I have a backend device with an MQTT client connected to an open source MQTT broker (Mosquitto). On the other hand, I have many interface devices (PC, tablet, mobile) with a graphical interface, also associated with a broker. Thus, here Mosquitto works as a point of communication between the backend device and external devices and only transmits messages between them. Sometimes the amount of data transferred can be quite high (for example, 1 MB / min).
One backend device + many interface devices - this is one installation. I need to prepare the infrastructure for thousands of such installations, operating simultaneously. Therefore, my service must be very dangerous. My company uses Azure cloud solutions, so I started to learn this solution, and I have to admit that I'm a little confused. I read that I need to use an IoT hub, but it will require an MQTT Gateway to work with MQTT devices. On the other hand, if I understand this well, Gateway should work on some virtual machines, so I lose the scalability of my solution. I'm right? Now, if I need to support 100 thousand. Or 500 thousand. Devices, will I need another virtual machine? Another thing is that I need to integrate all this with some webservice (for managing internal and external devices), so I need some connection between webservice and the MQTT broker ...
Before I started playing with Azure, I thought that I just launched the MQTT brokerage service, and, by magic, it will be very vulnerable and will be able to serve thousands of devices.
Can someone explain to me how to bite this?
kappa source share