AI-Enabled Cloudless Home Ecosystem A Decentralized Architecture for Enhanced Security and Intelligent Control

DOI: https://doie.org/10.10399/JBSE.2025811222

Andrew Nitin Joseph, Naindeep Singh, KVS Balaji, Jasmitha JK, Mathew Jacob, Dr. Revathi V

Keywords:

Cloudless, Decentralized, Local Server, VPN, Home Automation, NAS Server, Ollama, Llama 3.1, AI Voice Assistant and pattern recognition.

Abstract:

This study suggests a decentralized, cloud infrastructure-independent home automation system, with local processing and decentralized architecture, aiming for security, privacy, and efficiency. With the use of a local server for automation task execution and data processing, the system reduces third-party cloud service reliance. The system is made secure with remote access through Tailscale VPN, using WireGuard encryption, to enable direct peer-to-peer communication between devices. Modular architecture with microservices allocates duties to automation control, NAS-based file processing, AI voice assistant, and real-time monitoring to ensure performance and capture faults. The local system's AI utilizes Llama 3.1 LLM model for voice commands and adaptive scheduling from past usage behavior. The user data is utilized by the AI module to suggest power-saving procedures, and the NAS server incorporated within offers secure local storage of personal and group files without recourse to external cloud storage solutions. Initial results show success in terms of low latency, consistent data processing, and increased security via local data processing and encrypted communication protocols. The provided solution effectively addresses the risks of centralized cloud-based designs while offering an efficient and scalable solution for future-proof smart homes in general. In conclusion, this research illustrates the practical benefits of combining high-level local artificial intelligence capability with a secure, decentralized home automation system.

PDF