Intent-based production network configuration automatic generation method and system

Abstract

The application provides an intention-based production network configuration automatic generation method and system, which comprises the following steps: obtaining a topology of a specified communication network and inputting a plurality of intention strategies to be executed, wherein the topology comprises network autonomous domain division; obtaining link information and basic information of each node in the obtained topology; obtaining required ports and priorities of each intention strategy; combining all intention strategies to obtain a conflict-free intention expression according to whether there is a port conflict between the intention strategies and the priorities of the intention strategies; dividing the topology into a global topology and an internal topology by taking the network autonomous domain as a basic unit; decomposing the intention expression, taking the priorities of the intention strategies as a target function, and setting a link constraint and a path uniqueness constraint, and then obtaining all paths in the topology that meet the intention strategies through linear programming; and generating all network configuration files of the topology through a network protocol of the production network according to the paths obtained by the intention, so as to ensure that the network state meets all intention strategies.

Description

Technical Field

[0001] This invention relates to the fields of deployment planning and configuration, operation and maintenance configuration technology for production networks such as telecommunications core networks and data center networks. Background Technology

[0002] In traditional network environments, to address the problems of inefficiency, error-proneness, high cost, and poor maintenance associated with manual configuration, manufacturers and academia have gradually adopted automated methods to implement various network functions, translating given network requirements into actual network configurations. Commonly used network functions include technologies for network configuration generation and updating.

[0003] Network configuration generation refers to the process of transforming given network requirements into actual network configurations. In recent years, many automated network configuration generation technologies have emerged, such as model checking, constraint programming, automated planning, and symbolic execution, to automatically synthesize network configurations. These technologies can significantly improve the efficiency and accuracy of network configuration and automate network management and maintenance.

[0004] Network configuration updates refer to the process of adjusting network device configurations due to business needs or routine maintenance. However, network configuration updates must ensure that the normal operation of other services is not affected during or after the update process. Automated network configuration update technologies include using automated scripts, automated tools, network programming interfaces, network automation platforms, continuous integration and deployment, and network security auditing to automate the updating and management of network configurations. These technologies can greatly improve the efficiency and accuracy of network configuration updates and automate network management and maintenance.

[0005] In practical applications, to provide a network operation environment that even non-professional network operators can quickly master, it is necessary to set simple, easy-to-use, and highly expressive input intents during network configuration generation and updates. Here, "intent" is a higher-level abstraction independent of the specific underlying network. It can be represented by an intent strategy graph composed of higher-level abstract nodes and edges, representing the network administrator's management intent towards the underlying network. For example, ... Figure 1 As shown, there are two office areas, Zone A and Zone B. The gray boxes represent network function nodes that must be traversed, including firewalls (FW), load balancers (LB), and other nodes. The goal is to enable communication between the two areas, specifying the port (Port), bandwidth (BW), and weight information (Weight). This intent is represented in a graph. The advantage of using intents is that it allows complex and large-scale underlying configurations to be represented with simple and few intent policies, thereby reducing the workload of network administrators and improving their efficiency.

[0006] In traditional network environments, configuration updates require additional installation and configuration before using automation tools, depending on the network scenario. This often involves preparing the correct device list, templates, and commands, which is highly inefficient.

[0007] Current network configuration generation and update technologies have the following shortcomings: 1) Parsing high-level abstractions of intents does not support generating configurations that can be directly executed by devices from different vendors (such as Cisco and Huawei) in the production network. Devices from different vendors often have different configuration commands and formats, and currently there is no direct work to derive configurations from intents to devices from a specific vendor. 2) Other configuration generation methods are tightly coupled with the underlying production network and do not support higher-level abstractions or descriptions closer to natural language. Therefore, administrators need relevant professional skills and a thorough understanding of the current network situation. This significantly increases the difficulty of network configuration management, and as the network scale increases, the workload required for input intent planning may double or even increase exponentially. For example, the input needs to clearly define the relevant paths and parameter designs of various protocols, and also requires configuration sketches closely related to the underlying network devices. Due to its complexity, it is not conducive to scaling to large-scale network applications. 3) Existing configuration consistency update methods usually do not consider how to automatically and consistently integrate with existing configuration files in the underlying network. Due to the complexity of the content and format of configuration files from different vendors, achieving consistent updates directly from intents to configuration files remains very difficult. Summary of the Invention

[0008] The technical problems to be solved by this invention mainly include two aspects: First, it addresses the issues of redundant input information, complex synthesis processes, and limited application scenarios in current production network equipment configuration, achieving full automation from simple, limited intent policies directly to configuration files that can be directly executed by specific network vendors' devices. Second, it solves the problem of conflicts between updated and existing configurations during network updates, leading to errors or fluctuations in the operational status of existing services in the network, by implementing a process of configuration update combining configuration file recognition and intent.

[0009] Specifically, to address the shortcomings of existing technologies, this invention proposes an intention-based automatic generation method for production network configurations, including:

[0010] The initial step involves obtaining the topology of the specified communication network and inputting multiple first intent policies to be executed at the start. The topology nodes include basic information about routing nodes, host nodes, and network function nodes, and the topology includes network autonomous system divisions. The first intent policy is an expression of network connectivity between network regions with network autonomous systems as the basic unit.

[0011] The information processing steps include parsing the topology information to obtain the link information and basic information of each node in the topology; and parsing the first intent policy to obtain the required port and priority of each first intent policy.

[0012] The intent synthesis step combines the first intent strategies to obtain a conflict-free intent expression based on whether there are port conflicts between the first intent strategies and the priority of each first intent strategy.

[0013] The topology decomposition path solving steps are as follows: First, the topology is divided into a global topology with the network autonomous system as the basic unit. The nodes and their connections within the network autonomous system are then used as the internal topology. Based on the global and internal topologies, the intent expression is decomposed. The priority of each first intent strategy is used as the objective function, and link constraints and path uniqueness constraints are set. Through linear programming, the intent paths in the topology graph that satisfy all first intent strategies are obtained.

[0014] The configuration file generation step generates all network configuration files for the topology based on the path obtained from the intent and through the network protocol of the production network, ensuring that the network status conforms to all the first intent policies.

[0015] The intent-based automatic production network configuration generation method further includes a consistency update step:

[0016] Obtain multiple second intent policies to be updated, obtain the network configuration file of the topology map, and generate basic network topology information. Divide autonomous systems by intra-domain protocol number and boundary protocol. Based on the second intent policies and the basic network topology information, generate the optimal path consisting of global topology path, internal topology path and composite path. Generate the updated configuration file of the topology map by using the optimal path through the gateway protocol of the production network, so as to configure the generated network to run with the second intent policy.

[0017] The intent-based automatic production network configuration generation method, wherein the consistency update step includes:

[0018] Step S1: Obtain multiple intent policies to be updated, obtain the network configuration file of the topology map, and generate basic network topology information. Divide autonomous systems by intra-domain protocol number and boundary protocol. Generate the optimal path based on the intent policies and the basic network topology information.

[0019] Step S2: Based on the optimal path and network configuration file, perform connectivity verification on the first and last nodes of the optimal path. If the verification passes, no update operation is required; if the verification fails, proceed to step S3.

[0020] Step S3: Obtain the host node, boundary node, and internal topology path node of the starting autonomous region of the optimal path, and test the connectivity of the internal topology path node and boundary node of the starting autonomous region and the connectivity of the host node and boundary node. If all are connected, proceed to step S4; otherwise, update the configuration of the nodes in the starting autonomous region to obtain the updated configuration file.

[0021] Step S4: Perform connectivity checks between the boundary nodes of the starting autonomous region and the tail nodes of the optimal path. If they are connected, no update operation is needed; otherwise, proceed to step S5.

[0022] Step S5: Perform configuration file identification on the boundary nodes of the starting autonomous region and determine whether there is configuration information of the network segment where the endpoint of the blocking path is located in the identification result. If so, update the boundary nodes of the starting autonomous region to obtain the updated configuration file, so as to perform incremental configuration on the boundary nodes of the starting autonomous region; otherwise, proceed to step S6.

[0023] Step S6: Obtain the host node, boundary node, and internal topology path node of the endpoint autonomous system of the optimal path, and test the connectivity of the internal topology path node and boundary node of the endpoint autonomous system and the connectivity of the host node and boundary node. If all are connected, proceed to step S7; otherwise, update the configuration of the nodes in the endpoint autonomous system to obtain the updated configuration file.

[0024] Step S7: Perform configuration file identification on the boundary nodes of the endpoint autonomous system, and determine whether there is configuration information of the network segment where the endpoint of the blocking path is located in the identification result. If so, update the boundary nodes of the endpoint autonomous system to obtain the updated configuration file, and perform global topology path reverse incremental configuration on the boundary nodes of the endpoint autonomous system to obtain the updated configuration file; otherwise, proceed to step S8.

[0025] Step S8: Perform reverse traversal text recognition on the remaining nodes in the optimal path, excluding the starting autonomous region and the ending autonomous region. If all intent network segments are matched, perform global topology path reverse incremental configuration on the remaining nodes to obtain the updated configuration file; otherwise, proceed to step S9.

[0026] Step S9: Replace the path in the starting autonomous region of the optimal path with the path selected in the initial configuration, and configure the global topology path in reverse order to obtain the updated configuration file.

[0027] The aforementioned intention-based automatic generation method for production network configurations, wherein the specified communication network is a telecommunications core network and a data center network.

[0028] This invention also proposes an intent-based automatic production network configuration generation system, which includes:

[0029] The initial module allows users to design their own network topology or obtain the topology of a production network and input multiple first intent strategies to be executed at the start. The topology nodes include basic information about routing nodes, host nodes, and network function nodes, and the topology includes network autonomous system divisions. The first intent strategy is an expression of network connectivity between network regions with network autonomous systems as the basic unit.

[0030] The information processing module parses the topology information to obtain the link information and basic information of each node in the topology; it also parses the first intent policies to obtain the required ports and priorities of each first intent policy.

[0031] The intent synthesis module combines the first intent strategies to obtain a conflict-free intent expression based on whether there is a port conflict between the first intent strategies and the priority of each first intent strategy.

[0032] The topology splitting and path solving module divides the topology into a global topology with the network autonomous system as the basic unit, and the nodes and their connections within the network autonomous system as the internal topology. Based on the global topology and the internal topology, the intent expression is decomposed, and the priority of each first intent strategy is used as the objective function. Link constraints and path uniqueness constraints are set, and through linear programming, the intent paths in the topology graph that meet all first intent strategies are obtained.

[0033] The configuration file generation module generates all network configuration files for the topology based on the path obtained from the intent and using the network protocols of the production network, ensuring that the network status conforms to all the first intent policies.

[0034] The intent-based production network configuration auto-generation system further includes a consistency update module:

[0035] Obtain multiple second intent policies to be updated, obtain the network configuration file of the topology map, and generate basic network topology information. Divide autonomous systems by intra-domain protocol number and boundary protocol. Based on the second intent policies and the basic network topology information, generate the optimal path consisting of global topology path, internal topology path and composite path. Generate the updated configuration file of the topology map by using the optimal path through the gateway protocol of the production network, so as to configure the generated network to run with the second intent policy.

[0036] The intent-based production network configuration auto-generation system, wherein the consistency update module includes:

[0037] Module 1: Obtain multiple intent policies to be updated, obtain the network configuration file of the topology map, and generate basic network topology information. Divide autonomous systems by intra-domain protocol number and boundary protocol. Generate the optimal path based on the intent policies and the basic network topology information.

[0038] Module 2: Based on the optimal path and network configuration file, perform connectivity verification on the first and last nodes of the optimal path. If the verification passes, no update operation is required; if the verification fails, execute Module 3.

[0039] Module 3: Obtain the host node, boundary node, and internal topology path node of the starting autonomous region of the optimal path, and test the connectivity of the internal topology path node and boundary node of the starting autonomous region and the connectivity of the host node and boundary node. If all are connected, execute Module 4; otherwise, update the configuration of the nodes in the starting autonomous region to obtain the updated configuration file.

[0040] Module 4: Perform connectivity checks between the boundary nodes of the starting autonomous region and the tail node of the optimal path. If they are connected, no update operation is needed; otherwise, execute Module 5.

[0041] Module 5: Perform configuration file identification on the boundary nodes of the starting autonomous region, and determine whether there is configuration information of the network segment where the endpoint of the blocking path is located in the identification result. If so, update the boundary nodes of the starting autonomous region to obtain the updated configuration file, so as to perform incremental configuration on the boundary nodes of the starting autonomous region; otherwise, execute Module 6.

[0042] Module 6: Obtain the host nodes, boundary nodes, and internal topology path nodes of the endpoint autonomous system of the optimal path, and test the connectivity of the internal topology path nodes and boundary nodes of the endpoint autonomous system as well as the connectivity of the host nodes and boundary nodes. If all are connected, execute Module 7; otherwise, update the configuration of the nodes in the endpoint autonomous system to obtain the updated configuration file.

[0043] Module 7: Perform configuration file identification on the boundary nodes of the endpoint autonomous system, and determine whether there is configuration information of the network segment where the endpoint of the blocking path is located in the identification result. If so, update the boundary nodes of the endpoint autonomous system to obtain the updated configuration file, and perform global topology path reverse incremental configuration on the boundary nodes of the endpoint autonomous system to obtain the updated configuration file; otherwise, execute Module 8.

[0044] Module 8: Perform reverse traversal text recognition on the remaining nodes in the optimal path, excluding the starting autonomous region and the ending autonomous region. If all intent network segments are matched, perform global topology path reverse incremental configuration on the remaining nodes to obtain the updated configuration file; otherwise, execute Module 9.

[0045] Module 9 replaces the path in the starting autonomous region of the optimal path with the path selected in the initial configuration, and configures the global topology paths in reverse order to obtain the updated configuration file.

[0046] The aforementioned intent-based production network configuration automatic generation system, wherein the specified communication network is a telecommunications core network and a data center network.

[0047] The present invention also proposes a storage medium for storing a program that executes any of the intent-based automatic generation methods for production network configurations.

[0048] The present invention also proposes a client for any intent-based production network configuration auto-generation system.

[0049] As can be seen from the above solutions, the advantages of the present invention are:

[0050] This invention proposes an intent-based automatic generation and consistency update system for production network configurations. The system can perform two functions: intent-based automatic configuration generation and intent-based consistency update.

[0051] The configuration generation section can generate network configuration files independently. Through the intent processing module and the configuration file generation module, it implements the distribution and synthesis of intent policies, ultimately realizing the process from intent to underlying configuration file distribution. The current protocol supports BGP, IS-IS, and OSPF. This significantly reduces the input workload, improves preparation efficiency, supports large-scale topologies, and can generate network-wide node configuration files conforming to intent policies within a reasonable timeframe.

[0052] The configuration consistency update component integrates network configuration files and intents. Administrators input intents, retrieve network configuration files from the configuration files, and generate network topology information. The intent processing module generates intent network segment and path information based on the intent and topology information. The intent incremental update processing module first processes blocking intents, then analyzes node traffic to determine whether to update node configurations, establishing connectivity for connectivity intents. The system achieves error-free and conflict-free configuration updates through consistency verification and merging of old and new configurations. Attached Figure Description

[0053] Figure 1 A diagram illustrating the intent strategy in the form of a single graph;

[0054] Figure 2 Configure automatic generation and updating of system diagrams for intent-based production networks;

[0055] Figure 3 A production network topology diagram with 38 nodes;

[0056] Figure 4 A global topology graph for solving path decomposition mechanisms;

[0057] Figure 5 Automatically generate application scenario diagrams for intent-based device profiles;

[0058] Figure 6A schematic diagram illustrating the application scenario of the consistency update method for existing configuration files and newly added intents;

[0059] Figure 7 Flowchart of the intention incremental update processing algorithm;

[0060] Figure 8 Here is a flowchart of the method for handling connectivity in an autonomous system.

[0061] Figure 9 Flowchart of the method for configuring global topology paths in reverse order incrementally. Detailed Implementation

[0062] While researching configuration generation and consistency updates, the inventors discovered that existing configuration generation and update technologies suffer from complex inputs and cannot be directly applied to production networks. In practical applications, especially for initial configuration and routine maintenance of large-scale networks, these methods remain time-consuming and labor-intensive. To alleviate daily workload and improve efficiency and accuracy, the inventors transformed the step-by-step management process into a fully automated one, supporting intent-based configuration updates. Through research on intent-based networks, the inventors found that this deficiency can be addressed by designing a method to directly parse abstract intent expressions into configuration files that can be directly executed by underlying network devices. This led to the proposal of an intent-based automatic configuration generation and consistency update system for production networks. The system's functionality consists of two parts: intent-based automatic configuration generation and configuration consistency update.

[0063] Intent-based automatic configuration generation automates the entire process from high-level abstract intents to low-level configurations, reducing workload and ease of use. This part automatically generates new, executable network configuration files for the underlying network based on input intents, achieving fully automated intent assignment to the underlying network. Its core steps include: administrators issuing intent policies; the system automatically combining intents to eliminate redundancy and conflicts in the intent policies; a topology splitting and path-solving mechanism is proposed to split the network topology into global and internal topologies, solving for all paths between two nodes for each intent, and then performing path optimization to obtain the optimal path for the intent, shortening the path-solving time; a configuration file generation method is proposed to automatically generate configuration files. The intent-based automatic configuration generation proposed in this invention currently supports the automatic generation of Cisco's BGP, ISIS, and OSPF protocol configurations, and can be extended based on the same principle to support the automatic generation of configurations from intent policies to other vendors' devices and more network protocol functions.

[0064] The configuration consistency update section integrates administrator intent policies with existing network configurations, satisfying administrator configuration update needs without impacting the service quality and security requirements of other network services. The configuration consistency update section first obtains the network configuration file, then processes it in the intent incremental update processing module based on the path information obtained from the intent. This module analyzes node traffic to determine if relevant nodes need updating. If the node traffic analysis passes, no update is performed. If it fails, the relevant node information needs updating. Generally, administrator intents can be divided into two categories: blocking intents (blocking routing communication between specified areas) and connectivity intents (configuring routing connectivity between specified areas). For blocking intents, if a node has an intent network segment, blocking can be directly set on the node. For connectivity intents, the intent incremental update processing algorithm is used for configuration updates. This module handles two types of nodes: nodes on the path obtained from the intent, and nodes on non-intent paths obtained from all paths in the global topology before path optimization. This is because the initial network may have policy settings that cause the current network to choose a certain path, not necessarily the path obtained from the intent. Based on the path generated by the intent, the algorithm finds nodes that can connect the network segments. The intent connectivity processing algorithm can handle various scenarios, identify and update configuration file information, and generate incrementally updated configuration files.

[0065] Ultimately, an intent-based automatic generation and consistent update system for production network configurations was implemented, which can effectively realize the process of generating new network configurations based on intents and support consistent updates of configuration files based on intents.

[0066] To make the above features and effects of the present invention clearer and easier to understand, specific embodiments are described below, and detailed descriptions are provided in conjunction with the accompanying drawings.

[0067] In the specific implementation of this invention, the operation is carried out according to the administrator's selection, and the system supports configuration generation and consistency update functions.

[0068] When initially building a completely new network (i.e., going live at launch), according to Figure 2 The process requires the administrator to input a network topology file and an initial intent policy, which is a standardized expression. The entire process is implemented through an intent processing module and a configuration generation module. During this process, information such as bandwidth, service chains, and ports are configured for intent settings. Multiple specified policies are combined into a large network that ensures simultaneous operation without errors, and the optimal solution is selected based on the chosen policy, ensuring that each policy can handle the desired intent. Intents that cannot be handled are notified to the administrator. After generating the specific configuration file, it is verified using relevant software such as Batfish or GNS3. Experiments have shown that the network formed by the generated configuration file satisfies the administrator's intent.

[0069] When network configuration needs to be updated due to new requirements or business adjustments during network operation, the system automatically obtains the current configuration file via SSH or direct download, taking topology information, intent policies, and the current configuration file as input. The system intelligently processes nodes along each path using algorithms, reconstructs the graph, stores relevant information, intelligently analyzes routing information for the intent network segment and the current node network segment, and intelligently identifies various situations using an intent-based incremental update processing algorithm. This allows for incremental configuration methods under different network conditions, ultimately generating the relevant incremental configuration file. Verifying the file against a previously unmodified file ensures that it does not affect the routing information of unaffected node autonomous systems and related nodes, while still fulfilling the relevant intents.

[0070] To address the aforementioned issues, an intent-based system for automatic generation and consistent updating of production network configurations is proposed. The system implementation process is detailed below. Figure 2 The topology information input on the left needs to include basic node information, including: router nodes, host nodes, and network function nodes (firewall nodes and load balancer nodes).

[0071] Automatic generation of intent-based device profiles.

[0072] The intent-based automatic generation method for device configuration files implements a network configuration management approach that directly distributes intents to the underlying configuration files. The input consists of intents and basic topology information, where the intent is an abstract policy built upon the topology information, with each network autonomous system (AAS) as the basic unit. Figure 1 As shown, ZoneA and ZoneB are abstract descriptions of the network, with each zone containing several routers, hosts, or network function nodes. Basic topology information describes the specific link information, autonomous system (AS), and maximum bandwidth limits of each entity node. The entire operation process is implemented through the intent processing module and the configuration file generation module. In contrast, existing technologies require not only basic topology and requirement information but also configuration sketches and routing announcements to ensure the synthesis process. The input for the intent-to-lower-level configuration file synthesis method is shown below. Figure 2 The entire process can be completed by inputting topology and intent information. A usage scenario process diagram is shown below. Figure 5 As shown.

[0073] This method first passes through the intent processing module. The intent processing module implements the entire process through the initial information processing module, the intent synthesis module, and the topology splitting and path solving mechanism. It obtains synthesized graph information and optimal path information, which are then processed by the configuration file generation module to generate the entire network configuration file. Figure 2 The process in ①.

[0074] The initial information processing module preprocesses multiple input intent policies by analyzing the topology file (containing link information and basic node information) and multiple intent information (such as...). Figure 1 The content of the strategy shown is parsed to obtain the information needed by the intent synthesis module. The intent synthesis module combines multiple intent strategies, handling port conflicts, priority order, etc., adding multiple input intent strategies one by one to a graph, ultimately combining them into a conflict-free composite graph. The topology decomposition solution path mechanism is visible. Figure 3 The topology shown is a user-input topology with 38 nodes. The system splits it into two parts, the first part as follows: Figure 4 As shown, this is called the global topology, also known as the non-intended operation node. The second part is... Figure 3 Each node within a dashed box (autonomous region) is called the internal topology. Based on these two points, integer linear programming is used to solve for the internal and global topology paths, setting priority as the objective function and setting link constraints and path uniqueness constraints to solve for the intended path. The configuration file generation module processes the path information from the previous step and solves according to different protocols. This method supports the Border Gateway Protocol (BGP) and the Interior Gateway Protocol (IGP), where the IGP protocol supports OSPF and IS-IS protocols, and the protocol is set based on the obtained optimal path.

[0075] By inputting intent and topology information, the system can generate the correct configuration file for a specified vendor within a reasonable timeframe. This system currently supports configuration file generation for Cisco network nodes. It can generate configuration files for 5000 nodes involving 100,000 paths within 20 minutes.

[0076] A consistent update method for existing configuration files and newly added intents.

[0077] Corresponding to the consistent update methods for existing configuration files and newly added intents Figure 2 In step ②, the input is the intent information. The scenario description is available below. Figure 6 The network administrator, based on the intent information of a running network, implements configuration updates through a configuration file acquisition module, an intent processing module, and a configuration update subsystem. The configuration file acquisition module automatically retrieves the network configuration file and generates basic network topology information, dividing autonomous systems (AS) using intra-domain protocol numbers and boundary protocols. The intent processing module generates composite graph information and intent path information based on the intent and topology information. The configuration update subsystem implements the entire process through the intent incremental update processing module, updating node configuration files. Corresponding application scenarios can be found from... Figure 6 Check.

[0078] The intention incremental update module uses the path information obtained from the intention processing module and the network configuration file obtained from the configuration file acquisition module to deliver the intention and generate a network incremental update configuration file, thus completing the entire incremental processing process. The core of the intention incremental update module is the intention incremental update processing algorithm. The input is the intention path and configuration file information. The algorithm verifies whether an update is needed by performing connectivity tests and path tests on the beginning and end of the intention path. If the verification passes, no action is required; if the verification fails, it indicates that the current network is unable to connect due to the configuration of certain nodes or the lack of configuration on certain nodes. Node text recognition and node traffic analysis are required for the nodes along the intention path, and configuration update operations are performed on the intention nodes. The intention incremental update module divides the path into five parts for identification and processing: the connectivity of the intention starting autonomous system, the connectivity of the intention path boundary starting point and adjacent points, the connectivity of the global topology nodes of the intention path, the connectivity of the intention ending autonomous system, and the connectivity of non-intention paths. These five parts cover all nodes involved in network connectivity, ultimately realizing the configuration and delivery of the intention and generating the network configuration file.

[0079] The intent path is decomposed into different node scenarios based on the topology decomposition path solving mechanism, and then identified and processed. The specific process includes (see details). Figure 7 ): Assumption Figure 3 The topology intent yields the path ["node2","node4","node5","node6","node16","node19","node20","node21","node23","node25","node28"] (referred to as path A). According to the topology splitting and path solving mechanism of the intent processing module, path A is a combination of the global topology path "node5, node6, node16, node19, node20" (referred to as path A1) and the internal topology paths "node2, node4, node5" (referred to as path A2) and "node28, node25, node23, node21, node20" (referred to as path A3).

[0080] (1) Intent origin autonomous region connectivity processing

[0081] First, the blocking traffic is configured at the autonomous system boundary node for the intent endpoint of the intent blocking. Figure 7 ① Blocking intent processing. Afterwards, the connectivity and path matching of the obtained intent path A between node2 and node28 are tested. If both pass, it proves that the intent already exists, and no further configuration is required. Figure 7 ② Connectivity check of the starting node of the intent path. If connectivity is not possible, node2 and its autonomous system boundary node5 need to be connected according to path A2. Figure 7The ③ intentional autonomous region connectivity processing method can be used in Figure 8 Examine its implementation process. After connecting node2 and node5, test on node5 and node28. Figure 7 ④ Connectivity detection between the autonomous system boundary node and the end point of the path. If the path is connected, then node2 and node28 are tested. If they pass, the process ends. If the path is not connected, the remaining parts of path A1 and A3 need to be processed, and then proceed to step (2).

[0082] (2) Connectivity processing of the starting point and adjacent points of the intended path boundary

[0083] Suppose the initial configuration on node5 blocks all network segments in ZoneD, while node6 has all network segments. In this scenario, performing a configuration file recognition on node5 will identify the corresponding configuration. Figure 7 ⑤ Boundary starting point and adjacent point text recognition. If node5 has configuration information for the network segment where the blocked path endpoint is located, only node5 needs to be updated, and the relevant configuration information should be deleted. Figure 7 ⑦ Incremental configuration of boundary starting point or adjacent point. The configuration of node6 is the same; if not all are matched, there are still some network segments that are not connected, proceed to step (3).

[0084] (3) Intended endpoint autonomous region connectivity processing

[0085] The intended endpoint autonomous system boundary node is node20. Testing on node20 and node28 yields the following results: Figure 7 ⑥ Intended endpoint domain connectivity detection.

[0086] If the test fails, the paths for node20 and node28 in ZoneD need to be configured according to the A3 path, corresponding to the... Figure 7 The ③ intention is to handle autonomous system connectivity, in which case path A3 will be connected.

[0087] If we perform text recognition on node20 of path A1 through testing, the corresponding... Figure 7 ⑧ Boundary endpoint text recognition. Assume node20 is connected to node28, and the node28 network segment is blocked. In this scenario, if the text match is successful, process path A1, corresponding to... Figure 9 . Figure 9 The processing procedure involves swapping the intended network segment routes between node20 and its neighboring node19, and then traversing the remaining nodes along processing path A1 in reverse order. Specifically, node19 performs a route swap along the interface of node16, while the remaining neighbors block the intended route to maintain the status quo of the connected autonomous systems. All nodes operate sequentially. Figure 7 The method for incremental reverse configuration of global topology paths (9). If matching fails, proceed to step (4).

[0088] (4) Intent path global topology node connectivity processing

[0089] At this point, the remaining nodes of path A1 are processed. By reverse-order analysis, the text recognition is performed by traversing the text in reverse order from node 19 of A1. Figure 7 The remaining nodes of the ⑩ intent path are analyzed. If the identification results contain configuration information for the network segment where the blocked path endpoint is located, such as a blocking operation at a certain position on node19 or node16, it indicates that the target intent network segment is blocked through this setting. The global topology path reverse incremental configuration method is then executed. Figure 7 ⑨. If the match fails, it means that the intent path and the path selected in the initial configuration file are different, and proceed to step (5).

[0090] (5) Non-intended path connectivity processing

[0091] At this point, none of the nodes in path A1 are configured to allow node2 to connect with node28. This indicates that path A1 does not match any of the paths between nodes 5 and 20. Figure 7 of Graph connectivity and text analysis of other path nodes. There are two scenarios here. Figure 4 If a blocking information exists on an equivalent path node or a non-equivalent path in the global topology, then path A1 is replaced with the initial global topology path obtained by the intent processing subsystem, and the global topology paths are configured in reverse order. Figure 7 of How to modify the path movement configuration strategy.

[0092] The intent incremental update module is used for each intent when intents are connected. If it's a blocking intent, execution can be done directly at the boundary node; if it's a connected intent, the intent path is obtained and processed. The algorithm uses the traffic analysis module to determine whether the obtained path nodes need configuration. If the initial nodes of the path are directly connectable and the path meets the requirements, no configuration is needed and the process ends directly; if they are not connectable, node configuration is required to satisfy connectivity. The intent incremental update processing algorithm analyzes node configuration under different conditions through configuration file text recognition and performs corresponding incremental updates, ultimately outputting the configuration file for the updated nodes to achieve configuration updates. Different configuration methods have been developed, each for... Figure 7 ③, ⑦, ⑨

[0093] The algorithm described above analyzes and identifies the intent and the current configuration file, comprehensively matching the text information of the involved nodes with the intent. This avoids modeling and analyzing the entire network, saving workload. Finally, the intent information is fused with the current configuration file to achieve incremental configuration, outputting the files involved in the configuration update. After verification, the network behavior conforms to the newly added intent.

[0094] The following are system embodiments corresponding to the above method embodiments. This embodiment can be implemented in conjunction with the above embodiments. The relevant technical details mentioned in the above embodiments are still valid in this embodiment, and will not be repeated here to reduce repetition. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the above embodiments.

[0095] This invention also proposes an intent-based automatic production network configuration generation system, which includes:

[0096] The initial module allows users to design their own network topology or obtain the topology of a production network and input multiple first intent strategies to be executed at the start. The topology nodes include basic information about routing nodes, host nodes, and network function nodes, and the topology includes network autonomous system divisions. The first intent strategy is an expression of network connectivity between network regions with network autonomous systems as the basic unit.

[0097] The information processing module parses the topology information to obtain the link information and basic information of each node in the topology; it also parses the first intent policies to obtain the required ports and priorities of each first intent policy.

[0098] The intent synthesis module combines the first intent strategies to obtain a conflict-free intent expression based on whether there is a port conflict between the first intent strategies and the priority of each first intent strategy.

[0099] The topology splitting and path solving module divides the topology into a global topology with the network autonomous system as the basic unit, and the nodes and their connections within the network autonomous system as the internal topology. Based on the global topology and the internal topology, the intent expression is decomposed, and the priority of each first intent strategy is used as the objective function. Link constraints and path uniqueness constraints are set, and through linear programming, the intent paths in the topology graph that meet all first intent strategies are obtained.

[0100] The configuration file generation module generates all network configuration files for the topology based on the path obtained from the intent and using the network protocols of the production network, ensuring that the network status conforms to all the first intent policies.

[0101] The intent-based production network configuration auto-generation system further includes a consistency update module:

[0102] Obtain multiple second intent policies to be updated, obtain the network configuration file of the topology map, and generate basic network topology information. Divide autonomous systems by intra-domain protocol number and boundary protocol. Based on the second intent policies and the basic network topology information, generate the optimal path consisting of global topology path, internal topology path and composite path. Generate the updated configuration file of the topology map by using the optimal path through the gateway protocol of the production network, so as to configure the generated network to run with the second intent policy.

[0103] The intent-based production network configuration auto-generation system, wherein the consistency update module includes:

[0104] Module 1: Obtain multiple intent policies to be updated, obtain the network configuration file of the topology map, and generate basic network topology information. Divide autonomous systems by intra-domain protocol number and boundary protocol. Generate the optimal path based on the intent policies and the basic network topology information.

[0105] Module 2: Based on the optimal path and network configuration file, perform connectivity verification on the first and last nodes of the optimal path. If the verification passes, no update operation is required; if the verification fails, execute Module 3.

[0106] Module 3: Obtain the host node, boundary node, and internal topology path node of the starting autonomous region of the optimal path, and test the connectivity of the internal topology path node and boundary node of the starting autonomous region and the connectivity of the host node and boundary node. If all are connected, execute Module 4; otherwise, update the configuration of the nodes in the starting autonomous region to obtain the updated configuration file.

[0107] Module 4: Perform connectivity checks between the boundary nodes of the starting autonomous region and the tail node of the optimal path. If they are connected, no update operation is needed; otherwise, execute Module 5.

[0108] Module 5: Perform configuration file identification on the boundary nodes of the starting autonomous region, and determine whether there is configuration information of the network segment where the endpoint of the blocking path is located in the identification result. If so, update the boundary nodes of the starting autonomous region to obtain the updated configuration file, so as to perform incremental configuration on the boundary nodes of the starting autonomous region; otherwise, execute Module 6.

[0109] Module 6: Obtain the host nodes, boundary nodes, and internal topology path nodes of the endpoint autonomous system of the optimal path, and test the connectivity of the internal topology path nodes and boundary nodes of the endpoint autonomous system as well as the connectivity of the host nodes and boundary nodes. If all are connected, execute Module 7; otherwise, update the configuration of the nodes in the endpoint autonomous system to obtain the updated configuration file.

[0110] Module 7: Perform configuration file identification on the boundary nodes of the endpoint autonomous system, and determine whether there is configuration information of the network segment where the endpoint of the blocking path is located in the identification result. If so, update the boundary nodes of the endpoint autonomous system to obtain the updated configuration file, and perform global topology path reverse incremental configuration on the boundary nodes of the endpoint autonomous system to obtain the updated configuration file; otherwise, execute Module 8.

[0111] Module 8: Perform reverse traversal text recognition on the remaining nodes in the optimal path, excluding the starting autonomous region and the ending autonomous region. If all intent network segments are matched, perform global topology path reverse incremental configuration on the remaining nodes to obtain the updated configuration file; otherwise, execute Module 9.

[0112] Module 9 replaces the path in the starting autonomous region of the optimal path with the path selected in the initial configuration, and configures the global topology paths in reverse order to obtain the updated configuration file.

[0113] The aforementioned intent-based production network configuration automatic generation system, wherein the specified communication network is a telecommunications core network and a data center network.

[0114] The present invention also proposes a storage medium for storing a program that executes any of the intent-based automatic generation methods for production network configurations.

[0115] The present invention also proposes a client for any intent-based production network configuration auto-generation system.

Claims

1. A method for automatically generating production network configurations based on intent, characterized in that, include: The initial step involves obtaining the topology of the specified communication network and inputting multiple first intent policies to be executed at the start. The topology nodes include basic information about routing nodes, host nodes, and network function nodes, and the topology includes network autonomous system divisions. The first intent policy is an expression of network connectivity between network regions with network autonomous systems as the basic unit. The information processing steps involve parsing the topology information to obtain link information and basic information about each node in the topology; The first intent strategy is parsed to obtain the required port and priority of each first intent strategy; The intent synthesis step combines the first intent strategies to obtain a conflict-free intent expression based on whether there is a port conflict between the first intent strategies and the priority of each first intent strategy. The topology decomposition path solving steps are as follows: First, the topology is divided into a global topology with the network autonomous system as the basic unit, and the nodes and their connections within the network autonomous system are taken as the internal topology. Based on the global topology and the internal topology, the intent expression is decomposed, and the priority of each first intent strategy is used as the objective function. Link constraints and path uniqueness constraints are set, and through linear programming, the intent paths in the topology graph that meet all first intent strategies are obtained. The configuration file generation step generates all network configuration files for the topology based on the path obtained from the intent and through the network protocol of the production network, ensuring that the network status conforms to all the first intent policies.

2. The intention-based automatic generation method for production network configuration as described in claim 1, characterized in that, It also includes a consistent update step: The system acquires multiple second intent policies to be updated, obtains the network configuration file for the topology map, and generates basic network topology information. It then divides autonomous systems using intra-domain protocol numbers and boundary protocols. Based on the second intent policies and the basic network topology information, it generates the optimal path, which consists of global topology paths, internal topology paths, and synthesized paths. The optimal path is then used to generate an update configuration file for the topology map through the gateway protocol of the production network. This configuration file is then used to configure the generated basic network topology information to run according to the second intent policies.

3. The intention-based automatic generation method for production network configuration as described in claim 2, characterized in that, The consistency update step specifically includes: Step S1: Obtain multiple intent policies to be updated, obtain the network configuration file of the topology map, and generate basic network topology information. Divide autonomous systems by intra-domain protocol number and boundary protocol. Generate the optimal path based on the intent policies and the basic network topology information. Step S2: Based on the optimal path and network configuration file, perform connectivity verification on the first and last nodes of the optimal path. If the verification passes, no update operation is required; if the verification fails, proceed to step S3. Step S3: Obtain the host node, boundary node, and internal topology path node of the starting autonomous region of the optimal path, and test the connectivity of the internal topology path node and boundary node of the starting autonomous region and the connectivity of the host node and boundary node. If all are connected, proceed to step S4; otherwise, update the configuration of the nodes in the starting autonomous region to obtain the updated configuration file. Step S4: Perform connectivity checks between the boundary nodes of the starting autonomous region and the tail nodes of the optimal path. If they are connected, no update operation is needed; otherwise, proceed to step S5. Step S5: Perform configuration file identification on the boundary nodes of the starting autonomous region and determine whether there is configuration information of the network segment where the endpoint of the blocking path is located in the identification result. If so, update the boundary nodes of the starting autonomous region to obtain the updated configuration file, so as to perform incremental configuration on the boundary nodes of the starting autonomous region; otherwise, proceed to step S6. Step S6: Obtain the host node, boundary node, and internal topology path node of the endpoint autonomous system of the optimal path, and test the connectivity of the internal topology path node and boundary node of the endpoint autonomous system and the connectivity of the host node and boundary node. If all are connected, proceed to step S7; otherwise, update the configuration of the nodes in the endpoint autonomous system to obtain the updated configuration file. Step S7: Perform configuration file identification on the boundary nodes of the endpoint autonomous system, and determine whether there is configuration information of the network segment where the endpoint of the blocking path is located in the identification result. If so, update the boundary nodes of the endpoint autonomous system to obtain the updated configuration file, and perform global topology path reverse incremental configuration on the boundary nodes of the endpoint autonomous system to obtain the updated configuration file; otherwise, proceed to step S8. Step S8: Perform reverse traversal text recognition on the remaining nodes in the optimal path, excluding the starting autonomous region and the ending autonomous region. If all intent network segments are matched, perform global topology path reverse incremental configuration on the remaining nodes to obtain the updated configuration file; otherwise, proceed to step S9. Step S9: Replace the path in the starting autonomous region of the optimal path with the path selected in the initial configuration, and configure the global topology path in reverse order to obtain the updated configuration file.

4. The intention-based automatic generation method for production network configuration as described in claim 1, characterized in that, The designated communication network is the telecommunications core network and data center network.

5. An intent-based automatic production network configuration generation system, characterized in that, include: The initial module allows users to design their own network topology or obtain the topology of a production network and input multiple first intent strategies to be executed at the start. The topology nodes include basic information about routing nodes, host nodes, and network function nodes, and the topology includes network autonomous system divisions. The first intent strategy is an expression of network connectivity between network regions with network autonomous systems as the basic unit. The information processing module parses the topology information to obtain link information and basic information of each node in the topology; The first intent strategy is parsed to obtain the required port and priority of each first intent strategy; The intent synthesis module combines the first intent strategies to obtain a conflict-free intent expression based on whether there is a port conflict between the first intent strategies and the priority of each first intent strategy. The topology splitting and path solving module divides the topology into a global topology with the network autonomous system as the basic unit, and the nodes and their connections within the network autonomous system as the internal topology. Based on the global topology and the internal topology, the intent expression is decomposed, and the priority of each first intent strategy is used as the objective function. Link constraints and path uniqueness constraints are set, and through linear programming, the intent paths in the topology graph that meet all first intent strategies are obtained. The configuration file generation module generates all network configuration files for the topology based on the path obtained from the intent and using the network protocols of the production network, ensuring that the network status conforms to all the first intent policies.

6. The intention-based production network configuration automatic generation system as described in claim 5, characterized in that, It also includes a consistent update module: The system acquires multiple second intent policies to be updated, obtains the network configuration file for the topology map, and generates basic network topology information. It then divides autonomous systems using intra-domain protocol numbers and boundary protocols. Based on the second intent policies and the basic network topology information, it generates the optimal path, which consists of global topology paths, internal topology paths, and synthesized paths. The optimal path is then used to generate an update configuration file for the topology map through the gateway protocol of the production network. This configuration file is then used to configure the generated basic network topology information to run according to the second intent policies.

7. The intention-based production network configuration automatic generation system as described in claim 6, characterized in that, This consistency update module specifically includes: Module 1: Obtain multiple intent policies to be updated, obtain the network configuration file of the topology map, and generate basic network topology information. Divide autonomous systems by intra-domain protocol number and boundary protocol. Generate the optimal path based on the intent policies and the basic network topology information. Module 2: Based on the optimal path and network configuration file, perform connectivity verification on the first and last nodes of the optimal path. If the verification passes, no update operation is required; if the verification fails, execute Module 3. Module 3: Obtain the host node, boundary node, and internal topology path node of the starting autonomous region of the optimal path, and test the connectivity of the internal topology path node and boundary node of the starting autonomous region and the connectivity of the host node and boundary node. If all are connected, execute Module 4; otherwise, update the configuration of the nodes in the starting autonomous region to obtain the updated configuration file. Module 4: Perform connectivity checks between the boundary nodes of the starting autonomous region and the tail node of the optimal path. If they are connected, no update operation is needed; otherwise, execute Module 5. Module 5: Perform configuration file identification on the boundary nodes of the starting autonomous region, and determine whether there is configuration information of the network segment where the endpoint of the blocking path is located in the identification result. If so, update the boundary nodes of the starting autonomous region to obtain the updated configuration file, so as to perform incremental configuration on the boundary nodes of the starting autonomous region; otherwise, execute Module 6. Module 6: Obtain the host nodes, boundary nodes, and internal topology path nodes of the endpoint autonomous system of the optimal path, and test the connectivity of the internal topology path nodes and boundary nodes of the endpoint autonomous system as well as the connectivity of the host nodes and boundary nodes. If all are connected, execute Module 7; otherwise, update the configuration of the nodes in the endpoint autonomous system to obtain the updated configuration file. Module 7: Perform configuration file identification on the boundary nodes of the endpoint autonomous system, and determine whether there is configuration information of the network segment where the endpoint of the blocking path is located in the identification result. If so, update the boundary nodes of the endpoint autonomous system to obtain the updated configuration file, and perform global topology path reverse incremental configuration on the boundary nodes of the endpoint autonomous system to obtain the updated configuration file; otherwise, execute Module 8. Module 8: Perform reverse traversal text recognition on the remaining nodes in the optimal path, excluding the starting autonomous region and the ending autonomous region. If all intent network segments are matched, perform global topology path reverse incremental configuration on the remaining nodes to obtain the updated configuration file; otherwise, execute Module 9. Module 9 replaces the path in the starting autonomous region of the optimal path with the path selected in the initial configuration, and configures the global topology paths in reverse order to obtain the updated configuration file.

8. A storage medium for storing a program that executes the intention-based production network configuration auto-generation method as described in any one of claims 1 to 4.

9. A client for use in the intent-based production network configuration auto-generation system of any one of claims 5 to 7.

Images