Microsoft's Plan to Bring Big Data to a Billion Users #strataconf

3 minute read
Virginia Backaitis avatar

When it comes to big data, Microsoft doesn’t think small. The company wants to bring big data insights to a billion users, said Eron Kelly, the general manager of product marketing for Microsoft’s Data Platform Group.

Needless to say, it’s a big goal, but there may be no company in a better position to achieve it. After all, Microsoft has a footprint of more than a billion Excel users and its Azure cloud service is enjoying triple digit revenue growth. Add red-hot Hadoop, big data and SQL Server to the mix and you’ve got the recipe for what could become a big data powerhouse.


Kelly said Microsoft’s big data offering, HDInsight, provides customers with a number of important advantages.

First, it’s highly scalable and works with structured, semi-structured and unstructured data and offers OLTP and OLAP capabilities as well.

Second, it can be leveraged by Lines of Business (LoB) on demand.

Third, it provides rich and engaging user experiences delivered via visualization tools that are intuitive and, even pleasant, to use. 

Fourth, these experiences are available via browser and/or mobile.

And finally, data can be leveraged through a hybrid model consisting of Microsoft/ Hortonworks HDP running on premise and on Windows and HDInsight running on Azure’s Platform as a Service (PaaS).

Experience Microsoft Analytics Today Via Power BI

Much of the advantage of using Windows Azure HDInsight is that users don’t have to concern themselves with technology. “The technology hides in the backend,” Kelly said.

Learning Opportunities

Microsoft’s Power BI, for example, which is in beta right now, allows Excel users to ask questions and to receive visualizations in return. Not only that, but it leverages Natural Language Processing (NLP), so that this can literally be done in plain, spoken English. You can preview it now.

A Change for Microsoft

Eighteen months ago, Microsoft made a strategic decision to build HDInsight following Apache Open Source guidelines. “HD Insight is 100 percent Open Source,” Kelly said. “We have donated over 25,000 lines of code to the Apache Hadoop project. They took over 65,000 engineering hours to write." Kelly added that Microsoft even has Apache Hadoop committers on staff. They have also invested in Apache Stinger and Hive.

Microsoft historians may find the aforementioned unbelievable.

A Few Use Cases

One of the earliest adopters of HDInsight is Microsoft Halo, a game which runs on xBox and also streams in the cloud. The game’s competitors often compete for cash when they play.

When Microsoft noticed relatively poor Halo players were suddenly winning, the company looked to big data and Hadoop to find out why. What it discovered is that these players had hacked the game (if you want the specifics as to how, ask) and made their characters headless and therefore harder to kill.

In another project, the city of Barcelona ran sentiment analysis on the services it provided. By running analytics, the city was able to predict where supplies of free pick-up and drop-off bicycles were too low.

As we learned yesterday, a number of large and mid-sized enterprises are afraid to commit to Hadoop. That may now change. Microsoft’s solid, trustworthy reputation combined with its powerful tools that offer a familiar feel, might alas give big data what it needs to become mainstream.