SQL Server 2008 R2 is near release. Briefings for analysts are underway, and right now the picture is a mess. Although seemingly relegated to an adjunct position compared to Office in Microsoft’s thinking, R2 adds a great deal (source:softwarekeep.ca):
- CEP, although tools and integration are unclear;
- Master Data Management (MDM), although apparently Microsoft believes end users will do it;
- the long-awaited Parallel Data Warehouse (PDW) – but not yet (release date TBD.)
Many other features are in there too – but the briefing I attended last week didn’t discuss them much.
SQL Server 2008 R2 is positioned in marketing slides within the “Information Platform” vision, offered for “Empowered IT,” Pervasive Insight,” and “Dynamic Development,” which imply audiences, although they are not identified as such. The “mission-critical platforms” that are supported are “desktop & mobile,” “servers and datacenter” and cloud. SQL Server 2008R2 PDW and SQL Azure are also positioned as “What’s New” in the presentation. In typical Microsoft presentation style, 3 themes are listed: trusted scalable platform; managed self-service BI; and IT and developer efficiency. (I’ve reordered them to align with the three implied target audiences noted above.)
Microsoft led with BI (in an Office context), focusing on the new PowerPivot offerings for Excel and SharePoint, supported by the VertiPaq in-memory database, which is only available through PowerPivot. From that perspective, it’s not part of SQL Server, and developer and IT staff won’t see it as such. The self-service aspect is based on “report component libraries” and reports as data sources via Atom feeds. Report objects are stored along with metadata, and available to users (I’m not sure about the mechanism.) These ideas are in keeping with the ongoing evolution of SQL Server as a provider of services to be consumed by other elements of the Microsoft portfolio – a very SOA-compatible architectural design. PowerPivot itself is used in the management of PowerPivot via dashboards that monitor user applications. But Performance Point Server – now Services – is unmentioned. It’s now a SharePoint, not SQL Server, play.
Sorely needed multi-server management is added with this release, with policy-based capacity controls and easier management of upgrades across multiple servers. This reflects the use of SQL Server by larger organizations with more sophisticated management requirements, and it’s a step in the right direction. Collecting these ideas into a coherent discussion for that audience, and briefing analysts who track and service that audience, needs to be on Microsoft’s planning horizon. This briefing was not it.
MDM was next, and the discussion focuses on empowering end users to do it – from match/merge (MS calls it merge/match), versioning, hierarchy management, to role security and workflows. This is a good description – but not of a user-driven process. Master Data Services, the SQL Server capability for delivering this, sounds like it’s hitting the right notes (although I don’t cover MDM, and haven’t discussed a deep feature analysis with colleagues who do yet.) But it’s not an end-user activity, and suggesting it is will not help credibility.
The next key addition is the Microsoft CEP engine, branded as StreamInsight, and also architected as a service. It relies on SQL Server for “reference data” and apparently has its own set of output adapters, whose relationship to SQL Server is not entirely clear – whether they are shared, use the same controls, are monitored the same way, etc. is not clear to me. More discussion required there. The briefing offered no detail on how programmers will work with it other than “declarative query and pattern specifications,” suitable for 5000 events per second and greater than 5 second latency, except in the high-end DataCenter Edition. (More on Editions below.) Which Microsoft design and execution tools and languages will be involved were not part of the briefing.
SQL Server 2008 R2 will be available in 3 editions, plus PDW. Standard runs on up to 4 processors, and can be licensed for one virtual machine (VM.) The Enterprise Edition runs on 8 processors, 4 VMs per license, and the DataCenter Edition on more than 8 with unlimited virtualization. Features are added as you move up the ladder: security and resource governance features, data compression, Master Data Services and PowerPivot Management are available only above the Standard Edition. Differentiation between the Enterprise and DataCenter Editions as presented in the slide (shown here) seems only to be on volume. The Parallel Data Warehouse was not discussed in any detail.
The pricing discussion added to the complexity; introducing a Web Edition and two Azure Editions, which went by very quickly with no discussion. The full price list will be available May 1; although some pricing was specified in the briefing slides, I’m omitting it until it’s complete and “official.” Launch plans and many of the topics above are discussed in a Boulder BI Braintrust (BBBT) blog I participated in, available here. What was clear was that PDW timing will not be announced for some time.
There is a great deal going one in this release, far more than was discussed at the briefing. And Microsoft’s communications efforts will no doubt ramp up in the weeks ahead. But there are only weeks in which to flesh out, target, and deliver a coherent story. Today, it’s evident that there is much work to do.
Disclosures: Microsoft is not a client of IT Market Strategy