Service-driven network resource management method and system

Abstract

The invention relates to a service-driven network resource management method and system, and the method comprises the following steps: S1, carrying out the semantic analysis of an input configuration requirement, and generating a network configuration demand according to the predefined configuration; s2, acquiring the network state of the equipment layer; s3, according to the network state of the equipment layer and the network configuration requirement, performing segmented routing calculation, and generating a network path configuration instruction; and S4, generating a network configuration file according to the network path configuration instruction, and issuing the network configuration file to the equipment layer to perform network configuration of the equipment layer. According to the method, a flexible and efficient coping strategy under multiple tasks can be effectively solved, and the predefined semantic configuration strategy is continuously iteratively updated by utilizing the generic interpretation capability of the large language model, so that rapid issuing of new requirements is realized.

Description

Technical Field

[0001] This invention relates to the field of edge network management technology, and in particular to a service-driven network resource management method and system. Background Technology

[0002] In recent years, various heterogeneous end-point networks, especially unmanned intelligent platforms, have emerged. As requirements become increasingly stringent, ensuring efficient collaboration and reliable communication of end-point networks and meeting communication capabilities in different scenarios requires more than traditional manual network configuration and control methods. These methods are no longer sufficient to meet the demands of highly dynamic communication and network access. Therefore, there is an urgent need to implement an efficient, flexible, task-driven automated network configuration solution.

[0003] Chinese invention patent CN202410968949.4 proposes a task-driven distributed network intelligent management system and method, which relies on task intent understanding, translation, and parsing, and then generates network control strategies based on deep reinforcement learning, which are then verified and deployed. However, it has three significant drawbacks: First, the task escaping of natural language is inflexible, mainly based on predefined task instruction templates, and the difference from traditional manual configuration can only meet the predefined task requirements, with low satisfaction for temporary new requirements; Second, it lacks a comprehensive consideration in resolving various task conflicts, only allowing multiple tasks to negotiate service quality degradation under insufficient network resources in multi-task situations to achieve a minimum transmission rate, without considering the planning of different transmission paths for different tasks.

[0004] In view of this, there is an urgent need for a business-driven network resource management method and system to cope with the flexible management of network resources under multiple tasks. Summary of the Invention

[0005] To address the technical problems existing in the prior art, the present invention aims to provide a business-driven network resource management system that can effectively solve the problem of flexible and efficient response strategies under multi-task conditions. By utilizing the generalization and decomposition capabilities of a large language model, it can continuously iterate and update predefined semantic configuration strategies, thereby enabling the rapid deployment of new requirements.

[0006] To achieve the above-mentioned objectives, this invention provides a service-driven network resource management method, comprising the following steps:

[0007] Step S1: Perform semantic parsing on the input configuration requirements and generate network configuration requirements based on predefined configurations;

[0008] Step S2: Obtain the network status at the device layer;

[0009] Step S3: Based on the network status and network configuration requirements at the device layer, perform segmented route calculations and generate network path configuration instructions;

[0010] Step S4: Generate a network configuration file according to the network path configuration instructions, and send it to the device layer for network configuration at the device layer.

[0011] According to a technical solution of the present invention, step S1 specifically includes:

[0012] The input configuration requirements are parsed by a semantic parsing layer to extract natural language configuration keywords; the configuration requirements are in the form of artificial speech or intent text.

[0013] Based on the predefined configuration, the configuration keywords are matched to extract task features and generate network configuration requirements.

[0014] The network configuration requirements are defined as task feature vectors, which include: target object, service type, priority, latency requirements, and bandwidth requirements.

[0015] According to one technical solution of the present invention, in step S1, the predefined configuration is formed based on multiple protocol specifications, including a structured template library and a natural language-to-configuration parameter mapping dictionary; the structured template library includes device attributes, resource relationship attributes, and network configuration attributes, wherein the device attributes include device identifier, device type, device model, device serial number, number of device ports, and device deployment location; the resource relationship attributes include device interconnection resource attributes and operation and maintenance management resource attributes; and the network configuration attributes include transmission control configuration, QoS control configuration, protocol control configuration, and service bearer configuration.

[0016] The natural language to configuration parameter mapping dictionary for The rules for mapping business requirements to network QoS indicators, where business requirements are obtained through extracted natural language configuration keywords.

[0017] According to one technical solution of the present invention, in step S2, obtaining the device network status includes:

[0018] Step S21: Dynamically obtain the link state parameters of the device layer through the TWAMP protocol, and embed the obtained link state parameters into the IPv6 extension header and transmit them with the data packet; the link state parameters include link latency, jitter, packet loss rate and remaining bandwidth;

[0019] Step S22: Obtain network device topology information and abstract the network topology into a graph structure, where nodes in the graph structure are network devices, edges are links, and weights are link states.

[0020] According to a technical solution of the present invention, in step S3, the segmented routing calculation specifically includes:

[0021] Step S31: Generate each link path based on the obtained device network status;

[0022] Step S32: Perform path matching based on the task feature vector and each link path to generate SRv6 paths, and generate network path configuration instructions based on the generated SRv6 paths.

[0023] According to one technical solution of the present invention, in step S4, network path configuration is performed, specifically including:

[0024] According to the network path configuration instruction, it is matched with the stored current network configuration file. If the current network configuration file matches the network path configuration instruction, the current network configuration file is sent to the device layer.

[0025] If the current network configuration file does not match the network path configuration instructions, a candidate configuration file is generated based on the network configuration requirements. The candidate configuration file is then subjected to compatibility testing. If the candidate configuration file passes the compatibility test, it is sent to the device layer.

[0026] According to one aspect of the present invention, a service-driven network resource management system is provided, comprising:

[0027] The semantic parsing layer is used to transform semantic parsing into network configuration requirements.

[0028] The intelligent routing layer is used to perform segmented routing calculations and generate network path configuration instructions based on the network status and network configuration requirements of the device layer.

[0029] The configuration repository is used to validate network path configuration commands and generate configuration files.

[0030] The device layer is used for network connections based on configuration files.

[0031] According to one technical solution of the present invention, the configuration repository includes a configuration storage and a configuration controller. The configuration storage is used to store the current configuration file and the structured template library, and the configuration controller is used to generate configuration files according to network path configuration instructions.

[0032] Compared with the prior art, the present invention has the following advantages:

[0033] This invention provides a flexible and efficient strategy for handling multiple tasks. By leveraging the generalization and decomposition capabilities of a large language model, it enables continuous iterative updates of predefined semantic configuration strategies, thereby facilitating the rapid deployment of new requirements. Simultaneously, this invention utilizes an intelligent routing method based on IPv6 segmented routing to plan transmission paths for different services, proactively avoiding conflicts and fully leveraging the advantages of limited bandwidth. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0035] Figure 1 This diagram illustrates the working principle of the service-driven network resource management system provided in an embodiment of the present invention.

[0036] Figure 2 This diagram illustrates the workflow of the service-driven network resource management method provided in an embodiment of the present invention.

[0037] Figure 3 A schematic diagram illustrating the structured template library according to the present invention;

[0038] Figure 4 The schematic diagram illustrates the mapping dictionary according to the present invention. Detailed Implementation

[0039] The description of the embodiments in this specification should be taken in conjunction with the accompanying drawings, which should form part of the complete specification. In the drawings, the shape or thickness of the embodiments may be exaggerated and may be indicated in a simplified or convenient manner. Furthermore, parts of the various structures in the drawings will be described separately; it is worth noting that elements not shown in the figures or not described in words are in a form known to those skilled in the art.

[0040] The descriptions of the embodiments herein, including any references to directions and orientations, are for ease of description only and should not be construed as limiting the scope of the invention. The following description of preferred embodiments involves combinations of features, which may exist independently or in combination; the invention is not particularly limited to the preferred embodiments. The scope of the invention is defined by the claims.

[0041] like Figure 1 As shown, this invention provides a service-driven network resource management system and method. The system comprises a device layer, an intelligent routing layer, a configuration repository, and a semantic parsing layer. The semantic parsing layer is used to generate network configuration requirements from semantic parsing. The intelligent routing layer is used to perform segmented route calculations based on the network status and network configuration requirements of the device layer, generating network path configuration instructions. The configuration repository is used to verify the network path configuration instructions and generate configuration files. The device layer includes various network infrastructure devices and links, used for network connections according to the configuration files.

[0042] The semantic parsing layer generates network configuration requirements and task segmentation path calculation requirements, which are then sent to the intelligent routing layer. The intelligent routing layer, in conjunction with the network status of the device layer, sends these requirements to the configuration repository. After verification, the requirements are sent to the device layer, thus realizing network resource management and configuration.

[0043] Preferably, the semantic parsing layer can perform daily task semantic parsing and training iterations using a large model deployed at the edge, forming a set of regular pre-configured parameters. The large model at the edge can be a lightweight open-source model such as Qwen-7B or DeepSeek-7B, suitable for lightweight scenarios. Predefined configurations are established based on protocol specifications such as IEEE 802.1Q and RFC 8949, creating a structured template library containing IP fragmentation strategies, QoS classifications, and routing protocol parameters (such as OSPF cost values). This is combined with a dictionary defining the mapping from natural language to configuration parameters (e.g., "HD video" → "bandwidth ≥ 20Mbps, latency ≤ 50ms"), and regular expression matching to extract task features. The large model outputs a task feature vector (e.g., service type: video transmission, priority: high, bandwidth requirement: 20Mbps), which is then used to generate candidate configurations using a template matching engine. The compatibility of new configurations with existing configurations is checked (e.g., IP address conflicts, bandwidth exceeding limits). Leveraging the semantic parsing advantages of the general-purpose large model at the edge, corresponding network configurations can be quickly formed, enabling automated and rapid deployment. Compared to predefined templates, large models can parse unstructured task descriptions (such as "prioritizing medical image transmission in a rainstorm environment") and generate dynamic combinations of configuration parameters.

[0044] Preferably, the configuration repository includes a configuration storage and a configuration controller. The configuration storage is used to store the current configuration file and the structured template library, and the configuration controller is used to generate configuration files according to network path configuration instructions.

[0045] like Figure 2 As shown, the present invention provides a service-driven network resource management method, comprising the following steps:

[0046] Step S1: Perform semantic parsing on the input configuration requirements and generate network configuration requirements based on predefined configurations;

[0047] Step S2: Obtain the network status at the device layer;

[0048] Step S3: Based on the network status and network configuration requirements at the device layer, perform segmented route calculations and generate network path configuration instructions;

[0049] Step S4: Generate a network configuration file according to the network path configuration instructions, and send it to the device layer for network configuration at the device layer.

[0050] This invention leverages the semantic parsing advantages of a general-purpose large-scale model on the edge and terminal sides, and extracts task features using predefined configurations combined with regular expression matching. By outputting task feature vectors from the large model, corresponding network configurations can be quickly generated, enabling automated and rapid deployment. Compared to predefined templates, the large model can parse unstructured task descriptions and generate dynamic configuration parameter combinations. Simultaneously, it employs a segmented routing calculation method to generate network path configuration instructions. Facing different task requirements, such as environmental awareness, disaster relief, node collaboration, and the need to transmit text, voice, video, or images, it can implement customized paths, such as the optimal path under the OSPF algorithm and bandwidth balancing under segmented routing. Time-insensitive, bandwidth-intensive processes are then routed according to a bandwidth balancing strategy to achieve efficient network traffic forwarding.

[0051] Preferably, step S1 specifically includes:

[0052] The input configuration requirements are parsed by a semantic parsing layer to extract natural language configuration keywords; the configuration requirements are in the form of artificial speech or intent text.

[0053] Based on the predefined configuration, the configuration keywords are matched to extract task features and generate network configuration requirements.

[0054] Among them, network configuration requirements are task feature vectors, which include: target object, service type, priority, latency requirements, and bandwidth requirements.

[0055] Preferably, in step S1, the predefined configuration is formed based on multiple protocol specifications, including a structured template library and a natural language-to-configuration parameter mapping dictionary.

[0056] like Figure 3 As shown, the structured template library includes device attributes, resource relationship attributes, and network configuration attributes. Device attributes include device identifier, device type, device model, device serial number, number of device ports, and device deployment location. Resource relationship attributes include device interconnection resource attributes and operation and maintenance management resource attributes. Device interconnection resource attributes include the interface status of peer devices, recording the physical / logical connection status between devices and reflecting the connectivity of the network topology; operation and maintenance management resource attributes include northbound protocols and remote transmission management protocols. Network configuration attributes include transmission control configuration, QoS control configuration, protocol control configuration, and service bearer configuration. Transmission control configuration includes IP fragmentation, DHCP address configuration, NAT translation, etc.; QoS control configuration includes QoS level, link COST value, etc.; protocol control configuration includes VLAN configuration, trunk ports, OSPF / IS-IS / BGP, SRv6 policy, etc.; and service bearer configuration includes network slicing, etc.

[0057] Natural language to configuration parameter mapping dictionary for The rules for mapping business requirements to network QoS indicators, where business requirements are obtained through extracted natural language configuration keywords.

[0058] Preferably, in step S2, obtaining the device network status includes:

[0059] Step S21: Dynamically obtain the link state parameters of the device layer through the TWAMP protocol, and embed the obtained link state parameters into the IPv6 extension header and transmit them with the data packet; the link state parameters include link latency, jitter, packet loss rate and remaining bandwidth;

[0060] By dynamically collecting key indicators such as link latency (bidirectional latency / 2), jitter, packet loss rate, and remaining bandwidth through the TWAMP protocol, and seamlessly integrating them with the extension header fields of the IPv6 protocol (such as SRH), real-time awareness of network status can be achieved.

[0061] Step S22: Obtain network device topology information and abstract the network topology into a graph structure, where nodes in the graph structure are network devices, edges are links, and weights are link states.

[0062] The network topology information is collected from network devices through the BGP-LS (BGP Link-State) protocol, and the network topology is abstracted into a graph structure, where nodes are network devices, edges are links, and weights are link state parameters (latency, bandwidth, etc.).

[0063] Preferably, in step S3, the segmented route calculation specifically includes:

[0064] Step S31: Generate each link path based on the obtained device network status;

[0065] Step S32: Perform path matching based on the task feature vector and each link path to generate SRv6 paths, and generate network path configuration instructions based on the generated SRv6 paths.

[0066] The semantic parsing layer outputs task feature vectors, which are matched with network state parameters. Combined with real-time transmission status of each link obtained after network state data collection and processing, paths for each link are generated. Based on SRv6 and IPv6 address extension capabilities, path information is encoded in the IPv6 header field to match network paths and services, generating SRv6 paths and enabling customized forwarding of service traffic. Based on the generated SRv6 paths, path configurations are distributed to each router node. Through a natural language-to-configuration parameter mapping mechanism, combined with SRv6 path calculation, end-to-end automated control of "task-configuration-path" is achieved. Simultaneously, the open-source large model deployed at the endpoint performs rapid parsing of network configuration tasks in a natural language context, combining predefined configurations to form a mapping and distribution of service requirements to network configurations, fully leveraging high dynamic efficiency.

[0067] Preferably, in step S4, network path configuration is performed, specifically including:

[0068] According to the network path configuration instruction, it is matched with the stored current network configuration file. If the current network configuration file matches the network path configuration instruction, the current network configuration file is sent to the device layer.

[0069] If the current network configuration file does not match the network path configuration instructions, a candidate configuration file is generated based on the network configuration requirements. The candidate configuration file is then tested for compatibility. If the candidate configuration file passes the compatibility test, it is sent to the device layer.

[0070] Finally, it should be noted that the above description represents a preferred embodiment of the present invention. It should be pointed out that although preferred embodiments have been described, those skilled in the art, once they understand the basic inventive concept of the present invention, can make various improvements and modifications without departing from the principles described herein. These improvements and modifications should also be considered within the scope of protection of the present invention. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present invention.

Claims

1. A service-driven network resource management method, characterized in that, Includes the following steps: Step S1: Perform semantic parsing on the input configuration requirements and generate network configuration requirements based on predefined configurations; Step S2: Obtain the network status at the device layer; Step S3: Based on the network status and network configuration requirements at the device layer, perform segmented route calculations and generate network path configuration instructions; Step S4: Generate a network configuration file according to the network path configuration instructions, and send it to the device layer for network configuration at the device layer.

2. The service-driven network resource management method according to claim 1, characterized in that, Step S1 specifically includes: The input configuration requirements are parsed by a semantic parsing layer to extract natural language configuration keywords; the configuration requirements are in the form of artificial speech or intent text. Based on the predefined configuration, the configuration keywords are matched to extract task features and generate network configuration requirements. The network configuration requirements are defined as task feature vectors, which include: target object, service type, priority, latency requirements, and bandwidth requirements.

3. The service-driven network resource management method according to claim 2, characterized in that, In step S1, the predefined configuration is formed based on multiple protocol specifications, including a structured template library and a natural language-to-configuration parameter mapping dictionary. The structured template library includes device attributes, resource relationship attributes, and network configuration attributes. Among them, device attributes include device identifier, device type, device model, device serial number, number of device ports, and device deployment location; resource relationship attributes include device interconnection resource attributes and operation and maintenance management resource attributes; and network configuration attributes include transmission control configuration, QoS control configuration, protocol control configuration, and service bearer configuration. The natural language to configuration parameter mapping dictionary for The rules for mapping business requirements to network QoS indicators, where business requirements are obtained through extracted natural language configuration keywords.

4. The service-driven network resource management method according to claim 1, characterized in that, In step S2, obtaining the device network status includes: Step S21: Dynamically obtain the link state parameters of the device layer through the TWAMP protocol, and embed the obtained link state parameters into the IPv6 extension header and transmit them with the data packet; the link state parameters include link latency, jitter, packet loss rate and remaining bandwidth; Step S22: Obtain network device topology information and abstract the network topology into a graph structure, where nodes in the graph structure are network devices, edges are links, and weights are link states.

5. The service-driven network resource management method according to claim 4, characterized in that, In step S3, the segmented route calculation specifically includes: Step S31: Generate each link path based on the obtained device network status; Step S32: Perform path matching based on the task feature vector and each link path to generate SRv6 paths, and generate network path configuration instructions based on the generated SRv6 paths.

6. The service-driven network resource management method according to claim 5, characterized in that, In step S4, network path configuration is performed, specifically including: According to the network path configuration instruction, it is matched with the stored current network configuration file. If the current network configuration file matches the network path configuration instruction, the current network configuration file is sent to the device layer. If the current network configuration file does not match the network path configuration instructions, a candidate configuration file is generated based on the network configuration requirements. The candidate configuration file is then subjected to compatibility testing. If the candidate configuration file passes the compatibility test, it is sent to the device layer.

7. A service-driven network resource management system, used to implement the service-driven network resource management method as described in any one of claims 1 to 6, characterized in that, include: The semantic parsing layer is used to transform semantic parsing into network configuration requirements. The intelligent routing layer is used to perform segmented routing calculations and generate network path configuration instructions based on the network status, network configuration requirements, and segmented path calculation requirements of the device layer. The configuration repository is used to validate network path configuration commands and generate configuration files. The device layer is used for network connections based on configuration files.

8. The service-driven network resource management system according to claim 7, characterized in that, The configuration repository includes a configuration storage and a configuration controller. The configuration storage is used to store the current configuration file and the structured template library, and the configuration controller is used to generate configuration files according to network path configuration instructions.

Images