• Distributed cyber–physical systems
• Fault Tolerance
• Fog Computing
• Mixed integer linear programming
• Smart environments

Nowadays billions of devices are connected to the internet. The Cloud Computing model, in which data generated by devices are all transferred to geographically distant Clouds for processing and storage, is a problem for latency-sensitive applications.
The Cloud Computing architecture is extended with an additional layer called the Fog Layer, which decentralize the computing resources available in large data centers by distributing them towards the edge of the network closer to the end-users. Such Fog layer enables the execution of the application logic on processing nodes, usually high performance embedded computing boards, enabling a new breed of applications and services which exploits a fruitful interaction between Cloud and Fog.
The Fog Computing model can be used to generate distributed cyber-physical systems ensuring that application latency requirements are adequately met at runtime.
The design of such systems requires careful planning considering how the highly dynamic and heterogeneous nature of Fog nodes makes failures a popular event. It is therefore essential to implement a fault tolerance mechanism in such systems.
We propose an engine based on a Mixed Integer Linear Programming (MILP) formulation which takes in input a planimetry of the environment and a description of the applications and, based on a definition of which components of the application are critical and a repository of available processing boards, identifies a suitable architecture of the fault-tolerance cyber-physical system and the corresponding applications’ workload distribution in order to minimize the architectural cost.