One of the fun things about working at a company that’s involved in an exciting open-source project like Hadoop is that, surprisingly often, you get invited to talk about it. On February 18th, I presented Hadoop (slides) to the folks at the Utah Java User’s Group. A little over one hundred people were in attendance. There was pizza, a talk about Android development, my talk about Hadoop, a raffle for door prizes (software licenses, pens, t-shirts, books, etc.; Cloudera contributed two signed copies of Tom White’s Hadoop: The Definitive Guide), and breakout sessions.
About half the audience had heard of Hadoop, but fewer than 10 people had ever written MapReduce jobs, so we went over the motivation and the basic ideas behind HDFS and MapReduce. I fielded a couple of great questions (how does Hadoop fit in with cloud offerings from Amazon and Google?) at the talk, and then, at the breakout session, about 10 of us (including folks from Overstock.com and Tynt) dug deeper. We discussed HDFS High Availability (a perennial favorite), using Hadoop for real-time serving (usually the answer is “don’t”), HBase, and Zookeeper.
The next day, Michael Finger of Overstock and I took some turns at Snowbird (snow day!). Thanks to Chris Maki and the rest of the UJUG crew for having me!