NetEye: a user-centered wireless sensor network testbed for high-fidelity, robust experimentation

Experimentation infrastructures for wireless sensor networks (WSNs) have been widely adopted over the past decade for hi-fidelity validation and rapid prototyping of applications and for discovering and modeling phenomena that have led to improvements in analysis and simulation capabilities. Nonetheless, the community is still looking for more heterogeneous WSN testbeds for purposes such as federation, experiment regression analysis, and improved testbed availability. In addition, most existing WSN testbeds lack robustness and predictability in the presence of failures, and they do not provide user-friendly support for scientific experimentation (e.g., quick repetition of experiments). We develop the NetEye testbed as a high-fidelity, robust WSN experimentation infrastructure that provides the ease of scientific experimentation. Consisting of 130 TelosB motes with IEEE 802.15.4 radios and 15 Linux laptops with IEEE 802.11a/b/g radios, NetEye provides aWSN deployment that enables high-fidelity experimentation with single-hop as well as multi-hop WSN networking, and the lab-scale deployment mimics network properties of building-scale and outdoor WSN deployments. Building upon the authors’ experience in developing and using WSN testbeds in the past years, NetEye also has enhanced features that enables easy, robust scientific experimentation. In particular, NetEye provide users with a simple, powerful web portal, which streamlines users’ experience in the whole lifecycle of scientific experimentation (e.g., experiment scheduling, status monitoring, and data exfiltration). NetEye employs a stable, fault-tolerant state machine to improve the robustness of the testbed in the presence of failures. NetEye also develops a health monitoring service “NetEye Doctor” that monitors hardware and software status; the real-time health monitoring information about the testbed is seamlessly integrated with the lifecycle of experiment scheduling, experiment status monitoring, and experiment data analysis for both robust experimentation and for informed experiment analysis. To address the log-data congestion problem in long-running, large scale experiments, NetEye employs a TDMA mechanism to ensure reliability in exfiltrating experiment data to users. To facilitate quick, repetitive scientific experimentation, NetEye also speeds up the mote programming cycle through fast job status transition and multi-thread-based mote programming. NetEye has been successfully running for over 3.5 years, and it has been integratedwith the federatedWSN experimental infrastructure KanseiGenie. NetEye is widely used by researchers from USA and Asia, and more than 10 experiments are executed through NetEye per day on average.