A Computationally-Efficient and QoS-Aware Data Offloading Framework for Biased Fog Networks

Fog computing alleviates the cloud-centric limitations of Internet of Things (IoT). However, in the dynamic landscape of fog computing, the uneven distribution of workload among fog nodes emerges as a substantial obstacle to both, data latency and network profit. To mitigate workload imbalances, data packet offloading offers a twofold benefit. The offloading fog node leverages latency satisfaction, while the recipient fog node gains a financial advantage by leasing out its available processing resources. Motivated by the aforementioned advantages, in this work, we propose a novel load-balancing method to maximize monetary gains without affecting the Quality-of-Service (QoS) constraints of the subscribed IoT users in a biased fog network. The proposed method introduces an Optimized Matching Theory (OMAT)-guided data offloading framework, employing many to many matching without externalities. The method returns a novel matching among disparate fog nodes thereby achieving uniform workload distribution. The obtained results demonstrate that the proposed method attains improved performance in terms of inverse latency, throughput, and non-matchings, when compared to existing methods in the literature.