Cloud Computing

Currently, the Internet of Things (IoT) paradigm is experiencing a great expansion, it is positioned at the top of the "Peak of Inflated Expectations" phase in Gartner's hype cycle [1]. This paradigm, with its broad definition, covers several areas: smart homes, smart metering, smart cities, smart factories, etc. With the convergence of this concept, the IoT applications may require, depending on the deployment scale (home, district, city, etc.), a varying number of equipment (servers, machines, cars, sensors, actuators, etc.). In a wide scale, a huge amount of data can be generated, which can cause various problematics related to processing, storage, scalability, quality of service, access, security, etc.

In our team, the QoS challenge constitutes one the main treated challenges. It is aimed at answering the requirements of IoT applications in terms of delay, availability, reliability, losses rate, etc. Proposed solutions to face this challenge are focused at the Middleware layer through the OM2M platform [2] and are aimed at to be deployed in an autonomic way depending on the context, based on the Autonomic Computing paradigm. Some of them are inspired from those used at the Network and Transport layers from the OSI Model. Other solutions are deployed in cloud-based infrastructure to benefit from the cloud computing paradigm properties, such as: services duplication (good fault tolerance / robustness), elasticity, migration, clustering, etc.

Cloud-IoT Architecture

In LAAS-CNRS, a new cloud platform entitled Cloud-IoT (the figure above) has been installed based on OpenStack [3]. It proposes an infrastructure as a service (IaaS) and is composed of a controller node, a network node, and a set of compute nodes. The Platform is composed of the main services for its efficient functioning (Keystone, Glance, Nova, Neutron, Horizon, etc.). This infrastructure is mainly dedicated to the demonstration of IoT challenges (including QoS challenge) and the validation of the proposed solutions.

[1] http://www.gartner.com/newsroom/id/3114217 (access in March 2016)

[2] http://www.om2m.org (access in March 2016)  

[3] https://www.openstack.org (access in March 2016)