From the Cloud to the Edge Towards a Distributed and Light Weight Secure Big Data Pipelines for IoT Applications
Feras M. Awaysheh
Abstract
Part of the broader development of Internet-of-Things (IoT) architecture for intelligent environments is the developing of sophisticated Edge communications that support the modern requirements of IoT-to-Cloud connectivity. Big Data (BD) and Cloud computing represent a practical and cost-effective solution for supporting IoT operations and advanced analytics. Such a vision involves developing data pipelines, facilitating BD flow from the network’s edge (e.g., sensors, actors, etc.) to the cloud data warehouse. However, security is always a concern among practitioners and developers alike. There are vital factors relating to performance impacts and security vulnerabilities that may emerge during the increased deployment of such a system. Highly secure integration of BD pipelines is a cornerstone in our ability to exploit modern IoT-to-Cloud applications. Given the high impact of BD pipelines security measurements on the system performance, it was not a subject of intensive analysis in the literature. In this study, we analyze the building blocks that support data pipelines as a commodity service for IoT-to-Cloud applications to address the previous research gap. A structured multi-layer security pattern method supporting Edge/Cloud architectures is presented. Data confidentiality was investigated in the complete data pipeline life cycle, i.e., allocation, transmission, and storage. Our study carried the examination of access control, wire encryption, and at-rest data encryption techniques impact on the overall performance. The analysis guides potential large-scale data analytics to model their infrastructure in a secure context using an integrated scheme, technologies, and frameworks. It also highlights a timely demand for lightweight security 51approaches that supports the widespread of BD pipelines. Our findings point out the critical need for future research in Edge Intelligence and Artificial Intelligence IoT (AIoT) for sustainable Edge integration in the Cloud. Finally, this study is bridging a knowledge gap between the existing BD pipeline security approaches and problems related to security impact on large-scale edge data processing performance, emphasizing the necessity of lightweight security techniques toward achieving this vision.