]> Huawei

101 Software Avenue, Yuhua District Jiangsu 210012 China maqiufang1@huawei.com

Huawei

101 Software Avenue, Yuhua District Nanjing, Jiangsu 210012 China bill.wu@huawei.com

Agentic AI YANG intent The industry today is actively exploring the application of agentic AI to autonomous network operations. However, there is little attention given to the impacts that the introduction of agentic AI may impose on YANG modeling, and how YANG can be better leveraged to support AI-driven autonomous operations. Based on some end-to-end intent translation workflow analysis that engage YANG models across service, network, and device layers, this document identifies some critical gaps when AI agents generate and operate on YANG data. The purpose of this document is not to substantially redesign YANG or define a new data model language. Instead, it aims to facilitate discussion on how existing YANG ecosystems can be better leveraged in the context of agentic AI, and how complementary mechanisms can bridge these gaps while preserving YANG’s interoperability. About This Document The latest revision of this draft can be found at . Status information for this document may be found at . Source for this draft and an issue tracker can be found at .

Introduction RFC 8969 establishes a framework for automating service and network management using YANG . Combined with network management protocols such as NETCONF and RESTCONF , YANG delivers deterministic and interoperable capabilities for configuration manipulation and state retrieval across service, network, and device layers. Recently, Large Language Models (LLMs) and AI Agents have emerged as new enablers for network automation. Multiple IETF activities are ongoing in this area, proposals include but are not limited to:

• The network digital twin and agentic AI based architecture for AI driven network operations described in , which deploys hybrid agent systems within the network autonomous domain, and allows agents to invoke network operational function modules and collaborate to solve non-trivial tasks.
• The Network Management Agent (NMA) concept presented in , which proposes that network management agents can be deployed at both orchestrator and controller layers, with A2A, A2C, and A2N interfaces enabling communication among agents to agents, controllers, and network elements.
• The applicability of MCP for Network Management , which explores how the Model Context Protocol can refactor network management operations and network capabilities as LLM-callable tools.
• The applicability of A2A to Network Management , which outlines how the Agent to Agent protocol can be leveraged to develop various rich AI driven network applications, realize intent-based network management automation in the multi-vendor heterogeneous network environment.

While the industry is actively exploring the application of agentic AI to autonomous network, little attention has been paid to how YANG modeling, as the very foundation of model-driven network automation, would be impacted, and if there is any critical gaps that remain unrecognized/unsolved at the intersection of probabilistic AI agents and deterministic structured YANG. This document analyzes two typical network operation workflows driven by hierarchical AI Agents: service provisioning/optimization and network troubleshooting. It identifies key gaps arising from the adoption of YANG data models in these scenarios. The purpose of this document is not to substantially redesign YANG or define a brand-new data modeling language. Rather, it seeks to adapt YANG for agentic AI. By identifying fundamental gaps, this document aims to facilitate community discussion on how to bridge the gap between them.

Conventions and Definitions The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 when, and only when, they appear in all capitals, as shown here. The document uses the following terms defined in and :

• data model
• service model
• network model
• device model
• network domain

The document uses the following terms defined in :

• Agentic AI

The following definitions are used throughout this document:

AI Agent:

An autonomous software system driven by an artificial intelligence reasoning engine (typically a large language model) that integrates a set of resources such as data and tools to perceive its operational environment, dynamically plan and execute decisions, and invoke actions to achieve specific goals. Unlike traditional deterministic or rule-based autonomous systems, an AI Agent exhibits non-deterministic reasoning, maintains runtime and historical state, and possesses the capability to dynamically decompose abstract high-level intents into a sequence of operational actions.

non-determinism:

The property whereby an AI Agent may produce different outputs (e.g., generate different configuration or give different network diagnostic conclusions) across multiple invocations given the identical input, due to the probabilistic nature of the underlying model, historical context drift, or non-producible reasoning paths.

A Reference Architecture and Typical Workflows This section uses a simplified layered agentic AI reference architecture compliant with and describes two typical workflows: the network service provisioning/optimization and the fault troubleshooting. The workflows cover scenarios from configuration generation to YANG modelled data consumption, which are essential to expose and derive the gaps in . Note that these flows are not the only possible ones. The intent flows shown here are just examples, i.e., typical workflows that illustrate how YANG models at the service, network, and device layers can be engaged in an agent-integrated network automation context.

Layered AI Agent Deployment Architecture This document targets a hierarchical agent deployment architecture as specified in . It is a simplified view that hides task agents, function modules or toolsets that are available for agents to invoke. The architecture may also use memory management (e.g., short-term and long-term memory) to maintain conversational continuity and enable contextual search. Refer to for the complete architecture design.

A simplified Hierarchical AI Agent Architecture

Workflow Example 1: Intent-Driven Service Provisioning/Optimization A complete end-to-end workflow focusing on AI agents generating configuration from high-level intents is defined as follows.

Step 1:

A network operator submits an end-to-end service request expressed in natural language including explicit service requirements. The intent may imply some constraints such as not affecting existing network services.

Step 2:

The orchestrator AI agent parses the natural language intent, maps it into the corresponding service model and parameters for user confirmation, and then decomposes it into structured configuration using network model.

Step 3:

YANG-structured data are encapsulated as message payload, and transmitted to network AI agents along with other context metadata via agent communication protocol like Agent-to-Agent (A2A).

Step 4:

The network AI agent converts received tasks into structured device model layer operations. The configuration is handed off to the network controller's Southbound Interface (SBI) and delivered via YANG-driven protocols such as NETCONF and RESTCONF. Validation checks may be performed before the configuration is delivered to network devices.

Step 5:

Network devices receive incoming configuration, execute changes, and report execution response.

Workflow Example 2: Network Troubleshooting

Step 1:

A network operator submits a high-level operational intent, requesting the network AI agent to continuously monitor the network, perceive emerging anomalies, perform active diagnosis, and conduct autonomous repair while reporting root causes and handling results.

Step 2:

Network AI agent grasps a holistic view of the network operational state across multiple network devices and YANG modules. Instead of flooding the Network AI agent with massive, raw streaming telemetry data which would instantly overwhelm the LLM's context window and incur prohibitive token costs, agent may leverage existing network data analytics components or tools to be aware of real-time network state. Network AI agent could detect network anomalies swiftly, which enables the prompt identification of potential issues before they escalate into major faults.

Step 3:

Once an issue or fault is identified, network AI agent diagnoses the exact cause and generate targeted repair solutions. These solutions may involve making configuration changes on relevant network devices, for example, when a failure is caused by misconfiguration.

Step 4:

The network AI agent autonomously executes the formulated repair actions by delivering corresponding YANG configuration changes to devices, after invoking the network digital twin to simulate the proposed repair solution successfully.

Gap Analysis

Gap 1: Expressiveness Limits In real-world network operations, operator intents often contain expectations and constraints, relaxation preferences, retry strategies, or fallback rules. Nevertheless, YANG is designed solely focuses on defining data structures and formats, syntax and semantic rules and static data dependencies. It lacks the ability to express dynamic and flexible operational logic, which creates a hard boundary for carrying intent and limits the decision space of multi-agent system. This Gap hinders step 3 of the service provisioning/optimization workflow (). To adapt to static YANG structures, orchestrator AI agents have to prune and simplify the original intent. Implicit operational requirements embedded in original intents, such as service continuity, high availability and security compliance, are likely lost during the transformation process. This premature semantic pruning deprives downstream network AI agents of the flexibility to make dynamic decisions, perform graceful degradation or apply fault-tolerant adjustments when network runtime status changes.

Gap 2: Token Efficiency and Structural Encoding Overhead Existing YANG serialization formats (e.g., XML and JSON ) are heavily optimized for either human readability or strict software parsing. When adapted to agentic AI environments, these formats introduce significant redundancy, and thus consumes substantial LLM context window capacity and incurring significant token costs. This gap is explicitly exposed during step 3 of the workflow in and step 2 of the workflow in , where verbose encoding adds latency and cost without contributing semantic value. Highly compressed binary alternatives like CBOR mitigate bandwidth issues but introduce a fundamental incompatibility with LLM processing pipelines: LLMs operate exclusively on text token streams and cannot consume raw binary payloads directly. Currently, there is a clear gap for a token-efficient, semantic-preserving encoding format specifically designed for AI-agent consumption.

Gap 3: Non-determinism Task Handling when tasked with ambiguous instructions, missing/unreliable data, deep reasoning, or executing multi-step decisions, the agent's behavior could inevitably introduce randomness and unpredictability. These unpredictable variations in Agent and LLM outputs can significantly impact the reliability and consistency of network operations. For the same high-level intent, an agent may generate different YANG instance data due to historical context drift or the probabilistic nature of the underlying large language models (steps 2 and 4 of the service provisioning/optimization workflow in ). Likewise, when analyzing network anomalies or incidents, multiple distinct diagnostic trajectories and explanations may emerge even for the exact same observed network operational state (steps 3 and 4 of the network troubleshooting workflow in ).

Possible Ways Forward This section explores several operational directions to try to bridge the gaps identified in and improve the AI-readiness of existing YANG ecosystem.

YANG-driven Operations as AI-Invocable Tools A promising direction for partially addressing gap in is to refactor YANG‑based network operations into AI‑invocable tools using the Model Context Protocol (MCP). Instead of requiring an AI agent to generate raw YANG instance data directly, which is prone to hallucination or exhibits non-determined behavior, the agent invokes an MCP tool with structured parameters. By shielding agents from low-level syntax and model complexity, this abstraction layer reduces the agent’s burden of understanding complex YANG schemas and mitigates the risk of generating invalid configuration data. The applicability of MCP to network management is discussed in . Industry open-source exploration is emerging towards this direction, e.g., gNMIBuddy provides a toolkit to wrap gNMI and OpenConfig YANG data models based network operations, designed primarily for LLMs with MCP integration.

Semantic Enrichment via Knowledge Graphs and Semantic Metadata Gaps identified in can be partially addressed by adding a semantic layer on top of YANG data models. Knowledge graphs (KGs) provide a machine‑readable representation of network knowledge, enabling AI agents to better understand the semantics, relationships, and constraints embedded in the network. IETF work in this area includes and , which correlate data from different network planes, e.g., management, control, and data planes and present a holistic view of network status. When a user expresses a high-level intent such as "check why the VPN tunnel is down", a KG‑enhanced agent can query the YANG-based knowledge graph to understand the relationships between relevant services and metrics. Furthermore, KGs can explicitly model concepts such as "preferred vs. optional", temporal KGs can model concepts such as "transient vs. persistent failure", which are currently absent from YANG, partially addressing . There is also other work in NMOP focusing on semantic enrichment that could serve as a foundation layer for agentic AI driven network anomaly detection. A set of ongoing drafts include:

• provides a reference architecture for knowledge based service disruption detection;
• defines an experiment for managing the lifecycle process of a network anomaly detection system, spanning across detection, validation, and refinement.
• describes common network symptom semantics across different network planes.

An AI agent gains a structured and machine-understandable vocabulary for describing, querying, and reasoning about network anomaly. The standardized anomaly semantic metadata provides the missing semantic hooks, and contributes to bridging the gap in .

Declarative Intent/Policy Overlay To bridge the semantic loss challenge specified in without breaking the backward compatibility with existing network infrastructure, this section outlines a complementary approach that has emerged from operational deployments, that is the use of a declarative intent and policy overlay. This approach does not modify YANG or propose new YANG statements. Instead, it introduces a logical layer above YANG that serves as the bridge between human-readable intent and machine-executable configuration. Take a natural language based service provisioning intent for example, Instead of forcing AI agents to directly translate loose, high-level operator intent into rigid, low-level structural service YANG leaf nodes, the agent could first compile the intent into a declarative intent/policy overlay. This overlay explicitly encapsulates invariants, business constraints, and dynamic behavioral policies that existing service models cannot naturally represent. The overlay is co-delivered alongside the generated concrete YANG instances to the lower-layer agents.

Token-Efficient YANG Serialization for Agent Contexts As noted in , the verbosity of YANG instance data, particularly in JSON/XML encodings, poses a practical barrier to cost-effective agent operations. One possible way forward is the development of a token-efficient YANG serialization mechanism tailored to agentic consumption.

Security Considerations The integration of agentic AI with YANG-based network automation introduces new attack surfaces and operational risks that differ from traditional deterministic network management. For example, AI-generated configuration risks are introduced by agent autonomous decision-making. Unlike manually crafted or scripted configurations, LLM-generated configuration edits may contain unintended misconfiguration or semantic deviations. Such errors can lead to service outages, policy violations, or enlarged attack surfaces. The human-in-the-loop mechanism and network digital twin simulation and validation could mitigate this concern by blocking high-risk unvalidated changes before configuration is committed. Furthermore, tool invocation and credential exposure risks arise when AI agents are granted access to network devices via MCP servers. Open-source MCP implementations enable LLM clients to execute read/write operations over managed devices. Unauthorized or malicious prompt injection may trigger abusive command execution, configuration tampering, or information leakage. Token-based authentication, transport encryption, and tool-level guardrail mechanisms (blocked command lists and restricted configuration scopes) are required to enforce access control.

IANA Considerations This document has no IANA actions.

References Normative References In many standards track documents several words are used to signify the requirements in the specification. These words are often capitalized. This document defines these words as they should be interpreted in IETF documents. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements. RFC 2119 specifies common key words that may be used in protocol specifications. This document aims to reduce the ambiguity by clarifying that only UPPERCASE usage of the key words have the defined special meanings. Informative References Data models provide a programmatic approach to represent services and networks. Concretely, they can be used to derive configuration information for network and service components, and state information that will be monitored and tracked. Data models can be used during the service and network management life cycle (e.g., service instantiation, service provisioning, service optimization, service monitoring, service diagnosing, and service assurance). Data models are also instrumental in the automation of network management, and they can provide closed-loop control for adaptive and deterministic service creation, delivery, and maintenance. This document describes a framework for service and network management automation that takes advantage of YANG modeling technologies. This framework is drawn from a network operator perspective irrespective of the origin of a data model; thus, it can accommodate YANG modules that are developed outside the IETF. YANG is a data modeling language used to model configuration data, state data, Remote Procedure Calls, and notifications for network management protocols. This document describes the syntax and semantics of version 1.1 of the YANG language. YANG version 1.1 is a maintenance release of the YANG language, addressing ambiguities and defects in the original specification. There are a small number of backward incompatibilities from YANG version 1. This document also specifies the YANG mappings to the Network Configuration Protocol (NETCONF). The Network Configuration Protocol (NETCONF) defined in this document provides mechanisms to install, manipulate, and delete the configuration of network devices. It uses an Extensible Markup Language (XML)-based data encoding for the configuration data as well as the protocol messages. The NETCONF protocol operations are realized as remote procedure calls (RPCs). This document obsoletes RFC 4741. [STANDARDS-TRACK] This document describes an HTTP-based protocol that provides a programmatic interface for accessing data defined in YANG, using the datastore concepts defined in the Network Configuration Protocol (NETCONF). China Mobile China Mobile China Mobile Huawei Orange Orange The application of Digital Twin technology in the networking field is meant to develop various rich network applications, realize efficient and cost-effective data-driven network management, and accelerate network innovation. This document presents an overview of the concept of Network Digital Twin (NDT), provides the basic definitions and a reference architecture, lists a set of application scenarios, and discusses such technology's benefits and key challenges. This document is a product of the Network Management Research Group (NMRG) of the Internet Research Task Force (IRTF). This document reflects the consensus of the research group. It is not a candidate for any level of Internet Standard and is published for informational purposes. Huawei China Mobile Telefonica China Unicom Daejeon University A Network Digital Twin (NDT) provides a network emulation tool usable for different purposes such as scenario planning, impact analysis, and change management. Agentic AI enables dynamic goal-driven execution and adaptive behavior and closed-loop autonomy. By integrating a Network Digital Twin into network management together with the Agentic AI, it allows the network management activities to take user intent or service requirements as input, automatically assess, model, and refine optimization strategies under realistic conditions but in a risk-free environment. Such environment that operates to meet these types of requirements is said to have AI driven Network Operations. AI driven Network Operations brings together existing technologies such as Agentic AI and Network Digital Twin which may be seen as the use of a toolbox of existing components enhanced with a few new elements. This document describes an architecture for AI driven network operations and shows how these components work together with network digital twin and Agentic AI capabilities. It provides a cookbook of existing technologies to satisfy the architecture and realize intent- based network management to meet the needs of the network service. CAICT China Mobile Huawei Cisco Huawei ZTE The evolution from Level 3 (assisted automation) to Level 4 (closed- loop autonomy) in Autonomous Networks (AN) introduces requirements for agentic capabilities, including intent-based reasoning, autonomous planning, and context-aware decision-making, and execution coordination, which transcend the static, rule-based logic of traditional network controllers. This document defines the concept of the Network Management Agent (NMA), a network management entity with autonomous task processing capabilities designed to bridge the gap between service intent and network operations. This document describes the role of NMA in network management and control architectures, and specifies how the NMA collaborates with existing network controllers to achieve Autonomous L4 without replacing or duplicating their functions. It further defines the reference architecture, deployment modes, and logical interfaces of the NMA, including Agent-to-User (A2U), Agent-to-Agent (A2A), Agent- to-Controller (A2C), and Agent-to-Network (A2N) interactions. Huawei Huawei Telefonica Deutsche Telekom Orange Research Nokia The application of MCP in the network management field is meant to refactor network management operation and network capabilities as tools and provide more agile and extensible architecture to expose these AI integration capabilities. This document discusses the applicability of MCP to the network management plane in the IP network that utilizes IETF technologies. It explores MCP for network exposure, multiple MCP server discovery, communication between Network Elements or between the Network element and the Network Controller/Network Gateway. Telefonica Deutsche Telekom Orange Huawei Huawei Huawei Nokia This document discusses the applicability of A2A protocol to the network management in the multi-domain heterogeneous network environment that utilizes IETF technologies. It explores operational aspect, key components, generic workflow and deployment scenarios. The impact of integrating A2A into the network management system is also discussed. The IETF has produced many modules in the YANG modeling language. The majority of these modules are used to construct data models to model devices or monolithic functions. A small number of YANG modules have been defined to model services (for example, the Layer 3 Virtual Private Network Service Model (L3SM) produced by the L3SM working group and documented in RFC 8049). This document describes service models as used within the IETF and also shows where a service model might fit into a software-defined networking architecture. Note that service models do not make any assumption of how a service is actually engineered and delivered for a customer; details of how network protocols and devices are engineered to deliver a service are captured in other modules that are not exposed through the interface between the customer and the provider. Daejeon University DONG-EUI University Huawei Everything OPS This document outlines the key objectives of introducing Agentic AI to the field of network management and highlights the fundamental issues with existing technologies that must be addressed to achieve these goals. It emphasizes the necessity for relevant groups within the IETF/IRTF and presents the core technological areas requiring standardization. The aim of Agentic AI is to facilitate a paradigm shift in which multiple autonomous AI agents collaborate to fully automate network operation, management and security. This document defines encoding rules for representing configuration data, state data, parameters of Remote Procedure Call (RPC) operations or actions, and notifications defined using YANG as JavaScript Object Notation (JSON) text. This document describes a YANG library that provides information about the YANG modules, datastores, and datastore schemas used by a network management server. Simple caching mechanisms are provided to allow clients to minimize retrieval of this information. This version of the YANG library supports the Network Management Datastore Architecture (NMDA) by listing all datastores supported by a network management server and the schema that is used by each of these datastores. Huawei Everything-Ops Swisscom Deutsche Telekom Bell Canada NTT Telefonica This document describes some of the problems in modern operations and management systems and how knowledge graphs and RDF can be used to solve closed loop system, in an automatic way. Discussion Venues This note is to be removed before publishing as an RFC. Source for this draft and an issue tracker can be found at https://github.com/mike-mackey. Orange Research EURECOM Orange Research Orange Research Orange Research TiDB Operational efficiency in incident management in networking requires correlating and interpreting large volumes of heterogeneous technical information. Knowledge Graphs (KG) can provide a unified view of complex systems through shared vocabularies. YANG data models enable describing network configurations and automating their deployment. However, both approaches face challenges in vocabulary alignment and adoption, hindering knowledge capitalization and sharing on network designs and best practices. To address this, the concept of a IT Service Management Knowledge Graph (ITSM-KG) is introduced to leverage existing network infrastructure descriptions in YANG format and enable abstract reasoning on network behaviors. The key principle to achieve the construction of such ITSM-KG is to transform YANG representations of network infrastructures into an equivalent knowledge graph representation, and then embed it into a more extensive data model for Anomaly Detection (AD) and Risk Management applications. In addition to use case analysis and design pattern analysis, an experiment is proposed to assess the potential of the ITSM-KG in improving network quality and designs. Swisscom Swisscom INSA-Lyon INSA-Lyon This document describes the motivation and architecture of a Network Anomaly Detection Framework and the relationship to other documents describing network Symptom semantics and network incident lifecycle. The described architecture for detecting IP network service interruption is designed to be generic applicable and extensible. Different applications are described and examples are referenced with open-source running code. Huawei Swisscom Swisscom INSA-Lyon Network Anomaly Detection is the act of detecting problems in the network. Accurately detecting problems is very challenging for network operators in production networks. Good results require a lot of expertise and knowledge around both the implied network technologies and the connectivity services provided to customers, apart from a proper monitoring infrastructure. In order to facilitate network anomaly detection, novel techniques are being introduced, including programmatical, rule-based and AI-based, with the promise of improving scalability and the hope to keep a high detection accuracy. To guarantee acceptable results, the process needs to be properly designed, adopting well-defined stages to accurately collect evidence of anomalies, validate their relevancy and improve the detection systems over time, iteratively. This document describes a well-defined approach on managing the lifecycle process of a network anomaly detection system, spanning across the recording of its output and its iterative refinement, in order to facilitate network engineers to interact with the network anomaly detection system, enable the "human-in-the-loop" paradigm and refine the detection abilities over time. The major contributions of this document are: the definition of three key stages of the lifecycle process, the definition of a state machine for each anomaly annotation on the system and the definition of YANG data models describing a comprehensive format for the anomaly labels, allowing a well-structured exchange of those between all the interested actors. Swisscom Swisscom INSA-Lyon Huawei This document explains the motivation for defining semantic metadata annotations to help testing, validating and comparing Outlier and Symptom detection systems. These semantic annotations can be supported by supervised and semi-supervised machine learning algorithms and enable data exchange among network operators, vendors and academia, making anomalies apprehensible for humans. The proposed semantics uniforms the network anomaly data exchange between operators and vendors to improve their Service Disruption Detection Systems.

Acknowledgments The authors would like to thank the following individuals for reviewing and providing valuable inputs to this document (listed in no particular order):

• Joe Clarke
• Mohamed Boucadair
• Roland Schott
• Luis M. Contreras
• Oscar Gonzalez de Dios
• Benoît Claise
• Kent Watsen
• Thomas Graf