Security is becoming increasingly relevant for the secure and resilient operation of industrial production facilities. In particular, attacks, e.g. with malware such as ransomware, pose an existential threat to companies of all sizes and domains. In addition, security is increasingly becoming a selling point for components and systems used in industry. A key element in the field of security is security assessments based on a risk evaluation by the manufacturer of the components or the operator of the systems. However, these are currently carried out manually by experts, which makes them time-consuming and resource-intensive, as well as heavily dependent on the individual experience of the respective expert. Small and medium-sized enterprises (SMEs) in particular are often unable to afford their own experts with the necessary security expertise. In addition, the required experts are often not available on the market. Within the scope of the proposed project, a software- and machine-learning-based support of risk assessments for industrial production facilities shall be developed in order to relieve experts of routine tasks, to reduce the costs for security risk assessments, to increase the degree of automation and to ensure a uniform quality of the assessments by formalizing the expert knowledge.

Security assessments to evaluate the security of industrial production facilities are always based on a risk assessment by the manufacturer of the components or the operator of the systems. Threats, vulnerabilities and resulting risks are identified, analyzed, assessed and reduced to an acceptable residual risk through risk mitigation. Due to dynamic changes in machines and systems and information regarding threats and vulnerabilities, which form the basis for risk assessments, the examination of the security of a production system must be carried out at ever shorter intervals. However, the current analyses in the field of security are very time-consuming and resource-intensive, as they are currently performed exclusively manually by experts and are heavily dependent on the respective experience and specific prior knowledge of the experts. As a result, the results are sometimes not objective and the quality varies greatly.

The goal of the SUSI project is to provide software-based support for risk assessments for industrial production plants and components in the future in order to increase the degree of automation, relieve specialists of routine tasks and make faster decisions regarding safety. This also makes it possible to improve and standardize the quality of the results. 

In the project, a distinction is made between two use cases. In the first use case, "Secure Components," a solution for the secure development process of components is developed as a basis. In a second use case, "Safe Systems," a solution is developed at the system level, e.g., for production facilities. Here, software-based support and automation of monitoring of the current threat situation and vulnerabilities of the components used in the system will be developed to enable permanent and timely reassessment of security.

The added value of the SUSI project lies in a reduction of the effort required for security risk assessments in industry, thus relieving SMEs in particular. The use of algorithms from the field of machine learning (ML) to increase the degree of automation and thus relieve experts is very promising in this context. Formalization of expert knowledge also ensures uniform quality of such security risk assessments.