An adaptive multi-cloud offloading using hierarchical game-theoretic approach

Abstract

A Multi-cloud framework leverages resource-constrained user equipment by offloading its complex application task to divergent resources according to its on-demand requirements. An interaction among three entities, such as user equipment, root server, and application vendor, is studied for adaptive offloading in this work. A hierarchical three-stage game setting analyzes the best suitable offload policy for each user. A sequential game-theoretic model is used to find the user equipment’s offload demand, optimal resource allocation at the root server, and the app vendor’s optimal profit. This framework incorporates the computation and communication effects within the entities and the decision overhead of each entity. A closed-form solution, i.e., equilibrium, is achieved at each stage of the game, and the proposed iterative algorithm obtains the entire Stackelberg equilibrium. Simulation results show that the utilities of users, root server, and app vendor have been improved considerably due to the optimal mapping of offload demands to multi-cloud resources.