Everyone is talking about cloud computing these days.  As a spokesperson at a technology firm I am often asked to my opinion of the various models by employees, journalists and customers.  Perhaps the most confusing concept in the area of cloud computing is Platform as a Service or PaaS.  I recently completed reading Marc Benioff’s book Behind the Cloud, which offers excellent insights into one of the industry’s best known PaaS offerings – Force.com.  Benioff provided the best explanation I have heard for the rationale for PaaS.  In recent weeks I have been testing out a new explanation of PaaS based upon Benioff’s principles, but with an expanded version of the story.  Rather than focusing on APIs and development languages I have been taking the approach of explaining the business value of PaaS – specifically from the perspective of a startup.

During the dot com era there were thousands of startups that failed.  Many failed because they were attempting to compete in over-capitalized sectors. Others failed because of an irrational exuberance about the potential market growth that could be achieved.  Some were simply too late to the market.  Those funded in early 2000 never had a chance, because by the time they reached the market they were already tainted with the stigma of being a dot com.

Once the stock market began its long decline in March 2000 people began to pull away from all but the larger Internet companies.  There were many lessons learned from the dot com era especially on Wall Street.  One lesson learned that you do not often hear outside Silicon Valley has to do with the inefficiencies that existed in the dot com startup business model.  These inefficiencies delayed time to market for most startups and consumed unnecessary amounts of cash during early operations.

Each startup has to go through the same repetitive processes to set up its business operations.  They must select a payroll provider to perform biweekly funds transfers to employees.   Group benefits plans and retirement savings programs must be established.  Accounting systems for paying suppliers and collecting revenues must be implemented.  These are just a few of the repetitive tasks that each new company has to engage in to set up a business.

For Internet-centric startups there are an additional set of repetitive tasks that must occur in the product development and technical operations organizations.  Each company must select a data center provider at which to host their equipment.  Once a data center is selected, the startup will need to choose which hardware platform they wish to run their applications upon.  For Windows, Linux and Unix there are numerous server hardware options that can be chosen.

Beyond the data center, server hardware and operating system, a number of management systems must be established as well.  Security technologies such as firewalls, intrusion detection and anti-virus protection must be configured and deployed.  Monitoring technologies to identify problems with the health, performance or functionality of the applications must be selected.  And finally disaster recovery and backup technologies must be implemented to ensure continuity of operations in the event of a network or hardware failure.

Setting up data center infrastructure is both time consuming and expensive.  During the late 1990s, most startups chose to purchase their own hardware devices and software licenses up front.  Software as a Service and Managed Hosting models were just beginning to emerge at the time.   Most dot coms had irrational expectations about how quickly their revenues would grow.  As a result, they purchased more hardware and software than was required.  Some dot com era startups spent up to one third of their funding on capital expenditures related to technology infrastructure.

These investments, both in time and cost, to build out data center infrastructure were necessary, but not differentiating.  And unfortunately the need to build out infrastructure distracted the leaders at these startups from focusing on the truly differentiated and innovative parts of their business model – the new applications they were going to build.

Jeff Bezos and his team at Amazon.com were among the first to recognize the significant inefficiencies that startups faced with setting up data center infrastructure.  Having gone through this infrastructure setup process in the mid-1990s, Amazon.com understood the challenges faced by its peers in the Internet technology sector.  As a result, they decided to launch a concept called Infrastructure as a Service or IaaS.  With IaaS newly formed companies would no longer need to go through the time consuming process of selecting co-location space, hardware servers, security, monitoring and backup technologies.  Instead startups could now go online and configure a data center environment within a matter of minutes.  Provisioning computing resources and storage capacity became as easy as purchasing a book or a digital camera online.  Amazon branded it is IaaS platform Elastic Computing Cloud or EC2 for short.  Today, Amazon’s web services are used by thousands of Internet startups which recognized the value proposition of the IaaS model.

In my next post I will discuss how Amazon.com’s vision although compelling was incomplete.