“Securing Manufacturing Systems”

11/10/2023

Pendant la semaine du 9 au 13 octobre 2023, l’ICO a eu le plaisir d’accueillir le Dr. John Jeffery Prevost, Prof. à l’Université de Texas – San Antonio (États-Unis), qui a préparé un séminaire ouvert à tous intitulé “Securing Manufacturing Systems”, présenté au LIRMM (Montpellier) le 11 octobre et à l’UT3 (Toulouse) le 13 octobre, dans le cadre de sa visite scientifique en Occitanie.

Programme :

Abstract of the Seminar

As the executive director for the Open Cloud Institute (OCI), and the vice-president for secure cloud architecture for the Cyber Manufacturing Innovation Institute (CyManII), my research focuses on challenges related to cloud and edge computing and the Internet of Things. A recent thrust is in how to leverage advanced analytics to assist manufacturers without creating new attack vectors which could be exploited by bad actors. Often, the platforms used in manufacturing were created before security became a motivating force. These systems provide many pathways for an adversary to perform an exploit, resulting in manufacturing downtime or theft of intellectual property. The Cloud Lab for Engineering Application Research (CLEAR), a part of the OCI, performs research to assist CyManII in understanding fundamental properties of cloud and edge computing that lead to implementations that can be tested in facilities such as OakRidge National Lab’s Manufacturing Demonstration Facility. The goal is to provide a commercial pathway to move the research out of the lab and into the manufacturing plant as quickly as possible. We will be presenting two research activities in this area. The first is the evolution of an acoustic analysis framework that can ‘listen’ to the manufacturing activity for anomalous activity (eg. Denial of service, GCode replacement and direct register over right attack). We will then turn our attention to device level attacks on FPGA boards. FPGA’s are commonly used in manufacturing to provide for real-time control of the manufacturing process and are often unsecured. We will show our research in how we are using a side-channel attack pattern to determine cryptographic keys in an isolated block of the FPGA. This has applications to services such as cloud-hosted FPGAs being currently offered by companies such as Intel.