Contents lists available at ScienceDirect Journal of Network and Computer Applications journal homepage: www.elsevier.com/locate/jnca An intelligent power distribution service architecture using cloud computing and deep learning techniques Weishan Zhang a, ⁎ , Gaowa Wulan a , Jia Zhai b , Liang Xu a , Dehai Zhao a , Xin Liu a , Su Yang c , Jiehan Zhou d a Department of Software Engineering, China University of Petroleum, No. 66 Changjiang West Road, Qingdao266580, China b Science and Technology on Optical Radiation Laboratory, Beijing 100854, China c College of Computer Science and Technology, Fudan University, Shanghai 200433, China d University of Oulu, Finland ARTICLE INFO Keywords: LSTM MQTT Power distribution SVR Prediction Cloud computing XGBoost ABSTRACT Smart management of power consumption for green living is important for sustainable development. Existing approaches could not provide a complete solution for both smart monitoring of electricity consumption, and also intelligent processing of the collected data effectively. This paper presents a cloud-based intelligent power distribution service architecture, where an intelligent electricity box (IEB) is designed using Zigbee and Raspberry Pi, and a standard MQTT (Message Queuing Telemetry Transport) protocol is used to transfer monitored data to the backend Cloud computing infrastructure using open source software packages. The IEB provides cloud services of real-time electricity information checking, power consumption monitoring, and remote control of switches. The current and historical data are stored in HBase and analyzed using Long Short Term Memory (LSTM). Evaluations and practical usage show that our proposed solution is very efficient in terms of availability, performance, and the deep learning based approach has better prediction accuracy than that of both classical SVR based approach and the latest XGBoost approach. 1. Introduction Smart building aims to improve quality of life, and working quality by providing convenient and comfortable smart environment (Al- Fuqaha et al., 2015). Traditionally, smart building concentrates on using smart devices and equipment to build and to provide smart services. With the emerging requirements of green life and energy saving, more efforts should be put on managing power distribution and consumption in a smart building, in order to know in advance potential problems such as unusual power usage, power leakage, and so on. In order to effectively manage electric power systems, the first thing is to collect power usage status, which is achieved with smart metering in an intelligent electricity box (IEB) (Bahmanyar et al., 2016; Depuru et al., 2011). But these IEBs are not fully making use of existing Internet of Things protocols in order to provide effective control and management services, in different network environment. After the collection of the corresponding data, there raises the issue of how to effectively manage, store, process these data. Cloud Computing begins to be used as an effective solution for resolving this issue, as discussed in Grozev et al. (2016). A Smart Home Electricity Management System (SHEMS) using Cloud Computing was proposed in Garcia et al. (2013) to manage power usage data. However, these two didn't provide details on how cloud computing was used to realize scalable data collection, storage, and processing. We are looking for an effective solution scalable to handle large amount of real time collected data. At the same time, the solution should be affordable using open source software packages. On the other hand, dedicated data mining framework is emerging for electricity power consumption analysis using meter data (Silva et al., 2011). And some begin to use Cloud services to identify outage sources and fault localization, capturing peaks and repeat “offenders” of circuits and transformers, and so on as in Lang et al. (2016). However, the existing data analysis can not accurately predict power usage trend, the existing solutions can be improved by the latest deep learning techniques, as shown in our previous work on analyzing resource requests in a data center (Zhang et al., 2017). Considering all the above mentioned issues, this paper presents a comprehensive solution and architecture on intelligent power distribu- tion management services using existing network protocols, where a smart power distribution box is designed to take place of conventional http://dx.doi.org/10.1016/j.jnca.2017.09.001 Received 31 January 2017; Received in revised form 19 August 2017; Accepted 1 September 2017 ⁎ Corresponding author E-mail address: zhangws@upc.edu.cn (W. Zhang). Journal of Network and Computer Applications xxx (xxxx) xxx–xxx 1084-8045/ © 2017 Published by Elsevier Ltd. Please cite this article as: Zhang, W., Journal of Network and Computer Applications (2017), http://dx.doi.org/10.1016/j.jnca.2017.09.001