CaPTS scheduler: A context-aware priority tuple scheduling for Fog computing paradigm

With the increase in real-time latency-sensitive Internet of Things (IoT) applications, a huge amount of data is generated in the Fog-IoT paradigm. There is a need to schedule and execute this huge workload over Fog devices efficiently to support the increasing demand of these applications. But, Fog devices are resource-constrained in terms of processing/computing power, bandwidth as well as storage capacity which makes tuple scheduling a challenging problem. Moreover, due to the rise in IoT devices per application, a sharp increase in service response time, network congestion, and inefficiency in terms of energy consumption, and execution cost has been observed. Consequently, an efficient tuple scheduling algorithm is desirable that can reduce latency and network usage and optimize energy consumption and cost. Therefore, in this work, CaPTS scheduler: A Context-aware Priority Tuple Scheduling for Fog computing paradigm is designed and proposed. It takes into consideration various context-aware parameters such as task load of application, networking requirement, and data flow rate to set the priority of tuples and schedule them across Fog computing nodes while ensuring quick service response time and satisfying quality of service requirements of end-users. The CaPTS scheduler is implemented and evaluated using iFogSim toolkit on various performance metrics such as latency, network usage, energy consumption, and cost. Its performance is validated through a case study on the smart mining industry system. The results show that on an average the latency and network usage are minimized by 35.93% and 44.20%, while energy consumption and cost are optimized by 4.55% and 30.92%, respectively, in comparison with baseline techniques.

Automated summary

In this study, a context-aware priority tuple scheduling algorithm for the fog computing paradigm is proposed to reduce latency, optimize network usage, and improve energy consumption and cost. Experimental results demonstrate its feasibility and effectiveness in the smart mining industry system.