Using HTTP/3 and Containerized Local Cloud Deployments to Optimize Smart Home System Performance

The rapid growth and adoption of IoT devices in environments such as smart homes and smart businesses has introduced challenges related to latency, data privacy, and system scalability. But while cloud technologies have been introduced to address some of the challenges faced, reliance on geographically distant data centers can often lead to increased latency and reduced responsiveness especially in rural or geographic locations with poor Internet infrastructure. We propose an edge focused architecture named the “S.M.A.R.T. Home Cluster” that is built on Raspberry Pi 4B microcomputers and managed using K3S, a lightweight distribution of Kubernetes. Our system integrates HTTP/3, the new UDP based version of HTTP via Nginx as a reverse proxy and load balancer. This latest version of HTTP utilizes lighter weight communications using UDP and TLS 1.3 encryption for security. Our system also leverages Kube-Vip, which allows the user to assign an elastic IP address for Kubernetes clusters, allowing seamless failover. We performed benchmarking using Apache JMeter to send HTTP requests to our architecture. This revealed significant improvements in terms of throughput, latency, and power efficiency. Our findings demonstrate the benefits of localized, containerized architecture, which not only reduces the reliance on the external cloud systems, but also improves the scalability, and privacy of both smart homes and small business environments.