From a high storage SAN with a cluster of 30 servers to switching to AWS is indeed a globe-trotting step taken by Razorfish. Customer segmentation based on processing gigantic click stream data sets is at the core of Razorfish business model. As of 2010, Razorfish was collecting 25 TB per day, and with so much advent of social media and e-commerce, this number would have increased multi-fold by now. Building in-house infrastructure to analyze this click stream data will need expensive storage set-up with enough head room to handle seasonal spikes like holiday season, long weekend sales etc. Such business models rely on making key data available to customers on demand, and in order to support such a need will require on-demand infrastructure to collect/store/organize/analyze and share as Vogel points out in his talk. While data collection/storage is one piece of the puzzle, organizing & analyzing this data is where rubber hits the road. While AWS made storage possible through flexible and efficient AWS infrastructure that is both dynamic and cost efficient, platforms like easy map reduce in Hadoop helped Razorfish to organize and analyze the data. Razorfish is a good example of how to use cloud based web services to keep your operational cost in check, and pay as you use. And the best part of having this data on a cloud infrastructure like AWS is that it makes sharing the data easy.

Vogel’s example of sensors placed under the ocean sending the data directly to AWS infrastructure is definitely ground-breaking. Having said that most of the firms, esp in my industry, Banking, have an established infrastructure tied very close to the processes. So, getting banks to move their information from their private/secure infrastructure to a platform like AWS could be a hard sell. However, new startups like Razorfish can definitely take advantage of public/private clouds