Microsoft MVP Global Summit 2016, field notes

This year’s summit was held as usual. Almost as usual. Hyatt Regency Bellevue Hotel and all other Bellevue hotels from November 7 to 10 were filled with geeks and nerds.

But on November 8, I went into the central hall of Hyatt and suddenly realized that there were not geeks around, but normal-looking people, although with the same enthusiastic faces. A minute later, I realized that I was at a meeting of the Republican Party, and on the stage they announced that the Republicans had caught up with the Democrats. Everyone was screaming with delight, although no one had believed in victory ...

The summit was held under the sign of Azure, Microsoft Cloud Technologies. The dam collapsed, the skies fell, and almost all the sessions were about Azure. A couple of years ago, few programmers came across the development, installation, and maintenance of programs in Cloud. All this happened locally, near by, at workstations or on servers. But it seems that soon on the local computers there will no longer be large programs. Local computers will only be used as terminals during mainframe times. Spiral spiral.

Azure now has over a hundred different services . I hardly know even approximately what they all do and what they are responsible for. And every month there are more and more of them. One list of preview services is more than 20 pieces. The guys at Microsoft showed us new versions of services and many times it was Wow! We saw something that no one had done before. Microsoft in recent years has fused its industrial programming experience with fresh methods and ideas from the open source community. The guys are not rushing forward at all with the old, familiar speed of enterprise monsters and are holding on tightly.

Visual Studio Code - a light version of VS, runs on Mac, Linux, Windows, supports a lot of languages. And don’t worry about money, VS Core is free. I repeat, it is lightweight, geeky and similar to the works of JetBrain, but in no way to the work of Microsoft.

No one expects big changes from SQL Server, but what about the Data Lake Store or the SimpleStore, the Data Factory? We can say that it's time to try it all.
Try now to find on your server something that does not find its counterpart in Azure. Have you found? Unlikely. Now let's try the other way around. What is the result?

Came, it's time to begin to deal with the prices of cloud services. Already not expensive, and it makes sense. By the way, Price Calculator is now not an example easier and more understandable.
I do not campaign for Azure. In this segment there is almost a monopolist - Amazon, and Google is approaching. Understand for yourself what is better for you.
But I state my feelings from the point of view of a programmer working in his specific niche, with his own technologies. You may have a completely different view, absolutely correct and completely opposite to mine.
And now I'm turning to myself. It’s time for me to begin to deal with prices, otherwise I’ll leave behind the engine, for which I am destined to run and run ...

But it’s better not to pay attention to the calculator, since all Azure services can be tested for free. Many services have an initial free level.

What about the fear of becoming addicted to Microsoft? Reasonable sets of programming languages ​​and operating systems are supported. It's good.

Want something that isn't there yet? I go to the Feedback site and post my idea, and if enough people vote for it, it will be implemented. I tried using Logic Apps, API Services as an example. Feedback works, oddly enough.

I am now interested in Machine Learning, and Microsoft is not in the last position in it. Azure ML Studio is an interesting thing. The editor works in a browser, and the models train and work on Microsoft servers. Ready-made models for speech, language, video, search are brought together under the heading of Microsoft Cognitive Services . If you are creating Neural Network models yourself, then the Cognitive Toolkit , known as CNTK, will help. And if you are a hard-core programmer, that is, the Distributed Machine Learning Toolkit with the trendy LightGBM . Do not forget about R Server , which is now supported by Microsoft.

In general, according to the results of the summit, we can conclude that now there are two directions in the trend. Cloud is already out of childish pants, ready for production and is starting to push the usual, standalone applications. Machine Learning is the hottest direction, much more to come, and it’s not too late to jump into the second car.