A Distributed Decision Making Model for Cloudlets Network: A Fog to Cloud Computing Approach

Abstract

During the last decade, most Cloud-based applications have been designed based on a centralized architecture, built around a Cloud server, which orchestrates operations, saves data and makes decisions. Nowadays, industrial applications impose more efficiency and reliability which cannot be ensured by a centralized architecture. In fact, industrial environments connectivity, latency and storage constitute major problems for such applications. Recent years have witnessed the emergence of fog computing as a solution, deployed to offer better performance with respect to latency, elasticity, reliability and inter-operability as compared to a centralized architecture. Such solution relies on the use of cooperating Cloudlets that are bringing Cloud Computing services closer to the end-users. In this paper, we propose a new model for a distributed decision making solution in a Cloud environment based on an autonomic fog computing approach relying on self-knowledge techniques. In our work, we have used autonomic Cloudlets deployed as intelligent nodes in an industrial environment. Such Cloudlets are able to process some given tasks in order to make local decisions which will be used later by the Cloud server. Cloudlets are capable to manage local constraints like connectivity, latency, and synchronization with the Cloud. Moreover, the network situations are analyzed with respect to stated constraints using a case-based reasoning approach. Indeed, given the industrial setting, the appropriate type of connection available is automatically (WIFI, Ethernet, GPRS, 3G, etc.) The proposed architecture relies on proven technical solutions with respect to coordination, cooperation, autonomous behavior, communication, etc. The proposed architecture has been designed and deployed for validation, in an industrial company.