April 15, 2014 | Written by: Frank Bauerle
Share this post:
I’m worried about how we’re going to keep up with our competitors. I’m worried about how we can keep up with the rapid changes in the marketplace. I’m worried about how we can contain costs while providing the appropriate level of service to our sales, marketing, engineering and human resource organizations.
Our business is very seasonal. We have huge spikes in demand during our busy season and a large percentage of our servers sit idle much of the rest of the year.
I’m worried about how can we scale up to support those peak demands. And how can we scale down during the slow times of the year? How can we expand to and support different geographies?
Our marketing department wants us to make changes to our applications on what seems like an hourly basis. They want to build new applications for our mobile customers.
I’m struggling with how we can quickly deliver more applications.
Trying to decide what to do…
I called a meeting of my top technical leaders and discussed my concerns. After a few minutes of uncomfortable silence, Alice, our local deep thinker and the head of our enterprise architecture team, spoke up: “We’ve been thinking about this. What about looking at cloud computing?”
We’ve learned that cloud computing is more than providing virtual server capacity on demand. It’s much more multi-faceted than that.
What if we were to look at building an integrated end-to-end development and operations model leveraging DevOps principles? This new DevOps model should be based on a platform as a service (PaaS) strategy which could take advantage of infrastructure as a service (IaaS).
My team has been looking at different platforms. Right after IBM announced their Codename: BlueMix offering, we signed up for the beta program. We’ve been building and deploying prototype applications that our marketing department would kill to have.
We’ve also been evaluating the IBM SoftLayer cloud platform. It’s got a tremendous set of capabilities to deliver both virtual and bare metal capacity on demand. And everything within SoftLayer is accessible through their application programming interface (API). We think that we could use SoftLayer to help with our seasonal demand problem and integrate it with our PaaS solution to build an environment where we could build/deploy quickly and scale dynamically.”
Wow! It sounded almost too good to be true.
Then, the group spoke up. Bob, our data center manager, was worried that he would no be able to provide the specific hardware and specialized infrastructure that we used to support our business and marketing applications.
Ted, the manager of development, asked if we were ready for a DevOps model. He was a bit unsure on whether the development and operations teams could come together to build applications that could monitor themselves and scale dynamically. He acknowledged the backlog that existed for his team and recognized that he needed to do something.
Carol, our head of operations, expressed a similar concern. She talked about the separation of duties between development and operations that we traditionally had and wondered how we could pass our annual audits.
Alice spoke up again “The world as we know it is changing. The pace of change is increasing. We need to be prepared to react to this increased change so we can better serve our customers. If we don’t adapt, we may end up out of business.
“All of your concerns are valid. We need to start looking at the problems a little differently now. Rather than build infrastructure specifically for solutions, rather than building applications that require us to purchase bigger and faster hardware to scale, we need to look at how we can scale our solutions using commodity hardware scaling horizontally.”
That was it—that was the end of the discussion. I closed the meeting by asking Bob, Carol, Ted and Alice to put together a strategy and plan on how we could accomplish this.
After adopting the cloud
The team came back in two weeks, and they were pretty excited. I was impressed with the amount of thought that had gone into what they came back with.
Bob, the team’s designated spokesman, started by thanking Alice for suggesting cloud computing. He began by acknowledging that we had all been thinking about the problem incorrectly. Instead of looking for reasons why cloud computing was not right, we needed to focus on how it would work for us.
We decided to start small by developing a proof of concept (POC) for one of the more critical applications that marketing had asked us to build. We agreed to evaluate the DevOps approach and to use the POC as a driver for change within our organization.
So, how do I feel? Am I still nervous?
Of course. But I am feeling much better about our future—and, I’m learning to love the cloud.
How do you feel about the cloud? Let me know in the comments below or on Twitter @FRBauerle.