How to stop people from causing bad cloud architectures

We’ve all watched bad decisions about technology architecture made right in front of our eyes. It's time to do something about it

How to stop people from causing bad cloud architectures
Thinkstock

I’ve been in some knock-down, drag-out battles over both the configuration and the use of technology. On one side, you have somebody with a very different opinion as to what technology should be used and how. On the other side, you know that you are right.

These days the battles are about which cloud provider to choose, what database to use, what devops tool chain to engage. So many new things fly at us each day and so many more choices have to be made that conflicts are a foregone conclusion.

What drives me nuts is that there is typically one right answer to the problem—that is, one set of technologies and configurations that are the most efficient. Other solutions typically won’t fail outright, but they will work at a much-reduced efficiency.

There will be no “I told you so” moments, just millions of potentially investable dollars lost during the next several years. I call it “a stupid tax.”

The most politically savvy people usually make the architecture calls, right or wrong; however, they are typically motivated by emotion, not logic. Perhaps they like the sales team from one vendor and therefore rate their technology much higher than others. They don’t consider how well it lives up to the business requirements other than a pass/fail. Will it work or not? That never should be the question.

Advertisement

How do you remove the negative effects of people on enterprise cloud architecture decisions? I’ve found a few things that work.

First, predetermine guidelines that everyone can agree on regarding frameworks for selecting any technology and the configuration of that technology. Agreeing to a logical process will typically determine the right answer; then it’s tough for anyone to suggest that you divert from that path.

In essence, you’re using their political savvy against them. It doesn't look very good to break rules that they helped make.

Second, and most difficult, you need to change the culture. If the organization's culture is to not stick your neck out for any reason, the people with the strongest personalities will run roughshod over those who are less assertive—and in many cases, the quieter people have the right answers. Making asserting yourself part of the internal reward system is a good first step, or tweaking the decision-making process to allow for equal input from all personality types. Changes in culture must come from the top.

Forthcoming challenges are not around finding technology that can solve problems, it’s picking the most effective technology to solve that problem. People are going to make those calls, so we need to work on the human side of the process.