Intent-driven data link network topology control system and method

Abstract

An intent-driven data link network topology control method and system, first, the commander inputs the task intent, and obtains the network requirements through the intent translation system, and then realizes node selection, network topology control and task coordination based on the requirements through the network topology control technology. In addition, when there is abnormal state information in the network and the intention requirements cannot be realized, the cluster head node can calculate the topology solution for the internal business plan generated by the abnormal information in the network. In this topology control system, a centralized structure is adopted in each subnet, and the cluster head controls the nodes in the domain. The cluster heads between subnets realize inter-domain communication through distributed consensus technology, taking into account the advantages and disadvantages of centralized and decentralized networking methods, and solving the security risks brought by complete centralization and the scale problem of completely decentralized networks.

Description

Technical Field

[0001] The present invention belongs to the field of communication technology, and in particular relates to an intent-driven data link network topology control system and method. Background Art

[0002] The data link network is a key system for realizing the deep connection between the command and control platform and the sensor platform, and is a core system for improving the agility and accuracy of the "detection, control, attack, and evaluation" combat loop in joint missions. In order to provide real-time situation information sharing and rapid combat capabilities, it is necessary to conduct communication demand analysis and scientific and meticulous network design planning based on specific mission plans before the mission action begins. In addition, if the configuration changes, mission requirements change, and network components are affected by external interference factors due to end devices joining or leaving the network, then how the network can operate effectively is also a question that needs to be considered. In order to effectively use the data link network, careful and rigorous data link network networking planning is required before the mission is implemented to evaluate the required information flow, data transmission rate, data encryption and connection, so as to plan an effective communication system in advance so that all nodes can effectively participate in the network.

[0003] Typical existing technical solutions are as follows:

[0004] Existing technology 1: Use distributed control to connect a group of nodes with equal status. The nodes have self-organizing functions. All nodes can have the functions of host and router. When the node acts as a host, it interacts with the user; when the node acts as a router, the node will dynamically update and maintain the routers in the network according to the corresponding routing protocol.

[0005] Prior art 2: Adopting an on-demand routing protocol based on clustering, cluster heads use routing protocols to communicate with each other, and nodes in the network are divided into several overlapping or separated cluster groups through a distributed clustering algorithm. Each cluster is managed by a cluster head or cluster head to manage all nodes in the cluster, that is, all information and behaviors of member nodes. The cluster head node discovers neighboring clusters through gateway nodes and establishes routes accordingly. Each node in the network periodically broadcasts Hello information to obtain information about nodes within its own one-hop and two-hop range and the local topology information of the network. Each node maintains a neighbor table containing link status (unidirectional link or bidirectional link) and neighbor node status. The cluster head not only maintains information about all members in the cluster, but also maintains node information of neighboring clusters (including neighboring cluster heads and gateway nodes that can reach neighboring clusters).

[0006] Through the above analysis, the existing problems and defects of the existing technology are as follows: the traditional distributed routing protocol does not have mobility management tasks, the protocol has poor scalability, and is more suitable for small and medium-sized networks. For cluster-based routing protocols, since the cluster head node is responsible for managing and maintaining the communication of the cluster nodes, when the cluster head node fails, it may affect the communication of the entire cluster, that is, the stability and reliability of the cluster head node will largely determine the stability and reliability of the entire system. At the same time, as the nodes continue to move, the maintenance and management of the cluster is much more complicated than the flat routing protocol.

[0007] The difficulty in solving the above problems and defects is: user intent is an abstract description. As the communication protocol between cluster heads increases with the scale of the network, it is difficult to manage the network topology based on intent. It is also difficult to plan the network according to user intent and realize the dynamic network communication guarantee capability oriented to user intent. Summary of the invention

[0008] In order to overcome the shortcomings of the above-mentioned prior art, the purpose of the present invention is to provide an intent-driven data link network topology control system and method, so as to reduce human-configured network errors, simplify the communication protocol between cluster heads, and ensure the normal operation of the network.

[0009] In order to achieve the above object, the technical solution adopted by the present invention is:

[0010] An intent-driven data link network topology control method comprises the following steps:

[0011] S101, the controller inputs the task intention, and after the intention is translated, the translation result is broadcast to all nodes in the network;

[0012] S102, the service terminal configures the service type and service parameters of the node according to the intent translation result;

[0013] S103, the task node divides the management domain according to the geographical location and elects the cluster head node;

[0014] S104, cluster head node initialization state, start leader election mode;

[0015] S105, with the leader of the election as the leadership node, shares the global network topology and coordinates each subdomain to jointly complete the task intention;

[0016] S106, the task is completed and the task node returns to the initialization state.

[0017] In one embodiment, in S101 , the task intention is input in the form of voice or text.

[0018] In one embodiment, the S101, the intention translation includes:

[0019] Through entity and relationship extraction, the key elements of intent are extracted. Figure 5 Save in the form of tuples;

[0020] Through parameter mapping, the Figure 5 Complete the tuple information;

[0021] What is meant Figure 5 The tuple has the following form:

[0022] <region> <attribute> <object> <operation> <result>

[0023] in <region>Indicates the selection of the node area, including location, shape, and range; <attribute>Indicates the selection of network mode, routing protocol and communication frequency band; <object>Indicates the selection of task nodes, including source nodes, destination nodes, control nodes, monitoring nodes, and computing nodes; <operation>Indicates network operations, including link opening, cancellation, and maintenance; <result>Describe the state that the user expects to achieve or avoid, including latency and bandwidth.

[0024] In one embodiment, the S102 intends to translate and set different service types for each node according to the current network status, such as attack, reconnaissance, etc., and also allocates different service parameters to each node, such as priority, transmission rate, bandwidth and other parameters, among which network status and priority may generate different parameters.

[0025] In one embodiment, the cluster head node election process in S103 is as follows:

[0026] Step 1: Receive the network requirements for intent translation and divide the nodes into different management domains according to the mission intent of the command personnel and the node location information;

[0027] In step 2, the cluster head node is elected by default as the control node in each subdomain.

[0028] In one embodiment, the leader election process in S104 is as follows:

[0029] Step 1: Initialize the state of the cluster head node and start the leader election;

[0030] Step 2: When the heartbeat time is exceeded, the cluster head node nominates itself as a candidate, votes for itself, and broadcasts voting RPC messages to other cluster head nodes to run for the leader. The voting method is first-come, first-served.

[0031] Step 3: Each cluster head node has two voting rights. If it receives a voting RPC message from another cluster head node, it will vote for the cluster head node.

[0032] Step 4: If the candidate obtains more than half of the votes, he / she becomes the leader and starts his / her term;

[0033] Step 5: After a leader appears in the cluster head node, the leader node periodically sends heartbeat messages to prevent re-election.

[0034] In one embodiment, the S105, sharing of the global network topology comprises the following steps:

[0035] Step 1: The client sends a request to update the network topology information to the leader node;

[0036] Step 2: The leader node receives the request and broadcasts it to each cluster head node;

[0037] Step 3: The cluster head node obtains the network topology information in the domain through the routing protocol and replies to the leader;

[0038] Step 4: The leader node summarizes the network topology information in each domain to obtain the total network topology and broadcasts it to other cluster head nodes;

[0039] Step 5: The cluster head node obtains the network topology information and replies to the leader node. When the leader node receives replies from more than half of the nodes, it indicates that the network topology information is updated. If the node does not receive replies from more than half of the nodes, it will continue to send requests.

[0040] In one embodiment, the S105 further includes topology management of the internal business graph, wherein the internal business graph refers to a state in which the task intent fails to maintain stability due to node or link failure in the network or when the network is disturbed, and the node or link is abnormal, resulting in the task intent failing to maintain stability;

[0041] Among them, when the leader node is abnormal, the topology management steps are as follows:

[0042] Step 1: When the leader node is damaged, the non-leader node will trigger a new round of leader election mode after the heartbeat timeout to obtain a new leader node;

[0043] In step 2, the new leader node sends a request to update the network topology.

[0044] Step 3: The cluster head node feeds back the network topology information in each domain to the new leader node, and the new leader node updates the global network topology;

[0045] Step 4: The new leader node reselects the path based on user needs;

[0046] When a non-leader node is abnormal, the topology management steps are as follows:

[0047] Step 1: When the common nodes in the subdomain are damaged or the current network configuration cannot meet the user's needs, the cluster head node does not receive a message reply from the common nodes in the subdomain within the set time, and determines that the network status is abnormal;

[0048] Step 2: The leader node sends a request to update the network topology;

[0049] Step 3: The cluster head node feeds back the network topology information in the domain and updates the network topology;

[0050] Step 4: The leader node reselects the path based on user needs.

[0051] The present invention also provides an intent-driven data link network topology control system, which is composed of a user input layer, an intent translation layer, a consensus layer, a network layer and a data layer;

[0052] The user input layer is used to receive the user's input intention and receive the current network status in real time;

[0053] The intention translation layer is composed of an intention translation module, which extracts keywords from the intention input by the user to form an intention translation module. Figure 5 Tuple, configures the business type and business parameters of the node according to the intent translation result, that is, converts the user intent into a logical strategy;

[0054] The consensus layer mainly initializes the cluster head node and starts the leader election mode. With the leader as the core, each subdomain jointly completes the task intention to realize the key data sharing between cluster heads.

[0055] The network layer specifies the authentication and propagation mechanisms for shared data;

[0056] The data layer stores and processes the shared data.

[0057] In one embodiment, the intent-driven data link network topology control system consists of two levels of nodes. The first level is a subnet system consisting of a business terminal and a cluster head node. The subnet system adopts a centralized structure. The cluster head node maintains the network topology management within the domain and updates the network topology information within the domain in real time. The second level is composed of cluster head nodes, which implement cross-domain network topology control based on a distributed consensus mechanism. By exchanging network topology information within each domain in real time, the management node can synchronously update the overall network topology information and coordinate the business terminal nodes between domains to communicate and cooperate.

[0058] The existing scheme does not introduce intent, and uses local network topology when generating solutions, so the solution is not the optimal solution. Compared with the existing technology, the present invention implements network topology control based on user intent to reduce network errors configured manually. At the same time, distributed consensus technology is used in network topology control to coordinate tasks between subdomains, simplifying the communication protocol between cluster heads. By real-time detection of internal business intentions such as network interference, the normal operation of the network is guaranteed. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is the intent-driven data link network topology control process provided by an embodiment of the present invention.

[0060] Figure 2 This is the intent translation architecture provided by the embodiment of the present invention.

[0061] Figure 3 This is the network topology sharing process provided by an embodiment of the present invention.

[0062] Figure 4 This is the intent-driven network topology control process provided by an embodiment of the present invention.

[0063] Figure 5 The embodiment of the present invention provides management of business graphs within a network.

[0064] Figure 6 It is an intent-driven data link network topology control architecture provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0066] Traditional data link networks are manually configured and maintained, which is prone to human errors. By introducing user intent, network topology management can be achieved based on intent. Data link networks need to be applicable to large-scale networks, so it is very important to plan the network based on user intent and achieve user intent-oriented and dynamic network communication assurance capabilities. Since the communication protocol between cluster heads increases with the scale of the network, and user intent is an abstract description, how to achieve network topology management based on intent is a technical challenge.

[0067] To this end, the present invention provides an intent-driven data link network topology control method and system. For ease of understanding, based on the common understanding in the art, the following additional definitions and examples are given for the terminology involved in the present invention.

[0068] The data link network system includes multiple nodes and multiple management domains, each of which contains cluster head nodes and ordinary nodes. Each cluster head node can collect network information within its own management domain. In the present invention, in order to realize network topology control, the cluster head nodes realize network topology information sharing and network topology control through a distributed consensus mechanism, and its concepts include:

[0069] Intent: refers to the use of text or other forms to describe the state to be achieved, without indicating how to do it.

[0070] Accuser: The accuser is the person who operates the system, that is, the user.

[0071] Intent translation: Convert the textual intent entered by the user into machine-executable network policies.

[0072] Business terminal: refers to various devices that execute network policies, such as vehicles, radio stations and other devices that can receive information.

[0073] Node area: The area range where task nodes are selected for networking.

[0074] Leader node: A leader node is selected from the cluster head nodes to lead other cluster head nodes to achieve network topology sharing.

[0075] Subdomain: An area managed by each cluster head.

[0076] like Figure 1 As shown, the control method of the present invention mainly includes the following steps:

[0077] S101: The commander inputs the task intention, and after the intention is translated, the translation result is broadcast to all nodes in the network.

[0078] S102: The service terminal configures the service type and service parameters of the node according to the intent translation result.

[0079] S103: The task node divides the management domain according to the geographical location and elects the cluster head node.

[0080] S104: The cluster head node is initialized and the leader election mode is started.

[0081] S105: The elected leader is used as the leadership node, which shares the global network topology and coordinates each subdomain to jointly complete the task intention.

[0082] S106: The task is completed and the task node returns to the initialization state.

[0083] In an embodiment of the present invention, Figure 2 As shown in the figure, an architecture of intent translation, that is, an implementation method, is given. The architecture consists of an intent input module, an entity and relationship extraction module, and a parameter mapping module.

[0084] The intention input module is used to implement the task intention input described in S101, and perform file error detection and correction on the intention input by the user, that is, the controller, and send the correct text to the service terminal. For example, the user can choose to input the intention in the form of voice or text.

[0085] The entity and relationship extraction module is used to extract entities and relationships, that is, to extract the key elements of user intention. Figure 5 The tuple is saved in the following format:

[0086] <region> <attribute> <object> <operation> <result>

[0087] in <region>Contains the selection of the node area, including position, shape, and range. <attribute>Includes options such as network mode, routing protocol and communication frequency band. <object>Contains the selection of task nodes, including source nodes, destination nodes, control nodes, monitoring nodes, computing nodes, etc. <operation>Indicates operations on the network, such as link opening, cancellation, maintenance, etc. <result>Describes the state that the user expects to achieve or avoid, including latency, bandwidth, etc. Figure 5 The detailed description of the tuple is shown in Table 1.

[0088] Table 1 Figure 5 Tuple basic expression

[0089] Tuple Basic Description Region position / shape / range / etc. Attribute Routing Protocols / Communication frequency band / etc. Object sourceNetwork / destNetwork / sourceNode / destNode Operation start / stop / add / delete / expansion-up / down / etc. Result bandwidth / priority / delay / packet loss / security / etc.

[0090] Parameter mapping module, used to convert Figure 5 Tuple information completion, such as latency, bandwidth, and priority.

[0091] Therefore, through the above-mentioned entity and relationship extraction and parameter mapping, the intent translation of the present invention is realized. After the intent translation, the network requirements of the controller are obtained. Based on this requirement, the present invention realizes node selection, network topology control and task coordination through network topology control technology.

[0092] In an embodiment of the present invention, in S102, the intent translation sets different service types for each node according to the current network status, such as attack, reconnaissance, etc., and also allocates different service parameters to each node, such as priority, transmission rate, bandwidth and other parameters, among which network status and priority may generate different parameters.

[0093] In the embodiment of the present invention, S103, the division of the management domain and the election process of the cluster head node can be described as follows:

[0094] Step 1: Receive the network requirements for intent translation and divide the nodes into different management domains according to the command personnel's task intent and node location information.

[0095] Among them, the network requirements for intent translation mainly include latency, bandwidth, link packet loss rate, network load, etc. The management domain refers to the scope of task node selection based on the intent requirements. The present invention is divided according to the principle of closest distance, that is, the selected node is the closest to the task scope. Task nodes are selected within the scope, and cluster heads are selected after the task nodes are selected. After the cluster head selection is completed, each cluster head and the node managed by the cluster head (or called an area) is a subdomain.

[0096] Step 2: Since the nodes in the network are divided into nodes with control functions and nodes without control functions, the election of cluster head nodes defaults to the control nodes in each subdomain.

[0097] S104: The cluster head node is initialized and the leader election mode is started.

[0098] In the embodiments of the present invention, reference Figure 3 , the election process of S104 leader can be described as:

[0099] Step 1: Initialize the state of the cluster head node, for example, number each node, increase the term by 1, and randomly assign heartbeat time to all cluster head nodes to start the leader election.

[0100] Step 2: When the heartbeat time is exceeded, the cluster head node nominates itself as a candidate, votes for itself, and broadcasts voting RPC messages to other cluster head nodes to run for the leader, where the voting method is first-come, first-served.

[0101] Step 3: Each cluster head node has two voting rights. If it receives a voting RPC message from another cluster head node, it will vote for the cluster head node.

[0102] Step 4: If the candidate obtains more than half of the votes, he or she becomes the leader and starts his or her term.

[0103] Step 5: After a leader appears in the cluster head node, the leader node periodically sends heartbeat messages to prevent re-election.

[0104] The above process provides a distributed consensus leader election mechanism in the cluster head nodes. Through this mechanism, the cluster head nodes can share the global network topology, which not only takes into account the centralized network and can obtain the optimal solution through the global network topology, but also has the advantage of distributed network anti-destruction when the leader fails.

[0105] In the embodiments of the present invention, reference Figure 4 , the sharing of S105 global network topology includes the following steps:

[0106] Step 1: The client sends a request to update the network topology information to the leader node.

[0107] Step 2: The leader node receives the request and broadcasts it to each cluster head node.

[0108] Step 3: The cluster head node selects the routing protocol between subnets according to the user's intention, obtains the network topology information within the domain through the routing protocol, and replies to the leader.

[0109] Step 4: The leader node summarizes the network topology information in each domain to obtain the total network topology and broadcasts it to other cluster head nodes. The broadcast will be sent to all nodes, but because the leader node will discard the packets it sends, it is actually broadcast to other cluster head nodes except the leader node.

[0110] Step 5: The cluster head node obtains the network topology information and replies to the leader node. When the leader node receives replies from more than half of the nodes, it indicates that the network topology information is updated. If the node does not receive replies from more than half of the nodes, it will continue to send requests.

[0111] Through the above process, the global network topology sharing is realized. Through this sharing mechanism, all cluster head nodes can obtain the global network topology and obtain the optimal solution based on the global network topology when performing link selection and path planning.

[0112] The intention-driven data link network topology takes the intention as the input of the system, and the intention includes the user input intention and the network internal business graph, wherein the network internal business graph refers to when the network state changes or there is abnormal state information in the network, and the existing configuration cannot meet the current intention requirements, the current network configuration needs to be changed to ensure the smooth implementation of the user's intention. That is, the cluster head node can calculate the topological solution for the internal business graph generated by abnormal information in the network. Based on this, in an embodiment of the present invention, S105 also includes topological management of the internal business graph, which refers to when the network has a node failure, a link failure that causes the task intent to fail, or the network is disturbed, the node or link abnormality causes the task intent to be difficult to maintain a stable state.

[0113] like Figure 5 As shown in the figure, when the leader node is abnormal, the topology management steps are as follows:

[0114] Step 1: When the leader node is damaged, the non-leader node will trigger a new round of leader election mode after the heartbeat timeout to obtain a new leader node.

[0115] In step 2, the new leader node sends a request to update the network topology.

[0116] Step 3: The cluster head node feeds back the network topology information in each domain to the new leader node, and the new leader node updates the global network topology.

[0117] Step 4: The new leader node reselects the path based on user needs.

[0118] When a non-leader node is abnormal, the topology management steps are as follows:

[0119] Step 1: When the common nodes in the subdomain are damaged or the current network configuration cannot meet the user's needs, the cluster head node does not receive a message reply from the common nodes in the subdomain within the set time, and determines that the network status is abnormal.

[0120] In step 2, the leader node sends a request to update the network topology.

[0121] Step 3: The cluster head node feeds back the network topology information in the domain and updates the network topology.

[0122] Step 4: The leader node reselects the path based on user needs, including latency, bandwidth, etc.

[0123] The above process realizes the topology management of the internal business map. Through this management, the current network status can be obtained, and the task intent can be monitored based on the current network status. When an abnormality occurs in a network node, the problem caused by the network abnormality can be discovered and solved to achieve normal operation of the network status.

[0124] The distributed consensus leader election mechanism of S104 of the present invention and the global network topology sharing of S105 together constitute the distributed consensus mechanism of the present invention.

[0125] In order to achieve the above object, the present invention provides an intention-driven data link network topology control system, the software architecture is as follows Figure 6 As shown in the figure, it is mainly composed of user input layer, intent translation layer, consensus layer, network layer and data layer.

[0126] The user input layer is used to receive the user's input intention. At the same time, in order to understand the current network status, the input layer will receive the current network status in real time.

[0127] The intent translation layer consists of intent translation modules. Its main function is to extract keywords from the intent input by the user, such as source address, destination address, task scenario, QoS requirements, etc., to form an intent translation layer. Figure 5 Tuple, obtains the translation result through parameter mapping, and configures the business type and business parameters of the node according to the intention translation result, that is, converts the user intention into a logical strategy.

[0128] The consensus layer is the key to the intent-driven data chain network topology architecture. Its main function is to initialize the state of the cluster head node and start the leader election mode. At the same time, with the leader as the core, each subdomain jointly completes the task intention, and finally realizes the key data sharing between cluster heads, such as network topology and network parameters. For example, the consensus layer can be implemented by the Raft algorithm.

[0129] The network layer specifies the verification and propagation mechanisms for shared data. This architecture broadcasts shared data to other management nodes, and the verification mechanism is to verify the correctness of the data.

[0130] The data layer stores and processes shared data, including the storage format of shared data, in order to prevent tampering and facilitate subsequent queries.

[0131] At the functional level, the data link network topology control system driven by the intention of the present invention is composed of two levels of nodes. The first level is a subnet system composed of a service terminal and a cluster head node. A centralized structure is adopted in each subnet system. The cluster head node maintains the network topology management in the domain and updates the network topology information in the domain in real time. Among them, the subnet system is divided according to the node distribution and intention. The node finds the cluster head for itself by the distance from the cluster head node or the energy of each cluster head and the number of nodes in the current cluster head node, and is included in the subnet of the cluster head. The second level is composed of cluster head nodes, which realize cross-domain network topology control based on the distributed consensus mechanism. The cluster heads between subnets realize inter-domain communication through distributed consensus technology. By exchanging the network topology information in each domain in real time, the management node can synchronously update the overall network topology information, and can coordinate the service terminal nodes between each domain to communicate and cooperate. Therefore, this mechanism can take into account the advantages and disadvantages of centralized and decentralized networking methods, and solve the security risks brought by complete centralization and the network scale problem of complete decentralization.

[0132] The above is only a specific implementation of the present invention, but the protection scope of the present invention is not limited thereto. Any modification, equivalent replacement and improvement made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention shall be covered within the protection scope of the present invention. The task node resumes the initialization state.< / result> < / operation> < / object> < / attribute> < / region> < / result> < / operation> < / object> < / attribute> < / region> < / result> < / operation> < / object> < / attribute> < / region> < / result> < / operation> < / object> < / attribute> < / region>

Claims

1. An intent-driven data link network topology control method, characterized in that: The steps include: S101, the commanding personnel inputs the task intention, and after the intention is translated, the translation result is broadcast to all nodes in the network; the intention translation includes: By extracting entities and relations, the key elements of the intent are extracted and saved in the form of an intent quintuple. Completing the information of the intent quintuple through parameter mapping; The intent quintuple is in the following form: <region> <attribute> <object> <operation> <result> in <region>Indicates the selection of the node area, including location, shape, and range; <attribute>Indicates the selection of network mode, routing protocol and communication frequency band; <object>Indicates the selection of task nodes, including source nodes, destination nodes, control nodes, monitoring nodes, and computing nodes; <operation>Indicates network operations, including link opening, cancellation, and maintenance; <result>Describe the state that the user expects to achieve or avoid, including latency and bandwidth; S102, the service terminal configures the service type and service parameters of the node according to the intent translation result; S103, the task node divides the management domain according to the geographical location and elects the cluster head node; S104, cluster head node initialization state, start leader election mode; S105, with the leader of the election as the leadership node, shares the global network topology and coordinates each subdomain to jointly complete the task intention; S106, the task is completed and the task node returns to the initialization state.

2. The intention-driven data link network topology control method according to claim 1, characterized in that: In S101, the task intention is input in the form of voice or text.

3. The intention-driven data link network topology control method according to claim 1, characterized in that: The S102, the intention translation, sets different service types for each node according to the current network status, and allocates different service parameters for each node.

4. The intention-driven data link network topology control method according to claim 1, characterized in that: The process of selecting the cluster head node in S103 is as follows: Step 1: Receive the network requirements for intent translation and divide the nodes into different management domains according to the mission intent of the command personnel and the node location information; In step 2, the cluster head node is elected by default as the control node in each subdomain.

5. The intention-driven data link network topology control method according to claim 1, characterized in that: In S104, the leader election process is as follows: Step 1: Initialize the state of the cluster head node and start the leader election; Step 2: When the heartbeat time is exceeded, the cluster head node nominates itself as a candidate, votes for itself, and broadcasts voting RPC messages to other cluster head nodes to run for the leader. The voting method is first-come, first-served. Step 3: Each cluster head node has two voting rights. If it receives a voting RPC message from another cluster head node, it will vote for the cluster head node. Step 4: If the candidate obtains more than half of the votes, he / she becomes the leader and starts his / her term; Step 5: After a leader appears in the cluster head node, the leader node periodically sends heartbeat messages to prevent re-election.

6. The intention-driven data link network topology control method according to claim 1, characterized in that: The above-mentioned S105, sharing of the global network topology comprises the following steps: Step 1: The client sends a request to update the network topology information to the leader node; Step 2: The leader node receives the request and broadcasts it to each cluster head node; Step 3: The cluster head node obtains the network topology information in the domain through the routing protocol and replies to the leader; Step 4: The leader node summarizes the network topology information in each domain to obtain the total network topology and broadcasts it to other cluster head nodes; Step 5: The cluster head node obtains the network topology information and replies to the leader node. When the leader node receives replies from more than half of the nodes, it indicates that the network topology information is updated. If the node does not receive replies from more than half of the nodes, it will continue to send requests.

7. The intention-driven data link network topology control method according to claim 1, characterized in that: The S105 also includes topology management of the internal business graph, where the internal business graph refers to a state where the task intent fails due to node or link failure in the network or the network is disturbed, and the node or link abnormality makes it difficult for the task intent to maintain a stable state; Among them, when the leader node is abnormal, the topology management steps are as follows: Step 1: When the leader node is damaged, the non-leader node will trigger a new round of leader election mode after the heartbeat timeout to obtain a new leader node; Step 2: The new leader node sends a request to update the network topology; Step 3: The cluster head node feeds back the network topology information in each domain to the new leader node, and the new leader node updates the global network topology; Step 4: The new leader node reselects the path based on user needs; When a non-leader node is abnormal, the topology management steps are as follows: Step 1: When the common nodes in the subdomain are damaged or the current network configuration cannot meet the user's needs, the cluster head node does not receive a message reply from the common nodes in the subdomain within the set time, and determines that the network status is abnormal; Step 2: The leader node sends a request to update the network topology; Step 3: The cluster head node feeds back the network topology information in the domain and updates the network topology; Step 4: The leader node reselects the path based on user needs.

8. An intention-driven data link network topology control system, characterized in that: It consists of user input layer, intent translation layer, consensus layer, network layer and data layer; The user input layer is used to receive the user's input intention and receive the current network status in real time; The intent translation layer is composed of an intent translation module, which extracts keywords from the intent input by the user to form an intent quintuple, and configures the business type and business parameters of the node according to the intent translation result, that is, converts the user intent into a logical strategy; The intent translation includes: By extracting entities and relations, the key elements of the intent are extracted and saved in the form of an intent quintuple. Completing the information of the intent quintuple through parameter mapping; The intent quintuple is in the following form: <region> <attribute> <object> <operation> <result> in <region>Indicates the selection of the node area, including location, shape, and range; <attribute>Indicates the selection of network mode, routing protocol and communication frequency band; <object>Indicates the selection of task nodes, including source nodes, destination nodes, control nodes, monitoring nodes, and computing nodes; <operation>Indicates network operations, including link opening, cancellation, and maintenance; <result>Describe the state that the user expects to achieve or avoid, including latency and bandwidth; The consensus layer is the initialization state of the cluster head node, and the leader election mode is turned on. At the same time, with the leader as the core, each subdomain jointly completes the task intention, which is used to realize the key data sharing between cluster heads; The network layer specifies the authentication and propagation mechanisms for shared data; The data layer stores and processes the shared data.

9. The intention-driven data link network topology control system according to claim 8, characterized in that: The intent-driven data link network topology control system consists of two levels of nodes. The first level is a subnet system consisting of business terminals and a cluster head node. The subnet system adopts a centralized structure. The cluster head node maintains the network topology management within the domain and updates the network topology information within the domain in real time. The second level is composed of cluster head nodes, which realize cross-domain network topology control based on a distributed consensus mechanism. By exchanging network topology information within each domain in real time, the management node can synchronously update the overall network topology information and coordinate the business terminal nodes between domains to communicate and cooperate. < / result> < / operation> < / object> < / attribute> < / region> < / result> < / operation> < / object> < / attribute> < / region> < / result> < / operation> < / object> < / attribute> < / region> < / result> < / operation> < / object> < / attribute> < / region>

Images