Metascale is one of many Sears subsidiaries except that it focuses on Big Data born out of running just about “every distributed system…every flavor of data warehouse…and enterprise system you can think of,” says Sears CTO and Metascale Phil Shelley.
Metascale focuses on open source and in particular, how to introduce Hadoop into large enterprises. What makes Metascale interesting is that it was hatched by a large enterprise consumer of IT, not an IT vendor (not that they, too, do not consume IT). About 75% of CIOs do not how to spell Hadoop or much less, know what it does, adds Shelley.
What I like about this video is that he offers real examples of using legacy data with Hadoop. The Hadoop apps pertain to price elasticity, increasing speed of price changes, inventory and improving the user experience. Hadoop taps historical data based in traditional data warehouses. One Hadoop application he mentions reduced 6,000 lines of Cobol to 400 lines of Pig, a Hadoop programming tool.
Think of Hadoop as an “eco-system,” not a replacement for data warehouses, he says. Data still flows into traditional data warehouses, but is off-loaded into Hadoop where the analytical applications are run. Most Hadoop applications use structured data, not the gobs of unstructured data that’s so hard to tap.
But not to worry: we’re only at “the beginning” of this evolution, he says.
Shelley claims Sears is one of the biggest Internet mobile commerce destinations. Century-old and sometimes struggling Sears? Who would have thunk it?
Watch the video: Shelley knows a lot more about Big Data than me.
Follow me on Twitter.