How DevOps Has Changed

To understand the current and future state of DevOps, we spoke to 40 IT executives from 37 organizations. We asked them, "How has DevOps changed since you began using the methodology?" Here’s what they said:

Acceptance

• DevOps used to be only for Silicon Valley, now it’s on everyone’s mind. How can it help our organization? It’s a mentality and methodology to fit your organization. How do we get started with the framework and the momentum? 
• Most notable is greater awareness. Instead of how different from Jenkins now aware of other tools, understand where and where the different tools in the system apply. Fluency and understanding has improved dramatically. People are understanding the importance of including the database in the process. Maturity as people figured out to get business requirements, write app code and move faster. Realize there’s more to the story of full-stack DevOps. The entire stack needs to be automated in order to be successful in speed and quality. 
• More acceptance as people are aware of the benefits and more availability of tooling and knowledge. AWS has added a lot of services over time. People are building tools and products with increased adoption of DevOps – more easily available and easier to adopt. 
• Still in infancy. Start-up SMB is using DevOps. Global 2000 is very spotty because ITIL processes are ingrained. Congealing around super developer and getting rid of ops isn’t going to happen, more clarity about the agility of application teams and enablement. Formally a buzzword, now clients looking at adopting DevOps. 
• A big change in the past year with maturity of tools, platforms, and thinking. People are embracing and implementing. People understand the benefits. Large organizations are embracing DevOps and new ways of thinking. May lag with legacy systems but seeing a desire to bring legacy processes into DevOps. 
• Four to five years ago DevOps was a mythical thing from Silicon Valley with Netflix, Facebook, and Google. Over the last few years more companies getting on stage and talking about how they’ve transformed from within the organization and individuals taking on more responsibilities and automating more. How to share more openly with others. Measurement and sharing is growing. Finance and telco organizations have their innovation labs to attract new people with a different mindset to change the way and the behavior of the traditional employees. Teams are figuring out how the company goes through digital transformation and how to build software and fit within the current organization and processes. They teach the rest of the organization how to do things differently for the future. 
• Broader adoption. Infrastructure as code is getting more attention. Culture adoption was the first wave and is now a constant. Now rolling into more tools with software and virtual hardware with infrastructure as code automation. Using automation in CI/CD for fresh environments and blue-green releases.
• Better understanding. Through conferences, articles, podcast the market is much more well informed about what you can accomplish though there may be some differences in understanding what every aspect means. People accept the concepts and want to know how to implement.

Public Cloud

• The emergence of cloud-native gave us the push. We'll always need DevOps. We needed to up our game when we began delivering code for running services. The entire team working on a well-understood process. Why we automate everything through Jenkins and adopting containers, how to instrument the cloud. We don’t want to take chances. Experiment with code to see if we’re going in the right direction. Eight different environments before you get into production. Now fewer environments.  DevOps simplifies the entire process. 
• 1) Change in attitude the rise of Kubernetes (K8s) in the public cloud. When the cloud begins to run K8s for you that’s a game changer. Having someone else run it reduces the difficulty. Can deploy large amounts of software on a platform that’s already set up. Because some tasks are automated there’s less to do. The reality is when these changes happen there will be job loss, but so far having it become easier creates more workloads. Tools are changing for those moving to K8s. Large cloud properties Intuit has gone all-in on K8s. Hiring a lot of people. 
• Tools becoming more sophisticated and the cloud environment. Used to be on-prem and nothing more. Now cloud and hybrid cloud makes it easy to spin up and spin down. Ultimately the cost of doing has gone down and the benefits have gone up. 
• Leading edge people are pushing the frontier with the public cloud and more API access and automation. Leaders continue to blaze a path forward. Most folks are nowhere near that but their recognition they need to do this to keep up with their customer-facing applications. Feel fear and urgency across the board. 
• The DevOps terms that started back in 2007 has only now being kicked into production for the majority of organizations outside the big websites. So the strategy is moving from early adoption to mass deployment only now.
• The greatest change in the last few years is not necessarily the concepts of DevOps, but how widespread attempts at adoption are and which parties are spearheading adoption. 

Containers and Kubernetes

• It’s a philosophy or thought process but a tactic alignment of DevOps and containers. It’s much easier to do virtualization of code and automation containers make it easier for developers to do work in an environment that looks like production. Containers enable a shift-left attitude. Dev community is hangry for K8s – it’s the standard for automation. K8s is the baseline and developers are jumping on the platform which is much more agile.
• Seen increased adoption. Also, see people coalescing around K8s for managing containers. Containerization is an evolution we’ve seen, and it creates a mindset where I have a configuration that I apply to my test environment and then apply to my production environment. Now build container in non-production and then move to production. VMs and containers, not or.
• More mature. More adoption in industry verticals with more conservative companies. What changed is the container movement. Docker and container orchestration more digestible for users in small and large companies. Ability to easily package software component to move in a CI/CD pipeline. Docker is not that incredible but the standard format of the images and the ability to put the image where ever you want makes DevOps and automation more feasible. K8s is a mini-cloud. These created an acceleration.
• The biggest change to DevOps has been the recent movement to containers. It seems like everyone is rolling out containers within the cloud environment. They’re the new hotness. The problem is many organizations aren’t using them strategically. Containers do many things well including acting as a cost-efficient artifact that makes development quicker. However, many businesses have found diminishing returns if they roll out too many containers.
• At the start, DevOps was primarily used as an operational philosophy. This philosophy was mostly associated with the Agile methodology for development and integrating development and operations functions within the same teams and queues. Treating infrastructure as code grew out of this. With the advent of cloud architectures -- and now containers -- it has become the default way that modern organizations approach working with infrastructure. Now DevOps is less focused on the basic requirements around infrastructure as code and more focused on how to use these concepts to provide continuous integration and delivery pipelines and ensure auditable automation is a component of every function of product delivery. This allows DevOps teams to assist with organizational cost optimization and compliance goals.
• Containerization has been a big change. We've also almost moved from Puppet to Kubernetes!

Other

• We're an early adopter of DevOps enabling the fast scaling of the business. Make sure tools, processes, and standard operating procedures are well understood. We support seeing operation metrics out of the box. Handle logistics of data movement and staying resilient across data centers.
• Since we began using DevOps, the biggest change has been the explosion of tools and infrastructure options in the market for automating the entire DevOps process. Years ago, DevOps was merely a concept of bridging communications and processes between Dev and Ops. Today, DevOps is seen more as a continuous pipeline from planning to production. To optimize the pipeline, and automate all steps in the process, a massive proliferation of tools has emerged. For deployments, it was first Puppet/Chef/Ansible and now Docker is taking over. Everything is done in the cloud. Local labs (physical and even Virtual) are a thing of the past. Cloud/SaaS has been a huge enabler of optimized DevOps pipelines.
• Rushing from Dev into Ops and releasing three a day versus three a month. This is causing SQL ingestion to get worse. 45% have information leakage, up 8% versus previous years. Adapting more DevOps and getting more vulnerabilities. 60% of detected vulnerabilities has never been fixed. 139 days to fix a vulnerability. 40 to 60% of enterprises don’t use any security testing.
• Partly greater mature adoption. How to CD software, eliminate friction, collaboration. Getting faster. The application was not changing. Seeing more non-deterministic applications. The application may change based on the data from sensors. How to test and address that sort of change. Infrastructure applying DevOps focus on CD apply AI/ML for testing and determining patterns. Can put inert features until feedback is provided. Analyze the data. Shifting from deterministic to non-deterministic application DevOps becoming more intertwined.
• Within DevOps more automation which is natural. Tools for automation more sophisticated and creating more value. DevOps went beyond R&D and more of a cultural issue. Evolves into a culture issue later in the adoption. This is a function of seeing where it can help them and fit in their processes.
• The overall principle of “infrastructure-as-code,” and of reducing the variability in how application developers configure their applications, has stayed pretty much the same. Where we've seen major changes has been on the development SDLC side, where we've adopted new technologies and tools, such as Jenkins, Docker, Kubernetes and more.
• My experience goes back to before people were even calling it ‘DevOps’, so I’ve seen a lot of changes. The biggest ones have been the creation of various frameworks like RightScale, Puppet, etc. and the switch from doing things manually to the more automated processes we see now.
• Since we began, automation has become the way to do most of DevOps. Many manual processes like testing, configuration and deployment have become automated with various tools. Also, with the advent of the cloud, DevOps automation has become cloud-centric now.