* What “enabling enterprise agility” means for IT
Companies consistently tell us they’re concerned with “enabling enterprise agility.” Before you dismiss this term as YASAB (yet another silly analyst buzzword), you might want to take a look at what the IT executives who use it mean.
“Agility” is the ability to respond quickly to market changes. Specifically, your company might want to:
* Buy another company.
* Divest itself of a portion of its own operations.
* Roll out a new product line, or market a product line to a new breed of customers.
* Terminate an underperforming product line, or get out of a market altogether.
What does any of this have to do with IT in general, or data centers in particular? Nothing, and that’s the point. Business executives make such decisions based on bottom-line analyses: “We spend X and we reap Y in return.” Operating details – such as the potential impact on IT – don’t usually factor into the decision.
But being able to execute flawlessly on such decisions requires an IT architecture that supports agility – and that’s where data center managers end up involved. The problem, in a nutshell, is that over the years companies have fallen into an approach to developing architecture that’s logical, comfortable – and wrong.
Many IT organizations believe the proper way to create an architecture is “top down.” First, you assess the business drivers of your organization. Then, you select and implement applications that help you meet those business drivers. Finally, you design an infrastructure that allows you to optimally support the applications and associated business drivers.
Makes sense, right? But there are two major flaws with this approach. First, it can lead to the “multiple-infrastructure” problem. If business application A runs optimally on Solaris, but business app B prefers HP-UX, you’ll end up with a heterogeneous platform environment – and heterogeneity increases cost. The second flaw is that it doesn’t allow for dramatic changes in the business drivers or applications. If your five-year plan makes certain assumptions, and those assumptions change in five minutes, you’re hosed.
Designing for agility therefore requires a “bottom up” approach, even though it may seem counterintuitive. Recognize up front that your organization probably has two goals: to consolidate infrastructure for operational savings, and to be able to respond rapidly to business changes.
Meeting both goals requires a concerted effort to standardize around a core set of architectural principles. These include:
* Open standards – for servers, storage, and networks, among other components.
* Virtualization – again, servers and storage are the easiest areas in which to implement this.
* Common platforms – open-source operating systems will likely lower cost the most.
* Designed-in management and security – don’t view these as afterthoughts; they’re integral elements of the overall architecture.
