Improve IoT App Performance & Backend Transition

I already have an IoT mobile app with its backend up and running; however, the user experience suffers from noticeable latency. My goal is to polish the codebase, trim anything that consumes resources unnecessarily, and move the entire stack from Google Firebase to a VPS that uses MQTT for real-time messaging.

What I need you to tackle
• Profile both the mobile client and the existing Node/JavaScript backend, pinpointing functions, background tasks, or logging that slow the app down, then refactor or remove them without breaking current functionality.
• Reduce round-trip time for screen loads, device data sync, and control commands to well under a second on a normal 2G connection.
• Stand up a secure VPS (Ubuntu preferred), install and configure an MQTT broker such as Mosquitto or EMQX, and harden the server (SSL/TLS, firewall rules, basic monitoring).
• Replace each Firebase feature we use today—authentication, realtime database, cloud functions, and push messaging—with an equivalent self-hosted or lightweight alternative, wiring it all through MQTT where appropriate.
• Update the mobile app code so all data flows through the new backend, test end-to-end, and publish a clean build.
• Deliver concise documentation: new architecture diagram, setup steps for the server, and a change log that highlights removed or optimized components.

Acceptance criteria
1. Cold-start to first data load ≤1 s.
2. Command round-trip ≤500 ms on 2G.
3. No runtime errors or feature regressions after Firebase removal.
4. Clear, reproducible server setup instructions delivered in Markdown.

The existing code is well commented and will be provided along with current Firebase rules and functions. Let’s streamline performance and give the app a lean, self-hosted backbone.