]> China Unicom

Beijing China hanmy12@chinaunicom.cn

China Unicom

Beijing China zhangnh12@chinaunicom.cn

China Unicom

Beijing China zhaoj501@chinaunicom.cn

Operations and Management Area ANIMA Internet-Draft ANIMA defined Autonomic Service Agent to build intelligent management functions into network devices, and could interact with each other through a standard protocol (aka GRASP).With the rapid advancement of Large Language Model (LLM)-driven AI technologies, there is now a potential opportunity to enhance the ASA to be AI-powered, thereby elevating the intelligence of device-built-in management functions to a whole new level.This document analyzes the impact of the AI-powered ASA, mostly from the perspective of the ASA communication protocol.

Introduction The ANIMA provides a vision of a network that configures, heals, optimizes, and protects itself. An ASA is defined in as "An agent implemented on an autonomic node that implements an autonomic function, either in part (in the case of a distributed function) or whole. proposes guidelines for the design of Autonomic Service Agents for autonomic networks. Autonomic Service Agents, together with the Autonomic Network Infrastructure, the Autonomic Control Plane, and the GeneRic Autonomic Signaling Protocol, constitute the base elements of an autonomic networking ecosystem. Large-scale network models have attracted much attention in the field of artificial intelligence in recent years. They integrate the advantages of network technology and LLMs and show great potential in many fields. Especially for network operation and maintenance, it is demonstrating huge enabling potential and providing innovative approaches to solve increasingly complex network operation and maintenance problems. AI-ASA can achieve more intelligent management functions. Embedding AI-ASA into network devices can enhance operation and maintenance efficiency with LLMs. This draft analyzes AI-ASA vision and potential functions and describes the scenarios of AI-powered ASA Communication between Network Devices and Network Management Systems. The potential new requirements of GRASP are also discussed.

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.

Background

Definition of ASA In , ASA is a process that makes use of the features provided by the ANI to achieve its own goals, usually including interaction with other ASAs via GRASP or otherwise. Of course, it also interacts with the specific targets of its function, using any suitable mechanism. Unless its function is very simple, the ASA will need to handle overlapping asynchronous operations. It may therefore be a quite complex piece of software in its own right, forming part of the application layer above the ANI. Autonomic Service Agents, together with the Autonomic Network Infrastructure, the Autonomic Control Plane, and the GeneRic Autonomic Signaling Protocol, constitute the base elements of an autonomic networking ecosystem.

Emergence of AI-powered Agent Intelligent agent, as an important concept in the field of artificial intelligence, refers to a system that can autonomously perceive the environment, make decisions, and execute actions. It has basic characteristics such as autonomy, interactivity, reactivity, and adaptability, and can independently complete tasks in complex and changing environments. Intelligent agents can learn and make decisions. Al Agent, an automated intelligent entity capable of interacting with its environment, acquiring contextual informationreasoning, self-learning, decision-making, executing tasks (autonomously or in collaboration with other Al Agents) to achieve There are a few examples of AI Agents. A travel AI Agent that can help users search for travel destinations based on preferences, compare flight and hotel costs, make bookings, and adjust plans A loan handling agent that can help users take out a loan. The AI Agent can access a user's salary information, credit history, and then interact with the user to identify the right loan for the target use case the customer has in mind A shopping agent for clothing that can listen to user preferences and interests, look at prior purchases, and show users different options, ultimately helping a user find the right sports coat for an event AI Agent in 3GPP, an automated intelligent entity capable of interacting with its environment, acquiring contextual informationreasoning, self-learning, decision-making, executing tasks autonomously or in collaboration with other AI Agents to achieve a specific goal

The Vision of AI-powered ASA The AI-powered ASA provides more intelligent operation and management of network devices to achieve the Intention-driven network and Auto-driven network.

Scenarios of AI-powered ASA Communication between Network Devices

General The network devices to communicate with other network devices through anima's interface.

Possible Examples

AI Agent based Router for Automatic Congestion Relief In the automatic congestion relief use case, the traditional solution relies on built-in intelligent modules in devices to implement traffic rerouting via traditional protocols (BGP-LS/BGP-RPD). Device interactions are constrained by predefined protocol rules (e.g., policy triggering based on fixed bandwidth thresholds), lacking cross-device historical data sharing and AI model collaboration. Policy generation depends solely on local TOP-N traffic modeling, unable to adaptively optimize based on real-time traffic patterns. When AI-powered Agents are introduced into network devices, AI-powered ASA Communication can be established between devices. Devices extend BGP-LS to synchronize real-time link bandwidth and TOP-N traffic characteristics. The AI-powered Agents dynamically define congestion thresholds based on traffic data, replacing manual threshold configuration. Upon detecting congestion, devices use the GRASP protocol to negotiate AI-generated policies (e.g., dynamic adjustment of Multi-Exit Discriminator (MED) values) and route traffic precisely to lightly loaded links via the BGP Routing Process Daemon (BGP RPD). Reinforcement learning is applied to dynamically optimize policy parameters during this process.

AI Agent based Router for Automatic Network DDoS Attacks Defense With the evolution of attack forms, the Distributed Denial of Service (DDoS) Attacks present the features of short-term and high-frequency outbreaks, and the attack peak value keeps rising year by year, imposing an extreme challenge on the defense response speed. In response to the above attack problems, this document innovatively puts forward an edge defense architecture: deploy attack detection functions to end devices, achieve second-level flash defense against DDoS attacks via intelligent service traffic monitoring, and establish an autonomous network DDoS attack defense system. In the meantime, rely on the AI Agent based Router to support the second-level discovery and real-time interception of attack behaviors, so as to strengthen the network security barrier.

Scenarios of AI-powered ASA Communication between Network Management Systems and Devices

General The network controller communicates with other netwok devices by the anima interface or protocol.

Possible Examples

Coordinated IPv6 Monitoring In the current IPv6 end-to-end traffic monitoring scenario, traffic data collection and analysis rely on manual intervention, while the large volume of live network traffic data results in high resource requirements. When AI-powered Agents are deployed in network controllers and devices, AI-powered ASA communication can be established between IDC controllers and edge devices to enable hierarchical collaboration. The controller's AI-powered Agent module discovers network devices via the GRASP protocol, initiates multi-threaded real-time collection and monitoring of IPv6/IP traffic data, and performs preliminary analysis including flow pattern recognition and IPv6/IPv4 traffic ratio trending. Concurrently, the device-side AI-powered Agent collects customer traffic data, decomposes traffic distribution characteristics to identify high-value business scenarios, and synchronizes these insights to the controller via the GRASP protocol. The controller's AI-powered Agent integrates provincial-level traffic ingress/egress data to construct regional traffic matrices and uploads preliminary analysis results (e.g., internal IDC traffic distribution, inter-provincial link utilization) to the IPv6 end-to-end monitoring platform. The IPv6 end-to-end monitoring platform leverages multi-dimensional data models to conduct in-depth analysis on the uploaded traffic data and preliminary results, generating final operational decisions such as inter-provincial link bandwidth expansion plans and CDN node deployment recommendations. These decisions are then disseminated to the controller, which issues configuration instructions to the device-side AI-powered Agents via the GRASP API. Upon receiving the instructions, the device's intelligent module invokes relevant interfaces to adjust server resources and verifies operational effectiveness through self-monitoring threads.

Potential New Requirements of GRASP TBD

The interface and model extension for Prompt with AI agent TBD

Defination of Option for AI-ASA TBD

Security Considerations Uncertainty of Current AI Technologies.

IANA Considerations TBD

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 This document describes a reference model for Autonomic Networking for managed networks. It defines the behavior of an autonomic node, how the various elements in an autonomic context work together, and how autonomic services can use the infrastructure. Autonomic systems were first described in 2001. The fundamental goal is self-management, including self-configuration, self-optimization, self-healing, and self-protection. This is achieved by an autonomic function having minimal dependencies on human administrators or centralized management systems. It usually implies distribution across network elements. This document defines common language and outlines design goals (and what are not design goals) for autonomic functions. A high-level reference model illustrates how functional elements in an Autonomic Network interact. This document is a product of the IRTF's Network Management Research Group. This document specifies the GeneRic Autonomic Signaling Protocol (GRASP), which enables autonomic nodes and Autonomic Service Agents to dynamically discover peers, to synchronize state with each other, and to negotiate parameter settings with each other. GRASP depends on an external security environment that is described elsewhere. The technical objectives and parameters for specific application scenarios are to be described in separate documents. Appendices briefly discuss requirements for the protocol and existing protocols with comparable features. This document proposes guidelines for the design of Autonomic Service Agents for autonomic networks. Autonomic Service Agents, together with the Autonomic Network Infrastructure, the Autonomic Control Plane, and the GeneRic Autonomic Signaling Protocol, constitute base elements of an autonomic networking ecosystem. Five9 Cisco AI Agents are software applications that utilize Large Language Models (LLM)s to interact with humans (or other AI Agents) for purposes of performing tasks. AI Agents can make use of resources - including APIs and documents - to perform those tasks, and are capable of reasoning about which resources to use. To facilitate AI agent operation, AI agents need to communicate with users, and then interact with other resources over the Internet, including APIs and other AI agents. This document describes a framework for AI Agent communications on the Internet, identifying the various protocols that come into play. It introduces use cases that motivate features and functions that need to be present in those protocols. It also provides a brief survey of existing work in standardizing AI agent protocols, including the Model Context Protocol (MCP), the Agent to Agent Protocol (A2A) and the Agntcy Framework, and describes how those works fit into this framework. The primary objective of this document is to set the stage for possible standards activity at the IETF in this space. China Mobile Current advancements in AI technologies, particularly large models, have demonstrated immense potential in content generation, reasoning, analysis and so on, providing robust technical support for network automation and self-intelligence. However, in practical network operations, challenges such as system isolation and fragmented data lead to extensive manual, repetitive, and inefficient tasks, the improvement of intelligence level is very necessary. This document identifies typical scenarios requiring enhanced intelligence, and explains how AI Agents and large model technologies can empower networks to address operational pain points, reduce manual efforts, and explore impacts on network data, system architectures, and interfaces correspondingly. It further explores and summarizes standardization efforts in implementation.