HomeNetWatch
v0.6 in production

Enterprise infrastructure principles. Built for my house.

Thirty years of managing enterprise infrastructure, written down as operating rules. Coding agents built this platform from that institutional knowledge, and it runs on a real home network with the same principles a corporate network is held to.

Open live dashboard Read the source

Move your pointer to steer the radar. Click anywhere here to send a ping.

01 / EDGE HARDWARE

Raspberry Pi and Android TV sensors on a real home network

A Raspberry Pi measures both Wired Ethernet and Wi-Fi at the same time: HTTP and ICMP path checks, LibreSpeed throughput tests, Kismet wireless intrusion detection, on-demand LAN scans, and a custom Rust real-time UDP stream generator (Voice G.711 A-law/µ-law and Video). An NVIDIA Shield Android TV agent adds QR-onboarded Wi-Fi voice and throughput telemetry from the living room.

Architectural trade-off Heavy tests run spaced in a serialized queue, not as parallel streams that would peg the Pi's CPU, saturate the residential uplink, and create the very lag being measured.
Inspect live sensor Agent README ↗ Rust RTNP repo ↗

Advanced features you would expect in a corporate network.

PATH VIEW

BGP path correlation

Traceroute grouped into networks, set next to the BGP path public collectors see from the provider.

SPEED TEST · INTERNET CHECKS

Dual-link testing

A dual-NIC sensor measures Wired Ethernet and Wi-Fi at the same time: throughput, HTTP, ICMP, and path traces to the services the household uses.

AIRWAVES

Wireless intrusion detection

Kismet WIDS on a radio that listens and never joins a network: nearby Wi-Fi, clients, and wireless security.

SENSOR SECURITY

Encrypted sensor tunnels

Ed25519 device enrollment, pinned TLS, and dual self-healing WireGuard tunnels that keep measurement and access on separate planes.

REAL-TIME HEALTH

Synthetic voice and video

A custom Rust UDP generator sends G.711 voice and video streams over the link, queued so the test does not create the lag it reports.

SERVER HEALTH

Capacity forecast and backups

Disk forecast, encrypted backups, and the host the platform runs on, separate from the home network.

SECURITY

A server I run myself, built with cybersecurity first.

The platform is public, so the server behind it is managed like production: owned end to end, kept patched, and locked down in layers, from the sensor to the dashboard.

Security policy ↗
IN TRANSIT

Encrypted end to end

HTTPS with automatically renewed certificates, HSTS and a strict CSP, WireGuard tunnels to every sensor, and a pinned TLS key for enrollment.

IDENTITY

Every sensor proves who it is

Devices enroll with their own Ed25519 key and signed requests, and nothing joins until an admin approves it.

ACCESS

Read-only for guests

Visitors see live graphs through a guest role enforced on the server. Private details are removed in the API itself, not just hidden on screen.

PATCHING

Vulnerability tracking

Server Health checks the installed software against known CVEs every week, so unpatched packages are seen and fixed.

RECOVERY

Encrypted off-host backups

Backups are encrypted and kept off the host, and a disk forecast warns before the server runs out of room.

CHANGE CONTROL

Nothing ships unchecked

Every change passes tests and review before it reaches production, and the rules say what must stay private.

WHAT THE RULES REFUSED

No always-on language model reading logs. Deterministic correlation runs first.

Gemini is on demand, for a root-cause note or a fault walkthrough, and a guest session is read-only with the home address removed on the server.

SPEC.md ↗ House rules ↗ Open the platform
Evan You
Leader · Builder · Strategist · Mentor

Evan You

Infrastructure, cloud, networking, and cybersecurity. An IT degree from Rochester Institute of Technology, then thirty years managing enterprise infrastructure. As a leader, he has led teams that drove innovation with automation and AI, taking infrastructure engineers to a full DevOps way of working. That institutional knowledge became the written instructions coding agents built this platform from: the principles of a corporate network, applied to one house.

Away from this project: a five-year member of the Golf Mission Association, and a weekly podcast that turns Golf Genius match data into a match report, all built with AI as well.

LinkedIn ↗ GitHub ↗ Golf podcast ↗