MQTT Messages and Edge-AI on data
Engineers are familiar with MQTT, but what about securing MQTT - do you already know MQTTS? And why can MQTTS be related to AI or Node-RED?
Data Security
MQTT as an M2M (Machine to Machine) protocol is often used for retrofitting, i.e. modernizing a machine. Internal MQTT brokers, which accept and distribute (publish/subscribe) the machine messages, are often easy for users to access the data. A tool like the MQTT Explorer is enough to read and collect data. Reading along is the best way for an AI to learn.
"Yeah, but who wants to read our data?!"
Exactly, who wants this insignificant data that is really only interesting for us? "Nobody understands our data." In one way or another, the value of data continues to be significantly underestimated. There are companies that are already earning a few Euros with data. Strange, the data is actually worth nothing, isn't it? It is particularly popular in the AI world to have as much data as possible, when it comes to learning and finding new applications. Keep in mind that AI systems are particularly advanced in pattern recognition and therefore sometimes come across a pattern faster than humans.
More options, less power consumption.
In 2016 there was still 1.33 TFLOP per 30 watts (0.0443 TFLOP/W) "on edge", today in 2022 we are talking about more than 270 TFLOP per 20 watts (13.75 TFLOP/W) "on edge". This corresponds to approx. 300 times more computing power per watt. In other words, the effort is reduced and the possibilities increase. A serious assessment cannot be given here, but a look at the rapid development in the "AI on Edge" market, i.e. AI directly on the machine and in the factory.
MQTTS to secure the data streams
MQTT can be secured and regulate access to data via user and password. Ultimately, this is the principle that is also available for HTTP with HTTPS. Data can continue to be open and available. But consider carefully whether the value or worthlessness of the data has been well assessed. In case of doubt, we clearly recommend: users and securing the data streams. We are happy to help with the implementation or if you have any questions. Write to us!