Network Automation
Networking, i.e., the way of designing, building and operating networks, has changed a lot in recent decades, thanks to the introduction of programmability and virtualization of network equipment. These changes have enabled new forms of Network Automation, i.e., the development of new procedures to automate operations that today require time and high operating costs. We have therefore oriented the activity of the BONSAI lab to cover two new rising architectural paradigms, Software Defined Networking (SDN) and Network Function Virtualization (NFV), developing know-how and performing research in various application areas of these two paradigms. In the period from 2018 to 2021 these topics were investigated within the European Horizon 2020 project "METRO-HAUL" in which SDN and NFV were applied to develop a metro optical-network infrastructure able to support edge/fog computing and 5G. In the project, BONSAI was committed to various theoretical studies based on Machine Learning and to the development of a network-planning software module. Other studies have been carried out on the virtualized implementation of the 5G Radio Access Network (RAN) and on the topic of Software Defined Wide Area Networks (SD-WAN). The laboratory is equipped with two testbeds that allow to program SDN controllers and opensource NFV orchestrators and experiment with their use on physical networks of Openflow switches. In addition, both the teaching and the research activities on SDN are supported by the BONSAI Lab Optical Network Testbed.
Design and Management of Optical Networks
Optical networks are the backbone of today's Internet. In the BONSAI laboratory, problems related to optical networks and optical switching are studied. In particular, the activity on optical networking covers the following topics: network modeling, design and simulation, machine learning for automated network management, control plane, synchronization and time distribution. The switching activity focuses on: optical switching architectures, optical switching systems for metro and core applications, optical backplanes and optical interconnection for data centers. These issues are often at the service of socio-technological objectives such as the development of new platforms for Smart Cities and Industry 4.0, the energy efficiency and sustainability of the ICT sector, and the resilience of national critical infrastructures to natural, environmental and weather-based disasters.To support our activities on network modeling and simulation, a hardware testbed was set up equipped with optical network equipment such as long-distance transmitters at 10Gbit / s, wavelength-selective switches, optical amplifiers and high-capacity traffic generators.
Quantum Networking
Politecnico is part of a consortium (PoliQi) for the realization of an urban optical-network infrastructure capable of providing unconditionally-secure quantum keys, located in Milan. In support to PoliQi, in the BONSAI lab, we perform research activities to demonstrate open, large-scale, quantum communication networks and system architectures that distribute classical secret keys, or, more generally, quantum information, across multi-node quantum networks. Our activities span from mathematical modeling for resource allocation, to development of the key management system, to new cryptographic solutions in this area.
Edge Computing and Critical Infrastructures
Edge computing is a distributed computing model in which processing takes place close to where the data is generated and used. This model brings many benefits in terms of latency, reliability, and security at the price of greater management complexity. These advantages are particularly useful in critical infrastructure management (energy, transport, health, safety) and industrial control applications. The greater complexity stems from the large number of nodes with heterogeneous network technologies and in a difficult or hostile environment. In the Bonsai lab, we study the algorithms and technologies that allow the large-scale management of computing nodes with limited resources and heterogeneous networks. The laboratory is equipped with 5G network emulators and programmable switches. The research exploits advanced cryptographic techniques (homomorphic cryptography, smart contracts) and both traditional and machine learning-based optimization techniques.