Systems-of-Systems (SoS) result from the collaboration of independent Constituent Systems (CSs) to achieve particular missions. CSs are not totally known at design time, and may also leave or join SoS at runtime, which turns the SoS architecture to be inherently dynamic, forming new architectural
Configurations and impacting the overall system quality attributes (i.e. performance, security and reliability). Recently these systems have emerged in the form of Internet of Things (IoTs) and Cyber-Physical Systems (CPS), in a range of domains like health, transportation, energy, defence and natural disaster management. As such, these systems play a critical role in the success of socioeconomic and business processes, leaving an impact on human lives. it is vital to model and evaluate the impact of these stochastic architectural changes on SoS properties at abstract level at the early stage in order to analyze and select appropriate architectural design. Architectural description languages (ADL) have been proposed and used to deal with SoS dynamic architectures. However, these techniques lack the expressive power of syntax and formal semantics to model and reason about dynamic architecture, resulting configurations and their impact on system quality attributes.
This research proposes a model-based framework to specify and reason about the dynamic archiecture of SoS. The existing formalism of process algebra shall be extended by incorporation of semantics to deal with stochastic dynamic architecture at runtime. Formal models shall be transformed into stochastic models to generate possible system configuration state transitions. The proposed thesis validates the approach with simulation of benchmark SoS case studies with possible configurations and the impact of these structural changes on system quality attributes and ultimately provides alternative architectures to system architects to make better design decisions.
Conference Publications/ Presentations
Dr. Naeem Khalid Janjua, Dr. Syed MS Islam, Dr. Martin Masek
Please leave a comment about your rating so we can better understand how we might improve the page.