May 20, 2022
Back in 2007, the DevOps movement emerged as a solution to bridge the gap between software development and IT operations, and the inefficiencies it raised. With over 85% of organizations saying it’s important for their company to develop and begin production of new software quickly, DevOps became a no-brainer for organizations to embrace. Its methodology offered promises like accelerating agility in the enterprise, reducing the time spent fixing and maintaining applications, and improving speed to market.
As DevOps is entering its 15th year, more organizations are experiencing the benefits of the practice. However, going from siloed, outdated IT practices to an agile, collaborative operations is full of hurdles, challenges and failures. In fact, nearly 75% of DevOps initiatives are failing to meet expectations, according to recent Gartner predictions, due to issues around organizational learning and change.
As technology continues to evolve and new tools and innovations emerge, there are various elements of DevOps that organizations need to master in order to be an agile, resilient DevOps program -- and remain one. Adopting DevOps in an organization isn’t easy -- but the results pay off, if done correctly. The following common challenges that IT leaders and executives need to tackle in order to fully reap the benefits of DevOps.
Misalignment of Standards, Processes (and the Definition of DevOps)
One of the most overlooked aspects of establishing successful DevOps practices is defining the purpose of the shift in the first place. It sounds like a simple statement -- and it might be, but the path to adoption can differ greatly between teams looking to enable open innovation and those trying to shorten release cycles.
Today’s leadership and management teams are much more involved in the process, as they realized how critical of a role that IT operations plays in the success of today’s organizations. With this shift, users (such as the executive leadership teams) and the IT system professionals need to set expectations at the beginning of the development of their DevOps adoption process. This will help everyone to understand the constraints of the current ecosystem and recognize a solution that meets the needs of the entire enterprise.
If a financial institution was building a software product to compete with its fintech startup counterparts, for example, the leadership team needs to have conversations with the DevOps team to ensure the practice they develop is aligned with the purpose of the business objective, in this case rapid innovation. But it shouldn’t stop there – DevOps should serve to also connect various IT teams within an organization. So, for example, as security and privacy are critical elements of a quality software product, the IT security and DevOps teams need to connect on ways to ensure security is incorporated from the beginning of the product build. When it’s an afterthought, it can lead to bugs, late-stage security complications or much worse, which can turn into finger pointing and more siloes.
Cloud-Native Tools Outpacing Legacy IT Infrastructure
Nearly 75% of Fortune 1000 organizations are continuing to run their mission-critical applications on outdated IT systems like the mainframe. This can cause a lot of pain for organizations, including lack of scalability, elasticity, and extensibility – not to mention the exorbitant operating costs to maintain these legacy systems. COBOL, for instance, remains the most prominent language in these organization’s estate -- data shows that 45 of the top 50 banks, four of the top five airlines, and seven of the top 10 global retailers still rely on this approach, among others.
Today’s developers demand access to relatively new, cutting-edge tools (i.e., SASE, Docker, and Kubernetes, among many). Infrastructure as Code (IaC) has become an afterthought to modern makers, but very few cloud-native tools are available in legacy systems. No matter how tightly coordinated a DevOps practice is, they simply don’t have the ability to spin up a new mainframe with a few keystrokes. As cloud-based development tools have evolved, a growing divergence has arisen between capabilities that DevOps teams have in modern environments and those in legacy spaces. For organizations to embrace the advantages of DevOps and become truly agile, they need to liberate themselves from legacy infrastructure.
Lack of Talent and Knowledge
Acquiring and retaining quality cloud specialists and DevOps engineers has always been a challenge, and the “Great Resignation/Reorganization” has exacerbated this issue. In the world of mainframes, finding talent is a far larger problem. According to Reuters, 60.5% of COBOL developers are over the age of 45 -- and, nearly 20% of these developers are over the age of 55 -- meaning retirement is on their radar. A 2018 Forrester study also found that enterprises lost around 23% of specialist mainframe staff in the five years prior and 63% of open positions were unfilled at the time.
While this has enabled an opportunity for new and emerging IT talent to enter these posts, the reality is that a large majority aren’t familiar with procedural programming languages and mainframe architectures, and simply don’t have an interest in those skills due to a perception of job scarcity over time. The new generation was born from the digital age, taught the power of abstraction and cloud-friendly languages. In fact, a majority of enterprises say they’re concerned about having access to the right IT talent to maintain and manage their legacy systems.
With that, as engineering and IT leaders are developing and defining their technology infrastructure strategy, leadership needs to concurrently adapt and transform people strategies. Organizations need to begin investing in their people more, which is where IT skills training, upskilling and reskilling can come to play.
Will Enterprises Ever Get DevOps Right?
Organizations have a tremendous opportunity to embrace DevOps in the coming years – but leaders need to accept that most of the drastic improvements DevOps is famous for can only be realized with cloud-native concepts that simply do not exist on the mainframe.
About the Author(s)
You May Also Like