A federated fog-cloud framework for data processing and orchestration
Dapeng Lan, Yu Liu, Amir Taherkordi, Frank Eliassen, Stéphane Delbruel, Lei Liu · 2021
The fog computing paradigm has been proposed to alleviate the pressures on cloud platforms for data processing and enable computation-intensive and delay-sensitive applications in smart cities. However, state-of-the-art approaches mainly advocate either cloud-or fog-based data processing solutions, and they also lack a common framework for programming over the fog-cloud continuum. In this paper, we propose a distributed, fog-cloud data processing and orchestration framework, which is capable of exploiting the semantics of both fog platforms and the Cloud. Our framework can create on-demand process engine data flow (PEDF) spanning multiple device layers with various resource constraints. This will considerably help the developers rapidly develop and deploy data processing applications over the fog-cloud continuum. Our proposed framework is validated in a real-world scenario---IoT data streaming analytics for the smart green wall in a smart city---which demonstrates efficient resource usage and latency reduction.