Orchestration in a 5G-MEC Testbed for V2X Applications

Abstract

The Fifth Generation (5G) of mobile networks is revolutionizing connectivity,enabling faster data transmission and lower latency. 5G enhances the performance of the usage scenarios supported by the previous generation but also enables new ones, such as Ultra-Reliable Low-Latency Communication (URLLC), which is essential for various real-time Applications in today’s fast-paced world. From autonomous vehicles to Remote surgery, URLLC ensures minimal delay and utmost reliability, enhancing safety and efficiency. The requirements for URLLC in 5G networks are stringent, demanding ultra-low latency, high reliability, and scalability to support diverse applications. In parallel to the 5G networks, Multi-access Edge Computing (MEC) is rapidly gaining traction and undergoing standardization to meet the demands of applications such as URLLC. MEC consists of the Computing platform located in the proximity of the user at the edge of the network. MEC has the most prominent benefit of reducing latency and providing computation resources rather than relying solely on centralized data centers. Vehicle-to-Everything (V2X) communication, an integral component of smart transportation systems, heavily relies on URLLC. 5G and MEC ensure seamless and real-time connectivity between vehicles, infrastructure, pedestrians, and other entities on the road. The development of a 5GMEC testbed for testing V2X applications is essential to evaluate the performance and interoperability of edge computing solutions in dynamic vehicular environments and to explore the potential impact. The thesis contributions can be divided into three parts. First, the thesis focuses on understanding the technical details of existing frameworks and specifications provided by standardized bodies for building a 5G-MEC testbed for V2X applications. This involves closely examining protocols, interfaces, and architectural guidelines required by the standardization entities. Additionally, the thesis looks into the individual hardware, software, networking equipment, and communication interfaces needed for the testbed. By analyzing these technical aspects, the research activity aims to overview all the components to create a 5G-MEC testbed for V2X applications and how to tailor the available solutions (which can be open source) for these components to meet specific requirements of the testbed. Secondly, in a dynamic scenario such as V2X, the orchestration of 5GMEC applications is critical. A 5G-MEC system typically employs a distributed architecture where the network and computing resources are deployed in the proximity of the users. In such distributed environments, orchestration becomes essential to efficiently manage and coordinate the 5G-MEC applications across multiple edge nodes. This thesis aims to develop a testbed that includes the orchestration of MEC applications, which enables the effective allocation of data and network resources. The use of frameworks for the deployment, scaling, and life cycle management of MEC applications, such as Kubernetes (K8s), is explored. Furthermore, the thesis addresses the integration of K8s with the MEC framework, providing strategies for deploying orchestration in user mobility scenarios inherent to V2X communication. Third, the research activity delves into the complexity of the joint allocation of data and network resources in 5G-MEC systems, evaluating strategies that optimize multiple performance objectives, such as latency, throughput, and Quality of Experience (QoE). The multi-objective optimization is particularly important in the provision of services with heterogeneous requirements as 5G does. The evaluation of the strategies is performed in a 5G-MEC testbed for V2X application.