The popular software development methodology DevOps can be a bit confusing to beginners, especially when it encompasses other areas such as security (DevSecOps), business (BizDevOps), and the like.
So what is DevOps?
DevOps takes two previously separated teams – software development and IT operations – and turns them into one united front that creates secure code while speeding up the software development lifecycle. DevOps fundamentals include a collaborative and communicative culture, automated testing, releases and deployments, and frequent iteration. Another commonly used term in the DevOps space is DevSecOps, which refers to a DevOps practice with a specific emphasis on security.
What matters is what’s at the heart of the DevOps methodology – these four key principles that can improve your organization’s software development practice.
- Automation of the software development lifecycle
- Collaboration and communication
- Continuous improvement and minimization of waste
- Hyperfocus on user needs with short feedback loops
An examination of key DevOps principles
Roughly 15 years ago, the idea emerged to bring development and operations together in a seamless fashion. In 2009, the term “DevOps” was coined by Patrick Debois, who is considered one of the methodology’s primary gurus. DevOps includes a lot of the principles of Agile software development, but with a special emphasis on breaking down development and operations silos.
DevOps has continued to grow in popularity since that time, from small businesses to enterprises with legacy systems and nearly every size company in between. Like almost anything else, DevOps can adapt to an organization’s unique needs and environment, adjusting to what’s most important to the business.
As such, it’s possible to find many different flavors of DevOps, though, at their core, each has the following 4 must-know principles in place:
Automation of the software development lifecycle
The North Star for all DevOps teams is automation. Before DevOps, software development was a very manual effort requiring human involvement (and physical handoffs) at every stage of the process. All of this human involvement meant companies were lucky to update or release new code once a year, and many were on an 18- or 24-month release cadence.
Today so-called “elite DevOps teams” release code many times a day – and they’re able to do that largely because of automation.
To understand the power and importance of automation in DevOps, consider software testing, an often overlooked and unappreciated step that is regularly scapegoated for causing release delays. There’s no question software testing is critical; without testing companies risk releasing broken or actually even unsafe code.
But testing is perhaps the most hands-on of all the steps in DevOps, requiring people to write test cases, run myriad tests, analyze the results, and then circle back with developers for fixes. That’s all a long way of saying there’s a reason teams point to testing as the number one reason code isn’t released on time.
Enter automation and the idea that the most basic software tests could happen as the code is written. Test automation dramatically speeds up the entire process and frees software testers to look for potentially more damaging code quality issues.
Although testing is one of the most dramatic automation “wins” in DevOps, it’s far from the only one. Continuous integration automates the process of moving new code into existing code, while continuous deployment helps automate releases. And Infrastructure as Code makes it easy to automate the process of provisioning developer environments.
Collaboration and communication
A good DevOps team has automation, but a top-notch DevOps team also has collaboration and communication. The basic idea of bringing dev and ops together (as well as sec, test, stakeholders, etc.) hinges on teams being able to collaborate. And that can’t happen if there isn’t clear and regular communication.
This sounds like a deceptively simple principle of DevOps, but, like most things, the devil is in the details. Devs want to write code and move it along into the world; ops professionals focus on tools, compliance, and the cloud; and the security team wants to ensure the code is safe. Dev, ops, and sec don’t have the same priorities, might not speak the same “language,” and are likely to approach problem-solving from very different perspectives. A case in point: Dev and sec still don’t really get along, in part because the communication and collaboration gap remains wide.
It takes effort to bring teams together and often out-of-the-box thinking. And in one of those "chicken and egg" situations, teams need to communicate for successful DevOps, but DevOps itself can lead to better communication, and happier developers, according to findings in our 2021 Global DevSecOps Survey.
Continuous improvement and minimization of waste
Leaning heavily on earlier software development methodologies, including Lean and Agile, DevOps also focuses on reducing waste and continuous improvement. Whether it’s automating repetitive tasks like testing so as not to waste time, or reducing the number of steps required to release new code, well-functioning DevOps teams continue to measure performance metrics to determine areas that need improvement.
Expect teams to strive to continuously improve release times, reduce the mean-time-to-recovery, and number of bugs found, in addition to a number of other metrics.
Hyperfocus on user needs with short feedback loops
The final must-know DevOps principle is the importance of bringing the actual user into every step of this process. Through automation, improved communication and collaboration, and continuous improvement, DevOps teams can take a moment and focus on what real users really want, and how to give it to them. There’s no question that finding a way into the user mind is quite tricky, and teams can struggle to implement processes to achieve this.
The other difficult piece of this is that user feedback, once gathered, must be delivered quickly so time isn’t wasted. That’s why short feedback loops are critical, and why teams need to put energy into making them even shorter as time goes on.
Benefits of a DevOps model and practices
What happens if teams do DevOps right? In our 2021 survey, 60% of developers told us they were releasing code at least 2x faster, thanks to DevOps. Other benefits of a DevOps model include improved code quality, faster time to market, and better planning.
And for bonus points, survey takers told us that having a successful DevOps practice also made for happier developers, and there’s scientific data that shows happier devs are more productive.
What are some challenges of implementing DevOps?
DevOps can be challenging in the begining, particularly if it’s the first time being implemented within an organization. Here are some of the challenges of implementing DevOps.
-
Breaking down the silos. It may be difficult to break the mentality of development and operations being separate entities. Gather a basic understanding of the roles and responsibilities of a combination DevOps team.
-
Understanding the jargon. DevOps comes with a lot of shorthand, tech jargon, and SO many acronyms, like CI/CD. Take some time to study and remember that learning on the go is entirely normal and acceptable.
-
Migrating from legacy software. DevOps can be especially tricky for teams trying to migrate legacy software. Many tools designed for DevOps don’t work with mainframes (as one example) and integrations can be challenging even for experienced DevOps pros. It also doesn’t help that there’s a shortage of mainframe and other legacy programmers.
-
Too many tools. Teams spend so much time integrating and maintaining tools it’s getting in the way of actually developing, releasing and monitoring code. This is commonly known as the “toolchain tax.”
-
Taming the learning curve frustration. DevOps is complicated, and learning how it works is a marathon, not a sprint. Practice patience and grace with the team as implementation goes forward and rely on any resources available, such as help documentation and platform representatives – and sometimes plain old trial and error.
How to get started with DevOps in your organization
When preparing to get started with DevOps, the following preperation is required:
- Map out the goals behind DevOps implementation.
- Clarify the roles and responsibilities of each stakeholder involved.
- Start basic and grow with experience.
- Plam to automate as much as possible.
- Plan your toolchain (and remember, toolchains can always be altered).
- Set up regular progress checkpoints.
- Be prepared to constantly iterate (but after giving something enough time to prove that iterating is necessary).
What is the future of DevOps?
DevOps adoption and success experienced an enormous “jumpstart” thanks to the global pandemic. Teams moved past some of the cultural “how do we work together?” issues and matured into the “how do we adopt the right technologies?” mindset, based on results from our survey. Use of advanced technologies, including Kubernetes, DevOps platforms, and artificial intelligence (AI)/machine learning (ML) give hints as to what the future of DevOps looks like.
It’s safe to expect increased automation, smarter AI/ML-powered decision making (starting in places like code review and a more thoughtful choice of tools, such as continuing adoption of DevOps platforms to streamline the process.
