Cloud-Based Smart Energy Framework for Accelerated Data Analytics with Parallel Computing of Orchestrated Containers: Study Case of CU-BEMS

Abstract

This paper proposes a practical smart energy framework for data analytic on energy management system at Chulalongkorn University, called CU-BEMS. This serves as an example of demand-sided smart energy application that copes with the challenges of big data analytic and real-time processing needs. The framework is based on the divide and conquer paradigm to accelerate data analytics with parallel computing. The workload is containerized and deployed on the Kubernetes cloud facility of our internationally collaborated IoTcloudServe@TEIN playground. With this playground, the workload scalability and portability can be achieved. Applying the proposed framework, this paper reports on a practical data log analysis to determine the wasted energy consumption. Based on the experimental result, the wasted energy consumption of the whole data set of CU-BEMS's communication research laboratory area from March 2014 to August 2017 can be computed within 81 seconds by using 32 cores running in parallel. The framework is expected to serve as a basis template for further research ongoing at CU-BEMS and smart energy applications that can be computationally enhanced by data analytic pipelining with containerized services as orchestrated by Kubernetes.