Business Intelligence to get a boost from "Gemini" -- a BI adjunct to Excel.
In early October, Microsoft held its second annual Business Intelligence (BI) Conference. A heavy emphasis at this event was put on "Kilimanjaro," the code name for an interim release of SQL Server 2008 that is due in the first half of 2010 and will ostensibly be delivered in concert with Office 14. A key focus of Kilimanjaro is so-called "self-service analytics," which will be delivered in the form of an enhanced Reporting Services offering and, perhaps more importantly, a true BI adjunct to Excel, code-named "Gemini."
Does a self-service technology imply less work for developers and implementers? Does enabling users force a commoditization of the BI development and architecture skill sets? Does moving BI to the client deemphasize a proper server-based, enterprise approach? While these are all valid concerns, I think a close inspection of Gemini's capabilities -- and a consideration of the BI market -- show it will be more of a boon to BI developers than a hindrance.
Gemini provides several things to enable real analytics on the PC. The workstation solution marries an in-memory, client-side version of SQL Server Analysis Services -- which is a very exciting offering all by itself -- with Excel; the server-side solution marries Analysis Services with Excel Services and SharePoint. Both the client and SharePoint server versions of Analysis Services are fully programmable using the same APIs available for today's stand-alone server version.
Power to the Client
The Gemini client allows for the import of massive amounts of data -- over 20 million rows in the keynote demo -- and light extract, transform and load (ETL) on a user's PC. It then builds an Analysis Services OLAP cube and can quickly generate a PivotTable against it back in Excel. Excel 14 will add a new feature called "slicers," which essentially allows for multiple filtering dimensions to be added to a PivotTable through a very intuitive user interface. This strengthens the offering even further.
According to Amir Netz of the Analysis Services team, the Gemini client will not offer any APIs or other developer hooks. I do hope the team changes its mind on that. However, Excel is highly programmable through Visual Studio Tools for Office and, as previously mentioned, the client-side Analysis Services engine will be programmable as well. This means that users and developers can use the Gemini client as a design tool, and then developers can create apps around the generated cubes.
Many BI purists may protest that providing BI capabilities on the client is architecturally wrong-headed. But the fact -- and Microsoft's motivation and justification -- is that customers are already carrying out the equivalent tasks today, by keeping and manipulating huge volumes of data in spreadsheets. Try as they might, developers, IT departments and Microsoft itself can't stop users from creating these "spreadmarts."
But let's look beyond the reactive necessity of Gemini. Let's look at what takes place when a user wants to publish her self-service BI solution. She uploads it to a SharePoint server using Excel Services, and when she does, a few important things happen:
- IT can monitor the solution, through a very sophisticated dashboard
- IT can allocate more computing resources to the solution if usage statistics so dictate
- IT can evaluate the solution for possible full enterprise implementation
- Microsoft can sell more SharePoint Enterprise client access licenses (CALs)
All four of the above points are coups for Microsoft; not just the last one. All of them strengthen Microsoft's data center play by strengthening -- and drawing from the strength of -- its client PC franchise and the users that franchise serves. IT gets some insight and influence into otherwise "rogue" data analytics initiatives. Business users get more hands-on capabilities, and thus more autonomy, but they also get IT's buy-in. Developers and BI architects have more opportunity, because users are able to take their ideas further and really "get" BI.
Staying on Message
Microsoft using its client bona fides to strengthen and build a server franchise is certainly not limited to this scenario. Consider the recently announced Azure Services Platform for the cloud -- its strength and value is built on the foundation of the Windows client operating system.
And consider the runaway success of Office SharePoint Server. In fact, just consider the name: it starts with "Office," a suite of heretofore client business-productivity apps, and ends with "Server." Starting with the client and ending with the server is Microsoft's go-to market strategy, and it's working, both for it and for .NET developers.
The client market may be saturated, but it's still very important, as are Microsoft's credentials there. The server, the data center and the cloud may be growth frontiers, but a company and its developer stack don't make it in those arenas just by showing up. To be a player with new constituencies, Microsoft has made sure to continue to cater to its base. That such a strategy works in the software industry as well as it does in politics is no coincidence. Microsoft is waging what looks to be a winning campaign.
Andrew Brust is Research Director for Big Data and Analytics at Gigaom Research. Andrew is co-author of "Programming Microsoft SQL Server 2012" (Microsoft Press); an advisor to NYTECH, the New York Technology Council; co-moderator of Big On Data - New York's Data Intelligence Meetup; serves as Microsoft Regional Director and MVP; and is conference co-chair of Visual Studio Live!