Authors: Bo Cheng ; Jingyi Zhang ; Gerhard P. Hancke ; Stamatis Karnouskos ; Armando Walter Colombo
Internet of Things (IoT) technology provides new opportunities to build powerful industrial systems by connecting a large number of smart networked embedded devices. Devices within these industrial IoT (IIoT) or industrial cyber physical systems (ICPSs) can sense and control physical processes, make autonomous decisions, and communicate and cooperate, thereby collectively generating a massive amount of system data. By capturing, processing, and analyzing significant amounts of data from these devices effectively, industrial companies and organizations can manage their enterprise resources, optimize technical processes, understand the market demand, and develop business intelligence and analytics (BI&A). Due to poor scalability and low performance, many traditional computing technologies are inadequate for handling big data, which are characterized by the volume, velocity, variety, and veracity of the data (each of these characteristics applies to ICPS data). The volume of data will grow with the adoption of IIoT technology. The velocity, i.e., the rate at which data are generated, ingested, and processed, is crucial for decisions that feed back into the system to control real-time industrial processes. Since the ICPS consists of heterogeneous systems of systems, the variety of the data is also very high. The veracity (accuracy) of the data is also important in the cyber physical context as incorrect decisions made from low quality data could lead to physical disruption of industrial processes. Effectively handling these big data is an opportunity to generate added value and provide a business advantage. Cloud computing provides a promising solution for modern industrial systems and applications by hosting services with flexible computational and storage capabilities. The integration of IoT and big data technologies into the cloud will empower industries to build a more intelligent, flexible, and cooperative industrial environment. In some smart factories, for example, operators can track the entire production process, identify (often proactively) problem areas, and make informed. The usage of cloud technologies is increasingly penetrating industrial settings, and the term industrial cloud refers to the cloud building an industrial CPS and applications with IoT and big data technologies to distinguish it from the cloud for general purposes. This article discusses some of the main architectural issues related to collecting and handling big data for analysis linked to IoT and cloud technologies in the industrial context. The aim is to provide a high-level introduction view of this topic, underpinned with examples from popular frameworks, and discuss open research questions and future directions.