RELYonIT: Dependability for the Internet of Things
The Internet of Things will represent the backbone of modern society and will embrace a system of wireless networks delivering end users a plethora of attractive services and applications. The latter will rely heavily on the dependable operation of embedded wireless sensors and actuators despite their low-priced hardware, battery limitations, and resource constraints. Wireless systems employed, for example, to build smart cities, smart grids, and smart healthcare applications, represent a long-term investment and are therefore required to reliably convey information about the state of things and places, as well as to remain available for a prolonged time.
Failure in delivering the sensed values in a reliable and timely fashion or in minimizing battery depletion may result in high costs, insufficient user satisfaction, and physical damage to people or things. As developing such systems is challenging, we see few dependable IoT applications today.
Most of these wireless systems have very limited computation capabilities and their transmitting range is often restricted by their limited energy budget. Therefore, it is fundamental that they can form unsupervised networks that autonomously adapt to environmental changes in an energy-efficient fashion. This is however very challenging to attain, as the surrounding environment often affects the performance of networked sensors or actuators and their capability to meet application-specific dependability requirements.
Radio interference from surrounding wireless devices and electrical appliances may impair packet reception, reduce throughput, lead to high latencies, and cause retransmissions that may cause earlier battery depletion. The mobility of nodes in dynamic environments can affect the forwarding of information in multi-hop networks and cause long delays and network partitions. Even temperature variations can cause, among others, loss of synchronization and degradation of the wireless link quality, affecting the reliability and availability of networked embedded systems. Figure 1 shows that the normal on-board temperature fluctuations occurring during a day can indeed transform a perfect link (100% packet reception rate) between two wireless sensors into an almost useless one.
Figure 1: A high temperature has a strong negative effect on the packet reception rate
This vulnerability to the surrounding environment affects the design of IoT applications, and the vast majority of the latter are still non-critical, i.e., they are not able to give any dependability guarantee. How can developers create smart city solutions if parking spot occupancy and pollution concentration sensors are not operating as expected during the hottest times of the day or in the presence of radio interference in dense urban environments?
Solutions require proper test bed infrastructures
To solve this problem, there is a strong need to get a better understanding of the performance of networked embedded devices as a function of the environment. In order to obtain such a deep understanding and to come up with dependable solutions, we have extended existing test bed infrastructures to enable the repeatable playback of pre-recorded environmental conditions. In particular, we have designed low-cost extensions for wireless sensor network test beds that allow to (i) study the impact of temperature with only little hardware overhead (TempLab), and (ii) to create realistic and repeatable interference patterns without hardware overhead (JamLab). These test bed extensions allow us to rerun experiments under identical environmental conditions and hence play a crucial role for the investigation of network performance.
TempLab (see Figure 2) can accurately reproduce temperature traces recorded in outdoor environments with an average error of only 0.1°C, and preliminary experiments using TempLab have revealed several limitations of state-of-the-art communication protocols. For example, high temperature can drastically change the topology of a network and lead to network partitions, reduce significantly the performance of MAC protocols, as well as increase the processing delay in the network. JamLab provides simple models to emulate the interference patterns generated by typical appliances, as well as a playback capability to regenerate recorded interference patterns. Through a simple software upload, JamLab selects automatically the portion of the test bed nodes that reproduce radio interference and allows us to study how protocols can cope with and adapt to the existing interference autonomously.
We used these test bed facilities to devise accurate models capturing the impact of the environment on IoT hardware and protocols, and to develop new protocols that can be automatically configured to meet application-specific dependability requirements. In particular, some of our communication protocols anticipate and correct faults introduced by environmental conditions [1,2] and significantly help in increasing (i) the reliability of low-power wireless communications and (ii) the availability of the network.
More dependability, more business
These solutions have been developed in the context of the EU-funded RELYonIT project  and have been shown to significantly increase the dependability of IoT solutions to radio interference and temperature fluctuations. Industrial partners in the consortium used these technologies to develop a new line of dependable real-world applications involving different types of sensors and actuators. These applications include highly demanding monitoring scenarios such as smart parking systems and energy-efficient buildings. We expect that our solutions will pave the way for a broader range of dependable IoT applications in critical domains such as smart cities, smart grids, and smart healthcare.
Figure 2: Heating lamp on top of a wireless sensor node in our extended TempLab test bed
 C.A. Boano, M.A. Zúñiga, K. Römer, and T. Voigt. JAG: Reliable and Predictable Wireless Agreement under External Radio Interference. In Proc. of the 33rd Real-Time Systems Symposium (RTSS). December 2012
 C.A. Boano, K. Römer, and Nicolas Tsiftes. Mitigating the Adverse Effects of Temperature on Low-Power Wireless Protocols. In Proc. of the 11th Conference on Mobile Ad hoc and Sensor Systems (MASS). October 2014
 RELYonIT: Research by Experimentation for Dependability on the Internet of Things, http://www.relyonit.eu, November 2014
Carlo Alberto Boano (IEEE member since 2009) has been a research assistant at the Institute for Technical Informatics of TU Graz, Austria since September 2013. Previously, he was affiliated with the University of Lübeck, Germany and the Swedish Institute of Computer Science, Sweden. He received a doctoral degree with distinction from TU Graz in 2014, with a thesis on dependable wireless sensor networks. Dr Boano's research interests include the design and implementation of dependable networked embedded systems, with emphasis on energy-efficient and reliable low-power wireless communications.
Kay Römer is currently a Professor and head of the Institute for Technical Informatics at TU Graz. Before he held positions of Professor at the University of Luebeck in Germany, and senior researcher at ETH Zurich in Switzerland. Prof Römer received his doctoral degree from ETH Zürich in 2005 with a thesis on wireless sensor networks. His research interests encompass design, implementation, and evaluation of networked embedded systems, with current focus on increasing their dependability, efficiency, and scalability.
Thiemo Voigt leads the Networked Embedded Systems Group at SICS Swedish ICT, formerly Swedish Institute of Computer Science. He is also a Professor at Uppsala University from where he received his PhD. His research interests include networking, systems and security issues in wireless networks and the Internet of Things. He is a member of IEEE.
Sign Up for IoT Technical Community Updates
Calendar of Events
IEEE 7th World Forum on Internet of Things (WF-IoT) 2021
14 June-31 July 2021
Call for Papers
IEEE World Forum on the Internet of Things (WF-IoT) 2021
Submission Deadline: 22 March 2021
Artificial Intelligence-Based Systems for Industrial Internet of Things and Industry 4.0
Submission Deadline: 1 May 2021
Special Issue on Sustainable Solutions for the Internet of Things
Submission Deadline: 15 April 2021
Special Issue on Graph-powered Machine Learning for Internet of Things
Submission Deadline: 1 April 2021