Advance of card-computing turned measurement and control systems into distributed hierarchical ones. The availability of local computing power gives sensors intelligence and delegates data preprocessing to them. The tasks underlying no or only soft real-time constraints but demanding a higher computational capacity deployed to the cloud or even include external resources for the computations. These resources form a system, which called cyber-physical system. For the individual resources, and for the entire system, it is required to be fault tolerant.
One of the aims of the thesis is presenting the main definitions and context of dependability and security according to the fault-tolerant systems using heterogeneous resources. The thesis covers the details how the fault - error - failure mechanism acts, how can error propagation be prevented in case of system design and operation.
A new system designing paradigm referred to as fog computing can be applied to the design of cyber-physical systems. This methodology will be standardized in the future. According to this, the thesis includes a summary of the OpenFog reference architecture. The architecture will be important in the future according to the cyber-physical systems too because the OpenFog Consortium intends to create a unified standard for this.
After that, the thesis presents systematic designing process for the example of a pilot system is presented, which aims to solve a low complexity but practically relevant problem. It begins with the thoroughly specified requirements, then the functional designing follows, and at the end the functions allocated to resources. The designated system is a 3-layered system which combines a local card-computer, a local control machine and cloud service as heterogeneous resources.
The thesis details the advantages and disadvantages of the individual resource types, moreover it includes a recommendation for the candidate technologies.
The last chapter contains a short evaluation of the work and describes potential future work.