Edge computing
Processing data where it is produced — at the bedside, gateway, or device — instead of round-tripping everything to the cloud.
In one line
Edge computing moves computation next to the data source — a ward gateway scoring an ECG stream locally instead of shipping every sample to a distant data centre.
How it works
A hierarchy of compute: device (microcontroller running a tiny model — "TinyML"), gateway/edge node (a small box aggregating a ward's devices), and cloud (training, fleet management, long-term storage). The edge filters, compresses, and reacts: only events and summaries travel upstream. Wins: latency (alarm in milliseconds, not seconds), bandwidth (don't stream 250 Hz waveforms to the cloud), resilience (works through internet outages), and privacy (raw data can stay on-site).
Where it shows up in digital health
Real-time vital-sign analytics and early-warning scores at the bedside; ambient fall detection that must not depend on connectivity; rural/low-bandwidth deployments where the link to the cloud is the weakest component — a first-class concern for India-scale health IT.