TP3.4: Sensing on Demand

Subproject manager
Prof. Dr. Jörg Ott
Vittorio Cozzolino

Stationary sensor systems (parking garages, toll systems, cameras, weather stations, sensors for air quality and noise pollution) are increasingly used to register traffic and environment data. Furthermore, mobile devices (vehicles, mobile phones, tablets) are equipped with a set of sensors, which are able to deliver for example information about position and movement (and thereby during aggregation about traffic density and flow), but which are also able to register a set of additional parameters. All these sensors provide potentially a variety of information, which cannot be transmitted to the infrastructure (“Cloud”) because of the sheer data volume. We generally refer to such an agglomerate of sensors and devices as “IoT Ecosystems”.

One of the most comprehensive examples of IoT Ecosystems are Smart Cities: heterogeneous environments composed by a plethora of sensing networks and devices dedicated to different tasks. In this specific scenario we have multiple sensing sources/sensing devices (mobile/fixed, from user-centric devices to WSN) aiming to provide different information. It is not a trivial task to collect data from such wide range of disparate data sources; securely, efficiently and over a geographically disperse and possibly fragmented network [1].
Smart City
Figure 1: Smart City.
Even though it is crucial to push some of the computation toward the edge of the network, fostering the advance of fog computing, still the presence of a back-end infrastructure is mandatory. We are talking about Cloud computing.

Thus, the two critical and dominant technologies for realizing the ubiquitous communications vision are Cloud computing and IoT. The cloud can provide large-scale and long-lived storage and processing resources for the personalized ubiquitous applications delivered through the IoT as well as important back-end resources. However, cloud-based platforms stay far from the real nodes connected to them. On the other hand, device-centric technologies and applications, such as IoT, constitute part of a local to the users and distributed in nature infrastructure, where a lot of personalized, and also, vital, data comes from sensors and actuators. Such interaction paradigm also fosters end-users to be part of the data acquisition chain, becoming both exploiter of the service and provider of information. Figure 2 depicts a conceptual architecture overview where the interaction between the IoT and Cloud services is detailed.
Architectural Overview
Figure 2: Architectural Overview.
The aim of this subproject is to develop a virtualization environment able to let IoT devices register freely selected parameters (obviously limited based on the sensors available on the device) and, moreover, move part of the computation and data aggregation from the Cloud, to the Fog (edge computing). Hereby, the existing infrastructure can be virtualized and used in a more efficient way since not every service provider has to set up its own measurement infrastructure. A user can issue one-time or recurring orders to the platform through a service provider and specify, on which devices, at which time, how often they will be executed and how the collected data will be aggregated.

If you would like to get more information about this subproject, our work in general or if you are interested in a cooperation to develop a flexible, on-demand sensing platform for data acquisition and event processing, please contact Prof. Dr. Jörg Ott or Vittorio Cozzolino.

[1] HortonWorks DataFlow – Accelerating Big Data Collection and DataFlow Management
[3] Suciu, G., Vulpe, A. , Halunga, S. , Fratu, O. , Todoran, G. , Suciu, V. ,Smart Cities Built on Resilient Cloud Computing and Secure Internet of Things