Measuring Age of Information on Real-Life Connections

The Age of Information (AoI) has emerged as a crucial performance metric for evaluating data freshness in real-time networked systems, such as IoT-based monitoring and control applications. While most existing AoI studies remain theoretical or simulation-based, this paper bridges the gap by presenting a comprehensive experimental framework for measuring AoI over real-life Internet connections using UDP. We address key practical challenges, including clock synchronization between transmitter and receiver, and propose a novel estimation technique based on round-trip time (RTT) that eliminates the need for synchronization while maintaining accurate AoI computation.

Our methodology is validated using a real-world testbed spanning from METU (Ankara) to a remote echo server in Istanbul, where UDP packets containing time-stamped data are transmitted, echoed, and analyzed to estimate instantaneous and average AoI. We detail methods for AoI computation at both transmitter and receiver sides, discuss the impact of synchronization bias, and explore alternative measurement strategies using ACKs. The empirical results confirm theoretical predictions—such as the U-shaped AoI vs. packet rate curve—and uncover new insights into the behavior of AoI under real Internet infrastructure conditions, including the effects of packet losses and queue dynamics.

This work provides a first-of-its-kind practical demonstration of AoI measurement over heterogeneous networks, offering a valuable framework for researchers and practitioners aiming to deploy age-aware communication protocols in real-world scenarios. Our approach enables robust AoI monitoring in delay-sensitive systems without the need for tightly synchronized clocks, supporting more scalable and energy-efficient IoT deployments.

For more information, please click here!