After reading about the throttles and quotas of Azure IoT Hub, I am struggling to understand how an IoT Hub can support millions of devices (Microsoft's promise). Maybe you guys can help. These are my thoughts so far:
Quote from Microsoft doc
Device connections:
120/sec/unit (for S2), 12/sec/unit (for S1). Minimum of 100/sec. For example, two S1 units is 2*12 = 24/sec, but you will have at least 100/sec across your units. With nine S1 units you have 108/sec (9*12) across your units.
12/sec/unit - does it mean that if I only have one scaling unit I can have max. 12 simultaneous connections at the same time? Or does it mean that I can open max. 12 new connection per second?
What does "Minimum of 100/sec" mean?
IoT Hub supports AMQP 1.0. My understanding was that with this protocol I can have a permanently open device <--> cloud connection to be able to send/receive messages "in real time". It would not be possible to support more than 12 devices with a single scaling unit if only 12 simultaneous connections were allowed.