EN FR
EN FR


Section: New Results

Smart Cities

Participants : Nathalie Mitton, Valeria Loscri, Abdoul Aziz Mbacke.

Smart cities are a key factor in the consumption of materials and resources. As populations grow and resources become scarcer, the efficient usage of these limited goods becomes more important. Building on and integrating with a huge amount of data, the cities of the future are becoming a realization today. There are millions of sensors in place already, monitoring various things in metropolises. In the near future, these sensors will multiply until they can monitor everything from streetlights and trashcans to road conditions and energy consumption. In this context, effective strategies or solutions for refining data sets can play a key role. Based on these premises, we propose in [32] intelligent and adaptive filtering mechanisms as a service (FIlAAS) integrated in the VITAL-OS middleware and will show their feasibility and their effectiveness in the smart city context.

Connecting all these devices to a cloud encompasses the execution of many network tasks at the edge and in particular on constrained gateways by low computational resources capabilities. Moreover, these gateways have to deal with the plethora of disparate technologies available in the IoT landscape. To cope with these issues, we introduce a Lightweight Edge Gateway for the Internet of Things (LEGIoT) architecture [18]. It relies on the modular characteristic of microservices and the flexibility of lightweight virtualization technologies to guarantee an extensible and flexible solution. In particular, by combining the implementation of specific frameworks and the benefits of container-based virtualization, our proposal enhances the suitability of edge gateways towards a wide variety of IoT protocols/applications (for both downlink and uplink) enabling an optimized resource management and taking into account requirements such as energy efficiency, multi-tenancy, and interoperability. LEGIoT is designed to be hardware agnostic and its implementation has been tested within a real sensor network. Achieved results demonstrate its scalability and suitability to host different applications meant to provide a wide range of IoT services.

In parallel, we proposed a MOOC in the framework of the IPL CityLab project (See Section 9.2.1), whose working documents are available online [51], [52], [53], [54].