Getting a handle on Hadoop
I like to keep up with Microsoft's financial analyst meetings so I can get a sense of company strategy at a high level. Last week I was checking out the Financial Analyst Meeting 2013 and specifically Kevin Turner's message. (I found out there was a power outage during his talk, but it looks like the transcript contains everything he said.) Toward the end he goes into Microsoft's big-data strategy, and it got me thinking. Here's an excerpt from his message:
...When you look at our big data assets and you think about big data and what we do with big data here in the company, the ability to bet on Hadoop is at the heart of our big data story.
So what we've done with Hadoop is we've made a full bet. We're completely aligned with Apache on Hadoop. So as Hadoop evolves and changes, the ability to be aligned with Apache on that is something we offer that a lot of companies don't.
The second part of the big data story that I'd share with you is the ability to have Hadoop run on Azure. And being able to take Hadoop and run those clusters up in our cloud is something that we're really proud of.
The third thing I would tell you is we've got something in beta right now called Power BI. And on the Power BI front, the ability to have incredible, rich, visual presentations with graphs and three-dimensional visuals is something we offer very uniquely on the Power BI standpoint.
So when you look at the opportunity across big data, we think we're very uniquely positioned with SharePoint, with Excel, with SQL, and with the offerings that we've got in market because we also have a parallel data warehouse SKU available as well. We think that really gives us a very nice way and ability to really get into it.
After reading this I thought, "I don't even know how Hadoop works, on a technical level. I should ramp up on it, since the executives obviously think it's important." So I searched around a bit to find some ramp-up type information. I ran across this video interview of Mike Olson, CEO of Cloudera, called What is Hadoop?, which is helpful. In the first six or seven minutes of the interview he gives a nice, succinct description of Hadoop and how it crunches data.