Modeling and evaluating a complex edge computing based systems: An emergency management support system case study

The edge computing paradigm enhances the exploitability of cloud computing by providing, in principle, the means to overcome the limitations in terms of responsiveness, bandwidth needs, privacy and availability in critical applications. By moving some of the components of the system towards the physical location where results are timely needed, it is possible to support mission critical applications and back them up with the flexibility and the resource availability and scalability offered by the cloud, while keeping mission costs lower (and response efficiency higher) than classical approaches. The main challenge in merging cloud and edge components lies in their correct balance to allow for the best results at the lowest costs. This means that performance-oriented evaluation models are crucial in the design, deployment and execution phases of the system lifetime. In this paper we present a modeling approach for complex, critical-edge computing-based systems relying on the use of queuing networks, applied to a novel architecture aiming at supporting operations in case of medium or large-scale accidents that involve interoperability among responders during the emergency phase. The proposed architecture allows the coordination of fire brigade teams, equipped with sensors and augmented reality devices, to minimize mission problems and timely exploit local and external information, as well as supporting interoperations with other first responders. The results show that, even with a standard modeling approach, these systems are extremely interesting and show non-trivial behaviors. Security and dependability issues of the proposed architecture are discussed.