I was looking at the number of employment opportunities that had the word HADOOP in them. I found it fascinating that a new technology could be adopted by the industry so quickly, and show up on the employment websites. But why would Hadoop be considered by every enterprise? I don’t think it should be. Here’s why:
The perfect match for this technology are companies that want to do proactive forecasting, trending, fraud detection, fraud prevention and custom purchasing patterns on LARGE AMOUNTS OF COMPLEX DATA. The data could include test, click stream, log files, social media, documents and geographical data (to name a few). That, unfortunately, is NOT every company out there. While I do believe it is extremely proactive for EVERY organization to look at and understand Hadoop, data integration or master data may be a better place to start your data management initiative. BUT IT SOUNDS FUN, DOESN’T IT? HADOOP!
Does Hadoop fit into your corporate data architecture? If it does, there is a box (or component) in your data architecture strategy design document that would include something like Hadoop, and it MAY say “For Future Analytic Purposes.” There is data latency within the Hadoop environment that some companies may not be aware of. For instance, the data must be loaded into storage devices accessible by the Hadoop server. Programs must be set up, as well as an area for the results to be interpreted. So the data latency (for right now) could be an issue if your objective is proactive fraud detection and prevention.
I can’t wait for Hadoop to mature… WAIT UNTIL NEXT YEAR!