An agent service provisioning method, system, and storage medium
By defining RA-SRV resource records and a multi-dimensional optimization mechanism, the problem of low efficiency in agent service discovery is solved, and efficient and secure service matching and connection are achieved.
Patent Information
- Application Number
- CN202610673381.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-15
- Publication Date
- 2026-06-30
Smart Images
Figure CN122317154A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of computer technology, and in particular to a method, system and storage medium for providing intelligent agent services. Background Technology
[0002] With the rapid development of artificial intelligence technology, intelligent agent networks are gradually becoming an important direction for the evolution of the next-generation Internet. In this architecture, users, intelligent agents, and APIs constitute the core participating roles. Users express service needs through natural language, various service agents obtain resources and deliver services through API calls, and multiple agents collaborate to complete complex tasks. As a critical infrastructure of the Internet, the Domain Name System should theoretically undertake the core functions of intelligent agent addressing and service discovery, providing technical support for large-scale agent collaboration.
[0003] However, the existing DNS system was primarily designed to serve host-centric resource resolution needs. Its resource record types only support static queries based on domain names, failing to express dynamic semantic information such as agent service capabilities, interaction protocols, and load status. More critically, the current intelligent agent service ecosystem is highly decentralized and autonomous, with each service agent operating and maintaining itself independently, lacking a unified registration, discovery, and matching mechanism. When a user issues a natural language service request, the system cannot accurately map the user's intent to an agent with the corresponding capabilities, nor can it quickly locate the target service among a massive number of dispersed agents. Even when querying using traditional service discovery protocols (such as DNS-SD), only network layer address information can be obtained, failing to intelligently filter and optimize based on multi-dimensional factors such as service capabilities and service quality. This results in low service discovery efficiency, difficulties in resource coordination, and severely restricts the effectiveness of intelligent agent collaboration. Summary of the Invention
[0004] To address the aforementioned issues, this application provides a method, system, and storage medium for providing intelligent agent services.
[0005] The embodiments of this application disclose the following technical solutions: The first aspect of this application provides a method for providing intelligent agent services, including: Obtain natural language service requests sent by clients; Parse the natural language service request and extract service requirement keywords; Based on the service requirement keywords, the stored Domain Name System (DNS) resource records are retrieved to determine the target service agent. The DNS resource records are generated and stored based on the service information contained in the registration request sent in advance by the service agent. The connection information of the target service agent is returned to the user so that the user can establish a secure connection with the target service agent based on the connection information.
[0006] A second aspect of this application provides an intelligent agent service provisioning system, comprising: The acquisition unit is used to acquire natural language service requests sent by the client; The extraction unit is used to parse the natural language service request and extract service requirement keywords; The determining unit is used to retrieve stored Domain Name System (DNS) resource records based on the service requirement keywords, and determine the target service agent. The DNS resource records are generated and stored based on the service information contained in the registration request sent by the service agent. The return unit is used to return the connection information of the target service agent to the user, so that the user can establish a secure connection with the target service agent based on the connection information.
[0007] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the intelligent agent service provisioning method as described in the first aspect above.
[0008] A fourth aspect of this application provides a computer program product that, when run on a computer, executes the intelligent agent service provisioning method as described in the first aspect above.
[0009] A fifth aspect of this application provides a computer-readable storage medium storing instructions that, when executed on a terminal device, cause the terminal device to perform the intelligent agent service provisioning method as described in the first aspect above.
[0010] Compared with the prior art, this application has the following beneficial effects: By generating and storing core service information contained in the registration request pre-sent by the service agent as a Domain Name System (DNS) resource record, unified registration and centralized management of decentralized autonomous service agents are achieved, laying a data foundation for efficient discovery. Secondly, upon receiving a user's natural language service request, the request is parsed and service requirement keywords are extracted, transforming ambiguous user intent into explicit structured search instructions, thus solving the problem of accurately mapping natural language requests to corresponding service capabilities. Subsequently, based on the extracted keywords, stored DNS resource records are retrieved, enabling the discovery process to move beyond traditional static domain name queries and match based on semantic information such as service capabilities. This allows for rapid identification of target service agents with corresponding functions among a massive number of agents, significantly improving service discovery efficiency.
[0011] It should be noted that, in addition to using service capability tags to match keywords in user natural language requests to filter candidate service agents with corresponding functions, this application further introduces a multi-dimensional service optimization mechanism. This mechanism is based on the service overhead field, namely ServiceOverhead, included in the novel DNS resource record RA-SRV defined in this invention. This field has a fixed length of 12 bytes and is subdivided into three subfields: response latency, resource utilization, and call cost, each occupying 4 bytes. Response latency reflects the time required for the service agent to process the request, resource utilization indicates the current load level of the service agent, and call cost represents the economic cost of using the service. After initially filtering out a set of candidate service agents with corresponding service capabilities, the system's service matching and routing module will further read the service overhead information of each candidate agent and perform multi-dimensional comprehensive comparison and optimization of the above three indicators. Specifically, weight coefficients can be configured for response latency, resource utilization, and call cost according to business needs or system strategies, and the indicator values of each candidate agent are weighted and calculated to obtain a comprehensive service quality score. Generally, a lower overall score indicates better service quality. Therefore, the candidate service agent with the lowest overall score is selected as the final target service agent. If there are cases with the same score, a secondary decision can be made according to preset priority rules. Through this multi-dimensional optimization algorithm, this application not only achieves function-based service matching but also realizes optimal routing based on service quality factors such as performance, load, and cost. This effectively avoids resource overload or service quality degradation problems that may be caused by single-function matching, and significantly improves the service provision efficiency and user experience of the agent interconnection network. Attached Figure Description
[0012] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0013] Figure 1 The message structure of the RA-SRV extended DNS resource record provided in the embodiments of this application; Figure 2 This is a diagram showing the relationship between the Agent functional modules provided in the embodiments of this application. Figure 3 A schematic diagram of a workflow provided for an embodiment of this application; Figure 4 A flowchart illustrating an intelligent agent service provisioning method provided in this application embodiment; Figure 5 This is a structural diagram of an intelligent agent service supply system provided in an embodiment of this application. Detailed Implementation
[0014] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.
[0015] To facilitate understanding of the technical solutions provided in the embodiments of this application, the technical terms involved in the embodiments of this application will be explained below.
[0016] DNS, short for Domain Name System, is essentially a distributed database and network addressing service used to translate human-readable domain names into machine-readable IP addresses, thereby supporting the interconnection of internet devices. As a critical infrastructure of the internet, DNS not only acts as a URL translator, but its resource record types can also extend to store storage metadata (such as SRV records). Therefore, it has the potential to become a unified service discovery and addressing hub in intelligent agent interconnection networks, and is the technological cornerstone for realizing large-scale distributed system collaborative work.
[0017] An intelligent agent refers to a software or hardware entity possessing autonomy, responsiveness, sociality, and initiative, capable of perceiving the environment, making decisions, and executing actions to achieve specific goals. In this application, intelligent agent specifically refers to a service intelligent agent, that is, a program entity deployed in a network environment that provides specific functional services to the outside world. Service intelligent agents possess the following core characteristics: Service capability encapsulation: encapsulating specific business functions (such as weather queries, data calculations, task scheduling, etc.) into callable service interfaces; Autonomous registration and publishing: being able to proactively send registration requests to the aggregation agent, submitting its own service capability description, connection address, communication protocol, and other information; Secure communication capabilities: supporting certificate-based authentication and encrypted communication to ensure the security of interactions with other entities; Collaborative work capabilities: being able to collaborate with other intelligent agents to jointly complete complex tasks that a single intelligent agent cannot complete independently. In the technical solution described in this application, the intelligent agent is both a service provider and a service consumer, achieving accurate matching and secure connection with user requests through the intermediary role of the aggregation agent.
[0018] To facilitate understanding of the technical solutions provided in the embodiments of this application, the background technology involved in the embodiments of this application will be described below.
[0019] As mentioned earlier, the ambiguity of natural language requests in existing technologies leads to inaccurate interpretation of user intent. The decentralized and autonomous nature of service agents, lacking a unified registration, discovery, and matching mechanism, results in inefficient service discovery and difficulties in resource coordination. Furthermore, the lack of a standardized and secure connection protocol poses risks to agent task collaboration and information exchange processes.
[0020] To address the aforementioned issues, the core service information of decentralized autonomous service agents is centrally registered and stored, forming a service resource pool that can be uniformly retrieved, thus laying the foundation for efficient discovery. Building upon this, fuzzy natural language requests from users are structurally parsed to extract precise service requirement keywords. These keywords are then used for matching and filtering within the unified resource pool, solving the problem of inaccurate intent parsing and transforming decentralized service discovery into a unified index query, significantly improving discovery efficiency. Finally, agents submit security certificates during registration. During the matching process, each service agent is associated with a unique index and stored in a dedicated database, achieving unified management of security credentials. After matching a target agent for a user, the RA not only returns the address and protocol but also retrieves and returns the corresponding security certificate through the certificate index. The user uses this certificate to authenticate the target agent and establish an encrypted communication channel.
[0021] It should be noted that the intelligent agent service provisioning method, system, and medium provided in this application can be applied to the field of computer technology. The above are merely examples and do not limit the application field of the intelligent agent service provisioning method, system, and medium provided in this application. Furthermore, the embodiments of this application may not limit the executing entity of the intelligent agent service provisioning. For example, the intelligent agent service provisioning method of this application embodiment can be applied to data processing devices such as terminal devices or servers. The terminal device can be an electronic device such as a computer or a personal digital assistant (PDA). The server can be a standalone server, a cloud server, or a cluster server composed of multiple servers.
[0022] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.
[0023] To efficiently store and manage the core business information of service agents within the Domain Name System (DNS), this application defines a novel DNS resource record type, named RA-SRV, compliant with protocol standards, in the DNS-based agent service provisioning solution. Its type value is fixed at 0x8A. Unlike existing SRV, TXT, or A records in the DNS protocol, the RA-SRV record is a custom resource record newly designed for agent service discovery scenarios, and is not a reuse or simple extension of any existing record type. Structurally, this record fully reuses the common fields of standard DNS resource records, such as Name, Type, Class, and TTL. However, within the RDATA field, a customized data structure is used to describe the core business information of the service agent, including service capability tags, protocol type, address type, connection address, certificate index, service overhead, and agent unique identifier. Through this newly defined RA-SRV record, the system can establish a unified data foundation for the service capabilities, communication protocols, network addresses, security credential indexes, and multi-dimensional service quality parameters of decentralized autonomous service agents, enabling multi-dimensional optimization and secure connections. It can be said that the RA-SRV new record type is the cornerstone of the entire technical solution. It is precisely by relying on this custom DNS resource record that the existing DNS infrastructure, which originally did not support semantic service description, can uniformly encode and store (response latency, resource utilization, and call cost) in the DNS system, thereby supporting the subsequent registration, discovery, and optimization of intelligent agent services based on natural language, and thus ensuring the efficient operation of the convergence agent (RA) system.
[0024] See Figure 1 , Figure 1 The message structure of the RA-SRV extended DNS resource record provided in the embodiments of this application.
[0025] In the basic fields section, the Name field of the RA-SRV record is used to identify the domain name corresponding to the service agent, such as weather.agent.com, using the standard DNS domain name encoding format; the Type field is fixed at 0x8A, indicating that the record type is RA-SRV; the Class field is set to IN by default, indicating the Internet address class; the TTL field is set to 3600 seconds, which can be quickly applied when the service agent dynamically updates the registration information; and the RDLENGTH field is used to record the total byte length of the RDATA field.
[0026] In the internal structure design of the RDATA field, this invention defines seven core information subfields in sequence. ServiceCapabilityTag is the service capability tag, using UTF-8 encoding, with a length ranging from 1 to 64 bytes. It identifies the Agent's core service capabilities and supports concatenating multiple tags using delimiters to express composite service capabilities. ProtocolType is the protocol type, occupying 1 byte. It uses enumerated values to quickly identify the interaction protocol; for example, 0x01 represents HTTP, 0x02 represents MQTT, 0x03 represents gRPC, and 0x04 represents WebSocket. AddrType is the address type, occupying 1 byte; 0x01 represents an IPv4 address, and 0x02 represents an IPv6 address. ConnectionAddr is the connection address field, containing the port number (2 bytes) and the IP address. The IPv4 address occupies 4 bytes, and the IPv6 address occupies 14 bytes. All data is stored in network byte order. CertificateIndex is the certificate index, occupying a 4-byte unsigned integer. It serves as the primary key for retrieving the RA module's built-in certificate database, used to quickly locate and obtain the complete security certificate of the service Agent. ServiceOverhead is the service overhead field, with a fixed length of 12 bytes. It is further divided into three subfields: response latency, resource utilization, and call cost, each occupying 4 bytes, to support multi-dimensional service optimization. Finally, AgentID is the unique identifier for the Agent, using a 16-byte UUID format to ensure global uniqueness and distinguish different service Agent entities.
[0027] The Rendezvous Agent (RA) is the core hub of the intelligent agent service supply system described in this invention, responsible for processing user requests and registering, storing, retrieving, and maintaining Agent identification information. Its overall architecture and interactions with external roles are as follows: Figure 2 As shown, Figure 2 This is a diagram showing the relationship between the Agent functional modules provided in the embodiments of this application.
[0028] On the service supply side, the registration management module is responsible for receiving registration requests sent by service agents. It verifies the legality of the submitted core service information, such as service capability description, protocol type, connection address, and security certificate. After the verification is successful, the module stores the complete certificate in the certificate database module and generates the corresponding certificate index. On the other hand, it encapsulates information such as service capability tags, protocol type, address type, connection address, certificate index, service overhead, and unique agent identifier according to the custom RA-SRV record format and stores it in the DNS resource record storage module, thereby laying the data foundation for subsequent service discovery.
[0029] On the user demand side, when a user issues a service request in natural language, the Natural Language Understanding (NLU) module first parses the request, converting it into structured data and extracting core service requirement keywords. Subsequently, the service matching and routing module performs a rapid search in the DNS resource record storage module based on these keywords, initially filtering out a set of candidate Agents with the corresponding service capabilities. Building on this, it combines the service overhead information carried by each candidate Agent in its RA-SRV record, including response latency, resource utilization, and call cost, to perform a multi-dimensional comprehensive comparison and select the best target service Agent. After determining the target, the service matching and routing module quickly obtains the complete security certificate of the corresponding Agent through the certificate index provided by the certificate database module, and finally encapsulates the target Agent's connection address, communication protocol, and security certificate together and returns them to the user.
[0030] To ensure security throughout the entire interaction process, the security verification module performs preliminary verification of the service agent certificate during the registration phase and provides necessary security verification information after service matching is completed, enabling users to establish an encrypted communication channel with the target agent based on certificate authentication. Simultaneously, the logging and monitoring module records the system's operational status and key operations throughout the process, monitoring the performance and service availability of each module in real time, providing support for stable system operation and continuous optimization. Through the close collaboration of these modules, the RA system unifies distributed and autonomous service agents into a manageable and controllable identification system, transforms ambiguous natural language requests into precise structured search instructions, and embeds security mechanisms into the entire process of service discovery and connection. This systematically solves the problems of inaccurate intent parsing, inefficient service discovery, and lack of interaction security in existing technologies.
[0031] The workflow of this application embodiment mainly includes the service agent registration process and the user natural language request processing and matching process. The detailed steps and interaction logic of these two processes are as follows. Figure 3 As shown, Figure 3 This is a schematic diagram of a workflow provided for an embodiment of this application.
[0032] During the service agent registration phase, such as Figure 3 As shown, the user sends "1-1: Registration Request" to the Service Agent. The RA performs "1-2: Information Verification, RA-SRV Record Generation and Storage". The Service Agent returns "1-3: Registration Successful" to the user. That is, the Service Agent first sends a registration request carrying core service information to the Aggregator Agent (RA). This information covers key elements such as service capability description, interaction protocol type, service overhead, unique Agent identifier, domain name information, security certificate, and digest. After receiving the registration request, the RA module immediately performs integrity verification of the service information and preliminary validity verification of the security certificate. After successful verification, the RA generates a unique certificate index for the Service Agent, then encapsulates all core service information into RA-SRV resource records according to a predefined format and stores them in the DNS resource record storage module, while simultaneously storing the security certificate in the certificate database module. After completing the above operations, the RA returns a registration success response to the Service Agent, marking the formal inclusion of the Service Agent into the unified service discovery system.
[0033] In the user's natural language request processing and Agent matching stage, such as Figure 3As shown, the user initiates "2-1: Natural Language Request". The RA performs the following operations: "2-2: Natural Language Parsing, Keyword Extraction", "2-3: DNS Query, Agent Matching", "2-4: Certificate Acquisition". The user provides "2-5: Agent Information, including address, protocol, and certificate". The RA then proceeds to "2-6: Establishing a Secure Connection". That is, the user first sends a natural language service request to the RA module. The RA module calls the natural language understanding module to parse the request and extract core service requirement keywords. Based on these keywords, the RA retrieves stored RA-SRV records from the DNS resource record storage module, initially matching a set of candidate service agents with the corresponding service capabilities. On this basis, the service matching and routing module performs secondary screening and multi-dimensional comparison based on the service cost information of each candidate agent, ultimately determining the optimal target service agent. Subsequently, the RA extracts the complete security certificate of the target agent from the certificate database module through the certificate index, encapsulates it along with the target agent's connection address and communication protocol, and returns it to the user as response information. After receiving the response, the user establishes an encrypted secure connection with the target service agent based on certificate authentication, thereby obtaining the required service.
[0034] The following example illustrates the natural language request processing and Agent matching phase of this application. See also... Figure 4 ,Should Figure 4 A flowchart of an intelligent agent service provisioning method provided in this application embodiment is included, the method comprising: S101. Obtain the natural language service request sent by the client.
[0035] Users input service requests in natural language through a client, such as an application, a webpage, or a smart device, for example, "Please check today's weather." This request is then sent to the RA system as input for further processing.
[0036] S102. Parse the natural language service request and extract service requirement keywords.
[0037] The RA system invokes its Natural Language Understanding (NLE) module to perform semantic analysis on received natural language requests, removing redundant information, converting it into structured data, and extracting core service requirement keywords. For example, it extracts the keyword "weather query" from "help me check today's weather." This step aims to eliminate the ambiguity of natural language, laying the foundation for accurate service matching.
[0038] S103. Based on the service requirement keywords, retrieve the stored Domain Name System resource records to determine the target service agent.
[0039] The service matching and routing module of the RA system retrieves stored RA-SRV resource records from the DNS resource record storage module based on the keywords extracted in S102. These records are created by the RA system when a service agent sends a registration request during the registration phase, encapsulating and storing its core service information, including service capability tags, protocol types, connection addresses, certificate indexes, and service overhead. During the retrieval process, the system first matches the ServiceCapabilityTag in the RA-SRV records using keywords to filter out a set of candidate service agents with the corresponding service capabilities. Then, it combines the ServiceOverhead field of each candidate record, which includes at least one of response latency, resource utilization, and call cost, to perform multi-dimensional comparison and optimization, ultimately determining the optimal target service agent.
[0040] S104. Return the connection information of the target service agent to the user so that the user can establish a secure connection with the target service agent based on the connection information.
[0041] The RA system extracts the connection address and communication protocol from the RA-SRV record corresponding to the target service agent, and quickly retrieves the corresponding complete security certificate from the certificate database module based on the CertificateIndex in the record. Then, it encapsulates the connection address, communication protocol, and security certificate into response data and returns it to the user. Upon receiving the connection information, the user authenticates the target service agent using the security certificate and establishes an encrypted communication channel according to the specified communication protocol, thereby obtaining the required service in a secure environment.
[0042] In one possible implementation, the response latency, resource utilization, and invocation cost of each candidate service agent are weighted and calculated, and the target service agent is determined from the candidate service agents based on the weighted calculation results.
[0043] Specifically, the service matching and routing module extracts the ServiceOverhead field from the RA-SRV resource records corresponding to each candidate service agent. This field contains three core metrics: response latency (representing the time required for the agent to process a request), resource utilization (representing the current load level of the agent), and call cost (representing the economic cost of using the agent's service). These three metrics characterize the real-time status of the service agent from the dimensions of performance, load, and cost, respectively. To comprehensively evaluate the service quality of each candidate agent, this step uses a weighted calculation method. First, weight coefficients are configured for each of the three metrics according to business needs or system policies. For example, the response latency weight can be set to 0.5, the resource utilization weight to 0.3, and the call cost weight to 0.2, with the sum of the three weights being 1. Subsequently, the values of the three metrics for each candidate agent are normalized and then weighted to calculate a comprehensive score.
[0044] The weighting coefficients can be dynamically adjusted according to different application scenarios. For example, in scenarios with high real-time requirements, the weight of response latency can be appropriately increased; in cost-sensitive scenarios, the weight of call cost can be increased. After weighted calculation, the comprehensive scores of each candidate Agent are compared. Generally, a lower comprehensive score indicates better service quality (since all indicators are cost-based, lower values are better). Therefore, the candidate service Agent with the lowest comprehensive score is selected as the final target service Agent. If multiple Agents have the same score, a secondary decision can be made based on preset priority rules, such as prioritizing the one with better response latency. Through the above weighted calculation and multi-dimensional comparison mechanism, this step can select the service Agent with the best comprehensive performance in terms of performance, load, and cost from the functionally matched candidate set, thereby achieving fine-grained routing of service quality and significantly improving user experience and system resource utilization efficiency.
[0045] The technical solution of this application will be illustrated below with specific examples. This embodiment takes a weather data query service based on natural language as an example to fully present the entire process of service agent registration and user request processing.
[0046] Suppose there exists an intelligent agent providing weather data services, hereinafter referred to as Weather Agent, which wishes to integrate its service capabilities into this system. Weather Agent first generates a registration request, which includes the following core service information: the service capability is described as weather data query, the interaction protocol type is HTTP, the service overhead parameters are set as a response latency of 50ms and a call cost of 1 cent, and its own security certificate digest. This registration request is sent to the aggregation agent RA system.
[0047] Upon receiving the request, the RA system's registration management module immediately performs a completeness check on the service information and a preliminary validity check on the security certificate digest. If the verification passes, the RA system generates a unique certificate index (CertificateIndex) for the Weather Agent and stores its complete security certificate in the certificate database module. Subsequently, the RA system encapsulates the Weather Agent's core service information—including the service capability tag "Weather Data Query," protocol type HTTP, connection address (IP address and port number), certificate index, service overhead parameters (latency 50ms, cost 1 cent), and the Agent's unique identifier—according to the predefined RA-SRV resource record format and stores it in the DNS resource record storage module. After completing these operations, the RA system returns a registration success response to the Weather Agent. At this point, the Weather Agent is officially integrated into the system and can be discovered and invoked by users.
[0048] A user needs to check weather information and sends a natural language request to the RA system via the client: "Help me check the latest weather data." The RA system's Natural Language Understanding (NLU) module performs semantic parsing on the request, removes redundant information, and extracts the core service requirement keyword "weather data query." Based on this keyword, the service matching and routing module retrieves the stored RA-SRV records in the DNS resource record storage module. By matching the service capability tags in the records, the system filters out a set of candidate service agents with "weather data query" capabilities. In this scenario, assuming there are multiple agents providing weather query services, including the aforementioned registered weather agent, the system further reads the service cost field in each candidate agent's record and compares indicators such as response latency, resource utilization, and call cost from multiple dimensions. After comprehensive optimization, the aforementioned weather agent is determined as the target service agent.
[0049] Subsequently, the RA system extracts the connection address (IP address + port number) and communication protocol (HTTP) from the RA-SRV record corresponding to the target weather agent. Based on the certificate index in the record, it quickly retrieves the complete security certificate for that agent from the certificate database module. The RA system encapsulates the connection address, communication protocol, and security certificate into a response message and returns it to the user. Upon receiving the response, the user authenticates the target weather agent using the security certificate. Once authentication is successful, an encrypted communication channel is established according to the HTTP protocol, and a weather query request is sent over the secure connection to ultimately obtain the required weather data service.
[0050] Compared with the prior art, the technical solution provided in this application has the following beneficial effects: To address the problem of inaccurate intent parsing caused by the ambiguity of natural language requests in existing technologies, this application utilizes the natural language understanding module within the aggregation proxy RA system to accurately parse user-input natural language requests and extract core service requirement keywords, thereby transforming ambiguous user intent into structured search instructions. Based on this, unified search and matching are performed on DNS resource records using keywords, achieving precise conversion from user intent to target service, significantly improving interaction efficiency and user experience.
[0051] This application extends the standard DNS protocol and defines a new type of RA-SRV resource record (type value 0x8A). While fully reusing the standard DNS message structure, it encodes and stores the core business information of the service agent in the RDATA field. This design enables seamless compatibility with existing DNS infrastructure without requiring large-scale modifications to the underlying network architecture, significantly reducing deployment difficulty and demonstrating promising prospects for industry adoption.
[0052] By introducing a ServiceOverhead field into the RA-SRV record, this application implements Quality of Service (QoS) routing that transcends function matching. This field includes three core metrics: response latency, resource utilization, and call cost. After initially screening candidate agents with the corresponding service capabilities, the service matching and routing module further performs weighted calculations and multi-dimensional comparisons based on the above metrics to select the target service agent with the best overall performance in terms of performance, load, and cost. This mechanism effectively avoids resource overload or service quality degradation that may result from single-function matching, improving the overall system's resource utilization efficiency.
[0053] This application embeds a security mechanism into the entire service discovery process. During the service agent registration phase, the RA system centrally manages and stores the security certificates submitted by the agents and generates a unique certificate index. During the user request processing phase, the RA system, while returning the target agent's connection address, simultaneously retrieves and issues the corresponding security certificate from the certificate database using the certificate index. The user side uses this to authenticate the target agent and establish an encrypted communication channel, thereby ensuring the confidentiality, integrity, and non-repudiation of subsequent task collaboration and information exchange processes, effectively solving the problem of missing security mechanisms in existing technologies.
[0054] In summary, this application, through the construction of a converged proxy RA system and the definition of a new DNS resource record, organically integrates intent resolution, service discovery, multi-dimensional optimization, and security authentication, forming an integrated closed-loop solution from user input to secure service delivery. This solves the three major problems of inaccurate resolution, inefficient discovery, and lack of security in existing technologies, and significantly improves the service provision efficiency and reliability of intelligent agent interconnection networks.
[0055] The above are some specific implementations of the intelligent agent service provisioning method provided in the embodiments of this application. Based on this, this application also provides a corresponding intelligent agent service provisioning system. The system provided in the embodiments of this application will be described below from the perspective of functional modularization. Figure 2 This is a structural diagram of an intelligent agent service supply system provided in an embodiment of this application.
[0056] The system includes: The acquisition unit 110 is used to acquire the natural language service request sent by the client; Extraction unit 111 is used to parse the natural language service request and extract service requirement keywords; The determining unit 112 is used to retrieve stored domain name system resource records based on the service requirement keywords and determine the target service agent. The domain name system resource records are generated and stored based on the service information contained in the registration request sent by the service agent. The return unit 113 is used to return the connection information of the target service agent to the user, so that the user can establish a secure connection with the target service agent based on the connection information.
[0057] This application also provides corresponding devices and computer storage media for implementing the intelligent agent service provision solution provided in this application.
[0058] The device includes a memory and a processor. The memory is used to store instructions or code, and the processor is used to execute the instructions or code to cause the device to perform the intelligent agent service provisioning method according to any embodiment of this application.
[0059] The computer storage medium stores code, and when the code is executed, the device running the code implements the intelligent agent service provisioning method described in any embodiment of this application.
[0060] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems or apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple, and relevant parts can be referred to the method section.
[0061] It should be understood that in this application, "at least one" refers to one or more items, and "more" refers to two or more items. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one" or similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, "at least one" of a, b, or c can represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, and c can be single or multiple.
[0062] It should be understood that the terms center, longitudinal, transverse, up, down, front, back, left, right, vertical, horizontal, top, bottom, inside, outside, etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention.
[0063] It should be noted that, unless otherwise explicitly specified and limited, the terms installation, connection, and linking should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.
[0064] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the statement "including a…" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0065] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.
[0066] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. An agent service provisioning method, characterized by, The method comprises the following steps: acquiring a natural language service request sent by a user; parsing the natural language service request to extract a service demand keyword; retrieving a stored domain name system resource record based on the service demand keyword to determine a target service agent, wherein the domain name system resource record is generated and stored based on service information contained in a registration request sent by a service agent in advance; returning connection information of the target service agent to the user to enable the user to establish a secure connection with the target service agent according to the connection information.
2. The method of claim 1, wherein, Before the step of acquiring the natural language service request sent by the user, the method further comprises the following steps: receiving a registration request sent by a service agent, wherein the registration request contains service information of the service agent; allocating a certificate index to the service agent, binding the service agent and the corresponding certificate index, and storing them into a certificate database; and encapsulating the service information according to a predefined domain name system resource record format and storing it into a domain name system resource record.
3. The method of claim 2, wherein, After the step of receiving the registration request sent by the service agent, the method further comprises the following steps: performing information integrity verification on the service information and / or certificate validity verification; the step of allocating a certificate index to the service agent comprises the following step: allocating a certificate index to the service agent if the verification is passed.
4. The method of claim 1, wherein, The step of retrieving a stored domain name system resource record based on the service demand keyword to determine a target service agent comprises the following steps: matching a service capability label in the domain name system resource record according to the service demand keyword, and taking a service agent corresponding to a domain name system resource record in which the service capability label is matched with the service demand keyword as a candidate service agent; determining a target service agent from the candidate service agents based on overhead information of the candidate service agents.
5. The method of claim 1, wherein, The overhead information is contained in a service overhead field of the domain name system resource record, the service overhead field includes response delay, resource occupancy rate and invocation cost, and the step of determining a target service agent from the candidate service agents based on the overhead information of the candidate service agents comprises the following step: performing weighted calculation on the response delay, resource occupancy rate and invocation cost of each candidate service agent, and determining a target service agent from the candidate service agents according to the weighted calculation result.
6. The method of claim 1, wherein, The service information includes service capability description, protocol type, connection address, service overhead, agent unique identifier and security certificate.
7. The method of claim 1, wherein, The domain name system resource record is an RA-SRV type record defined based on DNS protocol extension, the type value is fixed as 0x8A, and the RDATA field includes service capability label, protocol type, address type, connection address, certificate index, service overhead and agent unique identifier; wherein the service capability label is used to identify the service capability of the service agent and supports multi-label splicing, and the service overhead at least contains one of response delay, resource occupancy rate and invocation cost.
8. The method of claim 1, wherein, The step of returning connection information of the target service agent to the user comprises the following steps: returning the connection address, communication protocol and security certificate of the target agent to the user.
9. An intelligent agent service provisioning system characterized by, The method comprises the following steps: An acquisition unit is configured to acquire a natural language service request sent by a client; An extraction unit is configured to parse the natural language service request and extract a service demand keyword; A determination unit is configured to retrieve a stored domain name system resource record based on the service demand keyword, and determine a target service agent, wherein the domain name system resource record is generated and stored according to service information contained in a registration request sent by a service agent; A return unit is configured to return connection information of the target service agent to the user, so that the user establishes a secure connection with the target service agent according to the connection information.
10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores instructions, and when the instructions run on a terminal device, the terminal device performs the method for providing an agent service according to any one of claims 1-8.