A Novel Fog Node Aggregation Approach for Users in Fog Computing Environment

Abstract

In order to provide highly accessible and efficient computing capacities and services, fog nodes in the fog computing are geographically distributed near end-users. Normally, a single fog node cannot provide sufficient capacities to deploy services because of its own limited resources. Therefore, this paper proposes a novel approach aggregating resources of fog nodes: multiple fog nodes are assumed to be one resource node hosting virtual machine where services could be deployed. Our proposed approach can be decomposed into two phases: i) seeking suitable fog nodes, where we formulate the corresponding overhead computation models in terms of data transmission latency and rental cost, to balance the minimum data transmission latency and the low rental cost, and ii) achieving enough fog nodes, where we propose a novel Fog node Selection algorithm based on Simulated Annealing to meet resource requirements of a single service while minimizing the overheads of the end-users. Experimental results show that the proposed approach performs better than existing start-of-the-art peers in terms of data transmission latency and rental cost while meeting resource requirements of services.