By PeterCoffee | April 26, 2010
As we all head for Cambridge in May, my own main goal for our Cloud Computing conversations there will be a focus on what people can do with cloud services — not what they need to buy to build clouds of their own.
The cloud computing stories that really resonate with me are stories of things being accomplished, in weeks rather than months or years, and with graceful handling of huge startup peaks of activity before a system settles down to a healthy but much lower baseline workload. Stories like the Census Bureau’s system for handling its decennial tornado of temporary labor force management; like Starbucks building its Pledge5 web site to engage the American people in community service, and getting it deployed in time for President Obama’s inauguration. These are things that either were not getting done in time, or would not have been attempted at all, with only past legacy models of IT acquisition and deployment.
If it doesn’t bypass the delays of a capital budgeting cycle, it doesn’t seem to me that the label of “cloud computing” is being earned. If an organization doesn’t get enormous acceleration and a vastly broader scope of possibility, it seems to me that all we’re really seeing is a new implementation of outsourcing.
The label of the cloud means something — several somethings — but it shouldn’t just mean the latest label used to sell more stuff.