Novel denial-of-service attacks against cloud-based multi-robot systems

Abstract

The development of robotics technology is accelerated by the strong support from cloud computing. Massive computation resources and services from the cloud make modern multi-robot systems more efficient and powerful. However, the introduction of cloud servers to multi-robot systems can also incur potential Denial-of-Service (DoS) threats, where an adversary can utilize the shared cloud resources to degrade or bring down the robot systems. In this paper, we conduct a comprehensive study about this security issue. By analyzing different attack vectors in cloud-robotic platforms, we propose three new DoS attacks, which manipulate the network resources, micro-architecture resources, and function parameters respectively. We conduct extensive evaluations and case studies to demonstrate the feasibility and severity of our techniques. We alert the robotics community to these catastrophic attacks on the safety and performance of cloud-robotic systems, and encourage building better defenses for higher reliability, in addition to automation and intelligence.