Multiobjective Placement for Secure and Dependable Smart Industrial Environments

Cyber-physical systems allow to implement efficient, highly-automated, and green smart industrial environments. To this aim, computation is a critical asset to control machineries, process data, and run proper optimization strategies. In general, computing capabilities are provided by virtualizing resources of local devices or servers deployed in a remote datacenter. Thus, their management is crucial for the security of the cyber-physical ecosystem or to engineer dependable industrial environments. In this perspective, we introduce a placement mechanism to pursue various security and dependability goals, such as prevent information leakages, provide redundancy, counteract hardware outages or software aging, avoid privacy breaks, and mitigate wastage of resources and energy. Multi-objective placement actions are computed by solving suitable mathematical program- ming problems with competing objectives. Simulations performed using real workload traces showcase the effectiveness of the approach in comparison with bin-packing-based heuristics.