Today is another tentative to explain part of the Microsoft Azure catalog of solutions.
As I did write about the different flavors of containers in Azure, I feel that it’s time for a little explanation about the different ways of running you IoT solution in Azure.
There are three major ways of running an IoT platform in Azure : build your own, Azure IoT Suite and IoT Central.
There are some sub-versions of those, that I will mention as I go along but these are the main players. I have listed those in a specific order, on purpose :
There you have it, I actually do not have to write another word 🙂
Alright, some words anyway. At the far end of the figure, you have what has always existed in the cloud and before. If you want a software stack, you just build it. You will probably use some third-party software, unless you really want to write everything from the ground up. Let’s assume you will at least use a DBMS, probably a queuing system etc. You might go as far as to use some PaaS components from Azure (IoT Hub is a good candidate obviously, along with Stream Analytics). Long story short, you will have complete control over the stack and how you use it. But with great power… etc. It is a costly solution, in terms of time, money, people. And not only upfront investment, but you will also have to maintain all that stack, even provide your users with some kind of SLA.
Let’s say you are not ready to invest two years of R&D into your platform, and want to be able at least to get your pilot on track in a few days. Here comes Azure IoT Suite. It is a prepackaged set of Azure PaaS components that are ready to use. There are several use cases fully ready to deploy : Remote monitoring, Predictive Maintenance, Connected Factory. You can start with one of those, and customize it for you own use. Once it is deployed, you have full access to each Azure component and you may evolve the model to suit your own needs. There are some very good trainings, with devices simulators, available. You can start playing with a suite in a few hours, see the messages and data go back and forth. You still have to manage the components once they are deployed, even though they are PaaS, so the management overhead is rather limited. But it is your responsibility to operate.
At the other end of the scope, we have Azure IoT Central. IoT central is a very recent solution to help you start your IoT project. We have been lucky enough to discuss the solution early on and I have to admit I have been convinced very early on by the product and the team behind it. So, the point is you have a business to run, and you might not want to invest millions to build and run something that is indeed not your core business. Start your IoT Central solution, configure a few devices, sensors, users and applications, and you’re done. Pilot in minutes, production in a few hours.
And like a good SaaS solution, you do not operate anything, you do not even have to know what is under the hood.
To conclude, I’ll say that the SaaS, PaaS and IaaS subtitles on the figure were here to remind you that the same choice principles apply here as anywhere in a cloud world : it is a choice you have to make between control and responsibility.