No abstract

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 proof 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

Since the invention of the steam engine in the 18th century, innovation drove the development of industrial processes. The next industrial revolution will form an ecosystem of over 20 billion connected devices with unforeseeable influence to the gross domestic product by 2020, and connected assets will generate about 44ZB of data, which pose interesting challenges related to privacy, connectivity, scalability, and others. A current line of action that leads to this direction is the development of cyber‐physical systems; considered as the coupling of physical processes and the digital world, its influence in the next industrial revolution is essential. In this work, we discuss its implementation, taking the Fog computing paradigm into consideration. As a starting point, we are extending a standard‐compliant machine‐to‐machine communication architecture to support container‐based orchestration mechanisms to enable cyber‐physical systems to be programmable, autonomous, and to communicate peer‐to‐peer. As the primary field of application, we are considering Industrial Internet domains in general and Smart Factory environments in particular. In this paper, we present an architecture for Fog Nodes, as well a more in‐depth discussion on the orchestration system and programmable characteristics of the Fog Node. On the basis of a simulation model, we show the advantages of having a programmable Fog Node supported by an orchestration system. Finally, we open a discussion about our solution and its application in the field of Smart Factories.

No abstract

No abstract