Design and implementation of a 5G testbed in a virtualized environment

The new 5G network is one of the hot-topics of this century and promises great improvements in the world of mobile networks, network applications and the way society interacts with itself and technology. As this new network is a very complex and articulated system, the management and maintenance tools will also have to evolve accordingly. In this thesis, I attempt to introduce the concept of the digital twin, already imagined in the book Mirror Worlds by David Gelernter, to the context of the 5G network. The digital twin is a tool for supervision, control and analysis that combines the use of data, Machine Learning and simulation environments in order to create a virtual parallel system, faithful to the real one, that reacts to the same stimuli as its physical twin. This concept already exists in the engineering world in various forms. Thanks to the BNN (Barcelona Neural Network centre) research group, we have applied this idea to the 5G-core network, a concept that has already been theoretically explored by others but which to date has not found a real implementation. The following work consists of the design and development of a testbed in order to run simulations of plausible 5G network usage scenarios through simulators. The goal is to obtain datasets to train artificial intelligence models to predict certain aspects of network behavior