Next-generation mobile networks will rely on their autonomous operation. Virtual Network Functions empowered by Artificial Intelligence (AI) and Machine Learning (ML) can adapt to varying environments that encompass both network conditions and the cloud platform executing them. In this view, it becomes paramount tounderstand whyAI/ML algorithms made a decision, to be able to reason upon those decisions and, eventually, take further decisions related toe.g., network orchestration. In this paper, we present ATHENA, an ML-based radio resource scheduler for virtualized Radio Access Network (RAN) system. Our real-software implementation shows that the proposed ML-based approach can outperform the baseline solution. We discuss how additional re-orchestration actions can be taken by analyzing our scheduling decisions and learning from the past.

Network Function Virtualization (NFV) and Multi-access Edge Computing (MEC) are two technologies expected to play a vital role in 5G and beyond networks. However, adequate mechanisms are required to meet the dynamically changing network service demands to utilize the network resources optimally and also to satisfy the demanding QoS requirements. Particularly in multi-domain scenarios, the additional challenge of isolation and data privacy among domains needs to be tackled. To this end, centralized and distributed Artificial Intelligence (AI)-driven resource orchestration techniques (e.g., virtual network function (VNF) autoscaling) are foreseen as the main enabler. In this work, we propose deep learning models, both centralized and federated approaches, that can perform horizontal and vertical autoscaling in multi-domain networks. The problem of autoscaling is modelled as a time series forecasting problem that predicts the future number of VNF instances based on the expected traffic demand. We evaluate the performance of various deep learning models trained over a commercial network operator dataset and investigate the pros and cons of federated learning over centralized learning approaches. Furthermore, we introduce the AI-driven Kubernetes orchestration prototype that we implemented by leveraging our MEC platform and assess the performance of the proposed deep learning models in a practical setup.

Network Function Virtualization is a promising technology that proposes to decouple the network functions from their underlying hardware and transform them into software entities called Virtual Network Functions (VNFs). This approach offers network operators with more flexibility to instantiate, configure, scale, and migrate VNFs at runtime depending on the demand. On introducing these VNFs at the network edges (e.g., base stations), emerging use cases such as connected cars can be supported. However, in such an environment, efficient VNF placement and orchestration mechanisms are needed to address the challenges of continuously changing network dynamics, service latency requirements and user mobility patterns. The purpose of this paper is twofold. Firstly, we propose a neural-network model (i.e., a subset of machine learning) that can assist in proactive auto-scaling by predicting the number of VNF instances required as a function of the network traffic they should process. Based on the traffic traces collected over a commercial mobile network, the model achieves a prediction accuracy of 97%. Secondly, we provide an Integer Linear Programming formulation for placing these VNFs at the edge nodes with a primary objective of minimizing end-to-end latency from all users to their respective VNFs. Our results show an improvement in latency by upto 75% when VNFs are placed at the network edges.

Network Function Virtualization (NFV) has raised tremendous attention in the academic and industrial communities alike. The former have been attracted by the service centric orchestration challenges. The latter have found that, by decoupling network functions from the underling hardware, significant savings can be obtained by means of infrastructure homogenization and service automation. As opposed to the initial NFV solutions, which heavily relied on a centralized cloud model, the emerging Multi-access Edge Computing trend is calling for a distribution of computational capacity at the customers' sites. This change of paradigm could result in the deployment of tens of thousands or even millions of Points-of-Presence. In this article we discuss the fundamental challenges of deploying NFV in scattered environments, then we introduce the Light-MANO framework, a Multi-access Network Operating System converging SDN and NFV into a single lightweight platform for management and orchestration of network services over distributed NFV infrastructure. Finally, we report on a proof- of-concept implementation of LightMANO and on its evaluation.

Intent-based networking is a major component that will transform the manner in which the SDN/NFV-enabled future network infrastructures are operated. In particular, Intent-based networking is expected to play a major role in the multi-technological and software-defined 5G systems development roadmap. In this paper, we present the design and prototype implementation of an Intent-based mobile backhauling interface for 5G networks. Finally, we report on the empirical evaluation of of the proposed Intent-based interface over a small Enterprise WLAN. We also release the entire software stack under a permissive license for academic use.