| Title | Cybersecurity Modelling for SCADA Systems: A Case Study |
| Publication Type | Conference Paper |
| Year of Publication | 2022 |
| Authors | Cheng, Benny N. |
| Conference Name | 2022 Annual Reliability and Maintainability Symposium (RAMS) |
| Keywords | attack and countermeasure trees, compositionality, Computational modeling, cybersecurity, Data models, Internet security, maintenance engineering, MTTSF, pubcrawl, Random access memory, resilience, Resiliency, SCADA systems, SCADA Systems Security, Sensitivity |
| Abstract | This paper describes a cybersecurity model for Supervisory Control and Data Acquisition system (SCADA) using techniques similar to those used in reliability systems modelling. Previously, cybersecurity events were considered a part of the reliability events of a cyber physical system [1] [2]. Our approach identifies and treats such events separately as unique class of events by itself. Our analyses shows that the hierarchical model described below has the potential for quantifying the cybersecurity posture of a SCADA system, which goes beyond the usual pass/fail metrics that are currently in use [3]. A range of Mean Time to Security Failure (MTTSF) values as shown in the sensitivity studies below can capture both peacetime and wartime cyber risk assessment of the system. While the Attack and Countermeasure Tree (ACT) constructed below could be taken as somewhat simplistic, more detailed security events can be readily introduced to the ACT tree to reflect a better depiction of a cyberattack. For example, the Common Processing Systems (CPS) systems themselves can be further resolved into constituent components that are vulnerable to cyberattacks. Separate models can also be developed for each of the individual failure events, i.e. confidentiality, integrity, and availability, instead of combining them into one failure event as done below. The methodology for computing the MTTSF metric can be extended to other similar cybersecurity metrics, such as those formulated by the Center for Internet Security (CIS) [3], e.g. mean time to restore to operational status, etc. Additional improvements to the model can be obtained with the incorporation of the repair and restore portion of the semi-Markov chain in Figure 3, which will likely require the use of more advance modeling packages. |
| Notes | ISSN: 2577-0993 |
| DOI | 10.1109/RAMS51457.2022.9893946 |
| Citation Key | cheng_cybersecurity_2022 |
Cybersecurity Modelling for SCADA Systems: A Case Study