Publications

The transition toward Open Radio Access Networks (O-RANs) is reshaping how cellular infrastructure is deployed, managed, and optimized. This paper investigates the energy-aware joint placement and migration of cloud-native functions (CNFs) in an O-RAN edge cloud. We consider both a Single-CU-UP association model and a slice-aware Multi-CU-UP relaxation, in which distinct slice-flow groups of the same distributed unit (DU) may be assigned to different Centralized Unit User Plane (CU-UP) processing targets under one Centralized Unit Control Plane (CU-CP). For brevity, these scenarios are referred to as Single-CU and Multi-CU, respectively; Multi-CU never denotes multiple CU-CP associations. We formulate the problem as a Mixed-Integer Linear Program (MILP) that minimizes server, transmission, wake-up, and migration energy while satisfying server-resource capacities and one-way delay requirements over the F1 user-plane interface (F1-U) between each DU and its selected CU-UP in a fat-tree edge data center. To improve computational scalability, we also develop a deterministic k-means-based heuristic that approximates the MILP decisions without requiring repeated exact optimization. Over the evaluated 24-hour workload, the theoretical Multi-CU relaxation reduces modeled energy consumption by 5.7% relative to the Single-CU baseline. For the Multi-CU case, the proposed heuristic remains within approximately 9.7% of the proposed MILP, demonstrating a favorable trade-off between energy efficiency and computational tractability.

Segment routing (SR) involves routing requests through shortest paths or a limited number of segments, which are themselves the shortest paths between their endpoints. While this flexibility makes segment routing attractive for traffic engineering (TE) in IP/MPLS networks, it becomes challenging to update when some links become unavailable (e.g., maintenance) and routing decisions must adapt while limiting the number of configuration changes between two consecutive periods. We address the problem of dynamically selecting SR-TE configurations for multiple traffic demands with the objective of minimizing the maximum link utilization (MLU).In addition, a reconfiguration budget is imposed in order to limit the number of segments modified. Based on this representation, we develop a MLU column-generation optimization framework. It allows to jointly consider traffic distribution, SR configuration complexity, and temporal reconfiguration costs in time-varying IP/MPLS networks. The numerical results are obtained using realistic datasets provided by Orange, with topologies containing up to 1,263 nodes and 15,000 traffic demands. The optimality gap (i.e., accuracy) is below 5% in 80.45% of the 133 instance-time evaluations.

In 5G Core (5GC) networks, secure fault diagnosis should verify requesters and evidence while applying explicit policy checks to corrective actions. This letter proposes the Zero-Trust Fault Diagnostic Analytics Service (ZT-FDAS), which augments Network Data Analytics Function (NWDAF) services with request-scoped verification and adaptive responses. ZT-FDAS validates authorization, scope, integrity, provenance, and freshness of evidence before bounded LLM-guided graph traversal and deduplicated top-K ranking. A deterministic policy maps diagnoses and trust violations to monitor, step-up, restrict, or quarantine recommendations. An offline characterization of 1,049 resolved tickets assesses RCA graph construction, but not security effectiveness or diagnostic accuracy.

The transition toward Open Radio Access Networks (O-RANs) is reshaping how cellular infrastructure is deployed, managed, and optimized. This paper investigates the energy-aware joint placement and migration of cloud-native functions (CNFs) in an O-RAN edge cloud. We consider both a Single-CU-UP association model and a slice-aware Multi-CU-UP relaxation, in which distinct slice-flow groups of the same distributed unit (DU) may be assigned to different Centralized Unit User Plane (CU-UP) processing targets under one Centralized Unit Control Plane (CU-CP). For brevity, these scenarios are referred to as Single-CU and Multi-CU, respectively; Multi-CU never denotes multiple CU-CP associations. We formulate the problem as a Mixed-Integer Linear Program (MILP) that minimizes server, transmission, wake-up, and migration energy while satisfying server-resource capacities and one-way delay requirements over the F1 user-plane interface (F1-U) between each DU and its selected CU-UP in a fat-tree edge data center. To improve computational scalability, we also develop a deterministic k-means-based heuristic that approximates the MILP decisions without requiring repeated exact optimization. Over the evaluated 24-hour workload, the theoretical Multi-CU relaxation reduces modeled energy consumption by 5.7% relative to the Single-CU baseline. For the Multi-CU case, the proposed heuristic remains within approximately 9.7% of the proposed MILP, demonstrating a favorable trade-off between energy efficiency and computational tractability.

This paper presents a hierarchical multi-agent LLM architecture to bridge communication gaps between nontechnical end users and telecommunications domain experts in private network environments. We propose a cross-domain query translation framework that leverages specialized language models coordinated through multi-agent reflection-based reasoning. The resulting system addresses three critical challenges: (1) accurately classify user queries related to telecommunications network issues using a dual-stage hierarchical approach, (2) preserve user privacy through the anonymization of semantically relevant personally identifiable information (PII) while maintaining diagnostic utility, and (3) translate technical expert responses into user-comprehensible language. Our approach employs ReAct-style agents enhanced with selfreflection mechanisms for iterative output refinement, semanticpreserving anonymization techniques respecting k-anonymity and differential privacy principles, and few-shot learning strategies designed for limited training data scenarios. The framework was comprehensively evaluated on 10,000 previously unseen validation scenarios across various vertical industries.

Communications networks now form the backbone of our digital world, with fast and reliable connectivity. However, even with appropriate redundancy and failover mechanisms, it is difficult to guarantee “five 9s” (99.999%) reliability, requiring rapid and accurate root cause analysis (RCA) during outages. In the event of an outage, rapid and accurate RCA becomes essential to restore service and prevent future disruptions. This study evaluates three Large Language Model (LLM) methodologies — Fine-Tuning, RAG, and a Hybrid approach — for constructing a Root Cause Analysis (RCA) Knowledge Base from support tickets. We compare their performance using a comprehensive suite of lexical and semantic similarity metrics. Our experiments on a real industrial dataset demonstrate that the generated knowledge base provides an excellent starting point for accelerating RCA tasks and improving network resilience.

Ensuring the highest levels of performance and reliability for customized services in fifth-generation (5G) and beyond (B5G) networks requires the automation of resource management within network slices. In this paper, we propose PCLANSA, a proactive closed-loop algorithm that dynamically allocates and scales resources to meet the demands of diverse applications in real time for an end-to-end (E2E) network slice. In our experiment, PCLANSA was evaluated to ensure that each virtual network function is allocated the resources it requires, thereby maximizing efficiency and minimizing waste. This goal is achieved through the intelligent scaling of virtual network functions. The benefits of PCLANSA have been demonstrated across various network slice types, including eMBB, mMTC, uRLLC, and VoIP. This finding indicates the potential for substantial gains in resource utilization and cost savings, with the possibility of reducing over-provisioning by up to 54.85%.

Advanced Large Language Models (LLMs) have revolutionized various fields, including communication networks, sparking an innovation wave that has led to new applications and services, and significantly enhanced solution schemes. Despite all these impressive developments, most LLMs typically require huge computational resources, resulting in terribly high energy consumption.Thus, this research study proposes an end-to-end pipeline that investigates the trade-off between energy efficiency and model performance for an LLM during fault ticket analysis in communication networks. It further evaluates the pipeline performance using two real-world datasets for the tasks of root cause analysis and response feedback in a communication network.Our results show that an appropriate combination of quantization and pruning techniques is able to reduce energy consumption while significantly improving model performance.

Network operators are facing new challenges when meeting the needs of their customers. The challenges arise due to the rise of new services, such as HD video streaming, IoT, autonomous driving, etc., and the exponential growth of network traffic. In this context, 5G and B5G networks have been evolving to accommodate a wide range of applications and use cases. Additionally, this evolution brings new features, like the ability to create multiple end-to-end isolated virtual networks using network slicing. Nevertheless, to ensure the quality of service, operators must maintain and optimize their networks in accordance with the key performance indicators (KPIs) and the slice service-level agreements (SLAs). In this paper, we introduce a machine learning (ML) model used to estimate throughput in 5G and B5G networks with end-to-end (E2E) network slices. Then, we combine the predicted throughput with the current network state to derive an estimate of other network KPIs, which can be used to further improve service assurance. To assess the efficiency of our solution, a performance metric was proposed. Numerical evaluations demonstrate that our KPI prediction model outperforms those derived from other methods with the same or nearly the same computational time.