Rewind to Q4 2007. At the time, Amy Katz and I were building two IT channel blogs (later acquired by Penton in 2011). We were in stealth mode, trying to figure out this somewhat new world of blog-driven content -- the platforms, the content systems, the ad systems, the revenue models. In many ways it was our first attempt at DevOps. Well, sort of.
As WikiPedia puts it: "DevOps is a software development method that emphasizes communication, collaboration, integration, automation, and measurement of cooperation between software developers and other IT professionals."
The ultimate goal: Continual software innovation -- delivered in rapid fashion -- over and over again. Instead of "praying" and hoping that rapid innovations work, you basically know they will because of proper processes, testing, etc.
It's sort of like a never-ending concert road show that gets better and better at each city stop.
Blast From the Past
In our case back in 2007, DevOps involved:
- The underlying platform providers: Folks like Rackspace (IaaS), WordPress (content management) and MySQL (database management) constantly delivering code enhancements.
- Our own development team: Working on initial design specs, overall development and the launch of our sites.
- IT operations: In a way that was me. I'm no CIO. But I had some basic knowledge of how we wanted the overall system to work.
- Business operations: That was Amy. She was focused on monetizing the system in effective, innovative ways.
What we did right: All of our efforts paid off with a successful business launch in January 2008 -- two sites went live with barely a hitch. By May of 2008, we launched a third site. By December 2010, we replaced the third site with an alternative offering. By August 2011, Penton acquired our business. (We exited in May 2014.)
DevOps Hits and Misses
Did we really "master" DevOps during our journey as an independent company from January 2008 to August 2011? The answer, alas, is not really. Some would say, "not at all."
- We continually enhanced our sites with the latest WordPress software plugins, etc.
- We were always on the most current cloud infrastructure, database and content management infrastructure, thanks to our Rackspace, MySQL and WordPress relationships.
- We really didn't have a "process" for testing new innovations. I often shot from the hip on new content directions, and the underlying platform plumbing we'd need. The result: Sometimes untested enhancements went live on our sites just to "see" how they would perform. Twenty minutes later, I'd be dialing Rackspace for emergency help...
- We didn't "kill" antiquated features fast enough. There was no cradle-to-grave software strategy for certain functions.
- Many "new" features or functions required "manual" processes. And manual processes don't scale in a small, fast-growth business.
Still, the overall DevOps industry mindset didn't take hold until late 2008 or early 2009, so I'm not all that hard on myself.
What We Learned
That journey taught us plenty. My current line of thinking around DevOps and small businesses that want to drive innovation:
1. Automate Everything: I've said it before and I'll say it again. If you can't automate it, then don't do it.
2. Don't Be the Bottleneck: Everyone in an organization should help to drive innovation. One person can be the gatekeeper for those innovations -- but make sure he or she doesn't turn into a roadblock that slows down progress.
3. If Capacity = N: Then test your application workloads at N+20. When we launched our previous company in January 2008, we didn't really have a feel for cloud server workloads, database stress points, traffic bursts, etc. Everything worked out ok. But we should have asked around and educated ourselves more about data workloads and scalability.
4. Less Is More: Truly elegant software design often takes a really complex technology and transforms it into an extremely easy metaphor for users. An example for the non-techies: In War Games, Matthew Broderick taught a supercomputer about global thermonuclear war through the simplicity of Tic Tac Toe... (Spoiler Alert: The click below essentially reveals the movie's ending.)
Now, apply the War Games example to your business. How can you (A) simplify (B) and automate to (C) deliver enhancements in a faster and faster way?
5. Demand Continued Innovation: The good news... Today's cloud services deliver continual enhancements far more rapidly than in 2007-2008. But small businesses building services atop those clouds can't rest on their laurels. They need a DevOps approach to continually fund and deliver their own innovations atop those platforms.
Good luck in your journey. And thanks for tracking ours.
Tic... Tac... Stay Tuned.
Subscribe: Want to receive our blog headlines in your inbox each business day (except when we go stealth on Fridays)? Then subscribe to our enewsletter. Thanks to those who already have.