A routing protocol for smart cities: RPL robustness under study

Abstract

Smart cities could be defined as developed urban areas, creating sustainable economic development and high quality of life by excelling in multiple key areas such as transportation, environment, economy, living, and government. This excellence could be reached through efficiency based on the intelligent management and integrated Information and Communication Technologies (ICT). Motivations. In the near future (2030), two thirds of the world population will reside on a city, thus increasing drastically demands on city infrastructures. As a result, urbanization is becoming a crucial issue. The Internet of Things (IoT) vision foresees billions of devices to form a worldwide network of interconnected objects including computers, mobile phones, RFID tags and wireless sensors. In this study, we focus on Wireless Sensor Networks (WSNs). The WSNs are a specific technology suitable to create Smart Cities. A distributed network of intelligent sensor nodes could measure numerous parameters and communicate them wirelessly and in real-time and makes possible a more efficient management of the city. For example, the pollution concentration in each street can be monitored, water leaks can be detected or noise maps of the city obtained. The number of applications with WSNs available for smart cities is only bounded by imagination: environmental care, sustainable development, healthcare, efficient traffic management, energy supply, water management, green buildings, etc. In short, WSN could improve the quality of life in a city. Scope. However, such urban applications often use multi-hop wireless networks with high density to obtain sufficient area coverage. As a result, they need networking stacks and routing protocols that can scale with network size and density, while remaining energy-efficient and lightweight. To this end, the IETF RoLL working group has designed the IPv6 Routing Protocol for Low-Power and Lossy Networks (RPL). This paper presents experimental results on the RPL protocol. The RPL properties in terms of delivery ratio, control packet overhead, dynamics and robustness are studied. The results are obtained by several experimentations conducted on two large WSNs testbeds composed of more than 100 sensor nodes each. In this real-life scenario (high density and convergecast traffic), several intrinsic characteristics of RPL are underlined: path length stability but reduced delivery ratio and important overhead (Fig. 1). However, the routing metrics, as defined by default, favor the creation of "hubs", aggregating most of 2-hops nodes (Fig. 2). To investigate the RPL robustness, we observe its behavior when facing a sudden death of several sensors and when several sensors are redeployed. RPL shows good abilities to maintain the routing process despite such events. However, the paper highlights that this ability can be reduced if only few critical nodes fail. To the best of our knowledge, it is the first study of RPL on such large platform.