| Popis: |
Real-time business practices require huge amounts of data directly from the production assets. This new thirst for accurate and timely data has forced the convergence of the traditionally business-focused information technology (IT) environment with the production-focused operational technology (OT). Recently, software-defined network (SDN) methodologies have benefitted OT networks with enhanced situational awareness, centralized configuration, deny-by-default forwarding rules, and increased performance. What makes SDNs so innovative is the separation between the control plane and the data plane, centralizing the command in the controllers. However, due to their young age, the use of SDNs in the industry context has not yet matured comprehensive SDN-based architectures for IT/OT networks, which are also resistant to security attacks such as denial-of-service ones, which may occur in SDN-based industrial IoT (IIoT) networks. One main motivation is that the lack of comprehensive SDN-based architectures for IT/OT networks making it difficult to effectively simulate, analyze, and identify proper detection and mitigation strategies for DoS attacks in IT/OT networks. No consolidated security solutions are available that provide DoS detection and mitigation strategies in IT/OT networks. Along this direction, this paper’s contributions are twofold. On the one hand, this paper proposes a convergent IT/OT SDN-based architecture applied in a real implementation of an IT/OT support infrastructure called SIRDAM4.0 within the context of the SBDIOI40 project. On the other hand, this paper proposes a qualitative analysis on how this architecture works under DoS attacks, focusing on what the specific problems and vulnerabilities are. In particular, we simulated several distributed denial-of-service (DDoS) attack scenarios within the context of the proposed architecture to show the minimum effort needed by the attacker to hack the network, and our obtained experimental results show how it is possible to compromise the network, thus considerably worsening the performance and, in general, the functioning of the network. Finally, we conclude our analysis with a brief description on the importance of employing machine learning approaches for attack detection and for mitigation techniques. |
