What happens when developers choose the databases?

We're seeing a broadening in the number of databases-- from a development perspective-- that people are managing. I think what's happening, one of the large trends that we're seeing, is that developers and development teams are making the technology adoption choices, rather than that choice being handed down to them from on high, from an IT department. And so what this is resulting in is just-- developers have got different tastes, and it's resulting in an explosion of the numbers of databases that developers are choosing to use. Those choices are made on a variety of bases. So for example, developers are choosing to use document databases like MongoDB because they perceive it's easier not to have to deal with schema-related issues during their development. It makes deployment easier. And they're bypassing what they might have chosen just a few short years ago in terms of relational databases, where it's seen as more of a heavy lift to have to deal with schema. That being said, there's still plenty of relational databases being chosen. One of the fastest-growing databases on the Azure platform, for example, are the relational databases, either Microsoft's Edge and SQL database, or the Postgres and MySQL variants. There are a variety of reasons to why developers might choose to use different types of databases, and even things that we may not have considered databases up until now. So developers are using Elasticsearch service inside AWS. And that technology is being used to be able to query and work with raw text and log files. And we might not have thought of that as a database, but it is. There a variety of other things, too, that are driving some of these choices. There are operational stores that people and developers are storing their data in. There are stores that we're using for decision support. There are a variety of use cases that are also driving these technology choices. We're sort of beyond just "system of record type" stores right now. I mentioned Elasticsearch. One of the other use cases is decision support, being able to support reporting and analytics, and a wide variety of choices. And then with that wide variety of choices comes all of the issues involved in getting data in and out of these stores, which is also a challenge that is new. If you have a variety of databases, you have a new set of problems that you didn't have when you just had one database. Looking forward to the future, I think this gets even more diverse. As we see more open-source projects get started or forked from existing projects in order to serve new use cases. As we see the cloud providers operating some of this open source and building their own data management solutions, in the future, we're just going to live in a world that's going to become more polyglot and more complex, which I think is an exciting place to be.