Last month, I shared the big news that Microsoft made two business intelligence (BI) announcements regarding Mobile BI and Power BI, a new Cloud Business Intelligence offering for Office 365 customers, noting a Public Preview would be available later this summer. That Public Preview will be available soon for anyone interested in evaluating Office 365 Power BI. You can get access to the Public Preview by registering at PowerBI.com.
After last month’s announcements were made, I was fortunate to be invited to personally participate in early previews and conversations with the Power BI Product Team. I’d like to share additional technical details, address commonly asked questions, and also provide a balanced technical perspective of what I saw, heard, and researched. If you have not read last month’s article, I suggest that you read that one first to get the full scoop on what is and what is not available in this initial Power BI release before continuing on with this article.
Fundamental Shift to Hybrid-BI
Keep in mind while reviewing Power BI that this release signifies the beginning of a huge fundamental shift of traditional, on premise Microsoft BI as we know it today to the cloud hosted Office 365 platform of the future. We are indeed seeing a Microsoft "Cloud First and Cloud Only" development philosophy in action with notable feature availability differences between on premise and cloud. The adoption of cloud and Office 365 will not be immediate and many enterprises will have a mix of on premise, Office 365 Power BI, and other cloud solutions. During the transitional period, new Hybrid-BI solution design patterns will arise. In doing research, I came across one of the first Office 365 Excel Hybrid-BI solution designs—Frontline Systems Solver.
Frontline Systems is the first known group to successfully migrate their Excel Solver Add-In from the old world of Office Add-In controls and VBA code to the modern Microsoft Windows Azure and Office 365 Apps for Office framework. They are pioneers in this new world of Hybrid-BI with Office 365 overcoming bleeding-edge technical design challenges with authentication, exchanging data between on premise Excel workbooks, cloud hosted Office 365 Excel workbooks, and remote application servers residing on the Microsoft Azure Cloud platform.
Pivotstream is another cloud hosted Microsoft BI platform solution that also has been an early adopter of Hybrid-BI design techniques. Pivotstream’s true beauty is in their much deeper, professional level, traditional Microsoft BI offerings in the cloud. There are other Hybrid-BI solutions in the market today. Most Hybrid-BI designs use classic VPN, Service Oriented Architecture designs, cloud data source connections, and OData to bridge on premise and cloud solutions together. As you play with the Power BI Preview, begin thinking about how you would design Hybrid-BI solutions to work with it in the future.
What to Expect and What is New
This initial Power BI Preview version is an exciting yet ongoing work-in-progress that will continue to evolve with more frequent feature release patterns than we have historically seen in the past. Some of the features shown in the Microsoft Power BI demos such as Natural Language Query will not be immediately available but will show up later during incremental version releases. Other features should continue to improve over time as the offering matures.
In digging deeper into the Power BI technical details, I learned that Power BI is limited to stand-alone, Office 365 hosted Excel 2013 PowerPivot files. There's no Multidimensional or Tabular flavor of professional SQL Server Analysis Services in the Office 365 Power BI cloud world per se. As it stands right now, Power BI is more of a Personal or Departmental BI offering. If Power BI can be successfully combined with hosted SQL Server Analysis Services for features like dimensional security, true direct connect, and partitioning for real-time reporting, then enterprise BI level roll outs may be viable. The supported Office 365 Excel PowerPivot file sizes have greatly increased from 10MB to 250MB. Considering PowerPivot data compression ratios, a lot of data can fit into a 250MB file.
There is also a significant, new feature that allows scheduled data extract transfers between on premise data sources and Office 365 via a Data Management Gateway. Power BI also has its own Management Dashboard feature that shows query and system-wide data usage. Similar self-service BI management, governance, and best practices that were needed with on premise Excel PowerPivot remain needed in the Power BI realm.
Power Query, formerly Data Explorer, now has Enterprise Data Search capability allowing users to locate and extract data from internal data sources published to a data catalog as well as the external data sources. Another key point mentioned by the Power BI Product team was additional data sources and data providers would be available in the Office 365 platform soon. A really cool new feature in Power BI is the capability to publish and share Power Query queries. Shared Power Queries conceptually sound a lot like SQL Server Reporting Services Shared Data Sets and allow for similar design, deployment, and consumption patterns. I have seen one blog covering usage of parameters with Power Query to date. I could see Power Query parameters being quite useful with shared Power Query queries since Power View does not accept parameters yet. Power Query parameters is an area that I am going to test during my evaluation.
Moving on to Natural Language Query, it will not be available right away. Future iterations of Power BI will allow for free-form questions to be typed into a speech bubble for instant data visualization. As is, the shown implementation looks a little odd. In talking with BI industry thought leaders, there does not seem to be much appreciation for the Natural Language Query feature or Power Map, formerly GeoFlow. Both are interesting and showcase well in sales demos. However, Power Map is not practical for most real-world BI use cases. Unlike Power Map, Natural Language Query might be genuinely much more useful and is something BI experts want to explore in a hands-on manner to better understand the real-world value.
Finally, Mobile BI and the capability to view Office 365 Excel Power View reports from the browser on any device with HTML5 support—this is a HUGE improvement and fantastic move away from dying Silverlight. I am truly disappointed that Mobile BI is an Office 365 Excel only feature. If you have on premise Microsoft BI: Excel Services, Reporting Services, or PerformancePoint and need mobile BI, there are browser-based and third party options out there. On premise Silverlight Power View is usually the biggest challenge unless you have a Microsoft Surface Pro tablet. I have been giving Mobile BI Today presentations with Power View disclaimers for several years now packed full of tips, tricks and warnings. I can’t wait to test out the new Mobile BI native apps for Windows 8 and Windows RT devices. Note that iOS Mobile BI native app is expected soon for iPad users. There are no Mobile BI native apps planned for Android.
That wraps up today’s news on the first Public Preview of Power BI. Additional deep dives into various components of Power BI are showcased on the Microsoft BI Product Team blogs: