Jun 2026· IEEE Conference on Network Softwarization· pp. 435-441· 0 citations· 20 references
Computer Science
Abstract
Configuring 5G networks through standardized exposure interfaces—the Network Exposure Function (NEF) and the Common API Framework (CAPIF)—remains operationally burdensome, requiring manual navigation of 3 GPP parameter schemas, OAuth2 authentication, and live network state. This paper presents a modular, intent-driven architecture that translates natural language operator goals into schema-compliant NEF API configurations using a compact, locally deployed Large Language Model (LLM), Qwen3-4B (4 billion parameters). A deterministic pre-classification layer reserves LLM inference for semantically complex requests, while a six-stage pipeline produces near-deterministic, conflict-aware JSON output. Two complementary validation paths are provided: a Standard Mode with closed-loop Quality of Service (QoS) feedback via a NEF emulator, and a Free5GC Mode that validates the same translation logic against a Free5GC-based experimental environment with real user-plane traffic. Demonstration across five vertical scenarios confirms end-to-end correctness in both the emulated NEF environment and the Free5GC-based experimental environment.
The 6G era introduces unprecedented complexity in managing heterogeneous, large-scale, and dynamic network infrastructures. These challenges are addressed by the concept of Intent-Based Networking (IBN), which has emerged as a promising paradigm for autonomous network management, enabling users to express high-level objectives that are automatically translated and enforced by the network. However, current IBN solutions remain constrained by rigid structured specifications and limited assurance mechanisms. This paper presents an overview of PhD research leveraging Generative AI (GenAI), specifically Large Language Models (LLMs), to address three fundamental IBN challenges: (i) intent translation, (ii) intent assurance, and (iii) GenAI operations in IBN systems. We propose a set of novel frameworks validated on real 5G/6G testbeds. Most contributions are supported by demos, datasets, and open-source implementations, which are referenced in the design section of each contribution. This PhD positions GenAI as a key enabler for advancing autonomous and user-centric 6G.
Large language model (LLM) applications deployed in regulated and high-risk environments face two coupled runtime risks: input-side prompt injection and output-side content compliance violations. Existing defenses are often either model-centric (alignment or fine-tuning) or application-local (keywords and regex), which limits auditability, policy agility, and cross-model portability. This paper proposes a gateway-centric framework that externalizes safety enforcement into a security side-chain. The framework integrates a unified LLM gateway with an external semantic guardrail service to perform bidirectional inspection before and after model invocation. We formalize the architecture with input-risk and output-compliance functions, define a comprehensive threat model and trust boundary, and present a reproducible evaluation protocol. To validate our approach, a prototype implementation based on LiteLLM and Prisma AIRS was evaluated on a rigorously balanced 2,000-sample adversarial dataset merged from diverse sources, including Deepset, wildjailbreak, and jailbreak-classification. Results demonstrate that the proposed architecture achieves an exceptional balance in detection capability, yielding an F1-score of 0.8322, a recall of 0.8531, and effectively maintaining a low False Positive Rate (FPR), thereby outperforming multiple state-of-the-art baselines including Llama Prompt Guard and Deepset DeBERTa. Furthermore, performance testing confirms practical deployability, showing an average added latency of only 320 ms under a 50-concurrent-user load. The paper concludes by discussing the side-chain advantages over model alignment mechanisms, guardrail bypass risks, and adaptive policy-evolution strategies for enterprise AI governance.
Integrated sensing and communications (ISAC) is moving from proof-of-concept demonstrations to system-level deployment in sixth-generation (6G) networks. Because sensing and communication share hardware, spectrum, and waveform resources, ISAC design now involves many tightly coupled choices, including waveform selection, sensing algorithm setup, resource scheduling, and deployment planning. This design space is already too large to manage well through manual tuning or isolated optimizers. This article introduces the \textit{Agent Compiler}, a large language model (LLM)-enabled compilation layer that translates high-level engineering intent into complete and executable ISAC system configurations. The Agent Compiler works in four stages: intent parsing, task decomposition, policy graph synthesis, and infrastructure mapping. It produces a verifiable intermediate representation called the ISAC Policy Graph (IPG). A runtime engine then deploys the compiled configuration and supports closed-loop adaptation at three levels: fast parameter updates, partial recompilation of affected subgraphs, and full workflow recompilation. The core design principle is strict time-scale separation: the LLM handles slow-loop strategic decisions, while proven algorithms retain real-time control in the fast loop. A UAV-assisted disaster rescue example illustrates the full compilation process. We also discuss open issues, including compilation latency, output reliability, constraint verification, and pipeline security, to guide future research.
Lijie Zheng, Xudong Zhong, Baoquan Ren et al.· 0 citations
As Software-Defined Networking (SDN) and Network Function Virtualization (NFV) enabled networks scale in size and complexity, monitoring and managing Service Function Chains (SFCs) under stringent latency and resource constraints becomes increasingly challenging. Although Deep Reinforcement Learning (DRL) is widely applied to SFC provisioning and Virtual Network Function (VNF) placement, enhanced network state monitoring is crucial to capture unexpected network conditions and guide DRL agents toward more adaptive decisions. In this context, Language Models (LMs) enable flexible, natural-language (NL)–based, query-driven network monitoring; however, directly processing complex multi-metric NL queries is computationally expensive and error-prone. This paper proposes an end-to-end (E2E) edge-based query translation pipeline that decomposes multi-metric NL queries into simpler single-metric sub-queries. Query decomposition is performed using a retrieval-augmented language model (RAG-LLM) and compared with a lightweight rule-based decomposition baseline. The resulting sub-queries are translated into Structured Query Language (SQL) using FLAN-T5. A cloud-only baseline, which directly translates NL queries to SQL without decomposition, is also evaluated. The results show that the rule-based edge pipeline achieves the lowest latency, reducing E2E latency by up to 78% compared to RAG-LLM and 18% compared to cloud execution under high workloads. Under increasing arrival rates for the largest workload, the rule-based edge pipeline maintains superior performance over cloud, reducing total E2E latency by 57% at $\lambda = 0.8$ . While RAG-LLM provides greater flexibility for unseen query patterns, both edge-based approaches achieve 100% NL2SQL accuracy with zero decomposition failures, outperforming the cloud-only baseline (95% accuracy).
Parisa Fard Moshiri, Xinyu Zhu, Poonam Lohan et al.· IEEE Transactions on Network...· 0 citations
Network slicing exposes powerful configuration levers for latency, reliability, and capacity in 5 G and emerging 6G systems, but the resulting design space is difficult to navigate for non-experts. Even when a slice configuration is deployed, operators still need to monitor traffic behavior, detect drift from design intent, and iteratively adjust baselines as usage evolves. This paper presents Agent Slicer, an agentic AI system that turns natural language descriptions of diverse deployment environments, including stadiums, hospitals, campuses, and factories, into a complete closed-loop network slicing workflow. A user describes the site in plain language; a retrieval-augmented generator grounded in a curated master document corpus synthesizes a standards-aligned slice plan; a digital-twin streamer and temporal convolutional network (TCN) classifier produce realistic KPI streams and slice labels; a suite of slice analysis tools detects deviations between planned and observed behavior; and a re-planning module proposes updated slice baselines when sustained drift is observed. All of these tools are orchestrated by a LangGraph agent that maintains shared state, coordinates planning, simulation, monitoring, and re-planning phases, and keeps humans in the loop at critical decision points. The system is an agentic architecture that is applied across multiple network slicing scenarios by specializing the master document corpus and tool parameters for each domain, while reusing the same LangGraph backbone.
Anshul Anilkumar Mundakatil, Anjan Phani Raj Narla, Akshil Anilkumar Mundakatil et al.· International Conference on...· 0 citations
Stateful P4 network functions introduce operational failures that emerge only under temporally ordered packet sequences and control-plane states. Existing temporal verifiers (e.g., P4TV) stop at logical verdicts, while dynamic testers (P4Testgen, CHIMERA) execute packets without temporal specifications, and both require operators to hand-author formal specifications. We present IntentP4, a formal-methods-aided pipeline that closes this loop: it translates an operator's natural-language intent into a P4LTL specification and then into a replayable multi-packet test case (packet sequence, control-plane rules, external operations, oracles), grounded throughout in compiler artifacts via a tool-queryable ProgramContext and gated by deterministic per-stage validators. On five stateful P4 programs spanning access control, monitoring, heavy-hitter detection, failure recovery, and load balancing, the P4LTL-to-test generator produces 10 scenarios, 89 packets, and 110 unified execution operations that pass eight consistency checks; on 11 specification subtasks, 4 strictly pass and 3 are semantically close; and an integrated BMv2/Mininet loop exposes runtime failures including a firewall policy-bypass manifestation and a missing multi-table control-plane entry under controller convergence.
Ruonan Feng, Mingming Zhang, Yu Jiang et al.· Proceedings of the ACM SIGCO...· 0 citations