Leaving aside technology, implementation, pricing and metal, on the human level I’m struck by how often cloud / flexible compute / call-it-what-you-will project approaches fall into one of two camps: Top Down or Bottom Up.
A print media company in the UK who F5 have worked with in the recent past provides a case in point for Top Down. Their approach to adopting cloud was driven by a new C-level executive. He came into the company and was determined to make his mark. Cloud services provided the vehicle for this to be done.
Decision to utilise cloud made, a flurry of activity ensued. His staff worked out how to move compute resource to the cloud without thinking too much about why this should happen. Lots and lots of ‘how do we do it’ meetings occurred – the existing environment was a complex one. Lots of time was spent working out how to port what already existed.
The results may or may not drive benefit to the company. It was a classic case of a project without objective: not even cost savings. Without objectives, there could be no clear communications of strategy. They started with the answer but neglected the question. The project may or may not work…
Bottom Down is much more focused on delivering specifics to the business. A large investment bank that has its HQ in the UK – a current F5 customer – is working on a project right now that is all about creating a flexible environment; to take resource to do specific workloads at specific times and to release that resource to be utilised elsewhere when required. The total amount of compute is therefore more efficient. This happens to be a private cloud project but the principles still hold.
Bottom Down tends to start with a realisation from IT architects that optimum ROI is best achieved by building a flexible compute environment. They therefore pick applications that work best in this scenario – use resources that are balanced against a genuine need. What they don’t do it try and build something cool from an IT perspective: they meet customer requirements with the budget they have.
The Russian Revolution of the title is a loose analogy which has some relevance here. While I’m not calling for a popular uprising to sweep away Tsarist CIOs, what would – in my experience – be a better business approach to adopting cloud would be less diktat, more collective decision-making. Flexible computing not fashion.