What's new in SQL Server 2016? Computerworld UK speaks to Microsoft about its latest database platform...
Update 02/06/16: SQL Server 2016 has now reached its general availability release date. This means that all customers will be able to access Enterprise, Standard, and Express editions of SQL Server. See here for more details on Microsoft's blog site.
And Microsoft took full opportunity to take another swipe at competitor Oracle, promising to support
"faster performing, more secure than Oracle, and at less than one-tenth of the total cost of using Oracle to run the same transactional, data warehouse, data integration, business intelligence and advanced analytics workloads ."
Read on for more information on SQL Server 2016...
The decision to
bring SQL Server 2016 to Linux has grabbed headlines recently, but Microsoft has also announced a host of new capabilities to lure businesses to the latest version of its database platform. This includes improvements to security, advanced analytics running inside the transactional database and the ability to move data into the Azure cloud.
Microsoft has also been intent on wooing customers from Oracle, offering
free licences to Software Assurance customers.
Computerworld UK spoke to Microsoft’s
Tiffany Wissner, senior director of Data Platform Marketing, about why businesses should be considering an upgrade to SQL Server 2016...
1. SQL Server 2016: Hybrid cloud capabilities with Stretch Database
Microsoft wants tighter integration between its on-premise server products and Azure, and has architected SQL Server to run more easily in a hybrid cloud environment.
Stretch Database forms part of this strategy, allowing transactional data to be seamlessly moved between SQL Server and the Azure cloud.
Wissner says that large volumes of ‘cold’ historical data can be stored and accessed at a lower cost than keeping it on a SAN or on tape storage, meaning that data can be kept online for longer.
“If you think of customer scenarios where there are large operational tables of customer transactions,
” says Wissner, "often when it gets beyond a couple years of data customers aren’t able to keep that in the same transactional data, because they want fast performance."
“With Stretch Database they are stretching that to the cloud, so they are leveraging our cost of storage in the cloud, it is backed up and geo-redundant and that data is also online.”
2. SQL Server 2016: R analytics and business intelligence
Last year Microsoft acquired Revolution Analytics – a company specialising in the R programming language, which is widely used for statistical computing and predictive analytics.
In January, Microsoft released a standalone product based on the acquisition – R Server – also making it available in Azure HDInsight, its cloud-based big data analytics tool and Azure Machine Learning. Now the technology has found its way into SQL Server 2016, meaning that data scientists won’t have to pull data out of SQL Server to conduct analysis in R. Instead they can run it in a sandbox within the core database itself.
“This is the first time we will have advanced analytics as robust workloads in SQL Server 2016,” says Wissner.
Fraud detection is one example. Wissner says this would typically require taking out historical customer data to run a fraud detection model before running the results to the core database. “Now you can actually run those models on your core database in real time,” she says.
3. SQL Server 2016: Faster processing with improved OLTP and Columnstore in-memory capabilities
The latest version of SQL Server builds on the in-memory capabilities seen in SQL Server 2014. Microsoft claims that SQL Server 2016's analytics query performance is increased a hundred-fold, while transactional workloads are 30 times faster.
It is also easier to use, bringing together its two in-memory tools. “In previous versions of SQL Server our in-memory Columnstore was doing very fast queries," says Wissner. "With this release we allow you to use the in-memory Columnstore with in-memory OLTP to get fast analytics and fast performance."
She adds: “Typically what people have had to do is run their mission-critical apps and their fast transaction database – and then they would pull data out or 'ETL' [extract, transform, load] into a data warehouse. We have really bridged that gap to have the real-time analytics. So it is bringing mission-critical intelligence together in the same database without having trade-offs in performance.”
4. SQL Server 2016: Always Encrypted security
Improving security was a big focus of the 2016 release, according to Wissner, with end-to-end encryption one of the major additions. Wissner says that
the Always Encrypted feature - developed by the Microsoft Research division - protects data when at rest, in motion and in memory.
“Today we have encryption at rest,
” she says. “With Always Encrypted you will actually encrypt it at the application.”
“The application will keep the encryption keys and the data stays encrypted when in transit to the database, in the data base and [when it] returns. So you actually query it as encrypted data.”
5. SQL Server 2016: SQL comes to Linux
Microsoft’s growing acceptance of open source technologies took another leap with the announcement that SQL Server would be
made available on Linux next year.
“This is something that our customers have been asking us for a while,” says Wissner.
SQL Server on Linux is currently in a private preview. Wissner says that 8,000 Linux users requested to take part in the pilot in the first week, and Microsoft has been working with a select few behind closed doors to test the software.
“It is something that is going to be a journey for us in terms of bringing support to Linux,” Wissner says.
“But that is part of why we wanted to announce the private preview early, so that we could have customers engaged and involved with us as they bring it to their Linux environments.”
However, the Linux version will not have the full range of features, at least from the start.
“First we are going to bring the core relational database capabilities around transaction processing and data warehousing,” Wissner says. These will be available sometime in the “mid-calendar year around 2017”.
“Then we will look at bringing additional capabilities and workloads based on customer feedback as to what they want for Linux environments.”