We need better real-time analytics capabilities. It's a bit challenging for us.

Moreover, there are some permission limitations.

In general, I am not really very satisfied with the tutorials that are out there. When Microsoft releases a new tool, technology — whatever it is — oftentimes it is not really very easy to get your hands on the insightful information and documentation, training courses, and other training materials. If you can find them, they may not explain what you need to know in a nice way. Often times they are a little bit fragmented. These user-oriented guides are something that should be better and released along with the products they are supposed to support.  

For example, we have servers in Azure called Azure Data Factory which I work with quite a lot. When a new feature or new release happens, finding the right documentation or resources that explain these features and how do you work with them is a little bit more difficult than it should be, in my opinion.  

There are probably a lot of extra features that might be considered to add to the scope of this solution. However, adding ports for different types of users may be one of the best. Certain users are advanced users and they can find their way around. But sometimes non-technical users or those that do not have a lot of technical background can find the complexity a little bit difficult to work with. Better handling of user gateways and privileges would be a benefit.  

Microsoft Analytics Platform System could have better support. Sometimes, it takes time to find support to tweak a model or get details.

The flexibility of this solution needs to be improved because you cannot make changes at every one of the different steps.

I think the biggest problem with the product is that it does a data ingest model, which is very expensive.

With the release of the 2019 version, I think Microsoft Stack has the capability for machine learning and in fact, it can now live in a Linux environment. I have had to deal exclusively with Microsoft technology for about 17 years working with the product and I personally have not deployed that capability, but I think that one thing is a big improvement in potential flexibility. If they can continue to do one important thing with each release like what Oracle is already doing, I think it is good.  

For example, if you do select star (*) from one table, Oracle returns the first 50 results. Microsoft will return all the results regardless of the number of rows in a table. I think these key features and functionality are something that Microsoft should improve because it makes sense how Oracle treats the customer queries. There are a few other improvements that can be made, but I can see key limitations in what Microsoft has in comparison to Oracle. They should concentrate on the most important features and add them.  

In my opinion, the standard technical support at Microsoft should be improved as well. It is not really helping a product to be noticed in the market if they allow support to just remain at the market standard. So I think it should be improved for clients who do not choose to pay for premium services.  

What I would like to see in upcoming releases is improvement in the machine learning and the AI to make it much easier for people to jumpstart their efforts. The foundation for this is probably already there the data platform, but making people able to do machine learning solutions and artificial intelligence very fast would help them have success and become more involved to learn about the technology. So, I would say to make that learning experience as short as possible and provide useful examples. Then that will help.  

Hybrid environments are complex to manage. We need to support customers frequently, even when they have done many training classes to ensure technology transfer.