Towards an emulated IoT test environment for anomaly detection using NEMU

Abstract

Abstract-The advent of the Internet of Things (IoT) has led to a major change in the way we interact with increasingly ubiquitous connected devices such as smart objects and cyberphysical systems. It has also led to an exponential increase in the number of such Internet-connected devices over the last few years. Conducting extensive functional and performance testing is critical to assess the robustness and efficiency of IoT systems in order to validate them before their deployment in real life. However, creating an IoT test environment is a difficult and expensive task, usually requiring a significant amount of physical hardware and human effort to build it. This paper proposes a method to emulate an IoT environment using the Network Emulator for Mobile Universes (NEMU), itself built on the popular QEMU system emulator, in order to construct a testbed of inter-connected, emulated Raspberry Pi devices.Additionally, we experimentally demonstrate how our method can be successfully applied to IoT by showing how such an emulated environment can be used to detect anomalies in an IoT system.

Edge Computing Simulators for IoT System Design: An Analysis of Qualities and Metrics

“…Computer simulation has been proven to be a suitable technique to use for investigating different aspects of complex systems, such as service configuration, resource placement, or management strategies in IoT systems. To this end, several simulators (simulation frameworks) were proposed that address specific requirements for the application in cloud computing, for example, modelling of physical networks and virtual topologies [16] or emulating IoT environments [17]. Additionally, there are a few simulators designed for the simulation of edge and cloud computing, for example, iFogSim [18], EdgeCloudSim [19], and IOTSim [20].…”

“…To investigate and compare the performance of different IoT architectures by means of simulation, and in terms of QCs relevant for the application at hand, quantitative metrics must be defined. As outputs, most frameworks focus on general network theoretical metrics, such as latency (end-to-end response time), network congestion, or energy consumption [11,17]. To investigate and assess edge-cloud architectures, as well as provide design decision support, the application of such standard metrics is not sufficient.…”

Analysis of Machine Learning Techniques for Anomaly Detection in the Internet of Things

“…Time-Series: IoT Testbed Dataset. As a second time-series dataset, we used our IoT testbed data generated by the process discussed in [16]. The testbed outputs room temperature values and those that are too low/high are flagged as anomalies.…”

A Requirements-Based Approach for the Evaluation of Emulated IoT Systems

“…Driven by the huge economic impact and benefits (estimated in $2.7 to $6.2 trillion until 2025 by Gartner [1]), the IoT is leading to an explosive growth in the number of Internet-connected devices worldwide. One estimate is that there will be more than 50 billion connected devices on the IoT by the year 2020 [2], [3]. While the criticality of the services provided by IoT varies by applications (which range from household devices to critical systems), the quality of all services can be affected by similar factors.…”

“…Hence, it is critical to properly test such IoT systems to fully understand their capabilities and to know the conditions under which they can operate effectively [3], [5]. Nonetheless, testing IoT deployments is a very challenging task [6] due to the intrinsic characteristics of a typical IoT system such as the vast number of devices involved and the complex nature of their inter-connectivity.…”