The recent explosion of Internet-of-Things (IoT) can be mainly attributed to technological advances in hardware that enable the  execution of complex processes, sensor technology and the design of appropriate operating  systems that support capabilities such as IPv6 connectivity for low-end  devices, data storage in relational databases, and routing protocols for  low-power and lossy networks. Despite the improvement and maturity of IoT, its rapid evolution has  created serious technological fragmentation with IoT systems isolated  from each other

Monitoring and managing an IoT system is a demanding and complex process for several reasons. First of all, an IoT network consists of a large number of heterogeneous devices, in terms of hardware, software as well as network availability (different communication technologies, such as WiFi, IEEE 802.15.4, LoRa etc.). Moreover, in several scenarios IoT devices reside in locations that are hard to access physically. Thus, it can be difficult to detect faults and perform maintenance. In addition, fault types can be of great variety, due to the diversity and complexity of IoT systems (e.g. sensor faults, hardware or software faults etc.). Finally, IoT networks are popular targets for cyberattacks, which often successfully exploit vulnerabilities emerging either from system misconfiguration due to personnel’s ignorance or inability to use strong cryptographic algorithms due to resource-constrained nature of IoT devices.

A Management-as-a-Service platform is able to offer an interoperable solution for monitoring the whole IoT network. The service consists of appropriate hardware and software for monitoring parameters across multiple layers, such as battery/energy level, device uptime, network statistics (e.g. lost packets per protocol), link quality (RSSI, LQI) etc. These parameters are collected by IoT devices and IoT gateways, are stored in the cloud backend and are processed through statistical methods and machine learning techniques for detecting and predicting IoT network faults. In addition, visualization tools offer a detailed overview of the network and its devices.

EPOPTIS has the following main objectives:

  • Design, implement, and evaluate an interoperable and secure platform for providing Management-as-a-Service for IoT networks.
  • Develop data collection and visualization techniques for heterogeneous data generated by IoT networks.
  • Utilize statistical methods and machine learning algorithms for processing collected data and automatically detecting and predicting faults in IoT networks.