Are we as IT professionals ever going to get to a point where we can say “No, I won’t just blindly implement this new widget because I saw an article on it in the WSJ. Tell me what you’re trying to do and we’ll see if it’s the right pathway to meet your goals?” 

Earlier this month John Foley over at Information Week (who is quickly becoming one of my favorite cloud commentators) wrote a piece about the federal government and the mandate from Federal CIO Vivek Kundra that agencies must move services to clouds. Foley wrote a little over a year ago  that the ROI model for going to clouds is vague and anecdotal at best, and not much has changed there. 

I recommend it because it’s a good read, along with a lot of the other pieces he’s written. I don’t necessarily agree with all the conclusions Information Week comes to, but it’s very informative and raises great questions. Foley’s fundamental point in the referenced article focuses around the adoption rate and feasibility of clouds for government organizations. The questions he brings to the table are fundamentally applicable to private enterprises as well.

Foley’s article made me think of my opening question, as I feel we’re seeing the same attitudes toward cloud cropping up that the industry felt over “Web 2.0″, “SaaS” and all of the other precursors to cloud (remember when deploying ERP was the trendy fad for EVERYONE?). The danger is that cloud will end up going the same way as all of those other “killer app” scenarios. When we get mandates from the top based on marketing hype, single use case scenarios, and vendor supposition, we end up with situations where these technologies are roundly declared to be failures.

This is tantamount to a CIO saying to his team: ”We must be at 80% iPad users in the next 12 months; it is after all, magical.” Forget use case, forget requirements, forget even cursory analysis, just go do it—it’s a “Fire, Ready, Aim” way to think.

Don’t get me wrong, I think that the promise of private cloud is real.

I’ve had plenty of experience working with businesses in the field who see real savings and efficiency from virtualizing and changing the way they operate. The problem is: it isn’t universal, and this is very much a YMWV (your mileage will vary) technology. Public cloud still suffers heavily from these three key roadblocks:

1. Bandwith cost
2. Security
3. Institutionalized thinking

You don’t just set a mandate to overcome these three things and assume it’s all going to work out (especially when you’re playing with tax money). 

Requirements ought to be the first step in the implementation process. Always, always, always. Enough with the technology mandates and let’s get back to the task at hand: understanding what technology helps our users best accomplish the goals their business has set out.

Photo credit via Flickr:chilsta