Website - www.thirasystems.com
Email me - gins@thirasystems.com
Follow me on twitter - @daveginsburg

Sunday, June 22, 2014

Big Data and Big Retail (23 June 2014 by gins)

Big Data and Big Retail (23 June 2014 by gins)

It is common to speak of the differences between 'bricks and mortar,' aka Main Street, and the Internet economy.   And much has been written about the ascendancy of Amazon and other Internet retailers, and the difficulties faced by the corner retail store and even category killers like Staples and BestBuy.   But why do the Amazon's of the world succeed?   

We live in the era of Big Data, where our lives are one with the cloud, where Siri is at our beck and call, and where retailers understand our deepest purchasing secrets.  The real face of the cloud are the massive data centers run by the likes of Apple and Google, or independent operators such as Equinix.   They look pretty cool, with aisles of servers, storage, networking gear, and the necessary power and cooling.  And, in the case of Google, a vision out of the Color Kittens.   But in the end, it is all just bits.



Overhead Data Center View



Google Data Center Racks

Now, what if the real secret of retail going forward is to just treat products as bits.  And remember, I'm not talking about items that have already crossed the chasm from physical to digital, such as videos or books.  The warehouse now looks much like a data center, and just as you have complete and storage, the whole fulfillment chain is organized in the same way.  Who cares if instead of a few microns of flash it is a few inches on a shelf.  

It is the era of 'Big Retail', where the corner store has been virtualized to a shelf in million square foot warehouse (much like a process on a VM in the networking world).  The latest Amazon distribution centers are the best examples of this, but I'd expect other retailers to follow suit.  In fact, the way it is described, there is no logical connection between a given product and a location in the warehouse... when it is time to ship, the system determines the 'closest' instance of the product.   No different than in a data center with data replicated closer to a given compute element.  The parallels are scary.  But in a good way.



Overhead Amazon Warehouse View

Amazon Warehouse Racks