If you’ve followed this blog, you’ve heard my expositions (rants, screeds) on the fragile and often tenuous relationship between application development and IT operations. My contention has been that these are two worlds divided by opposing cultures and goals.
To wit:
– Apps: freewheeling, ideological, collaborative, self-organizing, and organic
– Ops: policy-driven, command-and-control, focused on repeatability and efficiency
One wants speed, the other control. More often than not, control wins.
These are two very different but interdependent worlds separated by a cultural and process gap that adds cost and risk to application deployment and maintenance.
It also means that there is a whole lot of waiting — abiding, doctors’ office style waiting.
Historically, waiting on IT has been one of those immutable and universal costs of doing business. “No” was the watchword of IT. In fact, “no” was often dispatched with such frequency and alacrity that one might have questioned allegiances from time to time.
Of course, this is a dated and anachronistic view of IT. Times have certainly changed.
Most IT leaders are on a quest for a way to say “yes.”
Today, the goal is to transform IT from a centralized bottleneck to a self-service policy-based infrastructure that provides business lines with the capacity and services they want, while ensuring that IT doesn’t sacrifice the control and predictability they need.
The rise of cloud computing — most notably swipe-a-credit-card-style-cloud like Amazon EC2 — has been a useful context for making this transformation. While cloud computing may not be for every organization or every application, the self-service model has almost universal appeal, and it has helped shine a light on the IT bottleneck.
At the heart of this bottleneck is the fact that control always trumps speed when they’re presented as mutually exclusive options — which they always have been. This is because, traditionally, the application deployment process and the administration of IT policies have been a manual undertaking, where human labor is the key constraint. In this context, the incremental cost of deployment speed is too high to justify. It’s a model that simply doesn’t scale.
As a result, we’ve become accustomed to waiting on IT.
But the reality is that automation can reduce the incremental cost of deployment speed to near zero. This, combined with automated administration of policy, can give IT the control that they need while ensuring lines of business have the speed they so desire.
As a result, we’ve become accustomed to waiting on IT.
But the reality is that automation can reduce the incremental cost of deployment speed to near zero. This, combined with automated administration of policy, can give IT the control that they need while ensuring lines of business have the speed they so desire.
This is exactly what we’ll discuss in a webinar we’re hosting on Thursday, July 16th with our friends over at Active Endpoints. We’ll discuss how organizations can combine orchestrated process and automated deployment models toward the goal of self-service IT. We’ll discuss the key requirements for a self-service deployment model and you’ll even see a demonstration of a self-service deployment process in action.
It should be a great event—I hope you can join us: http://www.rpath.com/corp/bottleneck
