Abstract
Cloud computing (CC) is one of the most popular technologies which provides on-demand ubiquitous services to the geo-located end users. Such services are hosted by physical infrastructure deployed at massive data centers (DCs) at various geographic locations. For handling millions of service requests, DCs consume a large amount of energy which increases the overall operational expenditure, grid load, and carbon footprints. So, to handle these issues, the integration of DCs with renewable energy sources (RES) is one of the solutions used by the research community from past many years. However, with an increase in smart communities such as smart cities, smart healthcare, and Internet of things (IoT), the dependence of users on CC has increased many folds. Hence, to deal with the service requests and the data generated from smart devices (vehicles, IoT sensors, and actuators) locally, a recent paradigm popularly known as Edge computing has emerged. But, to execute various applications smoothly, there is a movement of high volume of data across different geographically separated nodes which create a huge burden on the network infrastructure. Moreover, the interoperability of various mobile devices is one of the major concerns for effective network infrastructure usage. Therefore, to deal with above challenges, an emerging paradigm known as software defined networks (SDN) can be a suitable choice. Hence, in this paper, an efficient scheme for energy management with sustainability (MEnSuS) of Cloud Data Centers in Edge Cloud Environment using SDN is presented. In the proposed scheme, a support vector machine-based workload classification approach is presented. Moreover, a two-stage game for workload scheduling for sustainability of DCs is designed. In order to achieve energy efficiency and optimal utilization of network and computing resources, different consolidation schemes are also presented. The proposed scheme is evaluated using Google workload traces and the results obtained prove the effectiveness. (C) 2017 Elsevier B.V. All rights reserved.