DESIGN AND IMPLEMENTATION OF A DIGITAL TWIN TESTBED FOR CARBON-AWARE CLOSED-LOOP NETWORKING

This thesis presents a closed-loop programmable network architecture that integrates Intent-Based Networking with environmental sustainability principles. Implemented on a real Cisco-based infrastructure using NETCONF, YANG, model-driven telemetry, InfluxDB, and Grafana, the system employs a graph-based digital twin to evaluate alternative network paths according to performance, energy consumption, and carbon-equivalent emissions. Experimental validation, conducted through both synthetic traffic and real RTSP video streaming under congested conditions, demonstrates the feasibility of automated network monitoring and reconfiguration while highlighting the trade-offs between performance, quality of service, and sustainability.