Abstract/Summary

With the increasing popularity of the Internet-of-Things (IoT), organizations are revisiting their practices as well as adopting new ones so they can deal with an ever-growing amount of sensed and actuated data that IoT-compliant things generate. Some of these practices are about the use of cloud and/or fog computing. The former promotes "anything-as-a-service" and the latter promotes "process data next to where it is located". Generally presented as competing models, this paper discusses how cloud and fog could work hand-in-hand through a seamless coordination of their respective "duties". This coordination stresses out the importance of defining where the data of things should be sent (either cloud, fog, or cloud&fog concurrently) and in what order (either cloud then fog, fog then cloud, or fog&cloud concurrently). Applications' concerns with data such as latency, sensitivity, and freshness dictate both the appropriate recipients and the appropriate orders. For validation purposes, a healthcare-driven IoT application along with an in-house testbed, that features real sensors and fog and cloud platforms, have permitted to carry out different experiments that demonstrate the technical feasibility of the coordination model.