An Intent-Driven Intelligent Routing and Forwarding System and Method

Abstract

This invention relates to an intent-driven intelligent routing and forwarding system and method, belonging to the field of computer networks. The system of this invention configures an intent-driven intelligent agent encapsulation module on the admission control device and a parsing execution engine on the computational forwarding node. The main forwarding nodes in the network architecture are divided into ordinary forwarding nodes and computational forwarding nodes. The computational forwarding nodes provide the encapsulation and computational environment for network packet intelligent agents in the network data flow, while ordinary nodes provide regular packet forwarding capabilities. In the method of this invention, after the network data flow encapsulated with intent-driven intelligent agents arrives at the computational forwarding node, it determines whether to execute the parsing execution engine based on the current network state, enabling the network data flow to proactively and intelligently control its own forwarding behavior. This invention encapsulates user intents, such as network requirements and routing and forwarding control code, into network packets, allowing them to be executed in routing nodes with computational capabilities, making the IP network more proactive and intelligent during the forwarding process.

Description

Technical Field

[0001] This invention relates to the field of computer network technology, and specifically to an intent-driven intelligent routing and forwarding system and method. Background Technology

[0002] Since its birth in the 1970s, IP networks have developed and become widespread. However, with the emergence of a large number of new services and applications, the simple forwarding of network packets in the network has gradually failed to meet the diverse needs of users.

[0003] Therefore, how to inject diverse user needs into the network and achieve network programmability has become an important direction for network development. This led to the emergence of the SDN (Software-Defined Networking) framework. However, this framework primarily separates the control plane and data plane to achieve network programmability, which cannot meet users' needs for direct network programming. Therefore, user intent-driven networks have been developed. Intent-driven networks are programmable and customizable networks with capabilities such as deep analysis of user intent, global awareness of network status, and real-time network configuration. The network can automatically transform, verify, deploy, configure, and optimize according to user intent to achieve the target network state and can automatically resolve network faults to ensure network reliability.

[0004] Existing intent-driven networks typically involve five steps: intent expression / collection; translation and verification: the system obtains higher-level business policies from end users and translates them into necessary network configurations, generating and verifying the final design and configuration to ensure correctness; automated implementation: the system can configure appropriate network changes on existing network infrastructure, accomplished through network automation or network orchestration; network status awareness: the system provides real-time network status to systems under its management and control, and is protocol and transport agnostic; assurance and automatic tuning: the system continuously verifies that business intents are achieved and can take corrective measures when required intents cannot be achieved.

[0005] The current implementation of intent-driven networks mainly involves integration with SDN. Upper-layer business applications orchestrate intents and then pass them to the controller via the northbound interface. The controller generates corresponding routing policies, which are then distributed to each forwarding node via the southbound interface. Network packets entering the SDN network are forwarded according to the given fixed policies. This architecture involves numerous interfaces, resulting in a long time consumption for the entire lifecycle of deploying user intents to the network. More importantly, it cannot fully, timely, and dynamically express the needs and intents of network users. Summary of the Invention

[0006] To address the problems existing in the current intent-driven networks, this invention proposes an intent-driven intelligent routing and forwarding system and method, discloses a message structure that supports intent-driven agent encapsulation, designs and implements an intent-driven agent routing and forwarding system, and provides an active routing method for network data flow to make decisions along the path in computing nodes.

[0007] In the network architecture used in this invention, the forwarding nodes are divided into ordinary forwarding nodes and computational forwarding nodes. Computational forwarding nodes can provide the encapsulation and computational environment for network packet agents in the network data stream, while ordinary forwarding nodes provide conventional packet forwarding capabilities.

[0008] This invention discloses a message structure that supports intent-driven agent encapsulation. This message structure combines an agent identifier (ID), user network transmission performance requirements, and routing control code into an intent-driven agent, which is then encapsulated in the IP header of a network data stream. The user network transmission performance requirements include one or more of network latency, bandwidth, and jitter.

[0009] This invention provides an intent-driven intelligent routing and forwarding system, comprising:

[0010] An intent-driven agent encapsulation module is configured on the admission control device. The intent-driven agent encapsulation module combines the network performance indicators and routing control codes obtained by the admission control device into an intent-driven agent and encapsulates it into the IP packet header of the network data stream.

[0011] A parsing execution engine is configured on the computation and forwarding node. The parsing execution engine includes an extraction module, a calling module, and an execution module. After receiving the network data stream, the computation and forwarding node determines whether the current network status meets the user's network transmission performance requirements. If so, it directly forwards the packets; otherwise, it starts the parsing execution engine. The extraction module extracts the intent-driven agent encapsulated in the network data stream to obtain the network transmission performance requirements and routing control code. The calling module calls the local network status information as input when executing the routing control code. The execution module executes the routing control code in the intent-driven agent based on the obtained network transmission performance requirements and local network status information to obtain a better forwarding path.

[0012] After the receiving control device encapsulates the network data stream with an intent-driven agent, it forwards it to the computing head node. The computing head node parses the intent-driven agent of the network data stream, calculates the current forwarding path, and adds it to the intent-driven agent of the network data stream. The computing head node then forwards the network data stream according to the current forwarding path. When the network data stream reaches the computing forwarding node on the current forwarding path, the computing forwarding node performs network status monitoring to determine whether to start the parsing execution engine. When the parsing execution engine is started, it executes the routing control code to obtain a better forwarding path that meets the user's network transmission performance requirements. It then replaces the current forwarding path with this better forwarding path and adds it to the intent-driven agent of the network data stream.

[0013] This invention provides an intent-driven intelligent routing and forwarding method, comprising the following steps:

[0014] (1) The user’s network transmission performance requirements and the routing control code that guides the network data flow to forward itself are combined into an intent-driven intelligent agent and encapsulated into the IP packet header of the network data flow to form a packet structure that supports intent-driven operation.

[0015] (2) Configure a parsing execution engine on each computing forwarding node. When the network data stream arrives at the computing forwarding node, the computing forwarding node judges the current network status. If the network transmission performance requirements cannot be met, the parsing execution engine is started. The parsing execution engine performs the parsing and execution of the intent-driven intelligent agent. If the current network status meets the network transmission performance requirements, the message is forwarded directly.

[0016] The parsing and execution engine parses and executes the intent-driven agent, including: extracting the intent-driven agent encapsulated in the network data stream, obtaining network transmission performance requirements and routing control code; calling local network status information; and executing the routing control code in the intent-driven agent based on the extracted network transmission performance requirements and the obtained local network status information to obtain a better forwarding path.

[0017] In step (2), the encapsulated network data stream of the intent-driven agent is forwarded by the admission control device to the computing head node. The computing head node parses the intent-driven agent of the network data stream, calculates the current forwarding path and adds it to the intent-driven agent of the network data stream. The computing head node forwards the network data stream according to the current forwarding path. When the network data stream reaches the computing forwarding node in the middle of the current forwarding path, the computing forwarding node performs network status monitoring to determine whether to start the parsing execution engine. When the parsing execution engine is started, the routing control code is executed to obtain a better forwarding path that meets the user's network transmission performance requirements. This forwarding path is then replaced with the current forwarding path and added to the intent-driven agent of the network data stream.

[0018] The advantages and positive effects of this invention are as follows: Existing intent-driven network implementations require a process involving upper-layer application intent orchestration → northbound interface → controller → southbound interface → policy distribution to control forwarding nodes. This entire process is time-consuming and cannot fully, timely, or dynamically express the needs and intents of network users. Compared to existing technologies, the system and method of this invention encapsulate the user's intent in the network packet header when data enters the network. Specifically, it encapsulates the user's intent, such as network requirements and routing forwarding control codes, into network packets, allowing them to be executed in routing nodes with computing capabilities. This makes the IP network more proactive and intelligent during forwarding. Compared to existing technologies, the system and method of this invention can carry more granular network user intents, utilizes numerous interaction interfaces, offers better real-time performance, and places control policies closer to the user end, making it easier to realize user network intents. Attached Figure Description

[0019] Figure 1 This is a diagram illustrating the message structure and encapsulation structure of the intent-driven intelligent agent of this invention.

[0020] Figure 2 This is a structural diagram of the parsing execution engine of the present invention;

[0021] Figure 3 This is an architecture diagram of the intent-driven intelligent routing and forwarding system of the present invention;

[0022] Figure 4 This is a flowchart of the network data flow path decision-making process of the present invention;

[0023] Figure 5 This is a flowchart of the network data flow configuration at the head node according to the present invention;

[0024] Figure 6 This is a flowchart of the network state detection and decision-making process of the present invention regarding whether to enable the agent parsing and execution engine;

[0025] Figure 7 This is a flowchart of the active computation routing algorithm of the present invention. Detailed Implementation

[0026] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0027] The intent-driven intelligent routing and forwarding system and method of the present invention encapsulates user intents, such as network requirements and routing and forwarding control codes, into network packets, allowing them to be executed in routing nodes with computing capabilities, thereby making the IP network more proactive and intelligent in the forwarding process.

[0028] This invention encapsulates an intent-driven intelligent agent—comprising user network transmission performance requirements and routing control code that guides the forwarding of network data flows—into the IP packet header, thereby forming an intent-driven packet structure. For example... Figure 1 As shown, the message structure of an intent-driven agent includes agent identification information (ID), network performance metrics, and routing control codes. Here, network performance metrics refer to the user's network transmission performance requirements. Network performance metrics include data such as network latency, bandwidth, and jitter. User network transmission performance requirements must include at least one network performance metric. Routing control codes are used to guide the forwarding of network data flows.

[0029] In this invention, the main forwarding nodes in the network architecture are divided into ordinary forwarding nodes and computational forwarding nodes. Computational forwarding nodes provide the encapsulation and computational environment for network packet agents in the network data stream, while ordinary forwarding nodes provide regular packet forwarding capabilities. Each computational forwarding node in this invention's system has the ability to parse and execute intent-driven agents. Each computational forwarding node is configured with a parsing and execution engine. After the network data stream arrives at the designated node, if the agent parsing and execution engine is enabled, the intent-driven agent is parsed and executed. Figure 2 As shown, the parsing execution engine for the computing forwarding node configuration of the present invention includes an extraction module, a calling module, and an execution module.

[0030] The extraction module extracts the intent-driven agents encapsulated in the network data stream, thereby obtaining network transmission performance requirements and routing control code for subsequent execution.

[0031] The calling module is used to retrieve local network status information, which serves as input for the subsequent execution of routing control code.

[0032] The execution module executes the routing control code in the intent-driven agent based on the network transmission performance requirements carried by the network data stream and the obtained local network status information, in order to obtain a better forwarding path.

[0033] like Figure 3 As shown, in the intent-driven intelligent routing and forwarding system of the present invention, an intent-driven intelligent agent encapsulation module is configured on the admission control device. After the network data stream generated by the terminal device enters the network, the admission control device obtains information such as the network requirements (latency, bandwidth, etc.) and routing control codes of the network data stream. The intent-driven intelligent agent encapsulation module assembles this information into an intent-driven intelligent agent and encapsulates it in the IP packet header. A parsing execution engine is configured on the computing and forwarding node. When the network data stream arrives at the computing and forwarding node, the forwarding strategy of the network packet routing intelligent agent is executed, enabling the network data stream to actively and intelligently control its own forwarding behavior.

[0034] The routing method proposed in this invention, which involves path-by-path decision-making for network data flows in nodes with computing capabilities, first involves the verified network data flow carrying relevant fine-grained network requirements (bandwidth, latency, etc.) entering the head node. The head node then allocates a basic forwarding path and adds an intent-driven agent based on these requirements. Next, the intermediate node's agent parsing and execution engine determines whether to activate based on network and data flow states. Once the network data flow reaches the designated intermediate node, if the agent parsing and execution engine is activated, the intent-driven agent is parsed and executed to obtain a better forwarding path that meets the user's network requirements. The overall process is as follows: Figure 4 As shown, the details are explained below.

[0035] (1) As Figure 5 As shown, after entering the network, the network data stream obtains information such as network requirements (latency, bandwidth, etc.) and routing control codes from the receiving control device, forms an intent-driven agent, and encapsulates the intent-driven agent into the network data stream.

[0036] (2) Calculation head node 1 drives the agent to perform calculations and obtain the forwarding path result based on the intent carried by the data stream. For example, if the network requirement information carried is {latency ≤ 100MS, bandwidth ≥ 10M / S}, the forwarding path calculated by calculation head node 1 is: calculation head node 1 → calculation node 2 → calculation node 5 → calculation node 6. The network data stream will be forwarded according to the forwarding path, and the next hop forwarded by calculation head node 1 is calculation node 2.

[0037] (3) The network data flow is forwarded according to the calculated routing path. Upon reaching computing node 2, network status detection is performed, and a decision is made on whether to enable the agent parsing and execution engine. For example... Figure 6 As shown, the node first samples and checks the local network status, including link bandwidth, latency, and packet loss rate. If these metrics fail to meet the business QoS requirements when the network data flow arrives at the node, the parsing and execution engine is triggered. This engine executes the intent-driven agent encapsulated in the network data flow and recalculates the forwarding path of the network data traffic. If these metrics meet the business QoS requirements, the data continues to be forwarded according to the forwarding path configured by the head node, directly forwarding it to the corresponding next hop.

[0038] (4) When the network data stream arrives at computing node 2 and computing node 2 detects that network metrics such as link bandwidth, delay, and packet loss rate cannot meet the service QoS requirements of the network data stream, it will trigger computing node 2 to execute the intent-driven agent encapsulated in the network data stream, which can obtain a path that meets the network requirements.

[0039] like Figure 7As shown, the implementation process of the active computation routing algorithm set in the intent-driven intelligent agent of the present invention includes: First, obtaining the network topology structure, then traversing the links in the network topology, constraining the links in the obtained network topology with network intent, that is, finding the node transmission links that meet the network requirement information as in (2) above, pruning the links that do not meet the network intent, and calculating the optimal weight path among the remaining links that meet the network requirements. In the embodiments of the present invention, the link weight can be latency, bandwidth, packet loss rate, etc. If multiple optimal weight paths are calculated, a load balancing algorithm is used to select a path as the current forwarding path of the network data flow.

[0040] Except for the technical features described in the specification, all other technologies are known to those skilled in the art. Descriptions of well-known components and technologies are omitted in this invention to avoid redundancy and unnecessary limitation. The embodiments described above do not represent all embodiments consistent with this application. Various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of this invention are still within the protection scope of this invention.

Claims

1. An intent-driven intelligent routing and forwarding system, characterized in that, include: An intent-driven agent encapsulation module is configured on the admission control device. The intent-driven agent encapsulation module combines the user network transmission performance requirements and routing control codes obtained by the admission control device into an intent-driven agent and encapsulates it into the IP packet header of the network data stream. A parsing and execution engine is configured on the computing forwarding node. The parsing and execution engine includes an extraction module, a calling module, and an execution module. The computing forwarding node can provide an encapsulation and computing environment for network packet agents in the network data stream. After receiving the network data stream, the computing forwarding node determines whether the current network status meets the user's network transmission performance requirements. If so, it directly forwards the packet; otherwise, it starts the parsing and execution engine. The extraction module extracts the intent-driven intelligent agent encapsulated in the network data stream to obtain the user's network transmission performance requirements and routing control code. The calling module retrieves local network status information as input for the execution of routing control code; the execution module, based on the obtained user network transmission performance requirements and local network status information, executes the routing control code in the intent-driven agent to obtain a better forwarding path.

2. The system according to claim 1, characterized in that, The intent-driven agent generated by the intent-driven agent encapsulation module includes an agent identifier ID, user network transmission performance requirements, and routing control code; wherein, the user network transmission performance requirements include at least one or more of network latency, bandwidth, and jitter.

3. The system according to claim 1 or 2, characterized in that, The aforementioned admission control device encapsulates the network data stream with an intent-driven intelligent agent and forwards it to the computing head node. The computing head node parses the intent-driven intelligent agent of the network data stream, calculates the current forwarding path, and adds it to the intent-driven intelligent agent of the network data stream. The computing head node then forwards the network data stream according to the current forwarding path. When the network data stream arrives at the computation forwarding node on the current forwarding path, the computation forwarding node performs network status monitoring to determine whether to start the parsing execution engine. When the parsing execution engine is started, the routing control code is executed to obtain a better forwarding path that meets the user's network transmission performance requirements. This forwarding path is then replaced with the current forwarding path and added to the intent-driven agent of the network data stream.

4. An intent-driven intelligent routing and forwarding method, characterized in that, Includes the following steps: (1) The user’s network transmission performance requirements and the routing control code that guides the network data flow to forward itself are combined into an intent-driven intelligent agent and encapsulated into the IP packet header of the network data flow to form a packet structure that supports intent-driven operation. (2) Configure a parsing execution engine on each computing forwarding node. When the network data stream arrives at the computing forwarding node, the computing forwarding node judges the current network status. If the network transmission performance requirements cannot be met, the parsing execution engine is started. The parsing execution engine performs the parsing and execution of the intent-driven intelligent agent. If the current network status meets the network transmission performance requirements, the message is forwarded directly. The computing forwarding node can provide the encapsulation and computing environment for network packet agents in the network data stream; The parsing and execution engine parses and executes the intent-driven agent, including: extracting the intent-driven agent encapsulated in the network data stream, obtaining network transmission performance requirements and routing control code; calling local network status information; and executing the routing control code in the intent-driven agent based on the extracted network transmission performance requirements and the obtained local network status information to obtain a better forwarding path.

5. The method according to claim 4, characterized in that, In step (1), the encapsulated intent-driven agent includes an agent identifier ID, user network transmission performance requirements, and routing control code; wherein, the user network transmission performance requirements include at least one or more of network latency, bandwidth, and jitter.

6. The method according to claim 4 or 5, characterized in that, In step (2), the encapsulated network data stream of the intent-driven agent is forwarded by the admission control device to the computing head node. The computing head node parses the intent-driven agent of the network data stream, calculates the current forwarding path and adds it to the intent-driven agent of the network data stream. The computing head node forwards the network data stream according to the current forwarding path. When the network data stream reaches the computation forwarding node in the middle of the current forwarding path, the computation forwarding node performs network status monitoring to determine whether to start the parsing execution engine. When the parsing execution engine is started, the routing control code is executed to obtain a better forwarding path that meets the user's network transmission performance requirements. This forwarding path is then replaced with the current forwarding path and added to the intent-driven agent of the network data stream.

7. The method according to claim 4 or 5, characterized in that, In step (1), the routing control code implementation process includes: first, obtaining the network topology, then traversing the links in the network topology, constraining the network intent of the links in the obtained network topology, calculating the optimal weight path for the links that meet the user's network transmission performance requirements, with the link weight being the set network performance index, and when multiple optimal weight paths are calculated, using a load balancing algorithm to select a path as the current forwarding path for the network data flow.

Images