There is a real buzz around the concept of cloud computing, being it private or public cloud, but little is written about the implications this new paradigm can have on the role of IT. It is actually the combination of cloud and a couple other trends that initiates the important transformation of IT we are starting to experience.
But let me start by the begin. About 10 years ago, we sponsored a study regarding changes in business processes in large enterprises. We asked business people to tell us the percentage of business processes that were changing one or more times a year. They told us that in large enterprises this was in the area of 40 to 45%. We then interviewed the CIO’s and asked them the percentage of their applications that changed once or more a year, and the answer was 5 to 10%. This results in a large delta, forcing the business to execute business processes with inadequate applications. That’s probably one of the reasons why Microsoft Excel is the most used management tool. Spreadsheets are built to complement the deficient applications. I’m sure we all have examples of this.
So, the question is really how we could speed-up the capture of business processes in applications. In a traditional environment, business people are explaining their needs to business analysts, who in turn translate those to IT language and transmit what needs to be developed/adapted to IT engineers. This double translation is often the source of misunderstandings and mistakes.
Business processes fundamentally consist in the execution of a series of elementary tasks intertwined with decision points. Looking further, we actually realized that most often changes are not in the elementary tasks themselves, but rather in the order and the number of elementary tasks that are involved.
What now if we could have the business people designing the business processes choosing amongst a collection of elementary tasks, while IT would focus on the development or sourcing of those tasks. We would separate elementary tasks and business processes in the same way we separated data and code 25 years ago with the invention of the database.
An approach, called service oriented architecture (SOA in short) actually focuses precisely on this. Elementary tasks are called services. Those services are maintained in a service repository. Service Lifecycle Management tools exist to govern the creation, sourcing, use and end-of-life of specific services.
As the younger generation of business people are computer literate, we could provide them with graphical and comprehensive business process management tools in which, using the service repository, they design their processes choosing from a collection of services.
IT would become the guardian of those services. They would build some of them, particularly the ones that provide the competitive advantage to the organization. They would source others from suppliers, or from the cloud. Indeed, there is one particular type of services, called web services, that lend themselves for being resident in the cloud. Are we there yet, not really, but things are evolving fast. The major drawback is that large software providers do not look at making available services by the piece. But the market will force that to happen. Start-ups increasingly provide interesting alternatives. The cloud and the associated “pay per use” approach make it an ideal tool for the provision of SOA type services to IT departments.
Such evolution is a complete change to the role of IT, who now becomes a supplier to the business, providing them with the functionality they need. The business people can now change their processes on the fly and adapt them to their own requirements, without needing to go to the IT department. This makes the company increasingly agile.