As more IT organizations get comfortable with the definition of cloud computing, the one thing that becomes apparent is that there is no one single best way to go about achieving it. In some cases, deploying application workloads on public cloud services makes the most amount of sense. In other cases, that application workload because of security, governance or performance needs to run on a dedicated private cloud that may exist locally on premise or in a data center managed by a third party.
In much the same way that application workloads need to be optimized for different computing platforms, IT organizations are going to need to embrace multiple forms of cloud computing. The degree to whether those instance of cloud computing are shared will vary widely. In addition, financial considerations related to paying for IT as an operating or capital expense are going to be just as important. What is clear is that cloud computing is not an all-or-nothing proposition.
For example, IBM has developed a series of IBM SmartCloud Enterprise services. Right now, those services can be accessed in a data center managed by IBM, which is deploying portable data centers (PODS) that allow customers to essentially pay for units of data center computing capacity on demand. But in a webinar that can be found here, Fausto Bernardini, IBM director and distinguished engineer for IBM Global SmartCloud managed services enablement, makes it clear that IBM also intends to deploy PODs at customer sites, which it will then remotely manage. That means that regardless of the physical location of the IT infrastructure, cloud computing is going to be a federated endeavor.
Bernardini makes the argument that IBM is one of the few truly enterprise-class cloud service providers that gives customers the flexibility needed to adopt cloud computing on their own terms. Customers need to work through elasticity, availability, performance, security, governance and geography issues before deciding what application workload should run where. And over time, as the requirements surrounding an application workload evolve, Bernardini says customers need to be able to move that application workload as needed.
In short, IBM is starting to make the case that cloud computing should be more about flexible IT delivery service models, of which IT organizations need to have the utmost control over now and into the future.