Cloud Bursts: Reality versus Hype!!I
The taxonomy of cloud computing consisting of private clouds, public clouds and hybrid clouds is by now well established in the cloud computing literature and discussion forums. Inspite of the fact that proponents of cloud computing are gung-ho about associated technologies such as cloud burst -- the technology that ultimately helps realize hybrid clouds, Gartner predicts that businesses will use private clouds as of now. Even Forrestor backed private cloud in its report earlier this year. Does that mean the end of technologies associated with cloud burst? In my opinion it is not!!! While it is true that we may be witnessing a slowdown in hybrid cloud adoption, yet I believe that enterprises will be amenable towards adopting the cloud burst technology in the coming days. As different lines of business within an organization continues to invest in building their own private clouds, we will witness a scenario where an enterprise has multiple private clouds. The private clouds could belong to different lines of business or it could belong to one line of business possibly spread across multiple geographies. Typically, the private cloud of a particular business line on one hand can remain under-utilized for a major fraction of the time and on the other hand a short surge in demand could result in over-utilized servers and reduced quality of services even though spare capacity is available in the private cloud belonging to the other business line. Using cloud burst technologies, it possible to create a shared and yet private cloud. Each of the different LOBs has a dedicated cloud so that the applications belonging to a particular LOB runs only in its cloud under normal circumstances. Though discrete, these clouds form a shared virtual infrastructure so that an application in an overloaded cloud can procure the computing resources from an under-loaded cloud. That means, during overload situations, an application requiring more computing resources can procure these resources from a cloud different from its current host cloud. Both these cloud, however belong to the same organization; albeit different lines of business. Scenarios such as this are likely to be more common in future and can be termed as internal cloud bursts. As enterprises adopt cloud technologies organizations are going to witness internal cloud burst more frequently than external cloud burst where applications move to an external cloud provider’s environment such as Amazon, Mosso or Rackspace when faced with scarcity of resources.


