Towards Service Orchestration for the Cloud-to-Thing Continuum

Abstract

The cloud-to-thing continuum is emerging today as a solution to the requirements of next-generation Internet of Things (IoT) systems that integrate an immense number of heterogeneous devices generating massive amounts of data with applications that require real-time processing, extremely low service response times, and improved reliability and security. The concept of fog computing aims to offload the cloud by moving services to the edge of the network, closer to IoT devices, where the cloud-to-thing continuum refers to a highly distributed, decentralized, and dynamic environment that spans from devices to the cloud. Service orchestration is key to continuously ensuring optimal service placement in this complex environment, both in time and space, in accordance with specific real-time application requirements, taking into account workload balancing, fault tolerance, and system stability. In this paper, we identify service orchestration technologies as enablers for the cloud-to-thing continuum and present a general architecture for service orchestration in the cloud-to-thing continuum. The architecture is implemented using Eclipse ioFog to demonstrate and evaluate the service orchestration strategies in an emulated network, and to explore the applicability of ioFog for real-world next-generation IoT solutions.