There is certainly no doubt that cloud computing is real today and that it brings significant benefits to a wide array of IT organizations. There is also no doubt that a lot of what gets written about cloud computing is more wishful thinking than reality. We will use this newsletter to separate some of the reality of cloud computing from some of the fiction that surrounds it.
When industry pundits write about cloud computing they love to discuss a technique that is often referred to as cloud bursting. Cloud bursting refers to the ability of an IT organization to dynamically add public cloud computing resources to existing internal resources to meet a spike in demand. This capability would allow a retailer's IT organization, for example, to build the infrastructure that it needs to support the business's average demand and to use one or more public cloud computing solutions to support peak demand.
If there is a significant gap between the retailer's average demand and its peak demand, and if the peak demand was somewhat transient, then we can see how this could theoretically save money. The problem is that while it is very easy to talk about cloud bursting, it can be very difficult to implement cloud bursting particularly if it requires the movement of huge volumes of data over the WAN.
Cloud bursting is one example of how an IT organization can use public cloud computing solutions to meet temporary needs. Another example is that the IT organization can use an Infrastructure as a Service provider such as Rackspace to support the IT requirements that are associated with an upcoming launch of a new product.
We recently gave a survey to over 300 IT professionals and asked them a variety of questions about their current and planned use of cloud computing. The results of that survey are contained in Jim's report entitled Cloud Computing: A Guide to Risk Mitigation. As part of that survey we gave the survey respondents eight possible drivers and asked them to indicate which two drivers would have the most influence on their company’s use of public cloud computing solutions. One of those drivers was that the use of public cloud computing would allow their organization to meet temporary requirements. In terms of overall importance as a driver of the use of public cloud computing, this came in dead last. The top two drivers, by a wide margin were that it would lower cost and reduce the time to deploy new functionality.
We believe that over the next year or two that new technologies will be developed to better enable complex, high-value added tasks such as cloud bursting. In the mean time, IT organizations will use public cloud computing solutions only if those solutions are lower cost than alternatives and/or they enable IT organizations to be more agile than the would be otherwise.