Effective Timing for Reserving Fog Computational Resources for Time-Sensitive Vehicular Applications

Fog computing has garnered significant attention, particularly for time-critical applications. Numerous techniques have been proposed to minimize offloading latency and optimize resource utilization. Unfortunately, these techniques are insufficient for safety-critical applications, which require deterministic/guaranteed access to Fog computational resources for clients at any time. Fog resources are perse limited, making the reservation of computation resources mandatory for time-sensitive applications. As resource pricing is dynamic, it is essential to carefully schedule when to make a needed reservation. This work proposes a novel approach for predicting the optimized timing to reserve a certain Fog computation resource with minimal costs. Experimental results with real-world traffic data reveal the effectiveness of our proposed technique for reducing resource costs compared to related reservation schemes. Additionally, our solution is frugal so that resource-constrained clients, such as a vehicle on-board computer, can easily operate it.