Understanding Energy Consumption in Real-World E-IoT Systems with CoAP, RPL, and 6LoWPAN

Jinyan Lu, Kirk Martinez, Alex S. WeddellLecture Notes in Electrical Engineering 2025 · DOI · OpenAlex

This paper provides a detailed empirical analysis of energy consumption in "sleepy" Environmental IoT (E-IoT) nodes using the standard IP-based protocol stack (CoAP, RPL, and 6LoWPAN). By measuring real-world power draws across different system states, the study demonstrates that optimizing communication windows and minimizing network maintenance overhead can significantly extend battery life—potentially reducing energy consumption by over 30% in high-frequency reporting scenarios.

Introduction

Remote environmental monitoring (E-IoT) requires sensor nodes to survive for years in the wild on small batteries. To achieve this, engineers use "Sleepy Networks," where nodes spend the vast majority of their time in a low-power state, waking up only briefly to sense data and communicate. However, a fundamental tension exists between reliability and efficiency. Standard internet-compatible protocols like RPL (for routing) and 6LoWPAN (for networking) require "maintenance" messages to keep the network structure alive. This paper addresses the gap between theoretical simulations and real-world hardware-level measurements of these protocol overheads.

Experimental Methodology

The researchers built a testbed using the Microchip SAM R30 Xplained Pro, a low-power platform featuring a sub-GHz radio (868 MHz), which offers superior range and obstacle penetration compared to 2.4 GHz bands. The software stack utilized RIOT-OS, implementing the GNRC network stack with CoAP, RPL, 6LoWPAN, and IEEE 802.15.4.

The experimental hardware setup including the SAMR30-XPRO node, Nordic Power Profiler, and Border Router.

High-precision current sensing was achieved using a Nordic Power Profiler Kit II and Agilent 34411A multimeters to capture micro-ampere level sleep currents and milli-ampere level radio spikes. A custom scheduling system synchronized nodes via an external high-accuracy RTC (DS3231) to minimize timing drift during sleep cycles.

Protocol Overhead and Energy Profiles

The study generated a primary dataset of energy profiles for specific IoT activities, categorized into SoC states (Standby/Sleep, Idle, Sensing) and protocol activities (Neighbor Solicitations, RPL Control Messages, and CoAP Request/Response cycles).

High-resolution power profile showing current spikes (mA) during active radio communication and protocol maintenance.

Key findings from these profiles include:

  • Sleep Dominance: For a node sensing every 10 minutes, the "Sleep" state accounts for over 60% of the total energy budget. While the current is tiny (132 µA with peripherals), the duration makes it critical.
  • Active Costs: Radio-heavy activities like CoAP responses and RPL DAO messages draw between 11 mA and 15 mA.
  • Peripheral Tax: Adding an SD card for data buffering increased the sleep current from a bare SoC baseline of 2.56 µA to 132 µA, highlighting a significant hidden cost in robust designs.

Optimization and Results

Many IoT implementations default to a 60-second "communication window" to ensure network maintenance tasks complete. The authors found this is often excessive and proposed a formula to calculate the minimum necessary wake time based on sensor reading frequency and CoAP payload size.

By shrinking the communication window from the default 60s to a calculated "safe" minimum (approx. 2–9s), the study achieved:

  • A 1.6% energy reduction for nodes communicating once per day.
  • A 30.86% energy reduction for nodes communicating every hour.

Based on these optimizations, a node sensing every 10 minutes with one daily sync using a standard 2600 mAh battery is estimated to last 712 days.

Conclusion

While sleep current is the primary driver of battery life, the "default" time spent awake for network maintenance is often excessive. Future work should focus on "Adaptive RPL," where the frequency of network maintenance messages scales based on environmental stability, further reducing unnecessary radio activity in multi-hop mesh topologies.

Summary updated

BibTeX
@inproceedings{lu2025understanding,
  author = {Jinyan Lu and Kirk Martinez and Alex S. Weddell},
  title = {{Understanding Energy Consumption in Real-World E-IoT Systems with CoAP, RPL, and 6LoWPAN}},
  booktitle = {Lecture Notes in Electrical Engineering},
  year = {2025},
  doi = {10.1007/978-981-97-4780-1\_14},
}
Figure