BRIGHT: Broker-Less Residential IoT Gateway Over HTTPS for Transmission

Abstract

The growth of the Internet of Things (IoT) has raised demand for home monitoring that is cheap to deploy, secure, and power-efficient, particularly where infrastructure is limited. Many current designs rely on broker-based middleware or continuous cloud synchronization, and the resulting overhead is often a poor fit for ultra-low-bandwidth, low-power deployments. This work presents a broker-less home monitoring architecture built around the Raspberry Pi Pico W, an Android application, and a lightweight Firebase backend accessed over HTTPS. We evaluate it against a locally hosted MQTT/TLS broker using identical hardware and sensing payloads. Across 30 independent trials per configuration, the broker-less HTTPS/Firebase design produced a significantly smaller packet size (<inline-formula> <tex-math notation="LaTeX">$247.8~\pm ~9.4$ </tex-math></inline-formula> B versus <inline-formula> <tex-math notation="LaTeX">$327.1~\pm ~10.8$ </tex-math></inline-formula> B, <inline-formula> <tex-math notation="LaTeX">$p\lt 0.001$ </tex-math></inline-formula>). Its estimated energy per update was also lower (<inline-formula> <tex-math notation="LaTeX">$8.74~\pm ~0.38$ </tex-math></inline-formula> mJ versus <inline-formula> <tex-math notation="LaTeX">$9.16~\pm ~0.41$ </tex-math></inline-formula> mJ, <inline-formula> <tex-math notation="LaTeX">$p\lt 0.001$ </tex-math></inline-formula>, <inline-formula> <tex-math notation="LaTeX">$d=1.06$ </tex-math></inline-formula>), but that figure is analytically derived rather than directly measured and amounts to a negligible runtime difference (< 0.03 h on a 2000 mAh battery at a 60 s reporting interval); we therefore give it far less evidential weight than the directly measured latency and packet-size results. The MQTT/TLS baseline, in turn, achieved substantially lower end-to-end latency (<inline-formula> <tex-math notation="LaTeX">$0.69~\pm ~0.08$ </tex-math></inline-formula> s versus <inline-formula> <tex-math notation="LaTeX">$1.34~\pm ~0.12$ </tex-math></inline-formula> s, <inline-formula> <tex-math notation="LaTeX">$p\lt 0.001$ </tex-math></inline-formula>). Taken together, these results place the proposed architecture in sparse, latency-tolerant residential monitoring for bandwidth- and infrastructure-constrained settings, where removing dedicated broker infrastructure outweighs the latency cost, rather than in time-critical alerting. All experiments were run indoors with up to four nodes. Claims about rural or infrastructure-constrained suitability rest on the architecture’s low bandwidth footprint and the Wi-Fi sensitivity results, not on outdoor field measurements. A Wi-Fi signal sensitivity analysis characterizes how latency and reliability degrade as connectivity weakens, and a security analysis identifies replay protection and per-device identity management as the main residual risks of the current prototype.