The good news is that SQL Server 2016 has been released and is now available. This version has been anticipated with some excitement since it eclipses previous releases in both the number of new features and the improvements and extensions to existing features. The better news, at least to developers like myself, is that Microsoft has made good on its promise and the SQL Server Developer Edition is now a free download. You will, however, have to log on using your Microsoft account.
The installation process is faster and simpler that previous versions, though “simpler” might not be the right word. The installation disk no longer includes either the SQL Server Management Studio (SSMS) or the SQL Server Data Tools (SSDT), although the installation program does provide download links. The Management Studio has also been made a free download; an SQL Server license is not required. SSMS 2016 can be downloaded here https://msdn.microsoft.com/en-us/library/mt238290.aspx and can be installed alongside of previous versions. However, the 2016 version should work with previous versions of SQL Server, although its use with unsupported versions of SQL Server is, well, unsupported.
You can download the SQL Server Data tools here: https://msdn.microsoft.com/en-us/library/mt238290.aspx, but you need to know what you want. As you are aware, with previous versions of SQL Server the installation program either integrated the data tools with a Visual Studio you already owned or installed them with the Visual Studio Shell. These are now two distinct options, each with its own download. If you already have some variation of Visual Studio 2015, you can download SSDTSetup.exe, which will install SSDT as an upgrade to your Visual Studio 2015 whether it is the Professional edition or one of the Express editions. If you so not have Visual Studio 2015, you can download an iso file from the same page that includes everything you need. Running the setup program from the iso file will install the Visual Studio Shell (isolated) and then install the SQL Server Data Tools.
The installation of Polybase, the set of services that integrates SQL with Hadoop and Azure data, requires Java. The many different options for downloading and installing Java can sometimes be confusing, at least if you are not a Java aficionado. As of this writing, jre-8u88-windows-x64.exe installs the most recent stable run time and is sufficient for a successful installation of Polybase.
The installation of the developer edition did not enable TCP/IP. If you are installing this edition you may wish to enable TCP/IP. This will be necessary if you plan to use the Polybase services. Clearly, you will need the network protocols to work with Hadoop and Azure data in any case , but without TCP/IP the Polybase services will hang on startup. No error, no timeout, just hang.
I am also pleased to report that there were no R installation issues, on a machine on which, a release candidate was installed. I attempted a full installation on a machine that had a CTP installation and was not surprised that R failed to install, as this was a problem with the release candidate installation as well.
As mentioned in the discussion about the Release Candidate, there is no longer any script required for the configuration of R. All you need to do is run the install program, enable external script using sp_configure, restart the server and the launch service, and it just works.
SQL Server 2016 is now out. The developer version is free and the installation is seamless, providing no excuse to get started mastering all the new features. This task is daunting, since the new features are both numerous and important; only the disappearance of SQL Server 2000 and the introduction of SQL Server 2005 compares in scope.
In SQL Server 2012, virtually all the new features seemed directed towards analytics and business intelligence. In 2014, the most exciting new feature, in-memory optimization, focused on transaction processing. In 2016 the boundaries have begun to blur and a new priority, real-time analytics, is emerging. Rapid transaction processing that supports fast analytical queries and data mining, is now the direction in which we are heading.