Formal model of Edge–Fog–Cloud-based data distribution

Abstract

Smart-city and IoT systems generate heterogeneous data whose latency, sensitivity and computational demands cannot be met efficiently by a single computing tier. This paper formalizes data distribution across a hierarchical Edge–Fog–Cloud architecture, represented as a labeled graph whose nodes are partitioned into edge, fog and cloud layers with distinct capacities and latency ranges. A layer-assignment function maps each data unit to a processing layer according to its latency requirement, sensitivity level, computational load and current resource utilization. Four categories of distribution criteria – temporal, semantic, security and resource – are formalized, together with a weighted multi-criteria score function and a dynamic adaptation mechanism, providing a theoretical basis for latency-aware, privacy-preserving data placement.