Digital Theses Archive


Tesi etd-09052019-150245

Type of thesis
E-mail address
Implementation and Experimental Evaluation Of a Virtualised 5GNetwork for URLLC Services
Scientific disciplinary sector
INGEGNERIA - Ph.D. Programme in Emerging Digital Technologies (EDT)
Membro Prof. MONTI, PAOLO
Presidente Prof. AZCORRA, ARTURO
  • 5G
  • 5G Connected Vehicles
  • 5G Core Network
  • 5G Functional Split
  • 5G NG RAN Performance
  • 5G Scalability
  • MEC
  • MEC applications
  • Reliability in 5G networks
  • Slicing in 5G networks
Exam session start date
5G is Coming. Here is how immersive smartphones, autonomous cars and billions of intelligent sensors could transform your life. <br><br>The fifth generation of cellular technology, 5G, will offer much more than faster phones. Its first phase, scheduled to begin in 2019, will improve smartphone responsiveness. Soon after, 5G may also start to enhance many other aspects of your daily life, from automotive safety to your entertainment and your view of reality. For instance, your 5G smartphone will be able to stream ultra-high-definition video at gigabit speeds, with enough consistency to enable immersive virtual reality experiences. In a subsequent phase, 5G will also enable ultra-reliable and low-latency communication. Autonomous vehicles and high-precision manufacturing, among other use cases, will need these qualities to thrive. 5G will allow autonomous vehicles to establish a direct link between each other that by passes network towers and diminishes urgent communications delays. This means a vehicle could instantly receive alerts to stop or change course if a collision seems imminent, even if the danger is beyond the range of its cameras and radar. In addition, 5G-enabled vehicles could connect with smart city infrastructure like traffic lights and street signs, improving traffic management. With 5G communications between autonomous vehicles, cars could share their intentions and know intended paths, allowing them to plan more safely and efficiently. Therefore, steering wheels could disappear, with cars redesigned as living spaces: basically couches or offices on wheels. With 5G-enhanced mobile internet, your car will be able to stream ultra-high-definition video, and even VR, while you travel. Passengers could easily work as they ride, perhaps changing the entire nature of their commutes. <br><br>The imagination of our future is therefore a networked society with unbounded access to information and sharing of data, which is accessible, everywhere and ever time for everyone and everything. Present wireless based technologies, like 3GPP LTE technology, HSPA and Wi-Fi, will be incorporating new technology components that will be helping to meet the needs of the future. Nevertheless, there may be certain scenarios that cannot be adequately addressed along with the evolution of ongoing existing technologies. The instigation of a completely new wireless based technologies will complement the current technologies which are need for the long term realization of the networked society <br><br>This thesis provides contributions to the state of the art regarding the architecture and performance of the LTE and 5G mobile networks, from the conceptual and the experimental point of view. The thesis covers both research and implementation aspects, especially because almost everything was developed in an experimental setup to develop and analyze a standard compliant 5G network. The main objective of the thesis is the development of a standard compliant 5G network and all the components belonging to it. In such a network, firstly a deep analysis of its performance have been performed. This is because the need to understand if some typical 5G services and applications, such as autonomous vehicles collision application and ultra-high-definition video streaming applications could be deployed exploiting the deployed 5G network. To achieve these goals, the starting point has been the investigation of one of the main open source 5G emulation platform, the OpenAirInterface emulation platform. After a first phase in which the platform has been analyzed and set in the TeCIP Institute laboratory, it has been extended with new features and capabilities and adopted to setup a standard compliant 5G network. A deeply analysis and experimental measurements of latency and jitter experienced in the 5G network have been performed. <br><br>The next objective was the introduction in the 5G network of the so-called Multi-access Edge Computing (MEC). Edge computing as an evolution of cloud computing brings application hosting from centralized data centers down to the network edge, closer to consumers and the data generated by applications. Edge computing is acknowledged as one of the key pillars for meeting the demanding Key Performance Indicators (KPIs) of 5G, especially as far as low latency and bandwidth efficiency are concerned. This objective is mainly realized thanks to the collaboration with EURECOM in Sophia-Antipolis during the period abroad. Exploiting the MEC, two applications as part of our contribution to the 5G-TRANSFORMER project related to the road safety and automotive infotainment have been deployed in the autonomous vehicles scenario. <br><br>The 5G network architecture will be heavily based on virtual network functions (VNFs). Thus, in a scenario where network functions are virtualized, both hardware and software failures assume the same importance, and their reliability shall be guaranteed. Similarly, reliability at service chain level is important to as-sure proper service availability features to application service platforms deployed by verticals. Therefore a study of the reliability in 5G networks has been also performed in the conclusion of my PhD. Different protection mechanisms have been proposed and experimentally analyzed for both 5G core and access networks. <br><br>Last but not least, this work also pursued a really challenging goal: we were able to demonstrate a 5G network slice deployment in the ARNO testbed by using the 5G-TRANSFORMER architecture and offer a mobile/edge connectivity service with virtualized functions.<br>