Age of Information Based Flow Control in Communication Networks

This thesis begins with a study of AoI through measurements on bandwidth-limited links in emulated and real-world applications, where its results reveal similarities between the empirical measurements with the theoretical results. Furthermore, this study analyzes AoI variations over TCP/IP links that use WiFi, LTE, 3G, 2G, and Ethernet. It follows by addressing the issue of synchronization errors in AoI measurements and a solution for it. To achieve a deeper understanding of AoI behavior in practical scenarios, this study extends to measure AoI related to IoT nodes over TCP, UDP, and Web Sockets protocols. Also, we present a deep reinforcement learning-based approach that can minimize the AoI without any prior knowledge about the network attributes. These observations provide the basis for the design of Age-Aware Application Layer Forward Error Correction (A3L-FEC), a role-based data flow controller in the application layer of OSI with the aim of improving data freshness by preventing age violation. A3L-FEC ensures the reliable transmission of fresh data by leveraging forward error correction and user datagram protocol. The protocol proposes two variations, A3L-FEC-FSFB and A3L-FEC-VSVB for different types of applications, where each one operates with different sample generation, transmission intervals, and adaptive blocklength selection. Evaluations of A3L-FEC under different network conditions and application requirements have been performed and its age violation control performance has been compared with TCP-BBR in the mininet emulator environment and ACP+ in MATLAB. The outcomes reveal the effectiveness of A3L-FEC in minimizing age violations and improving data freshness compared to TCP-BBR and ACP+.