Microsoft's SQL Server 2014 crowns Redmond's data platform

The next release of Redmond's flagship relational database is done. In-memory computing; cloud and hybrid scenarios; and enhanced data warehousing capabilities are the product's major new features.
Written by Andrew Brust, Contributor

As reported earlier today by my ZDNet colleague Mary Jo Foley, Microsoft SQL Server 2014 has been released to manufacturing, and will be generally available on April 1st.

Quentin Clark, Corporate Vice President in charge of Microsoft's Data Platform Group, detailed the release in a blog post today, which went live at 9am Pacific Time. Just prior to the publishing of that post, I interviewed Clark, and learned a few important things about this release.

First, Clark covered the hallmark of this release, Microsoft's new In-Memory Online Transaction Processing (OLTP) engine, code-named Hekaton. Unlike other in-memory technologies from Microsoft, which are designed for BI and analytics, Hekaton is geared towards transactional, line-of-business, system-of-record database scenarios, where it can accelerate performance by up to 30x (i.e. 3,000 percent). 

Kid in a column store

Another big, though less well-covered feature in SQL Server 2014 is the addition of Clustered Columnstore Indexes (CCIs) to the product. CCIs are a data warehousing feature which add a true column store mode to SQL Server Enterprise.

While SQL Server 2012 offered non-clustered columnstore indexes, the latter only worked alongside conventional row store tables, and were not updateable. Meanwhile, SQL Server 2014 CCIs are updateable and serve as the storage repository for tables that have them. While the Parallel Data Warehouse edition of SQL Server 2012 already had CCIs, it's the SQL Server 2014 release that brings the feature to the mainstream version of the product.

Sprawling platform

And that brings up an important competitive factor. While a few of Microsoft's competitors (most notably SAP) do offer in-memory OLTP capabilities, and most competitors offer column store capabilities as well, they by-and-large do so via separate products. Microsoft's strategy, meanwhile, has been to add these capabilities to its core database product, assuring wider deployment, reduced migration burdens and therefore, hopefully, better adoption.

The SQL Server sprawl encompasses physical deployment options as well. SQL Server 2014 offers rich integration with Windows Azure (Microsoft's cloud platform), ranging from hybrid scenarios where the database engine runs on-premises but uses Azure cloud storage for its data files; to disaster recovery scenarios where "Always On" availability group nodes are deployed across the on-prem/cloud spectrum; and the ability to run SQL 2014 completely in the cloud, using Azure virtual machines.

In with the new, but what about the old?

When you take inventory of DPG's work of late, the output has been impressively prolific. Hekaton; CCIs; Power BI (including Q&A, Power View, Power Map, Power Query and Power Pivot); HDInsight (the Hadoop 2.2-based version of which goes into general availability today); and the Tabular mode of SQL Server Analysis Services that was released with SQL Server 2012, have brought a ton of innovation to a platform that moved at a comparatively slower pace in the last decade.

But what about the rest of the SQL Server stack? What about SQL Server Reporting Services? What about the Enterprise Information Management (EIM) suite that includes Integration Services, Master Data Services and Data Quality Services? What about the StreamInsight Complex Event Processing (CEP)/streaming data engine? And, indeed, what about the core row store relational engine in SQL Server?

Clearly, even a company with Microsoft's resources needs to focus and can't spread too thin. But the market is nonetheless getting a bit worried that these "legacy" components of the data platform may become neglected, and zombie-like. Fair or not, customers worry that when Microsoft isn't adding to a product or technology, it may soon take it away.


This remains a real concern, but those in the Microsoft ecosystem should take note of several upcoming Microsoft (and affiliated) events which are potential venues for announcements that could address these concerns, at least in part. These events include Microsoft's sold-out Build conference in San Francisco, April 1-4; Tech Ed North America in Houston, May 12-15; Worldwide Partner Conference (it was at this event last year that Microsoft announced Power BI) in Washington, DC, July 13-17; and PASS Summit in Seattle, November 4-7.

And, speaking of events, Clark points out in his post that Microsoft will be hosting an online, SQL Server 2014 and Power BI-focused event at 10am Pacific/1pm Eastern on April 15th. Clark will be joined by Microsoft COO Kevin Turner and newly-minted CEO, Satya Nadella, in hosting the event. Seems to me it may be worth watching, both for the Microsoft data platform faithful... and even its critics.

Editorial standards