AI Insights
Showing results for 
Search instead for 
Did you mean: 

Infrastructure as code: Enterprise velocity without limits?



By:  Joe Panettieri

As the DevOps wave goes mainstream, many enterprise executives are wondering how to further speed continuous software delivery and associated application innovations. The answer increasingly involves infrastructure as code—one of the best ways to shift your legacy enterprise toward a modern, agile organization.Infrastructure as code.png

Similar to the early days of cloud computing and big data, emerging IT terms can mean different things to different people. Let's start with the absolute basics: Infrastructure as code represents the next generation of configuration management and automation. It gives DevOps even more control—and more speed and velocity—for continuous innovation. Ultimately, it involves stepping beyond agile software development all the way to platform as a service (PaaS).

Instead of taking my word for it, let's look at four different viewpoints. Then, we'll identify the common themes across all the viewpoints to pinpoint some conclusions.

View 1: The path forward

Check in with HP Enterprise, and you'll find the company's experts describing four stages of application deployments. They include:

  1. Semi-manual processes, where applications are deployed atop provisioned infrastructure with minimal automation. It's a decent start but you can't really scale.
  2. Procedural processes, which leverage a flow of sequential scripted steps. The good news: you can now fully automate the process, but you'll still get bogged down in day-to-day management of the system.
  3. The declarative stage, when the application is designed in a way that addresses an overall approach to infrastructure, monitoring, and more. This is the start of Infrastructure as code.
  4. latform as service PaaS. This also represents infrastructure as code. Here, the development and deployment environments become one, maximizing your automation, though you may face some challenges as you probe various automation methods.

HP Enterprise isn't suggesting that you completely favor one approach over another. Enterprise IT must continue to support all for deployment types based on specific project and business deeds. Still, the company thinks the semi-manual approach is on its way out. I certainly agree.

View 2: DevOps gains more control

You've heard the classic line, "Lead, follow, or get out of the way." Sure, CIOs and your executive team still need to lead. But increasingly, your IT developers are shifting from a "follower" role and more toward a "get out of the way" mindset.

As your executive team sets strategy and shares key marching orders, infrastructure as code allows DevOps to execute the orders faster than ever—without waiting on peer departments for assistance. Basically, you're removing barriers ("get out of the way") and empowering DevOps to increase its overall velocity.

As describes it:

One could argue that [infrastructure as code] is just configuration management. And in a sense it is, just fully automated. But there is another big difference. It also means that the developer can do it all. Whereas in the past all this orchestration and automation belonged to the sysadmin, now the developer at the very least has to advise, but in many cases fully take over the process.

View 3: The right tools are emerging

At first glance, your journey toward more automation requires a lot of programming. Alas, you've got to write a ton of scripts to get complete automation in place. Then, DevOps is finally empowered to write and deliver great applications. Right? Actually, that's wrong.

A range of alliances, partnerships, and business relationships allows legacy enterprise organizations to march toward infrastructure as code. One prime example involves HP Enterprise's relationship with Chef.

As The Register puts it:

What Chef's product does ... is to allow the setup of a server (or virtual machine) to be specified in a Ruby script, including what services are installed and the content of configuration files. A Chef agent on each server checks whether it complies with the latest script and updates itself if it is not. Since the code in the script determines the server setup, this is called "infrastructure as code."

But that's one piece of the larger puzzle. Within HP Enterprise's Datacenter Care effort, you'll find a range of approaches, tools, and guidance for infrastructure as code.

View 4: There are clear steps to getting started

Infrastructure as code isn't just an HP conversation. Plenty of industry pundits, analysts, and consultants are part of this broader industry movement.

Steps for getting started will certainly vary from one expert to the next. But there are some great one-page cheat sheets to help you really boil down the conversation for your executive leaders and DevOps team. A prime example is this piece from Dmitriy Samovskiy, a PaaS expert who also guides readers on their infrastructure as code journeys.

He sees the journey in two stages, from monitoring (involving seven levels of maturity) to deployment (involving five levels of maturity. Simple, concise, and potential blueprint to get you started.

Infrastructure as code: The bottom line

We've all witnessed certain IT hype cycles before. But infrastructure as code isn't a future wave filled with empty promises and vaporware (software that doesn't exist yet).

As you scan each of the four expert views above, you'll notice some common themes: Infrastructure as code is real today, backed by real tools and experts that speed the velocity of innovation. Legacy enterprises can either ride the wave or drown in a sea of manual processes.

Clearly today's market pressures mean developers have to code faster and smarter. Learn more about getting ahead of the curve by reading Get faster value from 'infrastructure as code'.

0 Kudos
About the Author