Developing computer systems in critical embedded systems is a complex task. Dependability analysis is a way to verify our system and get details about its reliability and availability properties. I have defined a textual language, which is capable of describing a system with component level dependability properties. This language can be transformed to stochastic Petri-net, which provides a formal representation of error occurrence, propagation and recovery. This Petri-net can be used with the PetriDotNet tool for system dependability analysis. I have developed a transformation method for mapping the textual model to stochastic Petri-net. Each component entity is transformed to a subnet, and these subnets are connected together through the shared places in the model.
In this Thesis I review the capabilities of the AADL and CONCERTO framework for dependability analysis and using these observations define the dependability modeling language. I also construct the language transformation to stochastic Petri-net, and provide a demonstrating example for the technical usage.
For demonstrating the concrete usage of the language I choose a Telecare system, which is a healthcare system. This can be used for monitoring a person with medical sensors, and to collect these data by an Android phone or tablet. The system has a central server, which stores data, and can provide various measurements for doctors or family members. In this system the dependability is a critical property, and dependability analysis is supported by my modeling and analysis method.