Scalable Fog and Cloud Computing for Industrial IoT

点击翻译按钮获取中文摘要

Networked sensor rich systems, i.e., "Industrial Internet of Things (IIoT)" systems, are being increasinglydeployed to support a diverse range of applications such as oil and gas asset monitoring, energy metering, farming, connected vehicles, and supply chain management. Performance and energy usage are key concerns in such systems. Specifically, providing timely responses is critical for many applications, e.g., intrusion and fault detection. Typically, sensors have limited battery life and do not have powerful computation capabilities to meet stringent performance requirements. This necessitates smart strategies tooffload computation between fog resources, i.e., compute resources residing close to the sensors, and more centralized cloud resources. Despite the rapid advancements in this area, there is a lack of tools and techniques that system operators can use to evaluate such strategies before deploying a sensor rich system. Consequently, operators can be saddled with expensive deployments that do not meet performance, scalability, and energy usage requirements. This project will address this gap by initiating the development of a framework called PROFESS that supports performance evaluation of sensor rich systems. PROFESS will exploit recent advances in container virtualization technology to develop a realistic, cost-effective and controlled experimentation environment where large scale systems can be emulated on commodity multi-core servers. Specifically, PROFESS will seamlessly use containers to mimic sensor devices as well as fog and cloud resources with various computation and communication characteristics connected through operator-specified network topologies. It will provide an environment where various automated control strategies to coordinate computation between fog and cloud resources can be developed and evaluated by an operator in a controlled manner prior to a real deployment. Optimized deployments facilitated by PROFESS can allow Canadian IoT operators to reap cost savings due to efficient sizing and scalability of their infrastructure as well as the reduction of financial penalties arising out of poor performance.