Image may be NSFW.
Clik here to view.On 12 April 2011 it was Yuri’s Night— the night we space fans celebrate Yuri Gagarin’s 1961 history-setting flight into space.  We were celebrating 50 years of manned spaceflight. On that same day in 2011, we reached the end of support for SQL Server 2005 SP4. On 12 April2016 we will reach the end of extendedsupport for SQL Server 2005. That means no more hotfixes, no help from Microsoft and no love for your data still living in SQL 2005 databases.

I’m hoping your organization is already on its way to upgrading and migrating data and applications to newer versions of SQL Server.  SQL Server 2016 is already being used in production by early access customers. No matter which version you will be migrating to, I want to share with you some of the features and perks you’ll have available to you now that you are moving away from a dead version. Of course there are hundreds of enhancements that have happened since 2005, but today I’m focusing on those that a data architect would want to use in a database design for enhanced performance, security and data quality.

Columnstore Indexes

If you are designing for data warehouse type solutions, this is the closest thing we have for a "turbo switch" for SQL Server. Columnstore Indexes achieve high compression rates since they store columns together instead of storing rows together. They also support much faster query performance for batch and aggregation queries. They typically achieve 10x performance increases, sometimes even more. This feature was introduced in SQL Server 2012, but you’ll want the advances to this feature that came with SQL Server 2014.

New Data Types

So many new data types have been introduced since SQL Server 2005, but the ones that really stand out for me are DATE, TIME, DATETIMEOFFSET, the geospatial types, and the deprecation of timestamp.

It wasn’t until SQL Server 2008 that we had access to data types that comprised only the DATE or TIME portion of a point in time. So we had to do all kinds of conversions just to strip out unwanted data (00:00:00). We also had to make room to store that unwanted precision. Storing millions of rows of unneeded zeros hurts performance, both operationally and for backup and recovery.

SQL Server 2008 also introduced DATETIMEOFFSET, which allows us to track data in context of its time zone. If you remember the days when meeting invites did not include this important piece of information, you’ll know why this is important.

The spatial data types GEOGRAPHY and GEOMETRY and have added a new and feature-rich way of tracking places, their geometry plus special features that make it much easier to answer questions like "which is the closest" or "is this address located in this neighbourhood".

SQL Server was always an oddball when it came to the data type TIMESTAMP. In other DBMSs, this data type was one that included date and time, to a very large precision. In SQL Server, TIMESTAMP is a type of row version identifier that has nothing to do with TIME. So data architects migrating from other DBMSs were often bitten when they used the wrong data type. Microsoft announced in 2008 that it was depreciating TIMESTAMP and recommending the use of ROWVERSION, which is similar (but not the same) in functionality.

Encryption

SQL Server 2016 currently includes support of Always Encrypted, a feature that does just that: it support the encryption of data from application to database and back, so that it is better protected than solutions that encrypt data once it is written to the database. I’m always reminding you that keeping your CIO out of jail is part of your job description, right?

Data Archiving

As our data gets bigger and bigger, the size of our databases is growing as well. That means that performance takes a hit. Developers want us to take shortcuts on data quality to improve performance because size matters. One of the ways to help manage data volumes is to move "cold" data to other storage locations. Starting in SQL Server 2016, we can stretch a database to Azure, which means that data that isn’t accessed as often can be stored in the cloud and retrieved when needed. This allows our hot data to be local and fast, while the cooler data is more economical to store and still there and your application doesn’t even have to manage this difference.

JSON Support

In SQL Server 2016 we are getting support for JSON processing.  This isn’t the same as a JSON data type like we have with XML, but a set of import and export features for providing relational data as JSON documents and brining JSON data into SQL Server.

One Last Thing…

As vendors withdraw support for their products, third party tool makers do so as well. If you are supporting older, out of support versions of databases, it’s likely that your data modeling, data quality and data integration tools are also dropping support for these solutions. You’ll be left supporting database systems without vendor support and without professional enterprise class modeling and design tools.  I know how hard it is to keep track of databases that my tools can’t connect with.  Don’t let sticking with an old version be the end of data modeling support for that data.

Show Your Data Some Love

These are just a tiny number of the types of features that will be available to you when you upgrade to modern versions of SQL Server. The advent of security, data quality and performance features are leaving your old solutions behind, putting your data at risk and leaving your customer data feeling unloved. Go fix that!

Image may be NSFW.
Clik here to view.