Engineering Real-Time Telemetry and Secure Clinical Messaging with SignalR

I’ve spent plenty of time wrestling with WebSockets in my career, but things get a lot more intense when you're dealing with live cardiac vitals and patient privacy. If you’ve ever wondered how to keep real-time data snappy without leaking sensitive location data, this is a great deep dive.

It walks through the architecture of Halkyone Clinical OS, showing how they use SignalR to bridge the gap between .NET backends and React frontends for mission-critical healthcare workflows.

• Isolation of concerns using three specialized SignalR hubs (ChatHub, TelemetryHub, and NotificationHub) to manage distinct data streams.


• Implementation of a 500m privacy radius that uses the Haversine Formula to obfuscate clinician coordinates to (0,0) when they get too close to a patient’s home.


• Real-time streaming of IoT telemetry, specifically focusing on sub-second synchronization for heart rate and SpO2 vitals.


• Optimistic UI patterns in the Next.js client that use temporary UUIDs before resolving to permanent database IDs after a successful WebSocket broadcast.


• Handling transactional "IsSeen" receipts to sync read-states across multiple devices simultaneously.


• Managing the transition from SQLite/Turso database commits to WebSocket fan-outs to ensure data consistency.


• Fault-tolerance strategies for maintaining persistent connections between a .NET emr-server and a React emr-client.

Building real-time features is easy, but managing state synchronization without creating race conditions or privacy leaks is where the real engineering happens.

Read the full article here:

https://erwinwilsonceniza.qzz.io/blogs/engineering-realtime-telemetry-signalr