Cooperative and Parallel Fog Discovery and Pareto Optimal Fog Commerce Bargaining

While fog resource discovery involves propagating resource requests to a huge number of fog nodes, fog commerce refers to the activity of buying and selling fog resources. This work devises: 1) the KM-gossip algorithm for bolstering fog discovery and 2) a bargaining mechanism for pricing fog resources. The KM-gossip algorithm is a generalization of the gossip algorithm. It uses $K$ broker agents (BAs) to cooperatively “gossip” requests among themselves, and in parallel, each BA relays the requests to $M$ fog nodes. While computational complexity analysis validates that the KM-gossip algorithm has logarithmic time complexity (i.e., it is computationally efficient), empirical results show that it significantly outperforms the existing gossip and flooding algorithms. Supplementing and complementing previous empirical results, game-theoretic analysis in this work validates that the bargaining mechanism generates Pareto optimal (economically efficient) solutions. The solutions also satisfy the famous Nash’s axioms that prescribe highly desirable properties of bargaining solutions.