computing resource requirements, storage access requirements, network transmission requirements, and service semantic requirements. The business type is AI inference, and the service semantic requirements mainly reflect the need for AI intelligent computing service capabilities, real-time response capabilities, and model data reading capabilities.

[0095] 2. Acquisition of candidate resource agent information The node coordinating agent discovers candidate resource agents based on the business requirement description and sends the business requirement information to the candidate resource agents.

[0096] Candidate resource agents acquire their own basic resource status information, static resource information, and lifecycle status information, and obtain historical service performance information and business-specific reputation information related to AI inference business from the shared knowledge plane.

[0097] Among them, the basic resource status information is used to describe the computing, storage and network resource supply capabilities that the resource agent can currently provide, such as available CPU capacity, available GPU capacity, available memory capacity, available storage capacity, storage access performance, available network bandwidth, network latency, packet loss rate and network jitter.

[0098] Static resource information is used to describe the long-term stable attribute information of resource agents, such as the deployment location of resource agents, hardware configuration characteristics, and the management domain to which they belong.

[0099] Lifecycle state information is used to describe the current operating stage of the resource agent, such as idle state, locked state, executing state, adjustment and migration state, degraded isolation state, or deregistration state.

[0100] The shared knowledge plane provides the historical service performance and business-specific reputation information of the resource agent in AI inference business, which reflects its historical performance in similar business scenarios.

[0101] 3. Service capability profile generation Candidate resource agents generate service capability profiles based on their own basic resource status information, static resource information, and historical service performance information.

[0102] For AI inference operations, the service capability profile primarily reflects the resource agent's AI intelligent computing service capabilities, real-time business processing capabilities, and storage access service capabilities. This service capability profile describes the resource agent's strengths and applicable scope in AI inference scenarios.

[0103] 4. Business compatibility analysis The resource intelligence agent performs business adaptation analysis based on the service semantic requirements of AI inference business and its own service capability profile.

[0104] When business requirements are primarily reflected in model inference computation, model parameter reading, and short response time, the resource intelligence agent focuses on analyzing the compatibility between its AI intelligent computing service capabilities, real-time business processing capabilities, and storage access service capabilities and the business requirements, thus forming business compatibility information.

[0105] This business adaptation information is used to characterize the service features and adaptability of resource agents in the current AI inference business scenario.

[0106] 5. Generation of contracting modes Based on business adaptation analysis, the resource intelligence agent further combines basic resource status information, lifecycle status information, and business reputation information corresponding to AI inference business to generate a business acceptance model.

[0107] When the resource agent has sufficient resource supply capacity, its service capability profile is highly compatible with the AI ​​inference business requirements, and its lifecycle state is idle or capable of providing stable services, a standard business service mode can be formed. When the resource agent is currently reserving resources, adjusting resources, or waiting for resources to be released, a delayed business service mode can be formed. When some of the resource agent's resource capabilities are limited, but it can still support the AI ​​inference business by reducing the service level, a degraded business service mode can be formed. When the resource agent is currently unable to form an effective business service mode, a temporarily unavailable service mode is formed.

[0108] 6. Generation and publication of undertaking capacity declaration The resource agent generates a service acceptance capability declaration based on the generated service acceptance mode. The service acceptance capability declaration includes the resource agent identifier, the current service task identifier, service capability profile information, service adaptation information, the sub-service reputation information corresponding to the AI ​​inference service, lifecycle status information, service acceptance mode information, and declaration validity period information.

[0109] The capability declaration is used to express the business service capabilities and methods that the resource intelligence agent can provide in the current AI inference business scenario.

[0110] The resource agent sends the generated capacity declaration to the node coordinating agent. The node coordinating agent receives the capacity declaration output by the candidate resource agent and provides it to subsequent service scheduling, task allocation, or service orchestration modules.

[0111] Subsequent business scheduling or task allocation modules can carry out business orchestration and resource organization based on the capacity declaration.

[0112] 7. Performance feedback update After the AI ​​inference task is completed, the system collects the performance feedback information for that task.

[0113] The performance feedback information includes task completion status, service quality status, business response status, resource operation status, user feedback, and business execution results.

[0114] Relevant performance feedback information is written into the shared knowledge plane to update the historical service performance and business-specific reputation information of the resource agent in AI inference business.

[0115] At the same time, the resource intelligence agent updates its service capability profile and lifecycle status information based on the business execution results, thereby forming a closed loop for business acceptance oriented towards AI inference business.

[0116] Example 3: An electronic device, comprising a memory and a processor; The memory is used to store computer programs; The processor is configured to implement the method described in Embodiment 1 when executing the computer program.

[0117] Example 4: A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in Embodiment 1.

[0118] Example 5: A computer program product includes a computer program that, when executed by a processor, implements the method described in Example 1.

[0119] In the above embodiments, the reference to "this embodiment" in the specification indicates that a specific feature, structure, or characteristic described in connection with the embodiment is included in at least some embodiments, but not necessarily all embodiments. Multiple appearances of "this embodiment" do not necessarily refer to the same embodiment.

[0120] In the above embodiments, although the invention has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory structures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed. The embodiments of the invention are intended to cover all such substitutions, modifications, and variations falling within the broad scope of the appended claims.

[0121] As will be understood by those skilled in the art, the computer-readable storage medium described in this embodiment allows for the implementation of all or part of the steps in the above method embodiments by computer program-related hardware. The aforementioned computer program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0122] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver, and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication between them. The memory is used to store computer programs, the communication interface is used to perform communication, and the processor and the transceiver are used to run the computer programs, so that the electronic terminal performs the steps of the above method.

[0123] In this embodiment, the memory may include random access memory (RAM) and may also include non-volatile memory, such as at least one disk storage device.

[0124] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0125] This invention can be used in a wide range of general-purpose or special-purpose computing system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices, etc.

[0126] This invention can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This invention can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0127] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for resource intelligent agent service acceptance in a computing power network environment, characterized in that: Includes the following steps: S1: The node coordinating agent receives business task requests initiated by business users, business orchestration systems, or task scheduling systems, performs structured parsing of the business tasks, extracts business requirement information, and forms a standardized business requirement description. S2: The node coordinating agent sends business requirement information to the candidate resource agent; the candidate resource agent obtains its own basic resource status information, static resource information and lifecycle status information, and obtains historical service performance information and business reputation information from the shared knowledge plane. The candidate resource agent generates a service capability profile based on its own basic resource status information, static resource information, and historical service performance information. S3: The resource intelligence agent conducts business adaptation analysis based on the correspondence between service capability profiles and business service semantic requirements to form business adaptation information for the current business scenario; S4: The resource intelligence agent combines the basic resource status information, business adaptation information, lifecycle status information and sub-business reputation information to conduct a comprehensive analysis of the current business scenario and generate a business acceptance mode. S5: The node coordinating agent receives the acceptance capability declaration output by the resource agent and provides the acceptance capability declaration to the subsequent business scheduling, task allocation or business orchestration module. S6: After the business is completed, the resource agent's performance in this business service process is recorded and updated. The performance feedback information is written into the shared knowledge plane for storage and maintenance, forming a historical service record under the corresponding business type. At the same time, the resource agent updates its own service capability profile and lifecycle status information according to the business execution results.

2. The resource intelligent agent service undertaking method for computing power network environment according to claim 1, characterized in that: The business requirement information mentioned in step S1 includes business type information, basic resource requirement information, and service semantic requirement information; by uniformly modeling the business requirements, a business requirement description model that can be understood and matched by resource intelligent agents is formed.

3. The resource intelligent agent service undertaking method for computing power network environment according to claim 1, characterized in that: The basic resource status information mentioned in step S2 is used to describe the current computing, storage, and network resource supply capabilities that the resource agent can provide; the static resource information is used to describe the long-term stable deployment attributes and hardware characteristics of the resource agent; and the lifecycle status information is used to describe the current operating stage and resource service status of the resource agent.

4. The resource intelligent agent service undertaking method for computing power network environment according to claim 1, characterized in that: In step S2, the shared knowledge plane establishes reputation records corresponding to different business types based on the resource agent's historical business execution results, performance feedback information, and service records. After a business execution is completed, the resource agent's performance result for the corresponding business type is written into the shared knowledge plane and used to update the reputation record for that business type. The historical service performance information and sub-business reputation information are maintained and provided by the shared knowledge plane. The historical service performance information describes the resource agent's historical service experience and performance in different business scenarios. The sub-business reputation information describes the resource agent's historical performance in different business scenarios.

5. The resource intelligent agent service undertaking method for computing power network environment according to claim 1, characterized in that: The service capability profile described in step S2 is used to describe the business service capability characteristics formed by the resource agent based on its own resource conditions and historical service performance. The service capability profile includes the following capability dimensions: AI intelligent computing service capabilities are used to reflect the level of resource intelligent agents' ability to perform artificial intelligence training and inference tasks; Real-time business processing capability is used to reflect the resource intelligence agent's ability to support low-latency, high-real-time business. Data processing service capability reflects the resource intelligence agent's ability to handle large-scale data analysis and data computing tasks. Storage access service capabilities reflect the resource agent's service capabilities in data storage, data retrieval, and data exchange.

6. The resource intelligent agent service undertaking method for computing power network environment according to claim 1, characterized in that: In step S4, when the resource intelligence agent has sufficient current resource supply capacity, high adaptability between service capability profile and business needs, and its life cycle state is in an idle state or a state that can stably provide services, the resource intelligence agent forms a standard business service mode, indicating that the resource intelligence agent can provide business services in a normal service manner. When the resource agent is currently reserving or adjusting resources, or waiting for resources to be released, a delayed business service mode is formed, indicating that the resource agent needs to adjust or reserve resources or wait for business to provide services. When the resource agent has limited resource capabilities, but can still support business by reducing the service level, a degraded business service mode is formed, which means that the resource agent's current service capabilities are limited and business services are provided in a restricted manner. When a resource intelligence agent is currently unable to form an effective business service method, a service mode that cannot be undertaken temporarily is formed, indicating that the resource intelligence agent is currently unable to form an effective business service method.

7. The resource intelligent agent service undertaking method for computing power network environment according to claim 1, characterized in that: The capability declaration mentioned in step S5 is used to provide the subsequent system with the service capability characteristics, business adaptation information, historical service performance and business acceptance mode of the resource agent in the current business scenario, so that the subsequent business scheduling process can obtain the current business service capability expression result of the resource agent.

Images