Toward a Data-Driven Cognitive Framework for Adaptive QoS in IoT

Abstract

Quality of service (QoS) is an increasingly important, stringent, and challenging requirement in the Internet of Things (IoT). Domain and context-dependent circumstances bring forth the need for a cognitive design to facilitate QoS that can learn and adapt as the requirements change and deployments evolve. The performance data that measure QoS metrics based on variations in different communication stack parameters/protocols is becoming available. The research literature highlights the impact of various configurable parameters, such as transmission power (TP) and packet size (PS) on metrics, such as reliability and delay. In this article, we propose the design of a cognitive framework for adaptive QoS in the IoT. The proposed approach is based on performance data from deployments of different types of communication networks forming IoT. The aim is to be able to adaptively configure communication parameters/protocols ensuring that the QoS constraints in evolving network scenarios are met.