Lightweight, Scalable and Secure Middleware for Service-Centric Context-Aware Applications

Abstract

Wireless sensor networks (WSNs) play a vital role in the Internet of things (IoT)-based application. IoT environment contains monitoring and collaboration from a wide number of remote or local sensor actuators, a device such as smartphones, robots and computers. In this paper, an efficient REST-based lightweight scalable and security approach proposed for providing secure data exchange without compromising the public quality issues. The communication scheme purely centred on the Extensible Messaging and Presence Protocol (XMPP) protocol. XMPP supports push-based notification that will be more proper functionality for any event-driven model for any sensor-based application. Based on these features, XMPP protocol, which integrated with the proposed approach, assures a scalable real-time event notification scheme. The framework is also adopting the light and portable Representational State Transfer (REST) web support features and consequently integrates sensor and intelligent things with the web. Here, the use of convention JavaScript Object Notation (JSON) format which will be the alternative for the verbose XML for data exchange. It also provides an encryption and data authentication required during transmission. The JSON is used along with the XMPP to support remote web services over wireless sensor networks. Further, the proposed approach is used to realize a secured event-driven smarter environment. The proposed approach validated by simulating the number of concurrent requests, and this measures their corresponding CPU and memory utilization along with the additional overhead because of encryption. The proposed approach consumes 54% less CPU and five times more memory (i.e. because of encryption) than the traditional HTTP requests. Besides, the proposed encryption mechanism REST-JSON with Bcrypt consumes 7% more time than REST-JSON MD5/SHA1 to serve the requests and provide improvised security against dictionary attacks which is common in MD5/SHA1.