Edge-Based Streaming Data Analytics Solutions

It is processing continuous streams of events, such as sensor readings and logs, on or near the devices that produce them, so filtering, aggregation and alerts happen before the data travels to a central cloud.

Cloud streaming platforms such as Kafka, Flink and managed streaming services run in data centres and handle very large aggregated streams. Edge streaming runs in constrained, often disconnected locations and sends only useful results upstream.

Usually yes. Edge streaming reduces and reacts to data locally, while a cloud platform handles aggregation across sites, long-term storage and model training. See the Streaming Analytics Platforms directory.

MQTT is the dominant protocol for device telemetry, and brokers with rule engines or local routing are often the first stage of an edge streaming pipeline.

The tools here range from single binaries that run on small gateways to agents that run on single-board computers and fuller platforms for industrial PCs. Check each project's resource requirements.

Start with the data source. For MQTT devices pick a broker with a rule engine, for simple collection use a lightweight collector such as Telegraf or Fluent Bit, and for filtering or windowed analysis on the device use a stream processor such as eKuiper or a flow tool such as Node-RED.