Telefonica Research and Development

Large language models are capable of solving complex math problems or answering difficult questions on almost any topic, but can they generate random street addresses for European cities?

Internet of Things (IoT) application providers rely on Mobile Network Operators (MNOs) and roaming infrastructures to deliver their services globally. In this complex ecosystem, where the end-to-end communication path traverses multiple entities, it became increasingly challenging to guarantee communication availability and reliability. Further, most platform operators use a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">reactive</i> approach to communication issues, responding to user complaints only after incidents have become severe, compromising service quality. This paper presents our experience in the design and deployment of ANCHOR – an <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">unsupervised</i> anomaly detection solution for the IoT connectivity service of a large global roaming platform. ANCHOR assists engineers by filtering vast amounts of data to identify potential problematic clients (i.e., those with connectivity issues affecting several of their IoT devices), enabling proactive issue resolution before the service is critically impacted. We first describe the IoT service, infrastructure, and network visibility of the IoT connectivity provider we operate. Second, we describe the main challenges and operational requirements for designing an unsupervised anomaly detection solution on this platform. Following these guidelines, we propose different statistical rules, and machine- and deep-learning models for IoT verticals anomaly detection based on passive signaling traffic.We describe the steps we followed working with the operational teams on the design and evaluation of our solution on the operational platform, and report an evaluation on operational IoT customers.

Quantum Communications promise advances in cryptography, quantum computing and clock synchronisation, among other emerging applications. However, communication based on quantum phenomena requires an extreme level of isolation from external disturbances, complicating the co-propagation of quantum and classical signals. The challenge is greater when deploying networks that are both heterogeneous (e.g., multiple vendors) and installed in production facilities, given that this type of infrastructure already supports networks loaded with their own requirements. Moreover, to achieve a broad acceptance among network operators, the joint management and operation of quantum and classical resources, compliance with standards, and legal and quality assurance need to be addressed. This article presents solutions to the aforementioned challenges validated in the Madrid quantum network during the implementation of the projects CiViC and OpenQKD. This network was designed to integrate quantum communications in the telecommunications ecosystem by installing quantum-key-distribution modules from multiple providers in production nodes of two different operators. The modules were connected through an optically-switched network with more than 130 km of deployed optical fibre. The tests were done in compliance with strict service level agreements that protected the legacy traffic of the pre-existing classical network. The goal was to ensure full quantum-classical interoperability at all levels, while limiting the modifications to optical transport and encryption and complying with relevant standards. This effort is intended to lay the foundation for large-scale quantum network deployments.

Abstract The current data landscape presents several challenges in data governance and integration. This paper presents the integration of the Data Fabric implemented in the Horizon Europe project aerOS, which integrates and unifies data available in the IoT-Edge-Cloud continuum, with the Data Space implemented in the RE4DY project in order to extend the capabilities of the Data Fabric by enabling the definition fine-grained data access policies and data monetization. This integration would promote collaboration with other participants in the Data Space based on the reuse of data assets from the Data Fabric by sharing them in a trusted environment and the integration of new data or services with the Data Fabric, thus contributing to the creation and integration of new data ecosystems.

This demonstration introduces Cyclic Redundancy Check for Enhanced Verification and Efficient Routing (CRC4EVER). We propose how the Residue Number System (RNS) --- a number system that employs a shared secret scheme distributed across network nodes (nodeIDs)--- can enable lightweight forwarding and proof-of-transit (PoT) in path-aware networks, relying solely on CRC operations. Our approach leverages an RNS-based source routing, where a routeID encodes the entire packet path. At each hop, the routeID is decoded via simple modulo operations, executed at line rate, by repurposing existing CRC hardware in programmable switches. Furthermore, the unique mapping between the routeID and its corresponding set of nodelDs provides intrinsic path verifiability via CRC-based hash operations. A proof-of-concept was implemented on programmable Tofino switches, demonstrating the feasibility of executing both routing and path verification at line rate, through table-free CRC operations.

Data imputation and data generation have important applications across many domains where incomplete or missing data can hinder accurate analysis and decision-making. Diffusion models have emerged as powerful generative models capable of capturing complex data distributions across various data modalities such as image, audio, and time series. Recently, they have been also adapted to generate tabular data. In this article, we propose a diffusion model for tabular data that introduces three key enhancements: (1) a conditioning attention mechanism, (2) an encoder–decoder transformer as the denoising network, and (3) dynamic masking. The conditioning attention mechanism is designed to improve the model’s ability to capture the relationship between the condition and synthetic data. The transformer layers help model interactions within the condition (encoder) or synthetic data (decoder), while dynamic masking enables our model to efficiently handle both missing data imputation and synthetic data generation tasks within a unified framework. We conduct a comprehensive evaluation by comparing the performance of diffusion models with transformer conditioning against state-of-the-art techniques such as Variational Autoencoders, Generative Adversarial Networks, and Diffusion Models, on benchmark datasets. Our evaluation focuses on the assessment of the generated samples with respect to three important criteria, namely: (1) machine learning efficiency, (2) statistical similarity, and (3) privacy risk mitigation. For the task of data imputation, we consider the efficiency of the generated samples across different levels of missing features. The results demonstrate average superior machine learning efficiency and statistical accuracy compared to the baselines, while maintaining privacy risks at a comparable level, particularly showing increased performance in datasets with a large number of features. By conditioning the data generation on a desired target variable, the model can mitigate systemic biases, generate augmented datasets to address data imbalance issues, and improve data quality for subsequent analysis. This has significant implications for domains such as healthcare and finance, where accurate, unbiased, and privacy-preserving data are critical for informed decision-making and fair model outcomes.

A method based on state-driven control for instructing disaggregated transponder local NETCONF agents to select and adapt the optimal Operational (OP) Mode is experimentally demonstrated in a partially disaggregated optical network.

Telco operators are using Software Defined Networks (SDN) and Network Function Virtualisation (NFV) to virtualise a wide range of network functions and link or chain them together to create, deploy and deliver network connectivity services. In such a distributed software networking environment, operators use Internet Protocol Security (IPsec) with Internet Key Exchange (IKEv2) to provide secure connectivity between container network functions running on different computing nodes in their infrastructure. This paper discusses the adoption of the Self-Sovereign Identity (SSI) model in IKEv2 for authentication purposes to avoid the high costs associated with identity management of IPsec endpoints using Public-Key Infrastructure (PKI) and X. 509 certificates, while preserving all the security features of the protocol. The paper presents a novel design of the IKEv2 message flow with Verifiable Credentials (VCs), its open source implementation as a fork of the strongSwan library, and the successful experimental validation.

Novel networking paradigms based on programmability and softwarization are deeply transforming the architecture and end-to-end management of next-generation mobile networks. This demonstration provides a step beyond in this transformation process by showcasing the full cloud-native deployment of a disaggregated and distributed 5 G mobile network from core to Radio Access Network (RAN) using open-source software. The distinguishing features of the proposed approach are that the orchestration process includes the use of a cloud-based Distributed Unit (DU) communicating through O-RAN Split 7.2 with a Radio Unit (RU) performing over-the-air transmission, the end-to-end configuration of slices with on-demand activation of core network function components, and the dynamic configuration of transport network flows to differently route the traffic of active slices.