Key points are not available for this paper at this time.
The development of Fog Computing technology is crucial to address the challenges to come with the mass adoption of Internet Of Things technology, where the generation of data tends to grow at an unprecedented pace. The technology brings computing power to the surrounds of devices, to offer local processing, filtering, storage and analysis of data and control over actuators. Orchestration is a requirement of Fog Computing technology to deliver services, based on the composition of microservices. It must take into consideration the heterogeneity of the IoT environment and device's capabilities and constraints. This heterogeneity requires a different approach for orchestration, be it regarding infrastructure management, node selection and/or service placement. Orchestrations shall be manually or automatically started through event triggers. Also, the Orchestrator must be flexible enough to work in a centralized or distributed fashion. Orchestration is still a hot topic and can be seen in different areas, especially in the Service Oriented Architectures, hardware virtualization, in the Cloud, and in Network Virtualization Function. However, the architecture of these solutions is not enough to handle Fog Requirements, specially Fog's heterogeneity, and dynamics. In this paper, we propose an architecture for Orchestration for the Fog Computing environment. We developed a prototype to prof some concepts. We discuss in this paper the implementation, and the tools chose, and their roles. We end the paper with a discussion on performance indicators and future direction on the evaluation of non-functional aspects of the Architecture.
Brito et al. (Mon,) studied this question.