Redesigning compound TCP with cognitive edge intelligence for WiFi-based IoT

摘要

This article explores how to enhance Compound TCP design to learn and fix delays in future WiFi-based IoT networks continuously. We adopt a cross-layer approach and design multiple intelligent access points (APs) with cognitive collaboration approaches to improve efficiency across industry 4.0 WiFi networks. Comprehending the efficiency of Compound TCP is difficult due to its hybrid congestion mitigation strategies and the interrelationships with wireless channel failures, medium access level contentions/collisions, and overflows at the APs. We are therefore designing a detailed analysis to investigate the efficiency of Compound TCP streams over WiFi-based IoT. We take into account all kinds of losses at different TCP/IP layers. Our conceptual model contains WiFi network constraints, such as the maximum number of retries and the AP buffer capacity, which affect transport-layer latency and quality. More specifically, we are developing an analytic model for dual-APs and highlighting multiple AP settings' benefits. Based on the developed theoretical model and findings, we show that cognitive radio approaches in Dual-AP systems can dramatically increase the performance of industry 4.0 WiFi connections and predict performance enhancements in multiple APs settings.