Open Access

ARTICLE

Unmanned Aerial Vehicle Multi-Access Edge Computing as Security Enabler for Next-Gen 5G Security Frameworks

Jaime Ortiz Córdoba, Alejandro Molina Zarca^*, Antonio Skármeta

Department of Information and Communications Engineering, University of Murcia, Murcia, 30003, Spain

* Corresponding Author: Alejandro Molina Zarca. Email:

(This article belongs to the Special Issue: Advanced Achievements of Intelligent and Secure Systems for the Next Generation Computing)

Intelligent Automation & Soft Computing 2023, 37(2), 2307-2333. https://doi.org/10.32604/iasc.2023.039607

Received 07 February 2023; Accepted 05 May 2023; Issue published 21 June 2023

Abstract

5G/Beyond 5G (B5G) networks provide connectivity to many heterogeneous devices, raising significant security and operational issues and making traditional infrastructure management increasingly complex. In this regard, new frameworks such as Anastacia-H2020 or INSPIRE-5GPlus automate the management of next-generation infrastructures, especially regarding policy-based security, abstraction, flexibility, and extensibility. This paper presents the design, workflow, and implementation of a security solution based on Unmanned Aerial Vehicles (UAVs), able to extend 5G/B5G security framework capabilities with UAV features like dynamic service provisioning in specific geographic areas. The proposed solution allows enforcing UAV security policies in proactive and reactive ways to automate UAV dynamic deployments and provisioning security Virtual Network Functions (VNFs) in the onboard Multi-access Edge Computing (MEC) node. A UAV has been ensembled from scratch to validate the proposal, and a raspberry-pi has been onboarded as compute node. The implementation provides a VNF for dynamic UAV management, capable of dynamically loading waypoints into the flight controller to address reactive autonomous flights, and an ML-based VNF capable of detecting image patterns. According to the security policies, the onboard VNFs can be dynamically configured to generate alerts to the framework and apply local reactions depending on the detection made. In our experiments, we measured the time it takes for the solution to be ready after receiving a security policy for detecting patterns in a specific geographical area. The time it takes for the solution to react automatically was also measured. The results show that the proactive flow configuration considering ten waypoints can be enforced in less than 3 s, and a local reactive flow can be enforced in around 1 s. We consider that the results are promising and aligned with other security enabler solutions as part of existing 5G/6G security frameworks.

---

Keywords

---