Traditionally, critical systems have been deployed in isolation, that is, in closed environments where the access control was easily managed. However, the increasing complexity and connectivity of these systems make them vulnerable to cyberattacks, malfunctioning or any kind of uncontrolled events. In this work, we propose a framework that is capable of automatically adapt its configuration for addressing the challenges of an environmental change. To this end, we model the critical system as a Petri net which is enriched with product lines that implement actions for diferent scenarios. The execution traces are then continuously monitored and provide information to the control logic responsible for achieving the critical system goals by means of the product lines.
Traditionally, critical systems have been deployed in isolation, that is, in closed environments where the access control was easily managed. Recently reports indicate that these systems are vulnerable to cyberattacks as well [1]. Security flaws may directly impact safety in critical systems. Current approaches and tools concerning security do not ensure their adequacy to industrial standards for safety level.
We combine the following concepts for our approach: 1) Petri nets as formal model, 2) Product Lines for designing, and 3) Runtime verification for system monitoring.
A Petri net will model the critical system, where each product line implements a specific configuration. Depending on the Key Performance Indicators (KPI) defined by the customer and the measurements the monitor extracts by simulation of the formal model, our new framework checks if the critical system will manage to achieve the user requirements by switching on/of the product lines on runtime.
We are basing our approach on the following previous works: • A framework for modelling product lines with Petri nets, using an eclipse plug-in, called Titan [2]. • A framework for abstracting data and datatypes as colours and hierarchies in Coloured Petri Nets [3, 4]. • A transformation of product Lines with Petri nets into Coloured Petri Nets [5]. • A language for the specification and runtime verification of systems [6]. safety and security requirements and system characteristics in a single picture. • automatically adapt the system configuration based on the simulation reports.
This work was supported by the Spanish Ministry of Science and Innovation under projects FAME (RTI2018093608-B-C31), MASSIVE (RTI2018-095255-B-I00) and the Comunidad de Madrid under project FORTE-CM (S2018/TCS-4314) co-funded by EIE Funds of the European Union. [1] T. Telford, Cyber attacks on rail network, Computing
Security . 3. Approach [2] E. Gómez-Martínez, J. de Lara, E. Guerra, Extensible Structural Analysis of Petri Net Product Lines, Trans.
In contrast to [2], our new approach includes a Trans- Petri Nets Other Model. Concurr. 15 (2021) 27–49. formation phase, which maps product lines to colours in [3] M. Westergaard, L. M. Kristensen, The Access/CPN Coloured Petri Nets [3]. Framework: A Tool for Interacting with the CPN
Then, the Simulation phase runs the model, which now Tools Simulator, in: G. Franceschinis, K. Wolf (Eds.), supports all the features that the Access/CPN engine Applications and Theory of Petri Nets, 30th Internaprovides such as timed information. tional Conference, PETRI NETS 2009, Paris, France,
The Monitoring phase, aimed at detecting concurrence June 22-26, 2009. Proceedings, vol. 5606 of Lecture and performance issues, analyses the simulation traces Notes in Computer Science, Springer, 313–322, 2009. via TeSSLa [6] in order to report performance and con- [4] K. Jensen, L. M. Kristensen, L. Wells, Coloured Petri currence reports. Nets and CPN Tools for modelling and validation
During the Assessment phase, the framework will au- of concurrent systems, Int. J. Softw. Tools Technol. tomatically turn on/of specific product lines in order to Transf. 9 (3-4) (2007) 213–254. achieve the target KPI. [5] E. Gómez-Martínez, J. de Lara, E. Guerra, Analysing Product Lines of Concurrent Systems with Coloured Petri Nets, in: Accepted in the 34th International
Conference on Software Engineering and Knowledge Engineering, 2022. [6] M. Leucker, C. Sánchez, T. Schefel, M. Schmitz, A. Schramm, TeSSLa: runtime verification of nonsynchronized real-time streams, in: H. M. Haddad, R. L. Wainwright, R. Chbeir (Eds.), Proceedings of the 33rd Annual ACM Symposium on Applied Computing, SAC 2018, Pau, France, April 09-13, 2018, ACM, 1925–1933, 2018.
We are currently working on two directions. Firstly, we are extending the PNPL modelling framework capabilities to support additional features such as timed information. Secondly, we explore how to automatically (de)activate product lines based on the monitoring of Key Performance Indicators (KPI). In particular, we are connecting Titan with 1) AccessCPN, the kernel of CPNTools, for simulation purposes; and 2) TeSSLa, a runtime monitoring engine, for providing information to the control logic responsible for managing the product lines.
In this work in process, our aim is to: • augment our framework for enabling software engineers to express and monitor performance,