Earlier this week, I came across an article that did a great job clarifying one of the major benefits of migrating infrastructure to the cloud. The article — written by Jason Fried of Basecamp — doesn’t talk directly about cloud platforms. Instead it addressed one of the ways in which companies can stifle their own forward momentum. Removing that blocker is a capability cloud technology offers companies.
Fried’s focus was on the way companies develop dependencies between different workers, teams, and departments. I’m sure you’ve all experienced a situation in which you couldn’t do your work until someone else had done their work. Your work depended on their work, and if they were blocked, so were you.
Fried argues that many companies make the mistake of organizing their workflows with artificial dependencies of this sort. You depend on me, and I depend on someone else; the CEO can’t make a decision until he receives a report from you, and you can’t write the report until you get some data from a supplier. If you think about it for a few seconds, I’m certain you’ll be able to identify dependencies like this in your own organization.
Sometimes, there’s nothing to be done about it. Some dependencies are inherent to the work we’re doing, but smart organization and tool choice can help people to work independently or in parallel. The right communications tools, for example, can help people collaborate asynchronously rather than wait on input from each other.
What’s the relevance to the cloud? If you’re a developer or an executive, you might already see where I’m going with this. With traditional IT infrastructure, the IT department was the gatekeeper of all IT resources. Those resources were expensive and fragile, and it was the job of IT to make sure business critical operations were given priority. Getting access to IT resources to test out new ideas or implement new features is a long process that frequently goes nowhere.
Even when you can get the go-ahead from the IT department, setting up dedicated hardware in the company’s data center can take weeks or even months if the infrastructure has to be bought. Innovation has IT as a dependency, and traditional IT is slow moving.
Cloud infrastructure, by virtue of its elasticity, flexibility, and on-demand deployment, remains a dependency, but it’s one that can be almost entirely under the control of the developer or exec in question. Cloud infrastructure is not expensive and fragile; it’s cheap and ephemeral. Spinning up a new server for a development project or to test a feature branch takes no time at all. In fact, the whole process can be automated so a developer can fire off a command in a terminal and have their code running on a server before they get back from the coffee machine.
The removal of unnecessary dependencies is key to innovation, and freeing infrastructure from dependencies that stymie innovation is a killer feature of Infrastructure-As-A-Service.