Service Provisioning in Vehicular Networks Through Edge and Cloud: An Empirical Analysis

Abstract

This is the accepted version of the original article published by IEEE.Abstract-Vehicular Cloud Computing is a network infrastructure paradigm that has been largely used in the vehicular systems landscape for improving drivers' experience. In particular, the higher computational resources made available by cloud computing technologies have helped in coping with the tremendous growth of data traffic exchanged within vehicular networks. However, the advanced development of such infrastructure, together with the relentless proliferation of services and applications characterized by heterogeneous and demanding requirements, has led to redefine the way in which cellular-based vehicular networks assist vehicular communications. As an example, Multi-Access Edge Computing (MEC) is an emerging network paradigm that can be exploited also in vehicular scenarios to foster a more effective and flexible service delivery. Although in literature the migration of vehicular systems towards a MEC-based approach has been already envisaged giving rise to the concept of Vehicular Edge Computing, a not fully investigated aspect is represented by the lack of experimental insights that shed light on the actual feasibility of this emerging network infrastructures. In this paper, we try to fill the gap in this respect by presenting an extensive empirical analysis performed through a vehicular system testbed. In particular, our work aims at providing empirical insights on the advantages that an edge cloud-based service provisioning can enable in comparison to a centralized cloud-based approach. Besides, by focusing only on the transmission of small-sized workload-i.e. with payload comparable to the one produced by In-Vehicle's sensors-this work also aims at evaluating the suitability of different application layer protocols (HTTP, CoAP, and MQTT) in this peculiar context. In the performance analysis, additional aspects have been also considered, including the impact of vehicle's speed as well as scalability issues.

AMSense: How Mobile Sensing Platforms Capture Pedestrian/Cyclist Spatiotemporal Properties in Cities

“…For instance, vehicles are typically not affected by strict memory, processing, storage, and energy limitations, which enables the integration of various sensors, wireless transmitters and processing components. In recent years, several advanced technologies such as lightweight virtualization and edge computing have been applied to smart vehicles to enable novel applications and dynamic service deployment [25], [31], [32].…”

Investigating Messaging Protocols for the Internet of Things (IoT)

“…There are a number of studies that have examined the performance and scalability of communication protocols. Some studies have found that the throughput in MQTT drops significantly as the number of clients' subscriptions increases [112], [113]. In [114], the authors compared the performance of DDS and MQTT.…”

An Intelligent Path Planning Scheme of Autonomous Vehicles Platoon Using Deep Reinforcement Learning on Network Edge

“…Nevertheless, offloading schemes [33], where all vehicles offload their tasks to the same VEC server, can limit the performance gain due to overload. Multi-Access Edge Computing (MEC) [34] is an emerging network paradigm that can be exploited also in vehicular scenarios to foster a more effective and flexible service delivery. Therefore, we combine edge computing and Q-learning into the platoon.…”