Message Queuing Telemetry Transport (MQTT) is a light-weight communication protocol between devices, and/or computer systems. It supports a Publish-Subscribe ("Pub/Sub") architecture. Therefore, it can easily support constrained connections networks like in IoT: A lot of participants, often with limited resources at each device with ever-changing network constellations.
The centerpiece is an MQTT broker that receives the data from sending devices (the "Publishers") and provides them to users, and systems that request them (the "Subscribers"). Effectively, this decouples senders and receivers, reduces complexity, and makes IoT set-ups manageable.
Read more...
State-of-the-art IoT networks provide data in streams: devices and systems continuously provide data messages. This "Stream" of data is in most cases issued at a high frequency and not only at an individual request.
Event Stream Processing applications are special applications that have similar logic to conventional analytics tools. However, time-based processing plays a far larger role, as well as handling the communication with data streams. Such systems need to be able to pick up the right messages from the stream, but also to feed resulting messages back into the stream of messages.
Read more...
"Edge Devices" are those computerized systems that remain local on a premise. In contrast, are computers, and systems that are sitting in the cloud means very far away from the actual place where data is collected and where feedback impulses need to go ("change temperature x to y degrees").
"Edge Intelligence" summarizes that some or all of the analysis of the data is done directly by edge devices instead of doing that on a remote/cloud system. Enabling faster reactions, with a higher degree of data privacy and higher reliability applies especially to production or security-related environments, where connection losses or latencies are not tolerable.
Read more...