Benchmarking 5G MEC and Cloud infrastructures for planning IoT messaging of CCAM data

TL;DR

This paper benchmarks 5G MEC and Cloud infrastructures to evaluate their performance in supporting IoT messaging for CCAM data, focusing on scalability and data type handling for vehicular applications.

Contribution

It introduces a novel benchmarking methodology for assessing MEC and Cloud platforms specifically for multi-type vehicular IoT data in CCAM scenarios.

Findings

MEC and Cloud platforms exhibit different scalability profiles.

Performance varies significantly with data volume and concurrency levels.

Benchmark results inform optimal infrastructure sizing for vehicular IoT applications.

Abstract

Vehicles embed lots of sensors supporting driving and safety. Combined with connectivity, they bring new possibilities for Connected, Cooperative and Automated Mobility (CCAM) services that exploit local and global data for a wide understanding beyond the myopic view of local sensors. Internet of Things (IoT) messaging solutions are ideal for vehicular data as they ship core features like the separation of geographic areas, the fusion of different producers on data/sensor types, and concurrent subscription support. Multi-access Edge Computing (MEC) and Cloud infrastructures are key to hosting a virtualized and distributed IoT platform. Currently, the are no benchmarks for assessing the appropriate size of an IoT platform for multiple vehicular data types such as text, image, binary point clouds and video-formatted samples. This paper formulates and executes the tests to get a benchmarking of the performance of a MEC and Cloud platform according to actors' concurrency, data volumes and business levels parameters.