Multi-Access Edge Computing for Automotive Applications

Multi-Access Edge Computing (MEC) moves the complexity of the network closer to the end user. With the advent of the fifth generation network (5G) this paradigm becomes very important because, together with new technologies, MEC enables a series of services and applications in which a real-time response in strictly required, such as: autonomous driving, robotics, safety etc. Moreover with new applications and Internet of Things the computation demand is destined to grow as well as the core network overload, edge computing can be a solution to face also this problem. This thesis goes into the actual state of the art defined by the European Telecommunication Standard Institute (ETSI) and looks at the several current projects of ETSI compliant MEC platform. They have been evaluated on the basis of cost and reliability to find the most suitable to develop and test automotive applications. The final choice led to OpenNESS, the Intel open source toolkit to create and deploy edge applications. However OpenNESS is a new software in experimental phase and has been tested on few types of hardware, this did not allow to obtain the complete result. To overcome the problem it has been realized a cloud machine with an architecture very similar to the OpenNESS one and through an experiment, the “In-Vehicle Entertainement” (IVE) it has been possible to find, in terms of latency, what is the best location to place this machine to retrieve multimedia contents from a user driving in Turin. The analysis of the collected data confirms that the best solution is the closest one. Finally it has been implemented from scratch a beta version of the See Through, a real automotive Use Case for which the implication of Multi Access Edge Computing is essential. With the knowledges get during the IVE experiment will be possible to run this application on the proper server to obtain the most MEC likely results until the ETSI compliant MEC platform will be available.